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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
f4ba7346-9a87-4673-94e4-dbb21f0eba6b | academic-resource-text-level-multi-label | 2203.10743 | null | https://arxiv.org/abs/2203.10743v1 | https://arxiv.org/pdf/2203.10743v1.pdf | Academic Resource Text Level Multi-label Classification based on Attention | Hierarchical multi-label academic text classification (HMTC) is to assign academic texts into a hierarchically structured labeling system. We propose an attention-based hierarchical multi-label classification algorithm of academic texts (AHMCA) by integrating features such as text, keywords, and hierarchical structure, the academic documents are classified into the most relevant categories. We utilize word2vec and BiLSTM to obtain embedding and latent vector representations of text, keywords, and hierarchies. We use hierarchical attention mechanism to capture the associations between keywords, label hierarchies, and text word vectors to generate hierarchical-specific document embedding vectors to replace the original text embeddings in HMCN-F. The experimental results on the academic text dataset demonstrate the effectiveness of the AHMCA algorithm. | ['Ang Li', 'Yawen Li', 'Yue Wang'] | 2022-03-21 | null | null | null | null | ['document-embedding'] | ['methodology'] | [-2.90732890e-01 -1.34925783e-01 -5.40476382e-01 -1.71691850e-01
-7.31886148e-01 -6.24140978e-01 4.71060783e-01 5.75998247e-01
-1.52722076e-01 1.82194680e-01 7.65030503e-01 -4.00386095e-01
-2.32270986e-01 -5.57834864e-01 9.18891653e-02 -4.59154993e-01
6.81722999e-01 4.59406555e-01 4.03919816e-02 1.52703971e-01
5.61004817e-01 1.29320860e-01 -1.35062444e+00 4.71142441e-01
1.02876556e+00 9.07264113e-01 6.98036188e-03 6.81687772e-01
-6.82214916e-01 1.11440206e+00 -5.91994107e-01 -3.22596937e-01
-4.33286577e-01 -1.45753160e-01 -1.04871571e+00 4.93395418e-01
4.52015430e-01 -2.08749726e-01 -3.87152195e-01 8.80827844e-01
2.21001983e-01 2.95689404e-01 1.19548154e+00 -1.21391630e+00
-1.48992312e+00 8.97755802e-01 -6.31579638e-01 -7.36052692e-02
6.99809790e-02 -5.97975791e-01 1.78675818e+00 -1.12294579e+00
3.67063969e-01 1.61309195e+00 4.31777507e-01 6.09941743e-02
-1.01047683e+00 -6.15651846e-01 3.74044120e-01 7.15645790e-01
-1.50107694e+00 1.73380688e-01 7.02502608e-01 -7.79467821e-01
6.12349570e-01 2.24266440e-01 3.98703486e-01 8.82876933e-01
3.87908876e-01 1.11644661e+00 7.13771164e-01 -5.64771175e-01
-1.31600723e-01 4.27489430e-01 1.09316444e+00 7.54937649e-01
1.03957519e-01 -6.06647789e-01 5.42989224e-02 -3.15410495e-01
1.54580265e-01 2.30561182e-01 3.77187855e-03 -1.79422826e-01
-7.95508206e-01 1.35723650e+00 1.98395476e-01 5.77366769e-01
-1.34011721e-02 -1.79357275e-01 8.86218727e-01 -1.42685790e-02
5.79729915e-01 4.69335973e-01 -2.73730516e-01 4.80308741e-01
-7.35869527e-01 -1.74528230e-02 2.82130212e-01 1.19252169e+00
5.02895057e-01 -4.90594693e-02 -7.01071858e-01 1.14385366e+00
7.27549970e-01 6.59769699e-02 1.06699455e+00 -8.55103254e-01
3.35024178e-01 1.13130689e+00 -3.41392815e-01 -1.08018124e+00
-3.43320727e-01 -4.59962904e-01 -6.70746148e-01 -8.27649117e-01
-4.76669818e-01 2.03173861e-01 -5.39724290e-01 8.68209004e-01
7.47015700e-02 9.91032198e-02 3.26983064e-01 2.40881428e-01
1.22343719e+00 9.97838020e-01 4.25844759e-01 -3.28836024e-01
1.61556423e+00 -1.51929510e+00 -1.16709530e+00 1.16161242e-01
1.27801120e+00 -8.32560599e-01 1.01758158e+00 5.95545880e-02
-5.09392798e-01 -8.17866385e-01 -8.29858124e-01 -5.17105341e-01
-7.82650709e-01 4.19417500e-01 2.30680823e-01 4.38833654e-01
-7.98586130e-01 -2.60103364e-02 -1.88862998e-02 3.82759385e-02
9.56975743e-02 -2.72248108e-02 -9.11173224e-03 -1.60138458e-01
-1.37407875e+00 4.36226726e-01 6.10901475e-01 -4.67181742e-01
-4.08597231e-01 -5.47118425e-01 -1.27919936e+00 4.98398483e-01
-1.12277158e-01 -1.00114815e-01 9.12754774e-01 -3.84216279e-01
-8.50656211e-01 7.09816396e-01 -3.17423373e-01 1.33026913e-01
-3.10146689e-01 1.49640754e-01 -5.12034655e-01 2.50760198e-01
5.47167420e-01 5.37924588e-01 4.37675208e-01 -1.38709641e+00
-9.11101580e-01 -3.89451653e-01 -3.76532853e-01 3.38816673e-01
-1.33007848e+00 1.52342483e-01 -3.47643882e-01 -8.49615991e-01
1.63736567e-01 -6.13230586e-01 1.16256982e-01 -8.01806271e-01
-5.28025627e-01 -1.36653268e+00 1.13259387e+00 -6.69135034e-01
1.64620483e+00 -2.05940700e+00 2.74333298e-01 -6.69813976e-02
5.55635870e-01 -1.49610311e-01 -1.88161850e-01 2.14036226e-01
1.50747716e-01 5.42815506e-01 6.51515365e-01 -4.97643352e-01
3.82775217e-01 1.74754709e-01 -5.35190880e-01 2.72335351e-01
-5.38171649e-01 9.78841186e-01 -6.94078624e-01 -1.28485405e+00
2.56729901e-01 2.53074765e-01 -2.72739530e-01 3.35907221e-01
-1.05849383e-02 -1.04960188e-01 -6.31366193e-01 6.04626417e-01
1.77428067e-01 -6.27787352e-01 1.20423086e-01 -4.76439714e-01
-1.57128572e-01 -1.60877541e-01 -6.91068888e-01 8.80130768e-01
-5.11326790e-01 6.51669562e-01 -3.96023661e-01 -7.46067047e-01
8.06634665e-01 6.17162526e-01 5.36574841e-01 -3.92158687e-01
6.30333662e-01 -1.26822710e-01 -3.71825427e-01 -5.49836934e-01
7.88593411e-01 3.10698271e-01 -5.24524510e-01 4.21163023e-01
3.54890853e-01 1.61953568e-01 2.96497464e-01 4.78445202e-01
5.71506858e-01 -2.45129794e-01 2.13840917e-01 -3.60509932e-01
7.52880991e-01 -1.67831093e-01 3.90324652e-01 5.21620274e-01
-1.15865074e-01 1.28124908e-01 3.94155473e-01 -3.48484695e-01
-8.83599997e-01 -4.12552983e-01 -5.52336395e-01 1.97776902e+00
1.31962806e-01 -7.66125560e-01 -3.92219096e-01 -8.52317035e-01
1.58453643e-01 8.80812466e-01 -8.10397208e-01 -2.74202704e-01
-3.00639540e-01 -4.50699985e-01 3.83072972e-01 6.08032823e-01
1.22365542e-01 -7.76302397e-01 2.22519383e-01 3.24932665e-01
-2.80468073e-02 -1.16773283e+00 -9.24001515e-01 1.20769151e-01
-3.58804494e-01 -9.24908221e-01 -6.44419909e-01 -1.42177391e+00
5.64052105e-01 3.77886802e-01 7.29938507e-01 -5.72572835e-03
-1.04971029e-01 4.52596098e-01 -7.42361665e-01 3.14508565e-02
-1.62885591e-01 4.14495051e-01 -9.03474987e-02 1.43246546e-01
4.52044725e-01 -9.52818338e-03 -1.83401942e-01 2.38382548e-01
-7.65251875e-01 -5.51597998e-02 2.21041471e-01 8.84612203e-01
3.81397009e-01 2.93910623e-01 4.29931134e-01 -7.52354860e-01
1.03695560e+00 -6.67453945e-01 -2.74574965e-01 6.64208889e-01
-1.10955524e+00 -9.62921605e-02 1.07691002e+00 -5.82060814e-01
-7.45553792e-01 -2.92075068e-01 -2.37446483e-02 -6.11260056e-01
-1.72202453e-01 6.01908147e-01 -1.80439427e-01 1.26201093e-01
-3.73373851e-02 4.85267013e-01 -8.23674560e-01 -5.09668767e-01
6.46952450e-01 1.37171948e+00 1.52032927e-01 -7.01402128e-01
3.01865488e-01 -1.78042978e-01 -1.55505180e-01 -6.81720376e-01
-1.47270012e+00 -8.88490200e-01 -8.36615562e-01 -2.11433262e-01
1.30067396e+00 -8.19001377e-01 -8.99936497e-01 -2.51402557e-02
-1.15480292e+00 1.54170141e-01 1.47841260e-01 2.21260533e-01
-7.02576945e-03 7.21069098e-01 -1.18448079e+00 -5.02725005e-01
-4.29420114e-01 -1.46827984e+00 1.12667000e+00 1.83972582e-01
-2.54782975e-01 -1.54156601e+00 -6.57522306e-02 7.29019105e-01
-1.61856294e-01 -3.27558786e-01 1.82513654e+00 -1.14981580e+00
2.16056004e-01 -2.95402020e-01 -5.19799650e-01 9.80998203e-02
1.91472799e-01 -4.25900705e-02 -6.72403693e-01 -4.90468353e-01
-3.16418529e-01 -6.07804298e-01 7.17470765e-01 2.54967719e-01
1.55275536e+00 -5.40505588e-01 -5.05980134e-01 3.94992769e-01
1.27904391e+00 4.10216093e-01 -1.18001781e-01 5.70915937e-01
1.33822143e+00 6.10900998e-01 4.42318946e-01 4.08192337e-01
9.86483514e-01 5.87553442e-01 1.58178642e-01 1.34811267e-01
2.12640226e-01 -9.78274271e-02 5.04628532e-02 1.77243555e+00
6.50297642e-01 -6.00132644e-01 -1.02361727e+00 5.65352798e-01
-1.58891225e+00 -4.74650919e-01 -4.80414063e-01 1.31525648e+00
9.39865112e-01 3.89416218e-02 -3.17311317e-01 1.19166106e-01
8.19431901e-01 1.78651527e-01 -3.76346856e-01 -4.12363201e-01
-1.31637212e-02 -3.44301194e-01 3.53631675e-01 6.11521721e-01
-1.26725841e+00 1.17638791e+00 6.18552351e+00 1.14313221e+00
-5.84663391e-01 2.75077075e-01 5.81847727e-01 3.68554831e-01
-5.12041450e-01 -8.24945718e-02 -1.48103976e+00 6.22735381e-01
9.88724172e-01 -5.12185216e-01 -2.22944006e-01 9.64063287e-01
-1.22430399e-01 6.56714439e-01 -8.11650276e-01 6.83133483e-01
5.14183044e-01 -1.21300983e+00 8.69428277e-01 9.84701812e-02
7.83686161e-01 -7.26344883e-01 2.05728784e-01 8.59005570e-01
8.36476147e-01 -7.67980754e-01 3.68997723e-01 7.16943219e-02
6.71125889e-01 -8.69192600e-01 7.67324150e-01 1.04349375e-01
-1.59277403e+00 -4.41160709e-01 -4.97370839e-01 4.75333959e-01
-2.05833003e-01 5.95588051e-03 -3.29199880e-01 5.54845989e-01
3.84489298e-01 1.06954110e+00 -9.81710792e-01 3.58405620e-01
3.16239372e-02 7.23638177e-01 6.62374258e-01 -2.76540369e-01
7.32430339e-01 -2.00723678e-01 -7.26983473e-02 1.21504867e+00
1.42964512e-01 -5.49030378e-02 8.16872597e-01 3.69429708e-01
-3.25796217e-01 8.43421638e-01 -1.40884832e-01 -2.18622923e-01
7.24351466e-01 1.35071564e+00 -6.95039630e-01 -7.38382936e-01
-6.14463985e-01 8.38784099e-01 3.16211641e-01 1.69453949e-01
-6.84419990e-01 -8.68202388e-01 -2.15034876e-02 -3.32618266e-01
1.14708975e-01 3.28525268e-02 -1.61835745e-01 -1.00182998e+00
-4.26759064e-01 -5.77356756e-01 8.07396054e-01 -8.20988536e-01
-1.28411806e+00 5.25693178e-01 -2.75545955e-01 -9.44842160e-01
2.10985705e-01 -5.80173671e-01 -3.40203524e-01 5.64590216e-01
-1.22864568e+00 -1.46600461e+00 -1.99823350e-01 3.10423255e-01
1.18657732e+00 -5.08006096e-01 1.06261432e+00 3.16426843e-01
-9.58307743e-01 8.53066802e-01 8.23486865e-01 4.16900337e-01
7.14946568e-01 -1.48657846e+00 -2.86863267e-01 1.94455400e-01
-1.99617557e-02 5.58283806e-01 2.11486965e-01 -6.48222864e-01
-7.96720982e-01 -1.46991324e+00 1.17512703e+00 -3.79154354e-01
1.01097190e+00 -2.35179663e-01 -1.00965703e+00 9.16009843e-01
7.68979430e-01 -3.83206546e-01 1.44522607e+00 9.86488611e-02
-4.50204432e-01 8.86464864e-02 -6.42010331e-01 6.53277576e-01
2.22216189e-01 -6.00137115e-01 -9.15892303e-01 7.79497504e-01
1.40753400e+00 1.64254516e-01 -1.54073822e+00 6.89333901e-02
2.73465455e-01 -2.60463338e-02 8.45924735e-01 -7.24901438e-01
7.00862408e-01 2.01889709e-01 -1.53362229e-01 -1.15860224e+00
-9.98898685e-01 2.72133917e-01 -2.74601132e-01 1.49074411e+00
5.28792031e-02 -3.40910822e-01 2.63671607e-01 2.01968148e-01
-2.79688448e-01 -8.67428362e-01 -6.33494794e-01 -4.01542008e-01
6.23015523e-01 1.41006336e-01 5.52391350e-01 1.63798606e+00
3.94161314e-01 9.70036685e-01 -1.70424700e-01 6.97218254e-03
7.56905675e-01 5.06208599e-01 4.15335074e-02 -1.71791029e+00
3.58973801e-01 -7.87572980e-01 -1.76059648e-01 -1.04455280e+00
1.10638285e+00 -1.34702790e+00 -4.08855259e-01 -1.71240783e+00
4.48926270e-01 -2.16218874e-01 -7.23938167e-01 5.00556827e-01
-5.22365093e-01 4.32029441e-02 -2.34562278e-01 6.38996720e-01
-1.29023015e+00 8.00558686e-01 1.03826737e+00 -6.47762537e-01
4.99975793e-02 -5.25747538e-01 -7.12586939e-01 4.76539522e-01
5.06105542e-01 -5.01328826e-01 -2.38142326e-01 -2.79187381e-01
-3.26383375e-02 5.27981147e-02 -3.88386041e-01 -5.98127306e-01
3.93450260e-01 -2.34088793e-01 5.50653696e-01 -7.85662293e-01
4.06261086e-02 -7.61133373e-01 -7.55932868e-01 2.91941732e-01
-1.11637902e+00 9.39626917e-02 -5.67457173e-04 2.55593747e-01
-2.46696427e-01 -5.87191582e-01 5.99858761e-01 4.33294065e-02
-6.02223814e-01 2.25596100e-01 -9.83098388e-01 2.07605772e-02
9.24368799e-01 1.53822765e-01 -3.18929106e-01 -1.70954056e-02
-7.78856516e-01 8.59163761e-01 5.05051315e-02 8.37963402e-01
6.50357544e-01 -1.72352374e+00 -6.23183191e-01 -3.20561118e-02
4.74024594e-01 -3.63428414e-01 2.19650790e-01 1.83585271e-01
-6.95154965e-02 9.48965132e-01 1.35687590e-01 -2.81451583e-01
-1.55888379e+00 6.98089421e-01 1.83597282e-02 -5.32612264e-01
-5.04096627e-01 5.50236762e-01 2.96623141e-01 -5.34083009e-01
5.78720689e-01 1.26452118e-01 -1.34350753e+00 6.34058297e-01
4.27521735e-01 3.11735988e-01 -2.42326453e-01 -9.79966044e-01
1.38213456e-01 8.26583982e-01 -5.82676530e-01 3.44278425e-01
8.76869380e-01 -3.53362650e-01 -4.67415690e-01 6.27469361e-01
1.62279880e+00 -1.51670769e-01 -3.79543781e-01 -4.55118388e-01
3.08313906e-01 -5.60894646e-02 5.75433910e-01 -5.98690696e-02
-8.47549438e-01 8.39514732e-01 5.97724915e-01 4.39535558e-01
5.13335168e-01 2.41776779e-02 8.10471892e-01 3.52799118e-01
-3.78558040e-01 -1.17640293e+00 4.38214272e-01 6.65701687e-01
4.02566493e-01 -7.96085417e-01 2.08806731e-02 -1.91516161e-01
-4.71196085e-01 1.25435126e+00 9.20070946e-01 3.25045615e-01
6.40397787e-01 -4.68170196e-02 1.24913342e-01 -1.95902362e-01
-6.35252655e-01 -6.36636540e-02 5.06597996e-01 -2.14727581e-01
6.13221526e-01 -2.96019129e-02 -3.69992942e-01 8.95784318e-01
4.67422530e-02 -6.53986216e-01 4.02718455e-01 7.02480435e-01
-8.43284547e-01 -9.47271585e-01 -3.98852527e-01 6.67690217e-01
-5.75432599e-01 -2.57946372e-01 -2.85812646e-01 2.18586996e-01
1.71857700e-01 1.13813901e+00 1.82871610e-01 -5.65751016e-01
1.30264148e-01 5.71700513e-01 -1.46762863e-01 -8.62448335e-01
-5.36025286e-01 2.30692476e-01 -4.45256203e-01 2.75393933e-01
1.93835776e-02 -4.02985245e-01 -1.14809871e+00 1.97702963e-02
-5.70763350e-01 9.18416619e-01 2.33096644e-01 1.02682781e+00
1.98217139e-01 1.16122234e+00 9.87166524e-01 -4.55331087e-01
-2.40299180e-01 -1.27670252e+00 -6.20969176e-01 5.09468555e-01
9.31475237e-02 -6.39459252e-01 -6.92778289e-01 1.42686352e-01] | [10.332283020019531, 6.678091526031494] |
7c5b3261-15d5-4630-bcac-4ad7d225b082 | data-efficient-direct-speech-to-text | 1911.04283 | null | https://arxiv.org/abs/1911.04283v2 | https://arxiv.org/pdf/1911.04283v2.pdf | Data Efficient Direct Speech-to-Text Translation with Modality Agnostic Meta-Learning | End-to-end Speech Translation (ST) models have several advantages such as lower latency, smaller model size, and less error compounding over conventional pipelines that combine Automatic Speech Recognition (ASR) and text Machine Translation (MT) models. However, collecting large amounts of parallel data for ST task is more difficult compared to the ASR and MT tasks. Previous studies have proposed the use of transfer learning approaches to overcome the above difficulty. These approaches benefit from weakly supervised training data, such as ASR speech-to-transcript or MT text-to-text translation pairs. However, the parameters in these models are updated independently of each task, which may lead to sub-optimal solutions. In this work, we adopt a meta-learning algorithm to train a modality agnostic multi-task model that transfers knowledge from source tasks=ASR+MT to target task=ST where ST task severely lacks data. In the meta-learning phase, the parameters of the model are exposed to vast amounts of speech transcripts (e.g., English ASR) and text translations (e.g., English-German MT). During this phase, parameters are updated in such a way to understand speech, text representations, the relation between them, as well as act as a good initialization point for the target ST task. We evaluate the proposed meta-learning approach for ST tasks on English-German (En-De) and English-French (En-Fr) language pairs from the Multilingual Speech Translation Corpus (MuST-C). Our method outperforms the previous transfer learning approaches and sets new state-of-the-art results for En-De and En-Fr ST tasks by obtaining 9.18, and 11.76 BLEU point improvements, respectively. | ['Sathish Indurthi', 'Sangha Kim', 'Houjeung Han', 'Chanwoo Kim', 'Nikhil Kumar Lakumarapu', 'Insoo Chung', 'Beomseok Lee'] | 2019-11-11 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 4.70912486e-01 1.79728903e-02 -1.40732467e-01 -3.96386653e-01
-1.58919704e+00 -6.90446377e-01 7.52053678e-01 -2.12046906e-01
-5.10350406e-01 7.83478439e-01 3.29860479e-01 -8.41350019e-01
5.09867489e-01 -2.61395007e-01 -9.63464439e-01 -5.00575602e-01
6.05159998e-01 9.23000991e-01 7.12035969e-02 -3.54841590e-01
-1.57660782e-01 -4.58157947e-03 -7.93599784e-01 7.96718240e-01
9.87690806e-01 6.25636458e-01 7.99764812e-01 6.88512027e-01
-3.93263191e-01 5.46817422e-01 -5.67693532e-01 -6.10759318e-01
-1.05858669e-02 -5.43968022e-01 -1.01052475e+00 -7.48552233e-02
-1.97123550e-02 -1.19469412e-01 -2.92987287e-01 7.99856603e-01
7.36395121e-01 8.50720331e-02 5.51833391e-01 -9.17705715e-01
-6.23672962e-01 8.91310990e-01 -4.62329060e-01 2.11421281e-01
3.07274640e-01 9.13671851e-02 7.56168306e-01 -1.19800889e+00
4.84819680e-01 1.39794195e+00 1.72301963e-01 7.68501699e-01
-1.08902514e+00 -5.99263370e-01 3.39278914e-02 1.76036611e-01
-1.08764923e+00 -1.04959273e+00 3.92092437e-01 -2.20189124e-01
1.37894666e+00 1.47835866e-01 1.01852760e-01 1.65193057e+00
1.52684629e-01 1.03994215e+00 1.19580984e+00 -6.93728030e-01
1.55579560e-02 3.09113145e-01 -2.74385154e-01 2.72436470e-01
-4.34553057e-01 -3.01470142e-03 -8.32719207e-01 3.67938504e-02
4.04516250e-01 -3.26719016e-01 -2.17348650e-01 1.74586266e-01
-1.63272798e+00 4.47887480e-01 -3.71161439e-02 3.93261164e-01
-3.73203486e-01 -2.29359463e-01 6.26441121e-01 7.07562149e-01
6.06294274e-01 -6.34735152e-02 -9.02056873e-01 -4.42756385e-01
-6.36844099e-01 -3.48110199e-01 6.51028037e-01 1.20134902e+00
7.32948601e-01 1.40351519e-01 -2.04948112e-01 1.23155582e+00
2.84301460e-01 8.64475429e-01 8.53585005e-01 -2.21948206e-01
1.44003165e+00 1.91967845e-01 -9.32927281e-02 -2.08002090e-01
4.60525118e-02 -4.81143355e-01 -8.45434666e-01 -5.40889621e-01
1.21688440e-01 -3.15434605e-01 -1.05384147e+00 1.74919391e+00
2.29971707e-01 3.16966698e-02 6.99923396e-01 5.84481180e-01
7.05106378e-01 1.17781389e+00 7.80772865e-02 -3.63895506e-01
1.27977049e+00 -1.43605220e+00 -8.43812406e-01 -6.39360726e-01
9.62738097e-01 -1.24298203e+00 1.20948708e+00 3.32162231e-02
-1.25043297e+00 -6.64912343e-01 -6.19612694e-01 -8.45302194e-02
-1.58274844e-01 5.90689600e-01 -3.38269919e-02 4.53530103e-01
-1.05623901e+00 2.34082997e-01 -9.34011579e-01 -6.09625340e-01
-2.07324605e-02 4.29240972e-01 -3.26865047e-01 -1.41815618e-01
-1.23076403e+00 1.12158096e+00 3.82702589e-01 1.96473315e-01
-1.16359007e+00 -2.78311491e-01 -7.70337403e-01 -2.59440634e-02
2.36435696e-01 -6.29739106e-01 1.71102178e+00 -1.25543559e+00
-1.99850726e+00 7.46280432e-01 -5.96330404e-01 -2.77261883e-01
4.00846362e-01 -2.17476279e-01 -5.47388971e-01 -9.49991345e-02
-1.33270442e-01 3.40278119e-01 8.45198452e-01 -1.03476751e+00
-7.41726518e-01 -3.82926077e-01 -1.44754902e-01 6.44489944e-01
-3.23870301e-01 4.35646117e-01 -3.49030226e-01 -5.86030424e-01
4.49136198e-02 -1.02351415e+00 1.07198060e-01 -8.10411394e-01
-2.49864757e-01 -1.56938493e-01 6.69891000e-01 -1.06880128e+00
1.06463575e+00 -2.12145567e+00 4.66503441e-01 -3.41867119e-01
-3.51208776e-01 5.48871398e-01 -4.67936665e-01 7.80695558e-01
1.24137476e-02 4.52534556e-02 -1.45314991e-01 -7.00791419e-01
-1.96529672e-01 2.37710297e-01 -2.57013410e-01 8.90948996e-02
1.68461457e-01 1.03979254e+00 -8.07283640e-01 -3.94027919e-01
1.90160125e-01 3.19083035e-01 4.01861295e-02 4.78541553e-01
-2.01570541e-01 9.01214063e-01 -3.89681548e-01 4.21867788e-01
3.20984393e-01 -6.51845783e-02 2.68237621e-01 -1.04835257e-01
-5.34863658e-02 1.01844871e+00 -4.93290722e-01 1.90685463e+00
-1.11315608e+00 3.98498505e-01 7.39655420e-02 -1.08763719e+00
1.03380024e+00 9.57841456e-01 9.24416035e-02 -9.21251476e-01
9.85202193e-02 6.40795410e-01 1.98141664e-01 -4.35447037e-01
3.13872099e-01 -2.20319614e-01 -2.95992512e-02 5.95130622e-01
2.41925552e-01 4.00322601e-02 -1.99385911e-01 -1.38456747e-01
9.53147173e-01 2.34648451e-01 2.11431712e-01 2.93627501e-01
7.43481576e-01 -4.24096882e-02 4.22601312e-01 2.46403009e-01
7.89880753e-02 3.98417622e-01 -9.36146919e-03 -1.33818939e-01
-1.08424020e+00 -8.83204877e-01 4.00656939e-01 1.24666655e+00
-2.94240713e-01 -2.87605613e-01 -8.60559464e-01 -7.42970586e-01
-6.21012092e-01 8.80459666e-01 -9.62974504e-02 -2.57240415e-01
-8.36882770e-01 -5.62602878e-01 6.99416637e-01 4.50641572e-01
4.61595327e-01 -1.10802555e+00 1.85872734e-01 4.60546166e-01
-8.81824017e-01 -1.51205635e+00 -7.66886652e-01 2.33466968e-01
-1.07015908e+00 -4.20474261e-01 -8.42903137e-01 -9.86817837e-01
5.64016759e-01 3.26649666e-01 9.24224019e-01 -3.81534457e-01
5.29400527e-01 1.02995988e-02 -5.75103402e-01 -1.82486579e-01
-1.15742874e+00 4.61710036e-01 9.92610082e-02 2.87359476e-01
3.19990844e-01 -3.97893459e-01 -1.06444836e-01 5.90065122e-01
-5.62738359e-01 3.86278868e-01 1.12881362e+00 1.04970920e+00
5.41755080e-01 -5.37846029e-01 7.24871457e-01 -6.43967450e-01
5.23109972e-01 -4.98383850e-01 -3.57041597e-01 6.41587794e-01
-3.03660542e-01 8.42439681e-02 8.31431210e-01 -7.61872470e-01
-1.35762525e+00 1.54220715e-01 -2.98400700e-01 -4.80048597e-01
-9.90097970e-02 6.58551514e-01 -3.98610920e-01 3.75222534e-01
5.53854108e-01 7.45581090e-01 -2.57851511e-01 -5.74531496e-01
2.69345582e-01 1.45204878e+00 3.69285822e-01 -6.27621472e-01
7.50061572e-01 -2.42847815e-01 -5.93154788e-01 -6.85388327e-01
-5.53001046e-01 -4.40249145e-01 -7.71185160e-01 9.38140526e-02
7.73233294e-01 -1.15970659e+00 -1.26548275e-01 5.58169544e-01
-1.54820311e+00 -5.77203214e-01 1.80545479e-01 7.50372171e-01
-7.50129879e-01 1.92764074e-01 -7.02094316e-01 -7.64602423e-01
-5.97764611e-01 -1.36375391e+00 1.28966057e+00 -3.15242887e-01
-4.29629488e-03 -9.69036162e-01 3.90307717e-02 7.36376584e-01
3.88081968e-01 -5.60744464e-01 1.00985932e+00 -9.03096497e-01
-4.46560562e-01 1.02719374e-01 -1.95656680e-02 5.07575154e-01
3.63584429e-01 -4.33615774e-01 -9.75374103e-01 -4.89320338e-01
-6.88572507e-03 -4.21821922e-01 3.69257510e-01 8.75928029e-02
5.22307694e-01 -4.48995322e-01 -2.67915994e-01 4.35580462e-01
9.75158870e-01 3.83430064e-01 5.63124657e-01 1.89126208e-01
5.91554403e-01 5.87657511e-01 7.87870884e-01 -6.76278695e-02
6.23138189e-01 9.67566013e-01 -1.67162940e-01 -1.03597574e-01
-2.91768581e-01 -4.72597241e-01 1.14634550e+00 1.72052598e+00
2.27763489e-01 -5.04989266e-01 -1.04979253e+00 5.14427423e-01
-1.88151932e+00 -5.14479339e-01 8.75189528e-02 2.31750393e+00
1.06393647e+00 -2.10793130e-02 3.29563171e-02 -2.26268589e-01
9.11646008e-01 -4.52790223e-02 -2.69509673e-01 -6.05982125e-01
1.34619232e-03 5.14257103e-02 2.41577953e-01 5.78274846e-01
-6.72494352e-01 1.47464621e+00 5.06737185e+00 9.46582019e-01
-1.28124881e+00 7.11642206e-01 5.13551652e-01 1.14104584e-01
-2.50054181e-01 1.34289712e-01 -8.59378576e-01 3.51287067e-01
1.57903361e+00 -2.00805500e-01 7.60457754e-01 4.91237879e-01
3.29138190e-01 3.43905330e-01 -1.32226121e+00 1.12141919e+00
5.02191335e-02 -8.69676232e-01 2.78535157e-01 -1.12965874e-01
5.36830783e-01 4.32874024e-01 -8.78981128e-02 5.96041262e-01
2.70591974e-01 -7.77611673e-01 7.77471960e-01 -1.86803341e-02
1.12170744e+00 -5.46530306e-01 8.26771200e-01 7.53299654e-01
-1.20418537e+00 1.25326708e-01 -3.42952400e-01 2.16389477e-01
3.89189482e-01 2.04894871e-01 -1.18003523e+00 9.39680934e-01
3.14475477e-01 4.92946684e-01 -2.43042842e-01 4.30020720e-01
-3.32332104e-01 8.35755944e-01 -1.99390143e-01 -6.47634938e-02
2.28247941e-01 -1.41701698e-01 5.01399755e-01 1.33741617e+00
6.03264749e-01 -1.86035067e-01 1.65452078e-01 5.00121534e-01
-3.55011165e-01 5.39090157e-01 -6.09842062e-01 -3.00018340e-01
5.44227719e-01 1.07851124e+00 -2.77929336e-01 -5.01722336e-01
-6.30509973e-01 1.34239161e+00 4.98844504e-01 5.34147203e-01
-6.29151881e-01 -4.99514453e-02 3.85688156e-01 -9.01080146e-02
1.28498614e-01 -4.18224037e-01 -1.77504998e-02 -1.29287171e+00
2.53010839e-01 -1.38658476e+00 5.20343818e-02 -7.54914939e-01
-1.01525557e+00 1.12394774e+00 -2.39714146e-01 -1.29007339e+00
-6.21694207e-01 -2.57523805e-01 -4.06479210e-01 1.25511193e+00
-1.62926316e+00 -1.58598769e+00 2.13204473e-01 6.97600305e-01
1.18774164e+00 -4.70118374e-01 8.40809107e-01 4.87443238e-01
-5.18851638e-01 7.74836361e-01 2.60235578e-01 1.41797498e-01
9.60914731e-01 -8.76918435e-01 8.95978272e-01 8.77317369e-01
3.07592720e-01 4.30969626e-01 2.48343959e-01 -7.03588009e-01
-1.56656468e+00 -1.17822611e+00 1.33192313e+00 -4.32904005e-01
6.68565392e-01 -6.80296302e-01 -7.66674578e-01 1.00573850e+00
3.64332587e-01 -2.86128938e-01 5.09867907e-01 9.87882093e-02
-1.48621395e-01 -1.22123636e-01 -7.26126254e-01 7.03607142e-01
9.41907108e-01 -8.61958444e-01 -6.97008550e-01 4.34082896e-01
9.01166558e-01 -5.48366785e-01 -7.14920402e-01 1.75549716e-01
3.16179484e-01 -3.18439364e-01 6.06174767e-01 -6.44718230e-01
2.50770986e-01 -1.31118760e-01 -3.89600247e-01 -1.76066208e+00
3.88092436e-02 -9.78633285e-01 -2.23779660e-02 1.36274707e+00
9.09155011e-01 -6.71302736e-01 1.57129139e-01 2.86401995e-02
-6.45539761e-01 -4.56420004e-01 -1.20672774e+00 -8.96814883e-01
1.43927217e-01 -3.25784147e-01 6.52008653e-01 8.99541080e-01
-5.95533438e-02 9.71237421e-01 -4.38523650e-01 3.38636935e-01
2.26829678e-01 -2.44883671e-01 8.73857021e-01 -7.67765641e-01
-3.96106780e-01 -1.52310267e-01 2.46840879e-01 -1.33996975e+00
4.21036094e-01 -1.23168349e+00 1.77587077e-01 -1.60184300e+00
8.63830745e-02 -4.12599742e-01 -2.05055892e-01 6.73278451e-01
-1.76273167e-01 -2.80342817e-01 1.07157834e-01 4.07867759e-01
-3.48865837e-01 7.44990170e-01 1.24051178e+00 -1.47012532e-01
-3.19532484e-01 1.64444655e-01 -2.09717542e-01 2.77681231e-01
8.44905674e-01 -6.23510957e-01 -4.16494161e-01 -1.07767606e+00
-1.52043074e-01 5.75499594e-01 -1.18375286e-01 -5.71260393e-01
2.74298251e-01 -1.07054971e-01 3.58570591e-02 -5.24119318e-01
4.03895557e-01 -6.64705396e-01 8.89729187e-02 1.97082266e-01
-3.32263529e-01 3.57301235e-01 1.92940906e-01 2.18428716e-01
-3.76617938e-01 -1.47769213e-01 6.67542458e-01 -1.59925893e-01
-3.22635442e-01 2.01847896e-01 -4.82577771e-01 -1.39522433e-01
6.46304011e-01 -1.51211570e-03 -3.47101063e-01 -4.66498256e-01
-6.26212239e-01 1.10944435e-01 1.35957643e-01 8.56773078e-01
5.51633894e-01 -1.25553584e+00 -1.08788466e+00 1.91231862e-01
2.81208783e-01 -8.93613696e-02 1.36411086e-01 1.07413948e+00
-4.93982155e-03 5.81088483e-01 -5.22726104e-02 -6.97928190e-01
-1.25133348e+00 3.92130405e-01 3.24823618e-01 -4.59135085e-01
-3.90552610e-01 4.73786265e-01 2.24690616e-01 -9.56368506e-01
1.66095287e-01 -1.56004846e-01 1.60914987e-01 -2.33742759e-01
3.39209050e-01 1.53050631e-01 4.28182632e-01 -8.53580892e-01
-2.98049957e-01 4.03070718e-01 -3.31037551e-01 -6.22648180e-01
1.11395955e+00 -6.28443420e-01 6.64412379e-02 5.96906006e-01
1.10056913e+00 -8.67150202e-02 -7.47512877e-01 -6.58748150e-01
7.77027309e-02 -8.98914039e-02 -1.19162546e-02 -1.07855058e+00
-6.94557786e-01 1.29215622e+00 4.94165331e-01 -4.71382678e-01
1.04615855e+00 9.89235342e-02 1.19593847e+00 5.65464556e-01
5.02912283e-01 -1.15851831e+00 5.44066429e-02 8.41616809e-01
8.87957454e-01 -1.26174235e+00 -7.14346349e-01 -2.09787846e-01
-8.03763628e-01 1.00623262e+00 4.74816412e-01 6.74565494e-01
8.10588226e-02 1.79665446e-01 4.55583632e-01 3.97282600e-01
-1.07940984e+00 -1.17541388e-01 2.55012751e-01 6.12085700e-01
6.31678224e-01 1.20882444e-01 -9.48283076e-02 4.63521630e-01
-1.91058025e-01 -1.03911012e-01 1.18911020e-01 8.60784233e-01
-1.89985529e-01 -1.56619310e+00 -3.61047596e-01 1.11570791e-03
-4.68973726e-01 -4.31589037e-01 -4.16530639e-01 4.35666651e-01
-3.77172023e-01 1.30126882e+00 -2.97742218e-01 -5.03878474e-01
4.77312446e-01 4.53755498e-01 3.96321982e-01 -8.64707530e-01
-7.63123810e-01 4.34601814e-01 3.59035403e-01 -2.31220767e-01
-1.92900211e-01 -6.13807261e-01 -1.09400892e+00 -9.32781696e-02
-4.47722286e-01 2.38665998e-01 9.86963153e-01 1.22799897e+00
4.16882396e-01 4.74390239e-01 9.11784112e-01 -5.93090177e-01
-7.92157292e-01 -1.47219920e+00 3.58807035e-02 1.69347733e-01
3.18565845e-01 -2.00671017e-01 -1.24585144e-01 1.64247960e-01] | [14.487459182739258, 7.2103190422058105] |
ab8a0386-c7e6-4796-bdc2-c0c80d33a182 | image-coupled-volume-propagation-for-stereo | 2301.00695 | null | https://arxiv.org/abs/2301.00695v1 | https://arxiv.org/pdf/2301.00695v1.pdf | Image-Coupled Volume Propagation for Stereo Matching | Several leading methods on public benchmarks for depth-from-stereo rely on memory-demanding 4D cost volumes and computationally intensive 3D convolutions for feature matching. We suggest a new way to process the 4D cost volume where we merge two different concepts in one deeply integrated framework to achieve a symbiotic relationship. A feature matching part is responsible for identifying matching pixels pairs along the baseline while a concurrent image volume part is inspired by depth-from-mono CNNs. However, instead of predicting depth directly from image features, it provides additional context to resolve ambiguities during pixel matching. More technically, the processing of the 4D cost volume is separated into a 2D propagation and a 3D propagation part. Starting from feature maps of the left image, the 2D propagation assists the 3D propagation part of the cost volume at different layers by adding visual features to the geometric context. By combining both parts, we can safely reduce the scale of 3D convolution layers in the matching part without sacrificing accuracy. Experiments demonstrate that our end-to-end trained CNN is ranked 2nd on KITTI2012 and ETH3D benchmarks while being significantly faster than the 1st-ranked method. Furthermore, we notice that the coupling of image and matching-volume improves fine-scale details as demonstrated by our qualitative analysis. | ['Eduard Zell', 'Oh-Hun Kwon'] | 2022-12-30 | null | null | null | null | ['stereo-matching-1'] | ['computer-vision'] | [ 3.32333893e-02 1.37255937e-01 2.95105964e-01 -4.98772919e-01
-6.84467554e-01 -5.11946261e-01 5.79229236e-01 2.66197920e-01
-6.36461794e-01 2.05128506e-01 1.87074497e-01 -1.78802073e-01
2.68403947e-01 -1.07189500e+00 -9.41318095e-01 -5.22412539e-01
1.54271066e-01 2.29919672e-01 7.22983658e-01 -3.05323482e-01
3.86645168e-01 7.48642743e-01 -1.62318528e+00 5.06085038e-01
6.29864216e-01 1.53038156e+00 2.13647112e-01 3.73245984e-01
-3.70991319e-01 4.10128683e-01 -1.27683416e-01 -3.82343262e-01
7.76030123e-01 6.43365309e-02 -8.54866564e-01 3.41190696e-02
1.10037708e+00 -6.46694422e-01 -3.18338037e-01 1.06712878e+00
4.52823728e-01 -6.16719872e-02 3.74713629e-01 -9.75912213e-01
-1.55716270e-01 2.71314472e-01 -7.25989044e-01 -3.84082720e-02
2.38911271e-01 4.11000162e-01 9.80706930e-01 -1.08488953e+00
6.78910792e-01 1.34705055e+00 8.44125390e-01 2.51647741e-01
-1.22345757e+00 -5.95914543e-01 2.92068034e-01 -5.77382408e-02
-1.13645172e+00 -5.75231537e-02 8.82183373e-01 -5.19541800e-01
1.22247326e+00 1.80594902e-02 8.28283489e-01 5.90688288e-01
5.61812893e-02 5.72683096e-01 9.83958066e-01 -1.14529200e-01
-1.74043439e-02 -2.15663373e-01 7.84062222e-02 8.32567811e-01
-9.14767943e-03 3.92256051e-01 -4.86305058e-01 1.20559976e-01
1.07899463e+00 1.64004758e-01 -2.56948024e-01 -5.89015961e-01
-1.21525788e+00 7.05742657e-01 1.05342197e+00 3.48807126e-02
-2.52992213e-01 4.30010051e-01 2.66311765e-01 2.00975627e-01
5.52700937e-01 2.51587689e-01 -5.64687669e-01 1.70511007e-01
-9.10621643e-01 3.64204973e-01 3.42369705e-01 9.17866170e-01
1.41200244e+00 -5.85127473e-01 -3.05065632e-01 6.32516980e-01
2.66455472e-01 3.19341272e-01 1.89110097e-02 -1.08590901e+00
6.92117929e-01 1.09426880e+00 5.44209853e-02 -8.43597353e-01
-6.02363646e-01 -5.22993505e-01 -8.31507206e-01 4.87935871e-01
6.57582104e-01 2.35716373e-01 -9.28001583e-01 1.44928694e+00
4.98375297e-01 -8.70625451e-02 -2.96731800e-01 1.04827011e+00
9.14300323e-01 4.01313215e-01 -2.04960167e-01 6.27859831e-01
1.41810369e+00 -1.28837943e+00 7.56478682e-02 -3.98909718e-01
6.89289987e-01 -8.63915324e-01 9.08514977e-01 1.43548921e-02
-1.36906838e+00 -7.71393895e-01 -1.10931826e+00 -9.10182357e-01
-4.54982877e-01 -5.29166684e-02 7.80508220e-01 5.64826913e-02
-1.32994199e+00 7.53554940e-01 -4.90019560e-01 -1.63469478e-01
6.41293228e-01 3.38823140e-01 -5.42495012e-01 -2.12743491e-01
-1.03233135e+00 6.61626101e-01 1.44393519e-01 2.55717069e-01
-5.22674322e-01 -9.57944989e-01 -1.15085065e+00 1.59223810e-01
-2.30593365e-02 -1.03864598e+00 1.02993810e+00 -6.69296980e-01
-1.12866080e+00 1.35423267e+00 -1.34133816e-01 -4.13646966e-01
8.80123556e-01 -6.72642812e-02 4.02832955e-01 2.36254647e-01
2.09172964e-01 1.22435880e+00 6.96692467e-01 -1.06999242e+00
-9.81895030e-01 -6.37839317e-01 2.78383642e-01 1.35131553e-01
1.84517987e-02 -4.13110256e-01 -7.45265424e-01 -4.43733841e-01
5.11293411e-01 -7.12918699e-01 -2.71224618e-01 5.54168582e-01
-2.78522700e-01 5.99969588e-02 5.24481297e-01 -4.99787420e-01
7.37301350e-01 -2.32099795e+00 5.81728481e-02 1.55789852e-01
4.69022721e-01 -6.67216331e-02 -2.56034315e-01 2.02998713e-01
-1.31897880e-02 -4.13916148e-02 -2.69519508e-01 -7.47304559e-01
-1.21894822e-01 -1.39866158e-01 -4.01293129e-01 4.32162225e-01
5.28867126e-01 1.19639373e+00 -8.32855582e-01 -2.37949654e-01
4.79781181e-01 5.80322802e-01 -9.05029297e-01 1.12683266e-01
-1.97100326e-01 3.45017791e-01 -3.25447321e-01 5.80104172e-01
1.23623204e+00 -2.91180789e-01 -3.68796080e-01 -4.82771337e-01
-5.03842831e-01 4.88139778e-01 -9.79204655e-01 2.14422059e+00
-6.84335649e-01 6.59709036e-01 1.75281718e-01 -7.28667736e-01
7.81595767e-01 -2.72687614e-01 2.26003453e-01 -1.02996731e+00
1.94406107e-01 4.00937378e-01 -2.81206757e-01 -7.49613345e-02
5.12822807e-01 2.22388998e-01 -7.76917040e-02 1.40253276e-01
1.42070144e-01 -4.91606683e-01 -7.40807503e-02 3.74206230e-02
9.92959857e-01 3.01506728e-01 -3.70551497e-02 -2.67933547e-01
6.09620929e-01 9.28693265e-03 1.98818877e-01 5.59658110e-01
-5.28328940e-02 1.00087106e+00 5.05525529e-01 -8.87459338e-01
-9.93436217e-01 -1.09824049e+00 -1.40722975e-01 7.14186132e-01
5.14482379e-01 -2.46240005e-01 -6.40310168e-01 -6.14149809e-01
2.26382315e-01 1.89724788e-02 -8.81497204e-01 7.63126463e-02
-7.10796356e-01 -3.12005371e-01 2.60690331e-01 6.72677159e-01
9.77793097e-01 -8.12096179e-01 -7.90496290e-01 2.10230872e-01
-3.72546799e-02 -1.32313085e+00 -7.17300594e-01 5.42878568e-01
-9.50971007e-01 -1.06486952e+00 -6.86650217e-01 -8.61476362e-01
6.68850124e-01 6.28776014e-01 1.32514071e+00 3.26864243e-01
-3.52920443e-01 -2.85469182e-02 -1.49400130e-01 -2.06448421e-01
3.66068929e-02 1.52348340e-01 -6.14499569e-01 -1.47096634e-01
1.10879943e-01 -8.44638526e-01 -1.03548157e+00 3.50815952e-01
-8.07610333e-01 5.45894504e-01 6.45813048e-01 7.37736523e-01
7.69206285e-01 -3.33813101e-01 -2.72623628e-01 -5.04131734e-01
-2.02008277e-01 2.32310668e-02 -9.97815371e-01 -3.40765938e-02
-2.73907185e-01 3.06300759e-01 4.37293380e-01 -5.48253208e-02
-7.49789476e-01 4.87779081e-01 -4.01885420e-01 -3.26184779e-01
-9.05983299e-02 -1.81053430e-01 -1.97988585e-01 -2.84289598e-01
4.52124059e-01 1.96252763e-01 2.02468503e-02 -5.20848870e-01
3.83120865e-01 2.25295454e-01 6.11200690e-01 -3.85597318e-01
1.06591201e+00 1.00744987e+00 1.48391604e-01 -3.62542450e-01
-9.86836135e-01 -5.23469567e-01 -8.30707848e-01 -5.24021238e-02
1.09319401e+00 -1.20707107e+00 -8.50935578e-01 7.86970735e-01
-1.55244219e+00 -4.36613619e-01 -4.32916224e-01 1.85593516e-01
-4.44652081e-01 1.00138783e-01 -5.36287367e-01 -5.60770519e-02
-4.70019132e-01 -1.36586475e+00 1.75340533e+00 1.93697304e-01
1.47610396e-01 -7.52926707e-01 -1.15829408e-02 2.92931169e-01
4.88190055e-01 1.76453814e-01 1.05755222e+00 -1.32168606e-01
-1.06288385e+00 -9.46504399e-02 -8.12530696e-01 3.18891019e-01
-1.85299009e-01 -1.80594876e-01 -1.38645792e+00 3.33609171e-02
-3.88402939e-02 -1.56944171e-01 1.26887596e+00 4.05572772e-01
1.23598337e+00 5.87823205e-02 -1.11203253e-01 1.20858049e+00
1.55718482e+00 -3.06611657e-01 7.20733285e-01 4.81346577e-01
1.00169873e+00 8.33132565e-01 4.05743420e-01 2.08970636e-01
6.53161883e-01 8.09413135e-01 8.89329255e-01 -5.71604431e-01
-3.88801128e-01 -3.34443867e-01 1.20707460e-01 2.36597806e-01
9.03468430e-02 2.13510051e-01 -7.64854252e-01 3.14072907e-01
-1.81867099e+00 -7.60969937e-01 -3.13716352e-01 2.26919603e+00
6.81241810e-01 2.76314020e-01 -5.96885756e-02 1.62403539e-01
4.29117024e-01 3.05457383e-01 -4.46100682e-01 -2.51798421e-01
-3.02340508e-01 3.30530196e-01 6.62255645e-01 7.40051985e-01
-1.15015626e+00 9.86007094e-01 4.75782442e+00 7.94243395e-01
-1.23594666e+00 -6.68075830e-02 7.31963575e-01 -2.05110013e-01
-4.96578604e-01 1.63590610e-01 -9.36526775e-01 1.91886038e-01
3.98782417e-02 3.85571241e-01 1.38602927e-01 7.58494377e-01
1.46785736e-01 -3.05598676e-01 -1.47613752e+00 1.20149565e+00
-3.84706080e-01 -1.76518178e+00 2.06123948e-01 3.00901026e-01
7.06319153e-01 3.97631884e-01 -1.10799976e-01 -5.31401969e-02
-1.64200038e-01 -9.64664519e-01 1.16953874e+00 2.51755744e-01
6.77441955e-01 -7.03868628e-01 7.02224553e-01 1.05026148e-01
-1.39543343e+00 2.48876698e-02 -3.70411634e-01 -2.64213413e-01
5.44568477e-03 9.66724575e-01 -3.29891533e-01 5.21837533e-01
8.50330532e-01 8.21811676e-01 -6.81716800e-01 1.03068173e+00
-1.97938621e-01 -2.96669930e-01 -2.21445635e-01 4.33114082e-01
5.48578918e-01 -1.34390235e-01 2.30481505e-01 1.15355372e+00
1.73225179e-01 -2.15565845e-01 -6.96053356e-02 1.07961535e+00
-3.59792680e-01 -1.87707886e-01 -5.60024500e-01 5.66159368e-01
2.57903278e-01 1.35835278e+00 -7.61638105e-01 -1.40709177e-01
-4.70852733e-01 1.26898360e+00 5.21025598e-01 7.12600052e-02
-8.36319745e-01 -3.48216504e-01 9.26811397e-01 4.04600501e-01
5.18626690e-01 -2.23174006e-01 -6.68923199e-01 -9.81891155e-01
3.50663662e-01 -3.33807647e-01 2.24342775e-02 -6.92360878e-01
-1.10503376e+00 6.95129395e-01 -4.22165364e-01 -1.35132110e+00
6.04471117e-02 -5.55144131e-01 -5.37182570e-01 1.21115088e+00
-2.10988069e+00 -1.05004656e+00 -8.68305683e-01 6.51885271e-01
3.50389302e-01 5.19220471e-01 6.01442814e-01 3.02212834e-01
-2.50048429e-01 4.44319457e-01 -4.12846386e-01 1.51276365e-01
6.49110734e-01 -1.18845999e+00 8.99306893e-01 7.21399307e-01
-1.43041998e-01 3.71980757e-01 1.57318920e-01 -2.73734361e-01
-1.17281032e+00 -1.10018623e+00 1.10958087e+00 -4.16706502e-01
3.17151070e-01 -6.39637530e-01 -7.36324549e-01 2.04701811e-01
-1.28303677e-01 4.08374935e-01 7.69369770e-03 -2.26299435e-01
-7.64602184e-01 -2.19505832e-01 -1.09843338e+00 5.41841805e-01
1.43380344e+00 -6.88396931e-01 -3.35302711e-01 1.15201145e-01
9.45305169e-01 -6.01494014e-01 -6.82377517e-01 4.12894070e-01
5.86948514e-01 -1.48660100e+00 1.25580382e+00 1.78266205e-02
8.20343912e-01 -4.95802343e-01 -1.71179518e-01 -8.28252792e-01
-2.53690630e-01 -4.05602574e-01 2.89309770e-01 9.95198846e-01
4.45066303e-01 -4.40559238e-01 8.13979506e-01 4.19738919e-01
-3.00330400e-01 -9.48778331e-01 -9.25929129e-01 -6.15498722e-01
5.50070852e-02 -6.12250090e-01 5.96585751e-01 5.02069056e-01
-5.75753033e-01 2.69240707e-01 1.15834378e-01 1.49361148e-01
6.68397427e-01 4.54675019e-01 9.49769199e-01 -1.28045654e+00
-2.01385126e-01 -7.84601152e-01 -5.47394991e-01 -1.60596955e+00
-1.84447169e-01 -8.96476269e-01 -2.54869033e-02 -1.53878665e+00
1.95959777e-01 -4.06559557e-01 1.00148715e-01 4.30824518e-01
-1.22316450e-01 5.53723991e-01 2.20662504e-01 1.98028326e-01
-2.62485415e-01 4.40079600e-01 1.44327426e+00 -1.43228427e-01
-2.69222647e-01 -3.17114890e-01 -5.33412755e-01 8.31973433e-01
4.59774345e-01 -3.30597043e-01 -2.97457039e-01 -8.89616191e-01
2.26581663e-01 -1.60424992e-01 9.15999353e-01 -1.08414865e+00
2.90015608e-01 2.78070211e-01 5.35950840e-01 -7.94501662e-01
3.00726086e-01 -8.60952258e-01 -2.28248194e-01 4.78452832e-01
-8.90050530e-02 -8.01716894e-02 2.13408932e-01 3.14157814e-01
-3.15993816e-01 2.13624716e-01 9.78624940e-01 -1.71563655e-01
-7.70813763e-01 6.68464780e-01 2.69923061e-01 8.46726969e-02
7.12337613e-01 -5.26254296e-01 -2.64674932e-01 -4.25929017e-02
-3.91054422e-01 3.19353580e-01 9.75109816e-01 3.65628332e-01
6.31734073e-01 -1.10783279e+00 -4.79957491e-01 4.68863279e-01
2.14102477e-01 7.11963236e-01 5.03982425e-01 9.40441668e-01
-8.99375856e-01 3.67221713e-01 -3.12365830e-01 -8.79615128e-01
-9.35227811e-01 2.61835724e-01 5.56366444e-01 -4.72624660e-01
-7.95406461e-01 1.22442853e+00 8.14859271e-01 -3.74813855e-01
5.70035517e-01 -7.82766342e-01 1.25354871e-01 1.02086291e-01
6.13776624e-01 4.08167131e-02 3.88816714e-01 -4.48098660e-01
-5.21941841e-01 1.12440658e+00 -2.86035482e-02 -7.70508125e-02
1.44550550e+00 -1.11394219e-01 -2.00337484e-01 -2.21030768e-02
1.65858078e+00 -2.86907069e-02 -1.76202142e+00 -3.05271119e-01
-2.06364334e-01 -5.00468314e-01 1.44290701e-01 -5.95302165e-01
-1.39676869e+00 1.16984010e+00 5.68061054e-01 -1.50132924e-01
1.20417202e+00 4.23657615e-03 9.57421422e-01 1.29502550e-01
4.63585705e-01 -8.43062222e-01 2.36502402e-02 7.10598469e-01
1.00423348e+00 -1.44035625e+00 -1.70837432e-01 -6.84721768e-01
-1.77557826e-01 1.19085884e+00 6.09888017e-01 -3.32910240e-01
6.37003601e-01 3.77549946e-01 -4.33931649e-02 -3.43477428e-01
-3.93652737e-01 -3.93942267e-01 4.27561760e-01 5.40679812e-01
2.75507420e-01 -3.51651639e-01 -4.61595468e-02 3.91948611e-01
-1.98870555e-01 -3.06405485e-01 1.11004703e-01 6.93418860e-01
-3.97489607e-01 -1.05342019e+00 -1.31523088e-01 2.75484603e-02
-9.39949006e-02 -4.85676587e-01 -3.59959275e-01 8.40448856e-01
4.53075171e-01 6.18287861e-01 3.95566016e-01 -5.21068394e-01
6.20829225e-01 -2.42398307e-01 5.95230997e-01 -5.60860515e-01
-8.85425746e-01 -1.44221717e-02 -1.44476518e-01 -1.28152740e+00
-2.82219112e-01 -2.34604880e-01 -1.38721049e+00 -3.56643915e-01
8.75784829e-02 -3.84639561e-01 8.03691447e-01 7.72703350e-01
4.72860694e-01 4.10161823e-01 6.57928407e-01 -1.44272077e+00
-2.00917125e-01 -6.38712049e-01 -3.69316965e-01 2.62324214e-01
4.47336197e-01 -5.25154948e-01 -4.88631517e-01 -3.17098022e-01] | [8.748842239379883, -2.4117431640625] |
7fd43ec0-788a-4f17-9ab5-a933631449a1 | dannte-a-case-study-of-a-turbo-machinery | 2201.03850 | null | https://arxiv.org/abs/2201.03850v1 | https://arxiv.org/pdf/2201.03850v1.pdf | DANNTe: a case study of a turbo-machinery sensor virtualization under domain shift | We propose an adversarial learning method to tackle a Domain Adaptation (DA) time series regression task (DANNTe). The regression aims at building a virtual copy of a sensor installed on a gas turbine, to be used in place of the physical sensor which can be missing in certain situations. Our DA approach is to search for a domain-invariant representation of the features. The learner has access to both a labelled source dataset and an unlabeled target dataset (unsupervised DA) and is trained on both, exploiting the minmax game between a task regressor and a domain classifier Neural Networks. Both models share the same feature representation, learnt by a feature extractor. This work is based on the results published by Ganin et al. arXiv:1505.07818; indeed, we present an extension suitable to time series applications. We report a significant improvement in regression performance, compared to the baseline model trained on the source domain only. | ['Giacomo Veneri', 'Valentina Gori', 'Luca Strazzera'] | 2022-01-11 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [ 5.51410139e-01 3.46606970e-01 -1.16120271e-01 -2.03753158e-01
-7.29029775e-01 -8.01530838e-01 8.66385102e-01 -3.22249010e-02
-2.84010142e-01 7.65226483e-01 -3.50523174e-01 -1.15119234e-01
-5.15152998e-02 -7.31775463e-01 -8.72781038e-01 -9.94628191e-01
-2.62987524e-01 5.63157439e-01 -4.04658429e-02 -3.15885484e-01
-6.87409565e-02 5.94928324e-01 -1.31195974e+00 1.10514894e-01
3.40624303e-01 1.39226127e+00 -5.15558831e-02 8.58970881e-01
3.18836242e-01 8.92330825e-01 -8.90742719e-01 -1.35945445e-02
7.24728465e-01 -3.22166175e-01 -5.97550571e-01 -1.77514002e-01
-3.35534103e-02 6.15946464e-02 -6.08577192e-01 9.05258000e-01
1.17066555e-01 1.76617369e-01 8.74326050e-01 -1.59817243e+00
-1.48247927e-01 3.19343597e-01 -1.32831916e-01 2.70354182e-01
1.83625787e-01 -4.54074107e-02 7.13056386e-01 -4.68591243e-01
6.28935456e-01 6.43087864e-01 5.27687848e-01 6.98351383e-01
-1.41386425e+00 -6.55158520e-01 -1.91082940e-01 -4.42531146e-02
-9.92887139e-01 -2.50311375e-01 1.22028899e+00 -4.88983393e-01
9.59100366e-01 7.86027685e-02 4.05256778e-01 1.65349925e+00
2.04697996e-01 5.33799529e-01 1.10417485e+00 -4.64248419e-01
6.85716629e-01 2.53705114e-01 -1.42952174e-01 1.25886634e-01
-2.20926508e-01 6.37897849e-01 -3.39125752e-01 -3.27792913e-01
6.43386841e-01 1.70621146e-02 1.65614814e-01 -8.16609085e-01
-8.87206793e-01 9.00009394e-01 3.31675291e-01 4.35378253e-01
-5.32375395e-01 -1.04659021e-01 7.58141518e-01 1.17297661e+00
6.61358654e-01 6.47056818e-01 -8.82416427e-01 -1.03144929e-01
-9.12178278e-01 2.73532420e-01 1.17909324e+00 7.42557883e-01
5.62659144e-01 1.78779319e-01 2.05311581e-01 2.67316520e-01
-2.41655767e-01 5.57896078e-01 8.18489611e-01 -7.56652057e-01
4.91039753e-01 2.78539687e-01 1.07322380e-01 -4.82163399e-01
-3.44623476e-01 -1.40279248e-01 -8.24615777e-01 5.34926057e-01
5.10029435e-01 -5.00914454e-01 -7.56350040e-01 1.64762056e+00
1.61206856e-01 6.59234881e-01 3.28633845e-01 6.46808803e-01
2.41740450e-01 6.01925492e-01 -1.38431750e-02 -2.57768363e-01
6.82782114e-01 -6.61422193e-01 -3.26665401e-01 -1.82828903e-01
4.64412659e-01 -4.39705729e-01 5.31725645e-01 7.52137899e-01
-9.15442705e-01 -7.19142973e-01 -1.39377701e+00 2.82640904e-01
-8.00011754e-01 -1.25626251e-01 2.47917801e-01 4.53684449e-01
-7.59381890e-01 1.04150641e+00 -8.55373204e-01 -1.93372995e-01
1.78200737e-01 5.09441197e-01 -5.59775770e-01 1.58297285e-01
-1.34997880e+00 9.98715341e-01 4.30462778e-01 -4.12567258e-01
-1.21800816e+00 -6.56822443e-01 -9.37446594e-01 -2.75219291e-01
2.81812280e-01 -3.03908288e-01 1.48725975e+00 -1.46501529e+00
-1.62300301e+00 1.07446849e+00 1.82567671e-01 -1.04411864e+00
7.73375750e-01 -1.83004647e-01 -7.53201127e-01 9.79891792e-02
-3.39095714e-03 6.30256236e-02 1.66375697e+00 -8.02505672e-01
-5.19829452e-01 -3.56509954e-01 -6.24372698e-02 -3.68128806e-01
-1.18998073e-01 -8.14486369e-02 3.27717483e-01 -8.16162646e-01
-2.94002265e-01 -9.75235760e-01 -2.44560197e-01 -3.20151299e-01
-2.53887832e-01 -2.36370787e-02 1.37255073e+00 -8.03414047e-01
7.89410293e-01 -2.25837851e+00 4.29788768e-01 4.61585045e-01
-5.63808382e-02 3.45027089e-01 -2.09085774e-02 5.74688613e-01
-8.44956100e-01 -1.79022819e-01 -4.99184459e-01 -4.07504499e-01
-8.76978412e-02 2.91934967e-01 -9.33741987e-01 8.29009533e-01
4.59844172e-01 6.77883208e-01 -8.24524403e-01 1.48413464e-01
4.06262606e-01 6.37326613e-02 -5.38526140e-02 5.62021017e-01
-4.82582450e-01 9.60449755e-01 -6.30252898e-01 2.81530291e-01
6.19760036e-01 1.59548789e-01 1.04139552e-01 2.88100481e-01
4.50360253e-02 3.11357468e-01 -1.13912094e+00 2.02816010e+00
-5.47372520e-01 5.78343451e-01 -3.30803320e-02 -1.78150880e+00
1.35503876e+00 6.17832363e-01 9.07534122e-01 -6.31804943e-01
1.76278412e-01 2.90111989e-01 -2.52212375e-01 -3.95421952e-01
8.47872049e-02 -2.77344823e-01 -4.95585144e-01 4.81776416e-01
4.69811201e-01 -4.52377915e-01 -1.24591202e-01 -2.19395593e-01
1.56498802e+00 5.95811963e-01 4.57973242e-01 -1.04730807e-01
6.04256094e-01 6.45264164e-02 2.44691029e-01 6.38889790e-01
-1.75678417e-01 4.14303660e-01 7.22752154e-01 -6.80712759e-01
-1.48173606e+00 -1.07853341e+00 3.66830938e-02 9.34417546e-01
-2.41091356e-01 -2.23022923e-01 -4.67069954e-01 -1.05638123e+00
1.80323094e-01 8.51633489e-01 -9.36508715e-01 -6.83903098e-01
-7.45602429e-01 2.78450418e-02 6.39647782e-01 4.55202609e-01
1.88278645e-01 -1.20224893e+00 -6.56039774e-01 3.12755316e-01
4.59601104e-01 -9.57250834e-01 -2.69510448e-02 8.38089645e-01
-9.50912952e-01 -1.00500345e+00 -3.48444968e-01 -5.53789794e-01
2.34692737e-01 -3.76328647e-01 1.09523189e+00 -5.08130670e-01
-2.64725268e-01 5.09298801e-01 -3.27909648e-01 -8.73569369e-01
-8.02618861e-01 3.33294958e-01 2.52383053e-01 6.82540908e-02
6.82136059e-01 -9.50719178e-01 -6.01655878e-02 5.84036112e-02
-9.10550177e-01 -5.32248139e-01 2.36550346e-01 8.61289501e-01
3.66219223e-01 1.99206054e-01 6.72002435e-01 -9.40378070e-01
3.09715450e-01 -8.75986040e-01 -9.16532993e-01 -8.70762467e-02
-5.11297703e-01 2.28324786e-01 1.25438333e+00 -6.69419527e-01
-7.30716765e-01 4.87661302e-01 1.25776410e-01 -1.04354358e+00
-4.63041455e-01 2.31200203e-01 -6.46263584e-02 8.31360370e-02
8.95905435e-01 2.03832507e-01 2.13639826e-01 -5.35216630e-01
1.99377403e-01 4.89039689e-01 7.74780095e-01 -6.13817811e-01
1.36760449e+00 3.20386440e-01 1.98573425e-01 -5.60555279e-01
-6.99910223e-01 -5.93039095e-01 -1.20122266e+00 1.96080342e-01
4.98096824e-01 -8.71108353e-01 -1.88932240e-01 4.54618126e-01
-9.33391750e-01 -5.04743159e-01 -1.01781380e+00 4.36240494e-01
-1.01309490e+00 -1.26645699e-01 -1.37871370e-01 -7.60102749e-01
1.79560855e-02 -5.13237894e-01 1.00257146e+00 2.33966783e-02
-1.03045925e-01 -1.30328333e+00 4.21992183e-01 -3.07653159e-01
3.50667208e-01 8.20108414e-01 6.57748342e-01 -1.49643731e+00
5.87722026e-02 -7.22336113e-01 3.87872636e-01 7.98982382e-01
1.19192705e-01 -2.17311054e-01 -1.36279356e+00 -5.34606993e-01
5.93642890e-01 -4.76104170e-01 7.54074037e-01 1.25719279e-01
1.27727866e+00 3.40770534e-03 -3.23364049e-01 4.36136156e-01
1.24273109e+00 3.72051835e-01 5.26874244e-01 3.21180731e-01
2.80785531e-01 4.41703320e-01 8.10214102e-01 5.44880211e-01
-2.73523599e-01 5.51520407e-01 5.36314845e-01 -1.18408985e-01
3.96828592e-01 -3.76766562e-01 5.60874164e-01 3.72421980e-01
1.70953758e-02 -2.45562978e-02 -7.93366611e-01 5.45891881e-01
-1.72353852e+00 -8.52507234e-01 2.97812372e-01 2.51449370e+00
5.64166427e-01 3.57893616e-01 4.75822121e-01 4.72829610e-01
6.06161237e-01 2.40251437e-01 -9.35629189e-01 -6.03858054e-01
3.80154774e-02 1.00370741e+00 6.77346230e-01 1.39126047e-01
-1.53240597e+00 5.58505237e-01 6.05670023e+00 4.72780496e-01
-1.35498846e+00 1.65226400e-01 1.98358893e-01 1.74441207e-02
2.49273822e-01 -7.70446435e-02 -1.98979333e-01 4.06588614e-01
1.67748415e+00 -4.61202532e-01 6.48101449e-01 1.13371432e+00
-1.16983809e-01 3.47135961e-01 -1.61675012e+00 8.14694881e-01
-5.26306555e-02 -8.65983188e-01 -6.81754291e-01 3.94334346e-02
6.10521674e-01 1.43178761e-01 1.49307713e-01 7.16985941e-01
3.35528880e-01 -1.14404023e+00 4.18881953e-01 5.39478838e-01
8.51383626e-01 -7.75859237e-01 6.17118537e-01 6.01277590e-01
-1.02320409e+00 -2.65771598e-01 -1.59012109e-01 -2.80281723e-01
-4.24315393e-01 2.85641432e-01 -9.91837204e-01 8.34977984e-01
4.69765842e-01 9.73993540e-01 -1.94431230e-01 5.81077397e-01
-2.14063361e-01 8.36183906e-01 -2.91173428e-01 3.46471041e-01
3.49686891e-02 -1.90727755e-01 8.06847274e-01 8.52902710e-01
2.96488255e-01 -3.56103837e-01 2.79055715e-01 7.23265648e-01
-4.47952859e-02 -1.53657332e-01 -1.28432798e+00 -1.74917802e-01
3.01197171e-01 1.07537556e+00 -1.44639120e-01 -1.49994165e-01
-5.33483267e-01 1.15932584e+00 2.10130185e-01 2.72658110e-01
-7.95730889e-01 -4.49835002e-01 5.53459227e-01 -7.22324550e-02
5.00266492e-01 -1.50867119e-01 -1.23418383e-01 -1.09321821e+00
1.46764144e-01 -7.59729028e-01 6.17652595e-01 -5.90625226e-01
-1.54802597e+00 3.83002460e-01 4.08244170e-02 -1.61228859e+00
-1.05180097e+00 -8.20374012e-01 -7.65748918e-01 1.19005013e+00
-1.38078034e+00 -9.72384870e-01 1.25552714e-01 8.70131791e-01
3.95259172e-01 -6.24829173e-01 1.20367837e+00 -8.46292078e-02
-2.30497062e-01 2.84511119e-01 3.56427550e-01 1.22370049e-01
6.90111339e-01 -1.53789401e+00 5.20449281e-01 6.94044411e-01
1.11211352e-01 5.10570072e-02 7.86875844e-01 -4.48162109e-01
-1.43088126e+00 -1.28620696e+00 7.20702887e-01 -6.68385923e-01
9.79970992e-01 -5.36652505e-01 -9.71930981e-01 1.05183780e+00
-3.90130952e-02 2.68939435e-01 4.05796140e-01 -2.30715126e-01
-4.65768278e-01 -1.86965466e-01 -1.48977005e+00 1.54440291e-02
5.11490762e-01 -7.71418631e-01 -9.92498934e-01 3.11556250e-01
6.76819980e-01 -5.32842994e-01 -1.11929166e+00 1.19064543e-02
1.70384839e-01 -5.63338578e-01 9.19093192e-01 -1.11015749e+00
5.24511933e-01 1.96233168e-02 -1.01023868e-01 -1.54492033e+00
1.93573877e-01 -8.73959422e-01 -5.62128305e-01 9.64755356e-01
1.44286960e-01 -8.19960415e-01 7.04513311e-01 6.00807250e-01
9.28444117e-02 -2.03838184e-01 -1.34193850e+00 -1.08233345e+00
5.08584976e-01 -5.56950152e-01 6.52614176e-01 9.64243770e-01
-1.31902710e-01 2.58699715e-01 -3.67703319e-01 2.23764241e-01
4.52553600e-01 6.53520301e-02 8.60498548e-01 -1.64562857e+00
-3.77972543e-01 -5.00160456e-03 -5.78869700e-01 -5.78523576e-01
5.85339665e-01 -8.81786168e-01 -1.42402142e-01 -4.55112696e-01
-6.08381748e-01 -3.11624736e-01 -7.73296475e-01 3.91850084e-01
4.46244180e-01 -1.45051360e-01 -8.95716622e-03 1.41452402e-01
-2.28968054e-01 2.91665643e-01 7.41852760e-01 -2.11101130e-01
-2.43712351e-01 6.60517454e-01 -2.01113850e-01 5.60870767e-01
9.57313836e-01 -7.67012596e-01 -3.52466464e-01 2.72591978e-01
-1.84158772e-01 5.07812023e-01 6.05297506e-01 -1.03547621e+00
2.76868224e-01 -1.29614204e-01 4.80893016e-01 -2.20341936e-01
4.14545476e-01 -1.31334555e+00 4.48065251e-02 5.51625490e-01
-3.97862315e-01 -1.07001804e-01 2.40690559e-01 6.91691518e-01
-2.60874152e-01 -4.75029588e-01 7.36232638e-01 -1.12412661e-01
-6.50353670e-01 2.65386343e-01 -7.14510977e-02 2.16793511e-02
1.26775849e+00 2.69526578e-02 -3.41326818e-02 -3.48991662e-01
-9.68340099e-01 -7.25014061e-02 3.08651805e-01 5.50422728e-01
2.78375089e-01 -1.28907132e+00 -7.30412543e-01 6.50515199e-01
1.58708289e-01 3.99646498e-02 -1.48509756e-01 4.10978705e-01
7.96402395e-02 3.62269044e-01 -3.89083862e-01 -3.30673009e-01
-8.22446346e-01 1.10586083e+00 4.95469868e-01 -5.63837230e-01
-6.78744018e-01 3.98239523e-01 -1.91871569e-01 -5.72059214e-01
5.45760207e-02 -2.41552487e-01 -3.94653156e-03 -9.13763642e-02
4.36147630e-01 4.00106043e-01 4.12394762e-01 -3.91096532e-01
-8.02950636e-02 2.10205019e-01 1.32717311e-01 -1.92883402e-01
1.59559262e+00 2.84123540e-01 2.49840721e-01 8.16169500e-01
1.41565537e+00 -8.03731903e-02 -1.37114000e+00 -6.25936568e-01
2.20790386e-01 -1.53157160e-01 -1.80734530e-01 -7.97587276e-01
-9.29831922e-01 7.45509088e-01 7.02257514e-01 7.42949903e-01
1.27125788e+00 -9.11917239e-02 4.91698325e-01 4.16624278e-01
4.14252311e-01 -1.18089342e+00 -2.54536234e-02 4.99192476e-01
9.82579291e-01 -1.22399819e+00 -3.01312268e-01 7.47128129e-02
-5.67862272e-01 1.43601799e+00 3.36215168e-01 -8.04202020e-01
8.26390505e-01 3.30489546e-01 -9.16649029e-02 -1.40988648e-01
-9.23328757e-01 -2.22589988e-02 2.72323698e-01 1.09169650e+00
-6.73663467e-02 8.61450881e-02 2.80975014e-01 7.19106555e-01
-2.51237839e-01 5.67422099e-02 3.39924693e-01 9.81979370e-01
1.13286279e-01 -1.41222501e+00 -3.47166896e-01 4.77694392e-01
-4.34784740e-01 2.86942989e-01 -5.06456435e-01 9.80316818e-01
-3.50675248e-02 7.78970122e-01 2.32592255e-01 -3.20471466e-01
5.85905433e-01 5.06942272e-01 3.05365115e-01 -5.08811891e-01
-7.55403936e-01 -3.10454011e-01 -1.56787589e-01 -4.18285042e-01
-3.61020654e-01 -9.71562326e-01 -1.00747657e+00 4.19154577e-02
2.49772236e-01 1.48218811e-01 9.21244383e-01 1.01385283e+00
1.40747670e-02 4.99307334e-01 1.36012006e+00 -9.38971817e-01
-9.25699949e-01 -1.01705694e+00 -9.54065919e-01 5.87224483e-01
7.74469614e-01 -6.70122206e-01 -5.01280427e-01 2.20899388e-01] | [7.850674152374268, 2.5315794944763184] |
7cd33eaf-dd1d-4e3a-ae31-bf194c1c849b | patch-based-3d-natural-scene-generation-from | 2304.12670 | null | https://arxiv.org/abs/2304.12670v2 | https://arxiv.org/pdf/2304.12670v2.pdf | Patch-based 3D Natural Scene Generation from a Single Example | We target a 3D generative model for general natural scenes that are typically unique and intricate. Lacking the necessary volumes of training data, along with the difficulties of having ad hoc designs in presence of varying scene characteristics, renders existing setups intractable. Inspired by classical patch-based image models, we advocate for synthesizing 3D scenes at the patch level, given a single example. At the core of this work lies important algorithmic designs w.r.t the scene representation and generative patch nearest-neighbor module, that address unique challenges arising from lifting classical 2D patch-based framework to 3D generation. These design choices, on a collective level, contribute to a robust, effective, and efficient model that can generate high-quality general natural scenes with both realistic geometric structure and visual appearance, in large quantities and varieties, as demonstrated upon a variety of exemplar scenes. | ['Baoquan Chen', 'Jue Wang', 'Xuelin Chen', 'Weiyu Li'] | 2023-04-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_Patch-Based_3D_Natural_Scene_Generation_From_a_Single_Example_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_Patch-Based_3D_Natural_Scene_Generation_From_a_Single_Example_CVPR_2023_paper.pdf | cvpr-2023-1 | ['scene-generation'] | ['computer-vision'] | [ 6.18110895e-01 1.45888910e-01 3.67552578e-01 -6.87039867e-02
-7.07056761e-01 -6.95648909e-01 9.55457151e-01 -3.60639930e-01
4.37887698e-01 6.60320997e-01 1.68263212e-01 -1.04044110e-01
-4.53859508e-01 -9.31721747e-01 -8.33070576e-01 -6.95164859e-01
-5.21543510e-02 4.84425455e-01 -1.19174220e-01 -4.57783788e-01
3.37994218e-01 9.25944090e-01 -1.98976219e+00 9.95410755e-02
7.78844476e-01 8.98850024e-01 3.64939660e-01 7.20622003e-01
-7.88765901e-05 5.44057667e-01 -2.75203973e-01 -2.80234605e-01
5.87437630e-01 -6.01931155e-01 -4.55969214e-01 8.78752112e-01
9.48412955e-01 -1.88968226e-01 -9.35603082e-02 6.21420264e-01
4.95212197e-01 1.16859324e-01 1.04261065e+00 -1.02271259e+00
-9.19626296e-01 -9.02706012e-02 -5.33362091e-01 -3.58982384e-01
4.58313763e-01 4.10393119e-01 9.21339750e-01 -1.15501702e+00
8.60344112e-01 1.24947321e+00 7.49992013e-01 5.06904840e-01
-1.76698947e+00 3.56480628e-02 9.83613655e-02 -1.79444775e-01
-1.45355690e+00 -5.43830156e-01 1.01487434e+00 -5.60117722e-01
7.23913550e-01 4.08913106e-01 8.95607054e-01 1.19128406e+00
2.09122851e-01 5.09703040e-01 1.19336450e+00 -5.56271195e-01
2.92909026e-01 1.07174635e-01 -4.26202416e-01 3.54457110e-01
1.59762070e-01 3.31201583e-01 -3.93185169e-01 -1.49347156e-01
1.15840936e+00 -8.68834555e-02 -1.34432092e-01 -9.96601999e-01
-1.05267334e+00 7.30218947e-01 4.50054348e-01 -1.24123460e-02
-5.15753508e-01 9.95480940e-02 -1.92655131e-01 3.09899766e-02
4.31779265e-01 6.88969910e-01 -2.05092803e-01 3.38870883e-01
-9.02773619e-01 7.41720140e-01 5.07691979e-01 1.44975436e+00
1.18491828e+00 2.62039214e-01 -2.16673151e-01 9.01602328e-01
1.29789919e-01 6.21981323e-01 -1.87221184e-01 -1.05129671e+00
-4.43104766e-02 4.77156043e-01 7.05365464e-02 -1.17149198e+00
-5.06521687e-02 -6.70012057e-01 -1.10002220e+00 4.52911109e-01
-4.45009731e-02 1.84653282e-01 -9.06308770e-01 1.61895299e+00
6.88819826e-01 8.31993073e-02 -9.33473334e-02 6.50480866e-01
7.29488790e-01 5.73890626e-01 -2.11463407e-01 -1.10856012e-01
1.05175316e+00 -7.02494562e-01 -2.23557562e-01 -1.68698162e-01
2.20226422e-01 -1.08989537e+00 1.22593439e+00 2.61334866e-01
-1.41222489e+00 -7.00975955e-01 -8.69537532e-01 -1.38829187e-01
-1.55692682e-01 -3.49701904e-02 7.10799932e-01 5.60869932e-01
-1.38482213e+00 3.55651170e-01 -1.57538354e-01 -4.61891949e-01
4.29236293e-01 -1.38161834e-02 -2.81971425e-01 -3.76810580e-01
-6.69000983e-01 7.95651495e-01 6.66809157e-02 1.55592456e-01
-1.10011053e+00 -1.08923924e+00 -9.20977354e-01 -2.12145030e-01
2.02141657e-01 -1.44364583e+00 9.83533680e-01 -7.55059481e-01
-1.40822601e+00 9.42373812e-01 -1.34391770e-01 -1.35119811e-01
5.38782358e-01 7.98015818e-02 2.37492211e-02 -5.33461347e-02
1.16284870e-01 9.16734517e-01 1.14046431e+00 -1.83951271e+00
-2.33192742e-01 -2.16796547e-01 2.47923546e-02 4.23271924e-01
2.60266274e-01 -5.40625632e-01 -1.07683413e-01 -7.37806678e-01
7.81263411e-02 -8.35700274e-01 -7.50309885e-01 1.14111550e-01
-5.10894418e-01 1.70368612e-01 8.90640974e-01 -1.03210084e-01
5.91808558e-01 -2.25316310e+00 3.65771860e-01 2.99049199e-01
2.06325009e-01 -9.38985124e-02 -3.64322662e-01 7.76158750e-01
-1.84038892e-01 4.12291922e-02 -3.81389171e-01 -4.75224942e-01
7.12994784e-02 1.91744074e-01 -4.16212678e-01 3.50036263e-01
5.92034698e-01 9.54734981e-01 -6.86701596e-01 -3.28475505e-01
5.58228910e-01 6.69742644e-01 -1.01918101e+00 4.47420388e-01
-5.55067003e-01 7.10849702e-01 -4.65354323e-01 4.69834149e-01
8.47320557e-01 -2.01864526e-01 -2.24118099e-01 -1.35595605e-01
-2.16917425e-01 1.90827418e-02 -1.29748642e+00 1.98535299e+00
-5.17351091e-01 3.46090525e-01 1.02685899e-01 -8.16373110e-01
1.10232937e+00 3.09515130e-02 3.95224929e-01 -5.83031416e-01
-8.75942037e-02 1.48413748e-01 -3.73778820e-01 -2.56137997e-01
7.15589941e-01 -3.50858539e-01 -6.76732808e-02 2.77342975e-01
2.12087438e-01 -9.60919797e-01 1.11533925e-01 2.92744368e-01
9.49260533e-01 3.54744554e-01 3.25926870e-01 -5.67748606e-01
2.94060737e-01 1.87537923e-01 1.46398857e-01 7.64231920e-01
3.76147479e-01 1.16848338e+00 1.02562591e-01 -4.81119812e-01
-1.53446090e+00 -1.10706556e+00 -2.70888060e-01 4.79188144e-01
5.72683476e-02 -2.62574792e-01 -5.65127969e-01 4.87058274e-02
-1.66798055e-01 6.16252482e-01 -7.21701205e-01 9.20950714e-03
-4.63333100e-01 -5.30629396e-01 6.18477091e-02 -1.97078064e-02
2.34643042e-01 -9.33106422e-01 -4.45298970e-01 2.41274104e-01
2.46567249e-01 -1.20974541e+00 -2.92167962e-01 -2.75387298e-02
-7.29325056e-01 -8.51818442e-01 -7.32737541e-01 -6.85490549e-01
7.69948006e-01 5.66036820e-01 1.51491368e+00 4.59976755e-02
-7.81580031e-01 7.26413786e-01 -3.14239234e-01 -1.09111786e-01
-5.18660486e-01 -1.06045790e-01 -2.50126183e-01 4.77678627e-02
-3.94209594e-01 -1.01393247e+00 -7.64573455e-01 1.07717641e-01
-1.13918686e+00 4.82078582e-01 6.52937829e-01 7.39904225e-01
9.22736108e-01 6.52918294e-02 3.43866229e-01 -9.02326047e-01
3.39780003e-01 -5.65598428e-01 -5.52387655e-01 6.16842173e-02
-3.37307900e-01 -1.91055208e-01 7.77944744e-01 -2.12117419e-01
-1.09959924e+00 9.21322778e-02 -1.20687820e-01 -3.33469093e-01
-6.43926382e-01 -4.44755927e-02 -3.47047091e-01 -2.14858830e-01
1.04836738e+00 5.31156480e-01 -8.60065669e-02 -2.67148703e-01
8.17611516e-01 9.12823603e-02 5.28282404e-01 -8.01031291e-01
1.18638158e+00 5.75703740e-01 4.81574029e-01 -1.07634151e+00
-6.25664771e-01 -1.37133479e-01 -8.04432094e-01 -1.50461823e-01
6.06204331e-01 -9.21849072e-01 -5.93451746e-02 4.52792555e-01
-1.13657498e+00 -2.71555513e-01 -6.47783518e-01 -1.35140955e-01
-1.09404898e+00 1.45859256e-01 -5.26308902e-02 -7.62503564e-01
5.63001772e-03 -1.06268501e+00 1.49302483e+00 9.41491779e-03
-2.92486012e-01 -1.09150958e+00 2.04326242e-01 2.01402187e-01
6.08796179e-01 6.68101370e-01 1.21200848e+00 2.78916985e-01
-1.12864041e+00 5.11049889e-02 -1.14546686e-01 1.97517246e-01
2.43775845e-01 2.87308007e-01 -1.04941964e+00 -1.07504301e-01
-5.78898191e-02 -1.99545994e-01 3.62133980e-01 5.89165628e-01
9.93294299e-01 -2.11034775e-01 -8.43618810e-02 8.92458558e-01
1.81013727e+00 -1.11945055e-01 7.58605063e-01 -1.51790559e-01
8.55103254e-01 6.78473890e-01 1.14016511e-01 5.87571204e-01
2.49139562e-01 7.72199810e-01 4.02080894e-01 -3.38472754e-01
-6.23663008e-01 -4.77779388e-01 -1.13513231e-01 5.43688238e-01
4.14346484e-03 -3.04921865e-01 -6.12401903e-01 6.03708506e-01
-1.52932739e+00 -1.21631598e+00 -2.60730296e-01 2.13101792e+00
6.11070991e-01 -3.42971742e-01 1.99390054e-02 1.08010054e-01
4.39082325e-01 2.20333800e-01 -4.81924504e-01 -3.49331439e-01
-5.74155867e-01 2.47450054e-01 1.80848852e-01 4.00766015e-01
-8.47385049e-01 9.67364669e-01 7.38275480e+00 1.08799922e+00
-8.15131009e-01 -3.13664496e-01 8.66841912e-01 -1.09721094e-01
-1.07032323e+00 2.44880438e-01 -7.13188410e-01 9.99473631e-02
5.50025761e-01 -1.23998292e-01 6.75621808e-01 8.50129068e-01
2.61263430e-01 2.46318802e-02 -1.16596520e+00 1.15163052e+00
2.02284336e-01 -1.82344997e+00 3.83562446e-01 1.41641051e-01
1.18061209e+00 -3.88279915e-01 4.20070946e-01 -1.14057280e-01
4.37947541e-01 -1.30326355e+00 9.66662169e-01 5.23381054e-01
8.35949659e-01 -6.75662041e-01 -3.20186578e-02 3.32068563e-01
-1.04820120e+00 3.38758737e-01 -4.65982616e-01 4.06513410e-03
3.74343514e-01 8.87876689e-01 -6.23045504e-01 7.48021662e-01
6.16357028e-01 8.41335654e-01 -4.78550613e-01 1.03906059e+00
-6.07910194e-02 7.05650747e-02 -2.37490714e-01 1.89586505e-01
2.95856595e-01 -4.29695964e-01 6.76372945e-01 8.72738063e-01
5.65658629e-01 1.50332868e-01 8.87795165e-02 1.32810581e+00
2.15307951e-01 1.50605574e-01 -1.37885273e+00 4.29790050e-01
4.18697447e-01 1.36045992e+00 -5.53069532e-01 -7.44722504e-03
-1.41062394e-01 8.75546873e-01 3.35208207e-01 4.73317891e-01
-6.17865622e-01 1.38591051e-01 7.06402719e-01 4.87024307e-01
5.19235671e-01 -5.69282949e-01 -4.32320237e-01 -9.16864038e-01
5.62829226e-02 -1.04526401e+00 -1.40698954e-01 -9.48067844e-01
-1.61302197e+00 6.54075623e-01 1.23414449e-01 -1.36415148e+00
-4.54263866e-01 -4.28090781e-01 -5.86578667e-01 7.84643412e-01
-1.46553802e+00 -1.43977976e+00 -4.01561350e-01 7.70241916e-01
4.51665282e-01 1.17008902e-01 9.28990006e-01 4.90895286e-02
-1.76040605e-01 2.59306312e-01 -1.05557740e-02 -5.35833418e-01
2.52182841e-01 -1.11015761e+00 5.53375781e-01 8.31856310e-01
2.13200241e-01 5.98741293e-01 7.19245672e-01 -3.15316081e-01
-1.84988785e+00 -1.22847033e+00 3.58242065e-01 -7.06410348e-01
3.25354755e-01 -5.59985578e-01 -5.07889509e-01 2.76341856e-01
1.58887610e-01 -9.77128223e-02 7.06216574e-01 7.52472179e-03
-2.61376023e-01 1.82995230e-01 -1.25236785e+00 9.30411279e-01
1.68506444e+00 -5.18639266e-01 -2.30462849e-01 4.10340905e-01
6.23924673e-01 -3.20702076e-01 -9.01390970e-01 4.23178285e-01
2.81026810e-01 -1.28213453e+00 1.37157428e+00 -3.63517016e-01
6.28506362e-01 -4.59627241e-01 -4.92196679e-01 -1.38672376e+00
-6.39729261e-01 -1.02732539e+00 2.50221610e-01 1.14414859e+00
2.58886218e-01 -2.79437155e-01 6.46599591e-01 4.80947852e-01
-4.11201656e-01 -5.78253984e-01 -7.77804315e-01 -7.68441796e-01
2.20939070e-01 -4.88712311e-01 8.17047536e-01 6.60107851e-01
-7.56102324e-01 5.21864653e-01 -5.65380275e-01 5.48614599e-02
9.05307949e-01 4.44723397e-01 1.34743202e+00 -1.14012980e+00
-3.49249244e-01 -5.14788389e-01 -5.66119730e-01 -1.24267018e+00
-1.36906877e-01 -5.52824199e-01 1.30127415e-01 -1.49122357e+00
1.14939235e-01 -9.06159401e-01 4.43393707e-01 -7.31582195e-02
-7.63952956e-02 4.24150735e-01 3.94948386e-02 1.74098238e-01
-1.50656804e-01 9.09781694e-01 1.78753483e+00 -2.39629019e-02
4.23183404e-02 -2.35681236e-01 -1.02064598e+00 4.16440606e-01
4.45073426e-01 -3.10352929e-02 -9.39924181e-01 -6.07769072e-01
2.41595492e-01 -1.73091993e-01 7.43009746e-01 -9.75831568e-01
-6.85190558e-02 -3.65934193e-01 4.33884740e-01 -6.04143620e-01
7.00890779e-01 -8.09249520e-01 7.33624876e-01 -1.60958275e-01
-5.04832752e-02 -4.96652126e-02 1.08672336e-01 5.12128234e-01
-1.34709224e-01 -4.15576110e-03 8.52782249e-01 -4.47133571e-01
-7.37244785e-01 7.49313414e-01 -3.27097159e-03 2.83961613e-02
1.18349195e+00 -6.91974103e-01 -1.47632882e-01 -3.73002350e-01
-5.21341085e-01 -3.88029009e-01 9.94798839e-01 2.61454463e-01
6.64213777e-01 -1.56253886e+00 -8.30681264e-01 6.13776565e-01
2.36815989e-01 4.62468922e-01 7.15108097e-01 2.90054411e-01
-5.58231413e-01 2.29315236e-01 -3.28551650e-01 -8.69292319e-01
-7.85098851e-01 6.24931931e-01 2.97356129e-01 1.58791721e-01
-6.81931436e-01 8.29744160e-01 6.55959129e-01 -6.19618237e-01
-2.08255112e-01 -1.36034593e-01 2.48969540e-01 -3.47933263e-01
2.68128783e-01 -5.41227348e-02 7.48681352e-02 -7.77354419e-01
-1.37870824e-02 1.00393951e+00 2.63122827e-01 -3.23839337e-02
1.40124846e+00 -1.21355467e-01 -8.01547542e-02 1.65348083e-01
1.11377537e+00 5.57993688e-02 -1.47663581e+00 -4.91939560e-02
-6.06812239e-01 -9.12404835e-01 -1.95140783e-02 -3.46611679e-01
-8.49516630e-01 6.87438965e-01 2.60475039e-01 3.58445853e-01
1.03918588e+00 1.34625718e-01 5.53653777e-01 1.10586081e-02
6.81853652e-01 -5.18979013e-01 3.22771631e-02 3.92863274e-01
1.19450283e+00 -9.62314427e-01 1.10783845e-01 -6.38695538e-01
-3.54566097e-01 7.59341955e-01 3.99208665e-01 -4.91556436e-01
6.19713664e-01 2.23271668e-01 -9.79348719e-02 -3.85495096e-01
-8.54417562e-01 -1.94014326e-01 4.39621657e-01 1.11431050e+00
3.36992860e-01 5.64641207e-02 2.30971813e-01 -1.55802503e-01
-4.12581742e-01 -4.61372286e-01 3.37781876e-01 8.00274432e-01
-3.07196438e-01 -1.31735492e+00 -3.05263609e-01 8.50959718e-02
1.76420972e-01 -2.48447239e-01 -3.42175961e-01 7.56993234e-01
2.45788634e-01 8.12819064e-01 7.61699155e-02 -1.85292140e-01
4.31630403e-01 -3.28108698e-01 8.70518625e-01 -6.60855949e-01
-2.18533799e-01 7.60972202e-02 -4.29268442e-02 -4.09422368e-01
-5.44306099e-01 -7.45758235e-01 -4.65789348e-01 -5.84274352e-01
-8.56250077e-02 -1.87880844e-01 4.34516519e-01 7.03749835e-01
7.76389122e-01 2.09453762e-01 9.69119310e-01 -1.42446995e+00
-3.09470356e-01 -6.85589612e-01 -6.06563032e-01 5.82218707e-01
2.17526570e-01 -8.08708429e-01 -2.42429018e-01 3.01654786e-01] | [9.250343322753906, -3.224567413330078] |
57181681-62b7-4fec-93d5-92dd7a9a5c57 | metaems-a-meta-reinforcement-learning-based | 2210.12590 | null | https://arxiv.org/abs/2210.12590v1 | https://arxiv.org/pdf/2210.12590v1.pdf | MetaEMS: A Meta Reinforcement Learning-based Control Framework for Building Energy Management System | The building sector has been recognized as one of the primary sectors for worldwide energy consumption. Improving the energy efficiency of the building sector can help reduce the operation cost and reduce the greenhouse gas emission. The energy management system (EMS) can monitor and control the operations of built-in appliances in buildings, so an efficient EMS is of crucial importance to improve the building operation efficiency and maintain safe operations. With the growing penetration of renewable energy and electrical appliances, increasing attention has been paid to the development of intelligent building EMS. Recently, reinforcement learning (RL) has been applied for building EMS and has shown promising potential. However, most of the current RL-based EMS solutions would need a large amount of data to learn a reliable control policy, which limits the applicability of these solutions in the real world. In this work, we propose MetaEMS, which can help achieve better energy management performance with the benefits of RL and meta-learning. Experiment results showcase that our proposed MetaEMS can adapt faster to environment changes and perform better in most situations compared with other baselines. | ['Benoit Boulet', 'Di wu', 'Huiliang Zhang'] | 2022-10-23 | null | null | null | null | ['energy-management'] | ['time-series'] | [-9.39271450e-02 -1.09799579e-01 -2.68403143e-01 1.54263511e-01
-3.88087273e-01 -8.50578323e-02 3.28301609e-01 1.80084392e-01
-9.73427817e-02 7.93970108e-01 7.05919638e-02 -1.84055433e-01
-3.72728050e-01 -1.23133981e+00 -2.95224190e-01 -1.04510307e+00
1.49190575e-01 -1.10430256e-01 1.08804196e-01 -3.57913077e-01
2.00523406e-01 4.09354180e-01 -1.71212065e+00 -1.74013868e-01
1.02872956e+00 1.01591671e+00 6.89348578e-01 4.10777569e-01
7.07748607e-02 6.20411694e-01 -4.52278316e-01 6.59772515e-01
8.03380609e-02 -5.42780519e-01 -7.45198786e-01 -1.45187125e-01
-8.07228923e-01 -1.10693514e-01 -7.29126111e-02 7.37495363e-01
8.24992597e-01 5.35594404e-01 2.88551092e-01 -1.13777971e+00
-1.02045149e-01 2.35667825e-01 -3.37285399e-01 4.80684936e-02
2.66649239e-02 9.52798352e-02 6.77298844e-01 -7.98184127e-02
-1.72836930e-01 7.91780114e-01 4.53161150e-02 4.64025408e-01
-9.20544982e-01 -6.36319041e-01 7.10123554e-02 7.71684587e-01
-1.18150139e+00 -3.22680116e-01 7.75062501e-01 1.37651101e-01
1.23000753e+00 5.38009048e-01 1.10133386e+00 5.56679189e-01
2.39616185e-01 6.77523911e-01 1.17954028e+00 -7.17473567e-01
8.25489819e-01 -2.69066226e-02 -6.50224268e-01 2.73595572e-01
1.11317575e-01 2.58494824e-01 8.84706751e-02 1.76237941e-01
3.40284020e-01 -3.90340686e-02 -2.49729306e-01 -2.81960249e-01
-8.79313767e-01 6.37510002e-01 8.56047034e-01 8.15356076e-01
-4.58912253e-01 3.81885469e-01 2.84661889e-01 -1.37157604e-01
1.39221609e-01 5.30709147e-01 -3.92452359e-01 -2.65805244e-01
-6.86594546e-01 -1.88455001e-01 4.32137311e-01 5.07896841e-01
6.11925066e-01 2.92784482e-01 1.83170214e-02 9.63681519e-01
3.66299152e-01 6.02184951e-01 7.54204273e-01 -1.14246690e+00
2.01516256e-01 9.70909238e-01 1.99533701e-01 -7.09782660e-01
-6.00169837e-01 -2.44594857e-01 -1.17935455e+00 9.71442610e-02
-3.75227183e-01 -1.43310921e-02 -4.09941554e-01 1.57690942e+00
3.99153531e-01 5.83277233e-02 -7.65780807e-02 3.91019553e-01
3.04148495e-02 1.11033857e+00 2.42985606e-01 -6.20844781e-01
1.01457477e+00 -7.05114067e-01 -8.22049081e-01 -1.96957275e-01
3.88160020e-01 -3.28110665e-01 7.63218522e-01 1.56351194e-01
-8.72681677e-01 -5.43165743e-01 -1.11670041e+00 6.57653689e-01
-5.58548272e-01 -1.06226601e-01 3.62970829e-01 8.19073379e-01
-9.03093874e-01 6.35568678e-01 -1.09363425e+00 -5.61238170e-01
2.27560863e-01 4.63916749e-01 2.95783460e-01 1.27546653e-01
-1.22044587e+00 1.08527994e+00 9.36570942e-01 2.11168438e-01
-5.56806684e-01 -3.51899773e-01 -4.98617083e-01 1.78721577e-01
5.76957226e-01 -5.31850755e-01 1.32260835e+00 -3.85866553e-01
-1.72992229e+00 -5.73770814e-02 2.33540908e-01 -3.59885961e-01
1.81623295e-01 -7.31632188e-02 -5.78187346e-01 -1.11686751e-01
-2.81790316e-01 2.81303525e-01 4.80065256e-01 -1.13427079e+00
-1.00667906e+00 -2.25322366e-01 -1.63604617e-01 1.69452414e-01
-6.89576685e-01 -4.20624286e-01 3.76999713e-02 -2.16318831e-01
-1.08947843e-01 -1.15472901e+00 -3.61077428e-01 -7.41529822e-01
-1.05679249e-02 -6.09360874e-01 1.12659466e+00 -6.60029769e-01
1.58624196e+00 -1.89275038e+00 9.61323828e-02 3.43978912e-01
-5.31087995e-01 4.25750315e-01 2.66150802e-01 7.32532561e-01
-1.34366164e-02 1.19541287e-01 -1.92135498e-01 1.50458753e-01
-9.63838492e-03 6.25635326e-01 1.67947710e-01 1.30311668e-01
-2.91920274e-01 6.65490389e-01 -9.24451888e-01 -3.05474758e-01
8.71174812e-01 3.24174374e-01 -2.46779144e-01 4.10931557e-01
-4.09546226e-01 4.33346689e-01 -8.00660968e-01 4.08594638e-01
2.15651318e-01 -3.13524097e-01 4.33566332e-01 -1.93611756e-01
-3.76871973e-01 -9.17552784e-02 -1.31046486e+00 1.34817410e+00
-1.18227339e+00 3.11722338e-01 -3.58203165e-02 -1.29485548e+00
6.94225907e-01 4.49548483e-01 1.03515387e+00 -1.36392212e+00
2.79128142e-02 1.88986704e-01 -1.81312904e-01 -6.74482942e-01
2.01954857e-01 8.56044292e-02 9.80394632e-02 2.64551967e-01
-5.84638596e-01 -4.44922447e-01 2.30789900e-01 -2.58084625e-01
1.06536746e+00 3.11852276e-01 6.68901980e-01 -2.29745343e-01
8.86501849e-01 -4.01874363e-01 5.62660694e-01 6.29726052e-02
-4.43091542e-02 -3.50707859e-01 -4.57632899e-01 -2.03275919e-01
-7.81032443e-01 -6.72342718e-01 2.35449627e-01 9.65565503e-01
1.04207210e-01 -4.86558229e-02 -7.73158371e-01 -2.81084776e-01
-2.64147103e-01 1.27610660e+00 -2.79490519e-02 -4.70080256e-01
-7.90617764e-01 -9.89276767e-01 -2.53098100e-01 4.90138173e-01
1.14140081e+00 -1.14516771e+00 -1.34665298e+00 6.34021342e-01
-5.31575024e-01 -8.75423551e-01 5.20041734e-02 2.78640062e-01
-8.51229966e-01 -1.13837802e+00 -3.32108527e-01 -4.76512432e-01
3.99522215e-01 2.54476577e-01 8.45987976e-01 2.73254782e-01
-3.70824963e-01 5.64026952e-01 -4.36539680e-01 -4.78274554e-01
-7.15312660e-01 4.22091037e-01 8.33714679e-02 -3.43627274e-01
-9.20822918e-02 -6.52369440e-01 -1.09264123e+00 4.86320108e-01
-8.02694440e-01 2.25784585e-01 6.22391224e-01 4.66446787e-01
3.71156991e-01 1.10674548e+00 8.47036421e-01 -3.12302202e-01
5.06169379e-01 -3.00466776e-01 -7.66517758e-01 6.87468469e-01
-1.13870037e+00 3.43758494e-01 6.93245411e-01 4.94451299e-02
-1.16766334e+00 -2.23862365e-01 -7.34248385e-03 3.98400694e-01
-1.43872797e-01 6.72280267e-02 -4.57026750e-01 2.02796564e-01
-6.08528778e-03 2.85811305e-01 -3.77100497e-01 -3.41566026e-01
1.06586292e-01 7.45954871e-01 1.54460758e-01 -6.37670219e-01
8.42247963e-01 7.42891505e-02 4.38568324e-01 -9.76590812e-01
-5.71656585e-01 -5.23151219e-01 -1.09963641e-01 -4.35126662e-01
8.05948496e-01 -6.36877179e-01 -9.92467284e-01 2.50732690e-01
-5.56777537e-01 -5.11262000e-01 -3.29292446e-01 2.86651492e-01
-7.14412272e-01 2.47826263e-01 -6.63247891e-03 -1.18178189e+00
-6.34493530e-01 -1.04683137e+00 5.22541344e-01 5.44256151e-01
-3.30049619e-02 -8.51353168e-01 2.60296911e-01 2.27754042e-01
7.56231129e-01 4.40485835e-01 1.12600768e+00 6.79427683e-02
-7.63630927e-01 1.02547649e-02 1.78902239e-01 6.87350392e-01
6.34314656e-01 -6.38196617e-02 -8.45857978e-01 -6.06620669e-01
1.50983989e-01 -8.11885968e-02 4.37476754e-01 3.32730025e-01
1.42822111e+00 -3.64052773e-01 -4.89466637e-01 8.01828206e-02
1.73734117e+00 7.53401399e-01 6.82604432e-01 7.90450633e-01
1.50387272e-01 4.01444703e-01 7.61364102e-01 7.34672606e-01
4.63254362e-01 6.31998360e-01 8.06712329e-01 5.52396930e-04
2.09694266e-01 -2.79253982e-02 3.23208094e-01 1.05624664e+00
-4.85217452e-01 -3.24263781e-01 -7.69227386e-01 3.94970536e-01
-2.00419235e+00 -1.09768820e+00 4.03639585e-01 2.10368395e+00
5.77484071e-01 -1.31624416e-01 -6.10866286e-02 6.29284859e-01
4.84034389e-01 2.08008915e-01 -6.85007513e-01 -1.45309374e-01
3.67386788e-01 2.54739553e-01 4.44228470e-01 -1.75143227e-01
-7.28987336e-01 3.38807940e-01 5.48900509e+00 7.68943787e-01
-9.82045412e-01 -2.42557395e-02 4.96203452e-01 1.89049111e-03
6.14335500e-02 -2.37902924e-01 -3.02402377e-01 6.85940206e-01
1.34756780e+00 -2.90449619e-01 8.94516706e-01 1.00845218e+00
8.24719429e-01 -5.64712465e-01 -9.53213394e-01 9.39576209e-01
-5.95655262e-01 -1.00532448e+00 -5.16588926e-01 2.09027052e-01
8.63713801e-01 -2.04104632e-01 -2.67594427e-01 4.03994173e-01
2.95072138e-01 -6.16781414e-01 3.70786190e-01 4.35924053e-01
1.48342028e-01 -1.24539578e+00 8.72699618e-01 7.68261731e-01
-1.68236232e+00 -6.14316165e-01 -1.73646688e-01 1.98354915e-01
2.21344888e-01 4.03712481e-01 -6.36099279e-01 8.62798631e-01
1.04242373e+00 2.01266229e-01 -4.90115941e-01 9.42235053e-01
-3.27054471e-01 5.76382637e-01 -4.33137745e-01 -4.48813558e-01
-7.10349483e-03 -3.30786556e-01 -9.21142623e-02 5.51987708e-01
5.77213585e-01 2.48007059e-01 3.52434963e-01 3.37935328e-01
1.69202805e-01 1.42169148e-01 -2.75250614e-01 -1.55194268e-01
5.11427045e-01 1.23117805e+00 -7.53452420e-01 -9.51779783e-02
-2.27632582e-01 5.75374722e-01 -1.23109825e-01 1.52867749e-01
-9.14564908e-01 -1.02035165e-01 4.98240024e-01 3.92526835e-02
1.53079063e-01 -3.21742654e-01 1.46166220e-01 -4.40066338e-01
-2.26833493e-01 -6.49963021e-01 3.92755061e-01 -5.44765174e-01
-7.79781759e-01 -7.12419220e-04 2.29699820e-01 -8.88737381e-01
-4.04179513e-01 -3.48132044e-01 -6.29502773e-01 3.22243482e-01
-1.75317621e+00 -7.79543459e-01 -5.17002225e-01 3.62246513e-01
8.20051670e-01 9.19044763e-02 9.72488999e-01 1.84287295e-01
-9.27657068e-01 -7.25584924e-02 4.93218184e-01 -4.93100911e-01
2.52442390e-01 -1.19566190e+00 -6.20276779e-02 6.95771396e-01
-2.47870192e-01 -4.39448878e-02 6.50978386e-01 -2.85493553e-01
-1.58631873e+00 -1.07471073e+00 1.64052352e-01 2.35648364e-01
3.66087645e-01 1.13997132e-01 -6.27445817e-01 4.98287268e-02
6.25457346e-01 -4.64692444e-01 6.34077430e-01 -2.78670311e-01
5.66166222e-01 -5.06090939e-01 -1.21584594e+00 5.10313034e-01
7.82914758e-01 -1.46217540e-01 -1.97165877e-01 2.27634639e-01
4.20567125e-01 6.27001971e-02 -9.72614765e-01 6.67389691e-01
3.46278816e-01 -9.63525653e-01 8.72764468e-01 -5.41794784e-02
-2.13456124e-01 -4.05010611e-01 -2.52690852e-01 -1.55084896e+00
-4.78570879e-01 -2.74110585e-01 -6.68498099e-01 1.38525963e+00
-1.46124512e-01 -6.58026993e-01 5.01015902e-01 6.30943716e-01
-3.50563228e-02 -8.32969010e-01 -1.01610434e+00 -8.90555620e-01
-2.98275918e-01 -3.13228756e-01 9.65181649e-01 4.82996911e-01
-1.81127205e-01 6.47361204e-02 -3.38045955e-02 3.04956585e-01
5.45166075e-01 6.04028963e-02 5.43422222e-01 -1.01387298e+00
-1.87508553e-01 -4.50494617e-01 -1.94885433e-02 -3.33686471e-01
8.25661272e-02 -5.35628796e-01 5.52643053e-02 -2.27786708e+00
2.16269121e-02 -3.41341138e-01 -7.90687919e-01 4.96976256e-01
-8.43109936e-02 -2.94012010e-01 2.25991324e-01 -1.23401240e-01
-4.54440176e-01 1.09981799e+00 9.66207206e-01 -2.73689121e-01
-3.19297552e-01 3.90662223e-01 -3.06550413e-01 6.81858361e-01
1.35211980e+00 -3.64331126e-01 -6.15745306e-01 -8.61468632e-03
2.76708812e-01 -1.35197610e-01 -2.97392141e-02 -1.45305216e+00
1.48284450e-01 -6.76046789e-01 3.18022400e-01 -5.83591163e-01
4.84311096e-02 -1.51603484e+00 6.25711441e-01 1.06302488e+00
1.81976840e-01 1.39757425e-01 1.10019147e-01 6.39398634e-01
1.17331639e-01 -8.58521983e-02 9.94108975e-01 -2.48928979e-01
-1.00673413e+00 -4.33034487e-02 -4.41637307e-01 -3.73627037e-01
1.51943219e+00 -1.63376763e-01 -3.83366384e-02 -1.10186480e-01
-2.38229871e-01 6.25190735e-01 2.64310569e-01 5.56227148e-01
3.45504344e-01 -1.25594556e+00 -1.57657266e-01 -1.71067566e-01
-1.54602438e-01 6.03010505e-02 3.41570199e-01 4.25736040e-01
-3.00587028e-01 3.59387279e-01 -2.28046272e-02 -4.06546086e-01
-1.12865043e+00 6.54040456e-01 4.19536054e-01 -5.96694171e-01
-4.12693620e-01 1.75200358e-01 -3.55573177e-01 8.55523441e-03
1.13243058e-01 -4.57952172e-01 -1.18554525e-01 -4.74059619e-02
3.27207774e-01 1.13485670e+00 1.78906918e-01 -1.30942672e-01
-3.92418772e-01 5.29104233e-01 5.13857841e-01 3.50493878e-01
1.61349511e+00 -3.05064648e-01 -1.22846499e-01 4.25966352e-01
7.08715141e-01 -3.81988674e-01 -9.44111049e-01 1.78510845e-01
4.56687152e-01 -1.96157947e-01 4.69103724e-01 -7.50275075e-01
-9.51457381e-01 5.35965800e-01 1.16596699e+00 6.10853016e-01
1.72374737e+00 -2.81017363e-01 9.55757976e-01 5.53431928e-01
8.84671628e-01 -1.83087444e+00 4.01732475e-01 -1.38324618e-01
6.66012466e-01 -1.20325851e+00 3.05135995e-01 8.28313529e-02
-1.97225720e-01 1.04238427e+00 5.88631451e-01 3.98387432e-01
5.69780350e-01 4.77203727e-02 -3.72677177e-01 1.76813155e-01
-4.56234097e-01 -1.83230922e-01 -3.82374749e-02 4.59935844e-01
1.76342562e-01 2.81612068e-01 -6.36840463e-02 -1.39236040e-02
1.46572456e-01 -5.12617156e-02 -9.26156268e-02 1.23971784e+00
-9.39699173e-01 -1.44249523e+00 -3.44334364e-01 2.69811481e-01
3.06467596e-03 5.53712785e-01 3.33812594e-01 8.85388911e-01
1.94568604e-01 1.24717665e+00 -3.92202318e-01 -2.54872404e-02
7.72880197e-01 1.10334583e-01 4.07485425e-01 -4.02655512e-01
-3.81022811e-01 -7.36888126e-02 -1.87663227e-01 -6.02922678e-01
-9.06313717e-01 -3.98701936e-01 -1.55387199e+00 -4.30823952e-01
-5.29727161e-01 2.41727203e-01 9.94514108e-01 1.10128546e+00
3.10849696e-01 1.21404254e+00 1.21224904e+00 -6.04937136e-01
-6.16196513e-01 -8.36185753e-01 -5.29000759e-01 -4.51716669e-02
-6.54622912e-02 -6.73332870e-01 -1.87934265e-01 -2.81601101e-01] | [5.629508018493652, 2.4268252849578857] |
7d90a3cb-3737-427c-8889-f875fa7f990f | towards-general-robustness-to-bad-training | null | null | https://openreview.net/forum?id=kz6rsFehYjd | https://openreview.net/pdf?id=kz6rsFehYjd | Towards General Robustness to Bad Training Data | In this paper, we focus on the problem of identifying bad training data when the underlying cause is unknown in advance. Our key insight is that regardless of how bad data are generated, they tend to contribute little to training a model with good prediction performance or more generally, to some utility function of the data analyst. We formulate the problem of good/bad data selection as utility optimization. We propose a theoretical framework for evaluating the worst-case performance of data selection heuristics. Remarkably, our results show that the popular heuristic based on the Shapley value may choose the worst data subset in certain practical scenarios, which sheds lights on its large performance variation observed empirically in the past work. We then develop an algorithmic framework, DataSifter, to detect a variety of and even unknown data issues---a step towards general robustness to bad training data. DataSifter is guided by the theoretically optimal solution to data selection and is made practical by the data utility learning technique. Our evaluation shows that DataSifter achieves and most often significantly improves the state-of-the-art performance over a wide range of tasks, including backdoor, poison, noisy/mislabel data detection, data summarization, and data debiasing. | ['Ruoxi Jia', 'Ming Jin', 'Yi Zeng', 'Tianhao Wang'] | 2021-09-29 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 1.64145738e-01 2.16423571e-01 -6.92751825e-01 -2.65452228e-02
-1.34793675e+00 -7.48071611e-01 5.41019738e-01 6.29282236e-01
-2.27013022e-01 9.37220037e-01 3.51942003e-01 -3.78006816e-01
-3.73752147e-01 -5.55191100e-01 -7.80822396e-01 -1.06351364e+00
3.49152684e-02 6.25025988e-01 -1.51770130e-01 -8.06223303e-02
6.84780359e-01 1.80886090e-01 -1.39096355e+00 1.27964199e-01
8.46913993e-01 1.09038544e+00 -3.39820594e-01 5.65688312e-01
2.09852129e-01 1.17544925e+00 -7.78808475e-01 -6.02397382e-01
4.46972072e-01 -2.93468714e-01 -1.03563321e+00 1.64375708e-01
1.30407989e-01 -3.44826430e-01 -6.94181249e-02 1.00602996e+00
5.37262678e-01 -1.98999960e-02 7.09322691e-01 -1.61831236e+00
-5.37450552e-01 9.33726192e-01 -5.21977484e-01 3.15857381e-01
1.77925542e-01 4.11416590e-01 1.36356091e+00 -3.83937508e-01
5.39171100e-01 1.11972034e+00 5.29248357e-01 6.05306625e-01
-1.41244721e+00 -5.36946058e-01 1.27520978e-01 1.60117909e-01
-1.12616599e+00 -7.17263162e-01 4.97690320e-01 -3.59685481e-01
4.61454451e-01 5.76572657e-01 1.61527912e-03 1.13831389e+00
5.38011119e-02 1.21289182e+00 7.96643436e-01 -1.83556184e-01
5.65594316e-01 2.88587511e-01 5.15034437e-01 3.08793843e-01
7.37309217e-01 1.95687875e-01 -7.39652574e-01 -8.33743811e-01
-1.12404050e-02 -5.77682704e-02 -2.73830831e-01 -5.07908583e-01
-7.55010068e-01 9.24272895e-01 -8.05492513e-03 -1.41385809e-01
-3.74796569e-01 6.47919849e-02 6.71681821e-01 7.64080524e-01
5.94830871e-01 1.00083995e+00 -7.32267022e-01 -1.60065413e-01
-5.18250585e-01 4.92475808e-01 9.62323248e-01 6.74417853e-01
4.64032382e-01 -9.19628069e-02 -3.41388285e-01 5.52327633e-01
-1.42554954e-01 1.42421633e-01 6.06629550e-01 -9.72223938e-01
7.53167033e-01 4.55630034e-01 4.23654109e-01 -6.39676452e-01
-3.75133932e-01 -4.29827452e-01 -7.44010270e-01 -1.00001529e-01
5.99911332e-01 -2.15548769e-01 -5.51188767e-01 1.71736205e+00
3.41467172e-01 -2.48667851e-01 2.64224648e-01 5.91300428e-01
1.96652174e-01 3.68371397e-01 -2.59425025e-02 -7.16730893e-01
7.68078923e-01 -4.84340727e-01 -5.50455213e-01 -7.61085674e-02
9.57558692e-01 -2.25461990e-01 1.13015461e+00 9.17824566e-01
-1.08262312e+00 1.95417814e-02 -8.70451272e-01 1.26193613e-01
-6.29667863e-02 -4.37127590e-01 5.32553196e-01 8.05323064e-01
-6.67954564e-01 8.41459930e-01 -5.36680043e-01 -3.24316174e-01
8.49645674e-01 2.71990865e-01 -1.75440639e-01 -6.36343285e-02
-9.22872305e-01 7.57868171e-01 5.13210654e-01 -6.20198011e-01
-1.17661667e+00 -7.79369414e-01 -3.01961929e-01 5.20976074e-02
1.04622471e+00 -8.10265601e-01 1.41650832e+00 -6.63398683e-01
-6.94394648e-01 7.43221939e-01 1.09022232e-02 -9.63573098e-01
7.35331237e-01 -4.07505125e-01 -1.77008003e-01 -1.60963669e-01
-6.09989418e-03 -1.52777597e-01 7.36178815e-01 -1.47358799e+00
-8.31257403e-01 -6.24462366e-01 -2.18445778e-01 9.21884850e-02
-4.77385432e-01 -1.59752563e-01 -7.69765750e-02 -6.83946729e-01
-7.86883533e-02 -6.87064707e-01 -4.61950749e-01 -4.25545812e-01
-8.71363759e-01 -4.57401574e-01 5.72451651e-01 -7.45806769e-02
1.72493780e+00 -2.05622244e+00 -2.20761254e-01 4.38374758e-01
3.85837823e-01 1.27871230e-01 1.63863763e-01 5.71601927e-01
-2.04387069e-01 6.99601531e-01 -2.48018742e-01 -1.94270819e-01
9.42945294e-03 2.57968068e-01 -7.74894178e-01 7.68168926e-01
1.33864274e-02 6.26619399e-01 -8.95646632e-01 -9.68530029e-02
-1.13851830e-01 -4.90794063e-01 -5.39878666e-01 3.00706297e-01
-2.91035086e-01 1.33050486e-01 -6.73676431e-01 6.77456081e-01
5.58015883e-01 -3.19779992e-01 1.74389854e-01 2.43981525e-01
2.14477539e-01 4.38557923e-01 -1.33350480e+00 9.95405853e-01
-5.21811377e-03 1.09493241e-01 1.09963957e-03 -1.29641807e+00
5.87630033e-01 6.43576011e-02 6.62664175e-01 -5.26945651e-01
7.10606948e-02 1.78752244e-01 -3.47305983e-01 -5.98864555e-01
5.28772414e-01 -1.02159150e-01 -3.87707472e-01 8.72887194e-01
-3.23321491e-01 1.61945179e-01 1.77097484e-01 4.24801171e-01
1.54137754e+00 -7.46169388e-01 7.98439443e-01 -3.84621114e-01
7.45404959e-02 1.25581607e-01 6.43272161e-01 1.43782079e+00
-2.30033338e-01 4.65386093e-01 1.15410244e+00 -3.94654304e-01
-1.00173807e+00 -8.75291646e-01 -1.26371354e-01 1.27616179e+00
6.53803125e-02 -5.07930398e-01 -6.76811814e-01 -9.67644930e-01
4.88248557e-01 9.49537933e-01 -6.30039752e-01 -5.28124452e-01
-1.41499639e-01 -1.15403986e+00 4.31078464e-01 1.34737611e-01
8.59299377e-02 -8.87709141e-01 -3.94757539e-01 2.52910942e-01
6.92053288e-02 -5.99976599e-01 -2.94742852e-01 7.07844615e-01
-9.13907707e-01 -1.46769881e+00 -2.18693670e-02 2.57167537e-02
4.70012963e-01 2.95890808e-01 1.38026190e+00 2.42628664e-01
7.73337260e-02 2.94223249e-01 -4.82884824e-01 -6.67242229e-01
-7.40018547e-01 1.80759415e-01 3.06301355e-01 -7.65244365e-02
4.27995443e-01 -2.75642008e-01 -3.11406940e-01 2.31639445e-01
-1.00254261e+00 -5.89757740e-01 3.86238754e-01 1.02963853e+00
4.95566785e-01 5.15128076e-01 8.39627922e-01 -1.48410988e+00
8.81656706e-01 -1.01116860e+00 -3.84119064e-01 4.26627666e-01
-1.11379552e+00 4.37527865e-01 9.21827435e-01 -1.15888342e-01
-7.85861135e-01 -2.52615977e-02 3.01798042e-02 -1.49394542e-01
-6.35639429e-02 3.25757027e-01 -4.74434108e-01 3.86719674e-01
1.14030600e+00 1.42117247e-01 -9.75308474e-03 -6.54659569e-01
3.40465844e-01 8.19740355e-01 6.39087498e-01 -7.65975237e-01
6.02470934e-01 3.87156397e-01 -1.81546703e-01 -4.89207655e-01
-1.15924299e+00 -5.61937332e-01 -1.26870304e-01 4.06808138e-01
6.70289546e-02 -7.67764211e-01 -8.72496605e-01 3.95465612e-01
-6.32216752e-01 -2.28410050e-01 -6.89560592e-01 -2.24922806e-01
-5.89193761e-01 3.11396748e-01 -2.43402824e-01 -1.13699782e+00
-3.89358461e-01 -9.87374127e-01 6.81759238e-01 -2.27933064e-01
-3.58282357e-01 -7.86389828e-01 -1.92718998e-01 3.59989643e-01
-6.97121695e-02 2.95941561e-01 1.00687814e+00 -1.28208303e+00
-6.25850439e-01 -3.01501691e-01 1.60470575e-01 4.09147382e-01
-1.29374772e-01 -1.05218850e-01 -1.09603226e+00 -4.20962244e-01
2.07162499e-01 -6.59444332e-01 1.22464335e+00 4.11142826e-01
1.83056986e+00 -9.60013330e-01 -2.73648024e-01 4.17147785e-01
1.32107449e+00 4.00018767e-02 4.00672466e-01 2.67014295e-01
5.19318342e-01 5.45547068e-01 7.84172833e-01 1.19062936e+00
2.07905650e-01 4.25051004e-01 7.21826851e-01 2.47668475e-01
5.79002976e-01 -4.03172553e-01 2.51116842e-01 -5.88516034e-02
5.12377679e-01 -5.63348293e-01 -8.49359870e-01 7.19535887e-01
-2.07448769e+00 -1.19694161e+00 -2.35779092e-01 2.71024156e+00
1.07687652e+00 5.01385152e-01 6.25187576e-01 4.29383427e-01
4.56412494e-01 2.44109575e-02 -9.96867239e-01 -3.59298259e-01
-2.47189105e-01 -2.67764628e-01 1.09679580e+00 1.71803370e-01
-1.19245470e+00 6.25840545e-01 6.90669394e+00 1.12421668e+00
-6.53792560e-01 2.95649339e-02 1.07814896e+00 -4.01186258e-01
-5.07403791e-01 -1.05307661e-01 -7.62751639e-01 6.03254318e-01
9.58900750e-01 -8.15630019e-01 4.77731735e-01 1.06213760e+00
2.21521899e-01 -3.32070559e-01 -1.44748604e+00 9.38092768e-01
-1.81181595e-01 -1.42971623e+00 -4.36336435e-02 2.55566001e-01
7.73662686e-01 4.34653014e-02 1.95263177e-01 7.22011328e-02
9.75646377e-01 -8.80493045e-01 7.17297792e-01 3.37934434e-01
4.96100307e-01 -1.00938451e+00 5.57856560e-01 8.10830653e-01
-2.96888024e-01 -8.20467234e-01 -3.58219594e-01 -1.47278700e-02
-2.85094261e-01 1.06644976e+00 -9.32916224e-01 3.89416009e-01
6.04027987e-01 5.66810250e-01 -6.90128863e-01 9.58819568e-01
1.91032380e-01 9.75520849e-01 -3.28370243e-01 1.41824573e-01
2.70640804e-03 3.16071540e-01 5.52186668e-01 8.73923004e-01
3.76788378e-02 -2.89364439e-03 2.39115492e-01 4.43861902e-01
-4.71715271e-01 -2.44811978e-02 -7.76515067e-01 -3.87652498e-03
8.63775790e-01 7.06236482e-01 -2.13860303e-01 -4.23505455e-01
6.54596388e-02 3.78492206e-01 2.31322840e-01 5.97338229e-02
-2.65581638e-01 2.62449533e-02 7.41269231e-01 1.78664982e-01
-1.64227560e-01 4.53362197e-01 -9.52070594e-01 -1.00123990e+00
1.35337517e-01 -1.17027271e+00 1.10751188e+00 -1.97580427e-01
-1.76545990e+00 -1.62138175e-02 -1.52323142e-01 -1.06044996e+00
-3.21496457e-01 -3.01282555e-01 -5.89075148e-01 5.42311013e-01
-1.05021143e+00 -3.79281312e-01 2.07103595e-01 5.27833879e-01
3.73685151e-01 -3.38084608e-01 5.26074529e-01 -2.25713715e-01
-7.45524406e-01 8.18376839e-01 7.52920747e-01 -3.63703519e-01
6.93758130e-01 -1.50210226e+00 3.94734472e-01 8.63881588e-01
-9.43241343e-02 5.41753590e-01 1.08348715e+00 -6.45871043e-01
-1.45505488e+00 -1.04529703e+00 7.90974498e-01 -8.27989221e-01
5.80155492e-01 -1.65494725e-01 -1.03214407e+00 4.26909328e-01
-2.18503609e-01 -1.21701404e-01 6.50982082e-01 4.08225000e-01
-4.07892942e-01 -4.44298655e-01 -1.46941817e+00 3.14262152e-01
9.64277387e-01 -2.22667202e-01 -4.86582965e-01 5.60902834e-01
5.84924996e-01 -1.42691016e-01 -5.02285004e-01 9.04867724e-02
2.69978583e-01 -1.12145269e+00 8.21209729e-01 -1.11048102e+00
4.40691322e-01 1.10489219e-01 -1.89060837e-01 -1.42891312e+00
-1.78325459e-01 -1.18221366e+00 -3.96325886e-01 1.26006961e+00
4.00998801e-01 -4.79339659e-01 8.92554462e-01 1.02471697e+00
1.83971643e-01 -6.56218588e-01 -8.61110210e-01 -8.04704368e-01
2.35169441e-01 -6.32352114e-01 7.94828832e-01 1.05199540e+00
4.03789729e-02 1.38219461e-01 -7.28323877e-01 -7.64890164e-02
9.25678253e-01 -9.96952727e-02 1.05788374e+00 -1.35586333e+00
-2.09648982e-01 -4.83096540e-01 -5.30990586e-02 -9.64796364e-01
-2.57938672e-02 -7.06238508e-01 -1.05865851e-01 -1.09359097e+00
3.01073223e-01 -4.72084373e-01 -6.27578855e-01 6.28124654e-01
-4.22374129e-01 -2.63384432e-01 1.91510487e-02 4.05248106e-01
-1.00855994e+00 2.21236303e-01 7.38452375e-01 7.87272602e-02
-3.88054371e-01 3.93428773e-01 -1.73597097e+00 5.25986910e-01
8.73714209e-01 -8.16392422e-01 -3.24270934e-01 7.05386773e-02
4.82334495e-01 1.00622214e-01 3.85840118e-01 -5.32271028e-01
9.75802317e-02 -5.77234149e-01 4.10144627e-02 -5.78909397e-01
-4.55580056e-01 -8.15045834e-01 -5.77684753e-02 4.48572814e-01
-9.11660731e-01 -1.26722321e-01 -3.21242839e-01 7.93559313e-01
1.67187542e-01 -3.25761497e-01 8.69225264e-01 -1.61166325e-01
-5.29214084e-01 2.31549770e-01 -3.31966318e-02 6.26832128e-01
1.09124756e+00 1.28070265e-01 -7.60507643e-01 -4.12267894e-01
-4.66306806e-01 5.60329914e-01 4.64956850e-01 3.13006401e-01
2.41978824e-01 -1.13488567e+00 -7.56313384e-01 1.65309086e-01
5.85601211e-01 -7.36910552e-02 8.76579732e-02 7.09380984e-01
2.34195188e-01 2.36346498e-01 2.01900125e-01 -3.19882303e-01
-1.12503242e+00 8.09056342e-01 1.71471328e-01 -4.17546540e-01
-3.94387245e-01 8.08100522e-01 1.15213044e-01 -2.15282524e-03
4.54059541e-01 -5.96491434e-02 1.25648141e-01 4.67194110e-01
6.87890887e-01 6.74526274e-01 3.33946407e-01 -1.37496933e-01
-1.80165559e-01 -3.73240739e-01 -4.14810389e-01 3.92339915e-01
1.31964827e+00 -2.29364783e-01 1.46993265e-01 4.63448256e-01
1.23828280e+00 -1.20385475e-01 -1.30329609e+00 -4.88455534e-01
4.40624654e-01 -7.02055395e-01 6.05297275e-02 -1.07759905e+00
-8.44227672e-01 5.15940070e-01 2.79638618e-01 8.97643685e-01
1.18371463e+00 2.43511006e-01 5.48377633e-01 6.32247031e-01
3.37557018e-01 -1.39281988e+00 -1.86082702e-02 1.48946717e-01
8.36680770e-01 -1.45050347e+00 1.66118458e-01 1.79184332e-01
-1.05125260e+00 7.16221392e-01 3.87689680e-01 6.65400252e-02
4.27647442e-01 2.15690792e-01 -2.48227850e-01 -1.16220161e-01
-1.31177330e+00 -1.02020919e-01 -7.16129020e-02 5.59482336e-01
-8.96361098e-02 1.60705820e-01 -2.07986623e-01 9.15043592e-01
-2.48142719e-01 2.76628304e-02 8.35092425e-01 8.36134613e-01
-6.82413399e-01 -1.07782388e+00 -5.79225123e-01 1.11918461e+00
-8.84953201e-01 1.46633629e-02 -6.54732406e-01 5.58029354e-01
-4.18045186e-02 1.20161402e+00 -1.79875046e-01 -4.45773870e-01
4.59495753e-01 7.53150955e-02 -1.30448565e-01 -4.86142248e-01
-7.00815558e-01 -5.60269207e-02 1.68288380e-01 -6.76880181e-01
-7.21454620e-02 -8.54222953e-01 -8.18077385e-01 -7.57292449e-01
-2.53477097e-01 2.43539825e-01 2.14153051e-01 9.49893951e-01
4.18192059e-01 2.18836647e-02 1.15379381e+00 -9.04087648e-02
-1.48648167e+00 -4.98830259e-01 -9.76260424e-01 7.50518620e-01
7.29748011e-01 -4.22259361e-01 -6.05519712e-01 -1.07127644e-01] | [8.705284118652344, 4.2943196296691895] |
079e51e7-175f-4882-880a-a2da4acd70df | describing-textures-in-the-wild | 1311.3618 | null | http://arxiv.org/abs/1311.3618v2 | http://arxiv.org/pdf/1311.3618v2.pdf | Describing Textures in the Wild | Patterns and textures are defining characteristics of many natural objects: a
shirt can be striped, the wings of a butterfly can be veined, and the skin of
an animal can be scaly. Aiming at supporting this analytical dimension in image
understanding, we address the challenging problem of describing textures with
semantic attributes. We identify a rich vocabulary of forty-seven texture terms
and use them to describe a large dataset of patterns collected in the wild.The
resulting Describable Textures Dataset (DTD) is the basis to seek for the best
texture representation for recognizing describable texture attributes in
images. We port from object recognition to texture recognition the Improved
Fisher Vector (IFV) and show that, surprisingly, it outperforms specialized
texture descriptors not only on our problem, but also in established material
recognition datasets. We also show that the describable attributes are
excellent texture descriptors, transferring between datasets and tasks; in
particular, combined with IFV, they significantly outperform the
state-of-the-art by more than 8 percent on both FMD and KTHTIPS-2b benchmarks.
We also demonstrate that they produce intuitive descriptions of materials and
Internet images. | ['Subhransu Maji', 'Mircea Cimpoi', 'Andrea Vedaldi', 'Sammy Mohamed', 'Iasonas Kokkinos'] | 2013-11-14 | describing-textures-in-the-wild-1 | http://openaccess.thecvf.com/content_cvpr_2014/html/Cimpoi_Describing_Textures_in_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Cimpoi_Describing_Textures_in_2014_CVPR_paper.pdf | cvpr-2014-6 | ['material-recognition'] | ['computer-vision'] | [ 4.31609839e-01 -3.58592272e-01 -2.47941256e-01 -4.46199298e-01
-5.30532479e-01 -7.77015448e-01 9.88337398e-01 -1.76656604e-01
2.73339719e-01 3.41490060e-01 7.75901452e-02 1.13502830e-01
-5.61267138e-01 -7.88332582e-01 -6.58788264e-01 -1.03694773e+00
-7.15112314e-02 6.12684667e-01 6.10621907e-02 -3.79885465e-01
3.38534474e-01 7.19641864e-01 -1.97487497e+00 7.40757227e-01
5.48893251e-02 1.98374832e+00 1.69788346e-01 3.10356617e-01
-2.55141199e-01 8.85125875e-01 -9.31860134e-02 -4.83327627e-01
3.74309421e-02 1.16616435e-01 -1.06813395e+00 4.20800149e-01
8.07029307e-01 2.29740702e-02 -1.82331070e-01 8.94898713e-01
-2.06286199e-02 -4.19844650e-02 1.06869137e+00 -1.04808450e+00
-1.16361570e+00 1.08570941e-01 -3.89668703e-01 -2.52608389e-01
3.22193563e-01 -2.91287184e-01 1.14077282e+00 -9.23504829e-01
9.28577363e-01 1.39457738e+00 5.68617821e-01 6.22789383e-01
-1.32780349e+00 6.26474768e-02 -1.52983233e-01 1.46958128e-01
-1.24109626e+00 -4.46053773e-01 6.14366829e-01 -6.35092080e-01
7.81364202e-01 7.08069623e-01 5.09990931e-01 1.52343547e+00
3.80786717e-01 8.32342148e-01 1.47443628e+00 -4.15474832e-01
5.01993261e-02 -9.74867716e-02 1.73704967e-01 8.61420870e-01
1.55275792e-01 -4.79610898e-02 -7.60077059e-01 1.84679648e-03
7.91660964e-01 8.32243264e-02 1.19951621e-01 -4.95239168e-01
-1.32882714e+00 4.01493937e-01 6.12876087e-04 2.02523857e-01
-2.67308950e-01 2.25689545e-01 4.16758746e-01 6.30351424e-01
6.56215549e-01 4.62202370e-01 -2.88371593e-01 -4.19112563e-01
-3.19584161e-01 2.80154794e-01 1.00978684e+00 1.16769314e+00
7.82681942e-01 -4.66249734e-02 -1.87148720e-01 1.32830250e+00
-2.28793979e-01 9.46091950e-01 1.76137894e-01 -8.18383396e-01
-1.59711391e-02 4.91160780e-01 1.53682232e-01 -1.30297613e+00
-2.08104715e-01 5.18656299e-02 -9.50318515e-01 6.79096803e-02
4.90548611e-01 8.19310308e-01 -8.40689778e-01 1.19733202e+00
9.15896073e-02 -3.81379634e-01 -3.56885761e-01 8.74535382e-01
8.80572140e-01 3.98276716e-01 -1.27329469e-01 4.73060012e-01
1.75322700e+00 -7.36302078e-01 -5.35315156e-01 6.57920986e-02
3.86453897e-01 -9.50271785e-01 1.29604828e+00 7.62683570e-01
-5.78523338e-01 -6.37015760e-01 -9.28044319e-01 -2.18416676e-02
-6.01957977e-01 1.51277483e-01 1.06845748e+00 5.52644730e-01
-7.83003807e-01 8.19083035e-01 -6.08900726e-01 -6.69045210e-01
5.66115618e-01 6.29872307e-02 -9.54235792e-01 -2.18003556e-01
-5.22019744e-01 7.33013749e-01 1.17411196e-01 -2.70912591e-02
-6.94802225e-01 -5.05960286e-01 -6.92230225e-01 -1.57160833e-01
3.84848714e-01 -4.65866625e-01 7.90279269e-01 -9.39299464e-01
-1.48861039e+00 1.41861618e+00 -1.09400630e-01 -9.80454460e-02
2.80297846e-01 -1.22608170e-01 -5.80220878e-01 2.12128863e-01
-1.26242012e-01 5.87533772e-01 1.11662817e+00 -1.19693208e+00
-4.08761472e-01 -1.62566394e-01 -2.12321058e-02 -1.63067654e-01
-4.23503667e-01 -1.60231382e-01 -3.74309123e-01 -8.90284061e-01
2.16367364e-01 -1.11067998e+00 4.03819591e-01 3.24701011e-01
-5.35153866e-01 -2.46506393e-01 9.46419418e-01 -3.25333357e-01
7.21609652e-01 -2.49011183e+00 2.25620747e-01 2.88559020e-01
3.35410237e-01 -3.42033207e-01 -4.28028435e-01 3.40774268e-01
1.00283347e-01 2.21050739e-01 1.51639998e-01 -4.38417904e-02
4.54484999e-01 5.76676548e-01 -5.09108782e-01 5.05381405e-01
4.64723736e-01 9.41143751e-01 -5.50323188e-01 -3.32339257e-01
3.61323565e-01 4.25733298e-01 -3.15858454e-01 8.87131318e-02
-2.60724247e-01 2.27900609e-01 -7.83894598e-01 1.11177683e+00
6.23964846e-01 -4.32505876e-01 3.18360627e-02 -4.70069498e-01
1.60553381e-01 -6.71772882e-02 -8.99393559e-01 1.38533688e+00
-4.51754063e-01 8.07973266e-01 -3.60002726e-01 -9.66973960e-01
1.28087270e+00 6.63235318e-04 5.86482882e-01 -9.47522581e-01
5.60294278e-03 3.29399407e-01 -2.95782119e-01 -4.56344664e-01
5.41170001e-01 -1.28016800e-01 -3.30281496e-01 3.08569998e-01
1.61662623e-01 -4.54953223e-01 5.78719787e-02 -4.42904383e-01
1.01524568e+00 2.59628296e-01 5.85915744e-02 -9.40353572e-01
3.97494882e-01 -4.12445366e-02 -5.58069441e-03 8.81997108e-01
2.72763431e-01 7.07887113e-01 6.71583176e-01 -1.21021092e+00
-1.34099388e+00 -1.25143838e+00 -6.28053069e-01 1.37249410e+00
2.79427290e-01 -2.84030586e-01 -6.51721239e-01 -2.47195825e-01
4.33557421e-01 4.92441170e-02 -1.24967015e+00 1.90191463e-01
-1.33813843e-01 -4.20327723e-01 4.37207699e-01 4.37214643e-01
5.18288791e-01 -1.02076292e+00 -2.63706386e-01 -2.12965179e-02
-1.05168030e-01 -1.23798835e+00 -3.38910311e-01 2.60723531e-01
-4.56399769e-01 -1.07477772e+00 -5.17598987e-01 -7.65230834e-01
4.31264848e-01 2.76198387e-01 1.42630780e+00 1.19666077e-01
-4.77230310e-01 5.98818600e-01 -6.58383369e-01 -3.45575303e-01
-3.11905175e-01 -8.65932778e-02 1.78399712e-01 4.30019885e-01
4.09122407e-01 -2.67437071e-01 -2.60424465e-01 7.83368647e-01
-9.77604330e-01 2.12196186e-01 4.25052196e-01 1.01428604e+00
1.01641488e+00 -1.16183706e-01 1.53036639e-01 -9.05380905e-01
2.94745445e-01 -3.14511776e-01 -3.60693455e-01 5.14220357e-01
1.86334047e-02 3.73804212e-01 5.56671262e-01 -6.23260260e-01
-7.77798891e-01 -2.19930544e-01 5.81352450e-02 -2.88726866e-01
-4.36800659e-01 1.30739883e-01 -5.35787083e-02 -5.62499821e-01
5.47066271e-01 2.91461319e-01 -1.04425605e-02 -7.97358751e-01
1.92953303e-01 5.79761088e-01 4.50574458e-01 -1.31928098e+00
5.22865176e-01 8.52110505e-01 4.19238657e-01 -1.21294451e+00
-1.00609422e+00 -2.24395663e-01 -6.73551202e-01 -2.09975764e-01
7.50194311e-01 -5.34338593e-01 -9.30946529e-01 6.92191243e-01
-7.87102759e-01 -3.28534365e-01 -3.25176030e-01 -2.99590062e-02
-1.15842438e+00 8.72793123e-02 -6.60727739e-01 -4.34849352e-01
-1.55594898e-02 -1.08943129e+00 1.68122935e+00 -2.55888849e-01
-4.09205705e-01 -1.09412634e+00 -3.42684627e-01 1.03129126e-01
6.73477888e-01 6.36367857e-01 1.39039540e+00 -3.93223286e-01
-4.45121408e-01 -1.24326888e-02 -4.23745692e-01 2.17097685e-01
4.73729193e-01 7.65008526e-03 -1.08786476e+00 -8.77580345e-02
-7.73118511e-02 -6.69118762e-01 9.34905887e-01 7.64817595e-02
1.82306588e+00 -3.18311781e-01 -2.90256143e-02 8.53630543e-01
1.36824310e+00 7.19495211e-03 6.23648226e-01 4.72079277e-01
6.69140220e-01 8.91474545e-01 4.58716065e-01 4.35146779e-01
-5.13094626e-02 1.02728856e+00 2.35997289e-01 -1.82718616e-02
-3.36988181e-01 -1.28248572e-01 -6.66978508e-02 8.97241414e-01
-5.13610661e-01 -2.27601662e-01 -8.55112553e-01 9.74230170e-02
-1.52577603e+00 -8.52659166e-01 -1.35597110e-01 1.93723631e+00
5.82635283e-01 5.37131615e-02 -1.01553172e-01 1.06235057e-01
3.98495555e-01 2.10180342e-01 -2.55421042e-01 -6.38243973e-01
-7.63881147e-01 4.69215214e-01 3.69523078e-01 -4.34311368e-02
-1.52345312e+00 9.68815804e-01 6.87363815e+00 1.28338671e+00
-1.19314694e+00 -2.18976438e-01 8.43734264e-01 3.03794742e-01
-3.28155637e-01 -4.38817173e-01 -4.20976013e-01 3.08257133e-01
5.18282533e-01 1.72828421e-01 7.04384923e-01 9.26487386e-01
-5.08472562e-01 7.54695013e-02 -1.38804817e+00 1.32505608e+00
-4.62712124e-02 -1.39646447e+00 5.08910954e-01 1.50811195e-01
6.42300069e-01 -3.24086875e-01 4.81527030e-01 -1.42062113e-01
8.80859047e-02 -1.47359383e+00 1.10356796e+00 8.53401005e-01
1.30240357e+00 -2.63165802e-01 5.44699073e-01 -3.59188616e-01
-1.24453402e+00 2.65967488e-01 -7.89492071e-01 1.38024951e-03
-4.12251383e-01 4.87804264e-01 -1.99190378e-01 5.65552711e-01
8.73646677e-01 1.14358592e+00 -5.99154115e-01 6.63459241e-01
1.97320461e-01 4.77732480e-01 -2.25364357e-01 -2.97787488e-01
4.20707315e-01 -1.53553829e-01 3.03405106e-01 1.31219482e+00
3.97261620e-01 -3.12079024e-02 -6.68655559e-02 7.70983100e-01
1.41032776e-02 7.27286264e-02 -8.62885356e-01 -2.65619069e-01
1.77515432e-01 1.22879744e+00 -8.01161349e-01 -2.05482617e-01
-2.80366391e-01 8.59798014e-01 2.37171367e-01 2.38282606e-01
-3.99392009e-01 -1.99888647e-01 8.85651886e-01 2.13154256e-02
4.41042751e-01 -2.16993690e-01 -2.24948406e-01 -1.11583853e+00
1.83877051e-01 -1.07076430e+00 -7.49459788e-02 -9.32686746e-01
-1.93330860e+00 7.45011449e-01 -1.44831240e-01 -1.31060994e+00
1.82715178e-01 -1.39766884e+00 8.30331892e-02 6.61942959e-01
-1.02785957e+00 -1.53305876e+00 -5.98721206e-01 5.86646259e-01
4.41025704e-01 -3.66682082e-01 1.21764266e+00 1.98063791e-01
-9.57506746e-02 2.81016499e-01 4.14261878e-01 8.29328150e-02
5.88421166e-01 -1.22998917e+00 6.43041015e-01 -1.14699356e-01
4.66451794e-01 3.91889036e-01 5.98457217e-01 -2.32715473e-01
-2.10641170e+00 -7.80987918e-01 5.03901601e-01 -8.05425167e-01
9.22703385e-01 -8.08153272e-01 -9.84578431e-01 2.33729526e-01
-3.85049462e-01 1.91246003e-01 6.13198102e-01 3.21313083e-01
-9.91779387e-01 -2.17738926e-01 -7.88558125e-01 3.87247324e-01
1.30688322e+00 -8.54233146e-01 -3.59663665e-01 5.39879739e-01
7.49939606e-02 -3.45536023e-01 -1.19757104e+00 3.77555758e-01
1.19765556e+00 -8.81020069e-01 1.06864953e+00 -8.64859462e-01
7.11765587e-01 8.36990327e-02 -8.36815536e-01 -1.13716972e+00
-6.44980729e-01 -3.64020795e-01 2.67336935e-01 9.83017087e-01
1.35254622e-01 -5.49493372e-01 6.80370390e-01 2.85259217e-01
1.60137489e-02 -7.73036480e-01 -8.55724931e-01 -1.26766765e+00
4.74983230e-02 -4.46079195e-01 8.00857067e-01 9.43172574e-01
-2.82879770e-01 -3.46971631e-01 -5.56492448e-01 -4.75337774e-01
6.84646964e-01 2.23606065e-01 6.64084017e-01 -1.34501922e+00
-1.95297733e-01 -7.72479355e-01 -9.31842208e-01 -1.05583036e+00
2.28734896e-01 -9.45910871e-01 -1.16595246e-01 -9.16558862e-01
4.40415025e-01 -9.01261210e-01 -3.50386441e-01 6.64065540e-01
3.16476405e-01 5.80991566e-01 8.94384310e-02 2.84110039e-01
-5.98492265e-01 2.85531729e-01 1.54627633e+00 -5.10774434e-01
5.56123197e-01 -4.22795802e-01 -6.04298353e-01 6.22661173e-01
4.15698826e-01 -2.85342008e-01 -1.09479286e-01 -5.72909653e-01
2.99674869e-01 -1.93922579e-01 6.87094510e-01 -7.84778059e-01
-2.07515627e-01 -3.97503346e-01 4.00545269e-01 1.95204653e-02
4.87387717e-01 -8.95824134e-01 4.66313750e-01 -6.31394684e-02
-2.25517899e-01 -1.18803307e-01 2.22964898e-01 4.94696468e-01
-3.86716843e-01 1.72233552e-01 6.90717399e-01 1.55650586e-01
-1.13611972e+00 3.85152519e-01 -2.11079344e-01 2.76482124e-02
5.89895785e-01 -5.25542617e-01 -7.25421011e-01 -1.97114080e-01
-4.86652553e-01 -4.80433941e-01 7.09449947e-01 8.07593286e-01
6.18702650e-01 -1.76264310e+00 -5.86390376e-01 5.47859430e-01
6.05853856e-01 -5.56593776e-01 2.69860238e-01 5.88600218e-01
-8.34638596e-01 4.04336244e-01 -7.49226868e-01 -8.83050561e-01
-1.00834191e+00 3.76316577e-01 2.18857259e-01 1.10138826e-01
-7.55661666e-01 7.93451965e-01 7.17390954e-01 -8.13612342e-02
7.55637437e-02 -3.67693812e-01 7.44354278e-02 -5.37130833e-02
4.90204424e-01 1.28151312e-01 2.05590844e-01 -7.14221537e-01
-3.39488238e-01 1.13752222e+00 1.08567767e-01 3.30393791e-01
1.55864918e+00 1.97937563e-01 -4.18132275e-01 7.12891042e-01
1.37097931e+00 -1.08433895e-01 -1.06960094e+00 -3.22806418e-01
1.91640750e-01 -7.14709640e-01 -3.77239197e-01 -7.11765051e-01
-7.74818182e-01 8.35290432e-01 3.62866938e-01 5.48787653e-01
1.04918063e+00 2.36248612e-01 4.56255972e-01 8.05491924e-01
1.00538599e+00 -9.07174408e-01 3.19022387e-01 7.41071284e-01
1.19847572e+00 -1.16241777e+00 -1.82708979e-01 -6.63108587e-01
-7.51791835e-01 1.44358134e+00 -5.96611947e-02 -3.45553011e-01
8.42651308e-01 3.88405949e-01 -2.37239972e-02 -4.14777488e-01
-8.85795116e-01 8.59986618e-02 9.35198128e-01 5.56812525e-01
3.67950141e-01 4.37073559e-01 2.72028178e-01 4.89459813e-01
-3.44295084e-01 -3.01939487e-01 1.22589283e-01 6.42030776e-01
-4.36884046e-01 -1.17713773e+00 -2.57852674e-01 8.16643238e-01
-3.81058484e-01 -3.29783298e-02 -5.02981663e-01 8.68058085e-01
-5.37287220e-02 6.81148648e-01 2.28450865e-01 -7.03795016e-01
3.48213613e-01 -9.66778025e-02 9.17378902e-01 -2.39569291e-01
-2.00505316e-01 -1.88809484e-01 3.16838920e-01 -6.80518270e-01
-5.81382871e-01 -4.77342963e-01 -2.94438511e-01 -7.32257783e-01
7.76020214e-02 -1.33453548e-01 5.19581735e-01 9.43682015e-01
3.91069874e-02 1.16833262e-01 5.17757356e-01 -8.38949323e-01
-3.88746619e-01 -7.56621838e-01 -1.29109645e+00 1.07613111e+00
2.37451464e-01 -1.26036239e+00 -1.17883161e-01 3.66430938e-01] | [10.223735809326172, -0.16886340081691742] |
3ce3e435-2504-4b29-83c7-d61928e0ea10 | convolutional-neural-networks-trained-to | 2302.03992 | null | https://arxiv.org/abs/2302.03992v3 | https://arxiv.org/pdf/2302.03992v3.pdf | Convolutional Neural Networks Trained to Identify Words Provide a Surprisingly Good Account of Visual Form Priming Effects | A wide variety of orthographic coding schemes and models of visual word identification have been developed to account for masked priming data that provide a measure of orthographic similarity between letter strings. These models tend to include hand-coded orthographic representations with single unit coding for specific forms of knowledge (e.g., units coding for a letter in a given position). Here we assess how well a range of these coding schemes and models account for the pattern of form priming effects taken from the Form Priming Project and compare these findings to results observed with 11 standard deep neural network models (DNNs) developed in computer science. We find that deep convolutional networks (CNNs) perform as well or better than the coding schemes and word recognition models, whereas transformer networks did less well. The success of CNNs is remarkable as their architectures were not developed to support word recognition (they were designed to perform well on object recognition), they classify pixel images of words (rather than artificial encodings of letter strings), and their training was highly simplified (not respecting many key aspects of human experience). In addition to these form priming effects, we find that the DNNs can account for visual similarity effects on priming that are beyond all current psychological models of priming. The findings add to the recent work of (Hannagan et al., 2021) and suggest that CNNs should be given more attention in psychology as models of human visual word recognition. | ['Jeffrey Bowers', 'Valerio Biscione', 'Dong Yin'] | 2023-02-08 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 1.08270928e-01 -1.46193087e-01 8.14764053e-02 -1.73780918e-01
2.81679958e-01 -8.75103652e-01 9.98590112e-01 3.57804239e-01
-7.99406409e-01 5.87638505e-02 5.98896384e-01 -8.14980328e-01
1.83314726e-01 -1.01647031e+00 -5.44441581e-01 -4.06113625e-01
1.41195476e-01 1.42453149e-01 2.83836275e-01 -4.59995002e-01
6.91545963e-01 8.79540324e-01 -1.69699669e+00 5.95941305e-01
2.42676049e-01 5.10505259e-01 2.34957963e-01 4.18212235e-01
-1.43579841e-01 4.78770107e-01 -4.71534193e-01 -3.50130886e-01
3.62211704e-01 -5.81598043e-01 -6.85386062e-01 -8.59832913e-02
7.74749637e-01 -3.26275766e-01 -6.67593837e-01 9.88324702e-01
5.62494457e-01 9.39133167e-02 6.03358150e-01 -5.31130016e-01
-1.64517403e+00 3.46044928e-01 -1.56922191e-01 9.42622125e-01
2.87466884e-01 3.77766848e-01 8.77426803e-01 -1.14813972e+00
2.55676210e-01 1.55518007e+00 6.13555610e-01 3.23987931e-01
-1.22953939e+00 -5.66530168e-01 1.55713305e-01 4.38092425e-02
-1.23925650e+00 -3.42450500e-01 8.65276158e-02 -5.80771089e-01
1.56418312e+00 -2.07301732e-02 1.08096755e+00 8.24603140e-01
3.36922407e-01 1.75322387e-02 1.65527749e+00 -8.57172430e-01
9.04438496e-02 5.65276332e-02 3.66423041e-01 3.97781253e-01
7.69212067e-01 9.14950669e-01 -6.88473940e-01 6.21071905e-02
1.44787478e+00 -8.94132778e-02 -6.96562305e-02 1.21852040e-01
-1.25330508e+00 7.82915592e-01 6.57706082e-01 6.61691010e-01
-4.31697965e-01 -1.59509783e-03 3.24368685e-01 2.96464682e-01
5.89905381e-02 7.32080460e-01 -1.40509099e-01 2.38165736e-01
-6.57124341e-01 1.60800740e-01 2.80396760e-01 4.80732322e-01
6.59608960e-01 4.01631624e-01 -3.86691988e-01 6.77321851e-01
3.21055710e-01 3.77667427e-01 9.90288317e-01 -7.35704303e-01
8.15807749e-03 3.12425226e-01 8.21391195e-02 -1.21887147e+00
-5.82033813e-01 -3.74096893e-02 -1.03046946e-01 5.09210885e-01
7.34361172e-01 3.06094766e-01 -1.21160364e+00 1.92036271e+00
-3.64928365e-01 -3.50085169e-01 -1.84868827e-01 1.06521964e+00
7.16739416e-01 5.01247048e-01 6.39527321e-01 2.36261375e-02
1.91474271e+00 -5.23917317e-01 -4.18395013e-01 -9.80339766e-01
6.83435559e-01 -8.92702997e-01 1.23441410e+00 1.90474346e-01
-1.22770977e+00 -1.00771391e+00 -1.20913148e+00 -3.16877067e-01
-8.88700306e-01 -2.58470625e-01 9.60340321e-01 1.15002310e+00
-1.73139036e+00 5.34037232e-01 -2.95855850e-01 -7.50694036e-01
2.21797392e-01 2.68456966e-01 -3.79575551e-01 -1.14567272e-01
-1.38731265e+00 1.54731202e+00 6.56132936e-01 1.03181250e-01
-6.30730748e-01 1.68863474e-03 -7.83634067e-01 1.29231349e-01
-2.95822024e-01 -5.74241877e-01 1.12069070e+00 -1.55272126e+00
-9.96722519e-01 1.46593213e+00 -2.30603412e-01 -2.01499656e-01
-5.32154024e-01 1.31038785e-01 -7.21373260e-01 3.37188840e-01
5.11905923e-02 9.96374369e-01 5.23961425e-01 -9.77593184e-01
-1.25942379e-01 -3.73372614e-01 2.33335912e-01 2.78460056e-01
-3.22476983e-01 5.19548714e-01 2.11377129e-01 -7.71581590e-01
3.51096451e-01 -7.12741435e-01 1.16343433e-02 4.18685451e-02
4.43977088e-01 -1.28572822e-01 4.44261506e-02 -6.37731969e-01
8.90115321e-01 -2.26893926e+00 -4.92104411e-01 2.37317737e-02
1.74417779e-01 6.39105201e-01 -4.54001129e-01 6.83336496e-01
-6.69292688e-01 3.77107650e-01 9.87143442e-02 3.01069438e-01
9.30703580e-02 3.25514525e-01 -4.86356020e-01 5.81965387e-01
3.62101734e-01 1.19682658e+00 -7.53339171e-01 -1.18430987e-01
8.35754573e-02 4.05274749e-01 -4.87968445e-01 -2.45668545e-01
1.15781732e-01 -2.09865063e-01 4.86108124e-01 2.27295026e-01
7.85323799e-01 -2.02339754e-01 4.15096641e-01 1.34111151e-01
-5.21712899e-01 7.35029638e-01 -8.66951048e-01 1.15560699e+00
-1.55708909e-01 9.32007670e-01 -3.05958390e-01 -8.54549885e-01
7.68835068e-01 4.76453334e-01 -5.55467248e-01 -1.24993587e+00
3.11491519e-01 1.12494923e-01 8.15242589e-01 -2.16126114e-01
6.12619042e-01 -7.01592028e-01 3.83489072e-01 6.01338089e-01
-2.68859733e-02 -5.93390390e-02 9.55864787e-02 1.74406126e-01
6.74291551e-01 -1.82579085e-01 3.79514217e-01 -5.15179038e-01
5.88759966e-02 1.02395620e-02 1.66044474e-01 8.10628176e-01
-3.22176337e-01 7.55132616e-01 2.42521569e-01 -5.87861598e-01
-1.09706056e+00 -1.05121458e+00 -2.29660794e-01 1.29791117e+00
-3.43392752e-02 -2.18066052e-01 -5.24545014e-01 1.56911954e-01
-2.00238183e-01 8.27858746e-01 -8.94282699e-01 -5.12344658e-01
-4.74811763e-01 -6.98421001e-01 5.69709480e-01 1.03217745e+00
-1.42540345e-02 -1.56918204e+00 -9.14737642e-01 1.56074077e-01
4.25139278e-01 -9.46831703e-01 -1.85435802e-01 4.37876046e-01
-1.00131524e+00 -8.10871124e-01 -4.93076175e-01 -1.27351880e+00
7.74898350e-01 8.58013630e-01 1.23634887e+00 3.54389578e-01
-3.65600109e-01 6.00613534e-01 -3.07145774e-01 -3.25239807e-01
-4.19793367e-01 -6.47231996e-01 5.14664724e-02 -4.64256167e-01
9.85853851e-01 -4.09839511e-01 -6.40612781e-01 2.64231443e-01
-1.21231639e+00 -2.93210179e-01 8.74497712e-01 7.14117348e-01
2.31690496e-01 -3.14826161e-01 2.91087210e-01 -4.27469611e-01
8.12000155e-01 -3.24588507e-01 -4.22848254e-01 -1.50168881e-01
-6.59440637e-01 -2.11261973e-01 3.31853271e-01 -5.42892277e-01
-8.77165258e-01 -3.04121703e-01 7.28021413e-02 1.51231706e-01
-5.02849817e-01 4.34972107e-01 3.12397331e-01 -3.68016690e-01
1.13350296e+00 3.37102354e-01 -7.19041526e-02 -2.48246133e-01
4.62382495e-01 2.34839603e-01 4.32207823e-01 -4.90244001e-01
5.86614132e-01 4.85641956e-01 -1.91005036e-01 -7.25529492e-01
-4.43455279e-01 -1.60724789e-01 -7.30105221e-01 1.16757043e-01
1.09846663e+00 -1.27473676e+00 -4.40482825e-01 3.02238822e-01
-1.24673843e+00 -3.07738304e-01 -2.86944270e-01 6.67093456e-01
-3.61803353e-01 5.28357685e-01 -9.38476861e-01 -7.06546545e-01
1.76664189e-01 -6.71899855e-01 6.63471282e-01 3.31710756e-01
-5.06195605e-01 -1.01206064e+00 8.68066102e-02 -1.96955696e-01
4.48192239e-01 -2.96400994e-01 1.17333126e+00 -6.86025023e-01
-3.11356187e-01 -2.03338027e-01 -5.75469911e-01 2.32799351e-01
-6.67842478e-03 -1.91069052e-01 -1.13774192e+00 -9.00282785e-02
2.82717556e-01 -3.48619550e-01 1.03384912e+00 2.88219720e-01
5.95814288e-01 -7.55414516e-02 -8.21628571e-02 2.29127184e-01
1.45095360e+00 4.60059226e-01 9.94474232e-01 3.63785028e-01
3.54495496e-01 7.54110456e-01 1.27061337e-01 -1.68580301e-02
1.37100592e-01 3.95088732e-01 1.91336691e-01 -3.75184864e-02
-4.20506567e-01 -3.17095309e-01 4.51711982e-01 5.58656275e-01
1.04435571e-01 -7.03754127e-02 -1.08134377e+00 9.14434373e-01
-1.14629102e+00 -1.13773298e+00 -4.95731831e-01 2.38077211e+00
7.07832515e-01 3.30642134e-01 1.90834358e-01 1.90919619e-02
8.49340975e-01 2.50971079e-01 -9.43765417e-02 -1.07346189e+00
-5.36592126e-01 7.27590501e-01 5.26712179e-01 4.81715947e-01
-6.51973069e-01 1.19401836e+00 8.21798706e+00 2.10131839e-01
-1.07786632e+00 1.86981991e-01 4.01716918e-01 1.51038229e-01
-2.89406091e-01 1.87789768e-01 -5.10192454e-01 1.84430450e-01
1.34313524e+00 -2.19228654e-03 5.77634752e-01 3.95413935e-01
3.08217481e-02 -3.44685495e-01 -9.77962732e-01 7.89023638e-01
1.70122534e-01 -1.07076907e+00 1.68157235e-01 5.80661707e-02
4.56527650e-01 -1.43891145e-02 3.66656721e-01 -2.25826930e-02
3.78650934e-01 -1.47302544e+00 9.95761216e-01 -1.97266005e-02
6.38248920e-01 -7.94799030e-02 2.35116065e-01 1.03213144e-02
-8.03305984e-01 -1.18127316e-01 -1.03967345e+00 -1.25314391e+00
2.32793298e-02 1.09971605e-01 -4.91018414e-01 -3.57826799e-01
5.50341070e-01 2.02524170e-01 -9.48605657e-01 1.02767134e+00
-4.67564732e-01 4.91771162e-01 -3.02168839e-02 -1.03369661e-01
3.43914986e-01 2.03488037e-01 -1.23449482e-01 1.36438990e+00
1.45503074e-01 3.38205218e-01 -3.34283292e-01 7.93305993e-01
4.20902312e-01 2.58625984e-01 -7.13401914e-01 -3.54948789e-01
7.02336669e-01 1.10940623e+00 -1.25328207e+00 -4.28179026e-01
-8.34747374e-01 8.85796964e-01 3.69356394e-01 3.50939274e-01
-4.08080935e-01 -1.52788430e-01 6.54566646e-01 3.46361279e-01
4.89304841e-01 -5.59534967e-01 -7.73550272e-01 -6.60725594e-01
-2.08229065e-01 -7.74527311e-01 1.14247598e-01 -1.16226363e+00
-1.26466835e+00 5.08147895e-01 -1.43268615e-01 -7.52614617e-01
9.69152600e-02 -1.44736028e+00 -7.27515817e-01 1.45613241e+00
-1.18937635e+00 -7.14128017e-01 -1.23220272e-02 5.51967859e-01
1.57046970e-03 1.78574190e-01 1.06576979e+00 -3.17327864e-02
-3.19841653e-02 3.14449936e-01 -2.59870499e-01 1.63014606e-01
9.24143851e-01 -1.09311724e+00 1.06405330e+00 9.39380050e-01
5.11453569e-01 1.30740106e+00 4.73186284e-01 -6.04964852e-01
-1.03293657e+00 -3.10880899e-01 1.23930860e+00 -7.86662281e-01
5.15346408e-01 -4.05226380e-01 -9.59875226e-01 6.83636367e-01
6.67595148e-01 -1.07121974e-01 7.30664074e-01 -8.14526230e-02
-1.01983011e+00 4.35101748e-01 -8.24307263e-01 6.47870421e-01
1.30709422e+00 -8.74128163e-01 -1.19995224e+00 4.45024520e-01
5.12498617e-01 1.00839725e-02 -2.64052540e-01 -2.73204803e-01
4.57908005e-01 -1.31976676e+00 1.21908617e+00 -9.13702011e-01
4.76109743e-01 -1.81101322e-01 -1.16938509e-01 -1.34246647e+00
-1.28006065e+00 1.26128733e-01 5.10127544e-01 7.65966117e-01
3.52843821e-01 -9.65545237e-01 3.31009030e-01 6.66378915e-01
-1.95026398e-01 8.31411332e-02 -7.46881366e-01 -9.23393965e-01
5.63256145e-01 -1.33974969e-01 2.76469409e-01 1.00480831e+00
1.77036837e-01 3.82964253e-01 1.48498744e-01 2.03718841e-01
-2.79063731e-03 -3.07771802e-01 -4.07332927e-02 -1.28850091e+00
-3.30836326e-01 -8.53848934e-01 -5.44286430e-01 -7.00946569e-01
4.29524146e-02 -1.00415611e+00 -1.92710355e-01 -1.54253757e+00
2.48377286e-02 7.44645894e-02 -3.91172767e-01 6.66224837e-01
-3.02033782e-01 6.18218124e-01 4.71612602e-01 3.21756452e-01
2.08047315e-01 -1.36282176e-01 9.63055134e-01 3.09525579e-01
1.54144734e-01 -7.50368297e-01 -1.49532020e+00 5.59681177e-01
8.24360192e-01 -2.02814713e-01 -2.39051923e-01 -8.14582407e-01
5.28559148e-01 -3.78616452e-01 7.78699875e-01 -9.95959640e-01
2.03724895e-02 -6.90735597e-03 1.01887321e+00 1.04478225e-01
1.65949181e-01 -6.24540925e-01 -3.88480984e-02 7.03069627e-01
-9.59093571e-02 7.41835475e-01 8.60984385e-01 1.89170256e-01
-1.00855254e-01 -4.54811215e-01 8.80028069e-01 -6.02699757e-01
-1.11348128e+00 -3.80378455e-01 -9.97839808e-01 2.18545739e-02
6.17975533e-01 -7.01472580e-01 -8.37729573e-01 -2.19672218e-01
-6.52814567e-01 -5.56768775e-01 7.71706998e-01 5.56157231e-01
4.46422726e-01 -1.47355103e+00 -4.08255041e-01 2.43048057e-01
7.37011433e-02 -9.08313692e-01 -2.07693912e-02 4.94884849e-01
-6.83874905e-01 7.88955986e-01 -7.48255432e-01 1.24695167e-01
-6.05261207e-01 1.26130199e+00 1.79012671e-01 4.20995384e-01
-2.95803875e-01 1.05693293e+00 7.92866230e-01 2.59231012e-02
-4.22459394e-01 -2.37871751e-01 -1.97030395e-01 3.28130901e-01
8.82642031e-01 1.83937456e-02 -2.61139143e-02 -8.67992640e-01
-5.35846949e-01 5.85481763e-01 -1.07172571e-01 -3.95490855e-01
1.02933323e+00 1.24648005e-01 -3.39856654e-01 4.65865552e-01
8.08473468e-01 2.55107395e-02 -8.87490451e-01 -1.47466078e-01
-1.60848349e-01 -5.22725582e-01 -1.50531024e-01 -1.02932632e+00
-6.30599320e-01 1.27230561e+00 7.55231380e-01 2.03067198e-01
8.71588171e-01 -2.07212269e-02 1.48574680e-01 4.18678634e-02
3.08099780e-02 -1.07831979e+00 1.96065173e-01 6.72644258e-01
7.85612106e-01 -8.45704377e-01 -5.62219834e-03 -2.26311177e-01
-5.50601482e-01 1.12240207e+00 7.22691655e-01 -4.19990003e-01
3.10041547e-01 -8.95455182e-02 2.21815869e-01 -4.71982986e-01
-4.81722385e-01 -6.22056842e-01 1.29051134e-01 8.77773345e-01
1.00061214e+00 7.35227298e-03 -7.99229145e-01 4.35343593e-01
-3.33338499e-01 -2.38735512e-01 5.48769832e-01 1.01121223e+00
-9.25019145e-01 -1.02396023e+00 -8.93662989e-01 2.82822818e-01
-2.81078279e-01 -8.19967806e-01 -6.96999967e-01 6.95013046e-01
2.21820459e-01 8.63005519e-01 4.74280059e-01 -1.74387202e-01
2.61656761e-01 5.93693852e-01 6.85271382e-01 -1.00439763e+00
-8.83516312e-01 -1.36140615e-01 -8.27371627e-02 -2.61576653e-01
-3.09262514e-01 -7.48536408e-01 -1.00650072e+00 -2.66979963e-01
-3.09593324e-02 -1.63040444e-01 3.06706548e-01 7.24315345e-01
2.55573750e-01 2.43087634e-01 -9.84141827e-02 -9.19183791e-01
-2.37137109e-01 -8.84345055e-01 -7.64205277e-01 4.02763993e-01
-1.92727178e-01 -7.36522913e-01 -2.88024485e-01 2.87485551e-02] | [10.336335182189941, 2.2362051010131836] |
1a23d03c-3840-4c6e-a872-701e46bad63e | hierarchical-reinforcement-learning-of | 2203.10616 | null | https://arxiv.org/abs/2203.10616v1 | https://arxiv.org/pdf/2203.10616v1.pdf | Hierarchical Reinforcement Learning of Locomotion Policies in Response to Approaching Objects: A Preliminary Study | Animals such as rabbits and birds can instantly generate locomotion behavior in reaction to a dynamic, approaching object, such as a person or a rock, despite having possibly never seen the object before and having limited perception of the object's properties. Recently, deep reinforcement learning has enabled complex kinematic systems such as humanoid robots to successfully move from point A to point B. Inspired by the observation of the innate reactive behavior of animals in nature, we hope to extend this progress in robot locomotion to settings where external, dynamic objects are involved whose properties are partially observable to the robot. As a first step toward this goal, we build a simulation environment in MuJoCo where a legged robot must avoid getting hit by a ball moving toward it. We explore whether prior locomotion experiences that animals typically possess benefit the learning of a reactive control policy under a proposed hierarchical reinforcement learning framework. Preliminary results support the claim that the learning becomes more efficient using this hierarchical reinforcement learning method, even when partial observability (radius-based object visibility) is taken into account. | ['George Konidaris', 'Kaiyu Zheng', 'Sreehari Rammohan', 'Shangqun Yu'] | 2022-03-20 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-6.87722266e-02 2.74191916e-01 1.83287695e-01 8.07044357e-02
2.77275473e-01 -4.57474411e-01 4.38656956e-01 -3.17914896e-02
-8.59299481e-01 1.08907998e+00 -3.47083867e-01 7.30432123e-02
-1.45529360e-01 -8.22015047e-01 -8.21899712e-01 -8.54794919e-01
-7.46971548e-01 4.75975752e-01 4.65222090e-01 -5.93140960e-01
3.70985180e-01 8.52260828e-01 -1.85839319e+00 -1.94325730e-01
5.82216859e-01 3.55663776e-01 7.17676044e-01 9.24552023e-01
7.58544505e-01 8.86155486e-01 -2.55110294e-01 2.96261549e-01
3.95647198e-01 -4.41240281e-01 -8.63603652e-01 3.43701243e-01
-3.15484196e-01 -3.85060042e-01 -1.72512159e-01 7.37666249e-01
2.23270595e-01 6.70570910e-01 7.79591441e-01 -1.27561092e+00
-3.33356589e-01 4.97026592e-01 -3.62895191e-01 1.86104268e-01
4.68596905e-01 8.05769265e-01 8.07993233e-01 -3.18923563e-01
7.06885517e-01 1.16395044e+00 4.16959643e-01 5.68158746e-01
-1.15411353e+00 -1.04292795e-01 1.41869396e-01 3.13643396e-01
-1.10764015e+00 -1.62826017e-01 3.91835153e-01 -3.48530710e-01
1.28831613e+00 -1.83826953e-01 9.93801892e-01 1.02837276e+00
6.41506374e-01 5.77592373e-01 1.08086574e+00 -3.34877014e-01
5.60834348e-01 -2.09662005e-01 -6.13026857e-01 6.60505295e-01
4.77187127e-01 4.76275951e-01 -2.30812550e-01 -1.20843705e-02
6.46570385e-01 -4.57737774e-01 -9.36140716e-02 -9.15844560e-01
-1.29415107e+00 5.84515810e-01 7.48552620e-01 2.67047733e-01
-7.61874378e-01 4.62479740e-01 2.74992436e-02 4.13141161e-01
-5.71701407e-01 8.89973104e-01 -3.48909855e-01 -3.14463288e-01
-2.52104729e-01 5.44169068e-01 9.00664687e-01 7.90124655e-01
5.64530134e-01 3.94295126e-01 6.24363422e-01 2.77648866e-01
2.52661556e-01 3.63263071e-01 2.54490018e-01 -1.37666047e+00
2.57711977e-01 4.21970248e-01 6.60530865e-01 -8.40457141e-01
-7.96419322e-01 -9.71290842e-02 -2.45361075e-01 1.08635938e+00
4.30149287e-01 -3.94685328e-01 -5.56961954e-01 1.82194304e+00
6.97963893e-01 -2.80018538e-01 4.44862813e-01 1.23643517e+00
5.39013790e-03 6.43928587e-01 6.42936528e-02 -2.42114607e-02
1.09233677e+00 -5.72543979e-01 -1.76613763e-01 -3.50593269e-01
6.98852241e-01 -1.89125478e-01 9.64970887e-01 6.05142355e-01
-1.06673801e+00 -4.48254466e-01 -1.26795042e+00 1.76623240e-01
-4.35140043e-01 -3.52177054e-01 5.52199423e-01 3.12974900e-01
-8.97308230e-01 8.16769719e-01 -1.29202116e+00 -1.02610898e+00
1.81377053e-01 4.44872171e-01 -3.90262395e-01 3.31192076e-01
-1.04253507e+00 1.42161226e+00 5.81653774e-01 2.13094071e-01
-1.34329689e+00 -2.90247221e-02 -6.31771982e-01 6.28556833e-02
3.65366906e-01 -8.21299434e-01 1.44354975e+00 -8.56253505e-01
-1.79173195e+00 6.91955626e-01 5.16050994e-01 -5.53225040e-01
6.89209580e-01 -2.86339194e-01 1.18233927e-01 4.36524808e-01
2.07129821e-01 1.03616822e+00 7.32993722e-01 -1.36665881e+00
-6.25557303e-01 -2.76742756e-01 5.25327563e-01 7.57704020e-01
2.61529565e-01 -5.09381056e-01 5.43848991e-01 -2.06020355e-01
-7.30999410e-02 -1.31555414e+00 -5.83804965e-01 1.47137970e-01
1.23921204e-02 -1.58473834e-01 7.88727403e-01 1.78102434e-01
2.38555074e-01 -1.75807106e+00 5.18493414e-01 1.21157675e-03
-3.07778776e-01 -7.32826069e-04 -1.27979726e-01 9.38241005e-01
2.41766885e-01 -1.84289590e-01 -2.61319220e-01 4.14774984e-01
-7.13932514e-02 5.27790308e-01 8.07734486e-03 6.34606600e-01
1.40026286e-01 6.51740432e-01 -1.14539170e+00 -3.36625785e-01
-1.33545492e-02 2.72049993e-01 -7.99909949e-01 2.06629962e-01
-1.81693956e-01 6.84472680e-01 -8.26171458e-01 4.89605755e-01
1.87499985e-01 -4.04435284e-02 2.53924668e-01 4.61641997e-01
-5.26768029e-01 2.72408593e-02 -1.20230591e+00 1.19007218e+00
-4.69570667e-01 3.91899168e-01 2.70705402e-01 -1.10632014e+00
5.83659232e-01 1.69922128e-01 5.04141212e-01 -5.95982909e-01
2.01743349e-01 1.41220197e-01 5.98808527e-01 -8.89283419e-01
5.05181313e-01 -3.70851994e-01 5.79285435e-02 3.75313967e-01
7.13498471e-03 -6.54303133e-01 3.47059995e-01 -8.96897987e-02
1.32196045e+00 7.02805758e-01 5.13636351e-01 -4.55935985e-01
3.58310580e-01 4.41239178e-01 3.19059908e-01 9.78929162e-01
-3.27796996e-01 1.09449640e-01 2.45313331e-01 -2.01106727e-01
-1.09494483e+00 -1.21013224e+00 2.62211487e-02 1.09534121e+00
7.03643382e-01 3.40369195e-01 -6.02034748e-01 -9.37151834e-02
9.53062847e-02 7.34107673e-01 -7.46333003e-01 -2.89406598e-01
-9.13136661e-01 -5.74207366e-01 2.00965807e-01 4.93800730e-01
4.40913260e-01 -1.74047983e+00 -1.90683067e+00 5.97494483e-01
1.25745043e-01 -8.24164808e-01 3.63582224e-01 7.17423141e-01
-9.05991316e-01 -9.77822423e-01 -5.08360863e-01 -6.70300603e-01
6.46043599e-01 1.59351692e-01 6.30188525e-01 2.73214251e-01
-4.22104478e-01 7.61040032e-01 -6.08611226e-01 -1.13649532e-01
-2.80766398e-01 -1.30716830e-01 4.97092545e-01 -5.70716977e-01
-1.79409906e-01 -6.17645621e-01 -7.45742440e-01 5.10915101e-01
-7.77407229e-01 -8.70007500e-02 4.89182830e-01 6.62314475e-01
9.95168537e-02 3.94683838e-01 3.90144408e-01 -1.15464620e-01
4.22083527e-01 -5.87945223e-01 -5.96186161e-01 -2.57662028e-01
4.60987650e-02 2.14372855e-02 7.08240986e-01 -7.07067072e-01
-1.04959846e+00 2.26539925e-01 1.31918639e-01 3.32126439e-01
-3.36062998e-01 4.12267983e-01 3.00186090e-02 -9.94690284e-02
8.43078554e-01 1.03698768e-01 -1.14934087e-01 8.91776159e-02
2.66018271e-01 1.48347273e-01 2.71334797e-01 -7.87589729e-01
8.77990603e-01 7.05462039e-01 4.25520360e-01 -1.14423645e+00
-2.91637424e-02 -1.43439695e-01 -7.82523870e-01 -4.41375732e-01
8.14633906e-01 -6.13532484e-01 -1.44076073e+00 4.04493570e-01
-8.34450066e-01 -8.82615745e-01 -2.74149626e-01 6.75255835e-01
-1.41354227e+00 1.20450698e-01 -3.83289069e-01 -8.52148890e-01
5.52708924e-01 -1.09949112e+00 7.27725923e-01 4.90834713e-01
-2.93056726e-01 -6.72477484e-01 8.66513103e-02 1.99963655e-02
3.40919197e-01 2.58917749e-01 6.75665855e-01 -2.42609710e-01
-6.80914044e-01 1.28729880e-01 3.60137731e-01 -1.68252259e-01
1.99732140e-01 1.58571914e-01 -4.72559094e-01 -6.59777999e-01
1.07374407e-01 -7.65759706e-01 3.47347409e-01 2.56074995e-01
4.40817654e-01 -2.28842750e-01 -4.27850783e-01 1.10036679e-01
1.37990665e+00 4.76838857e-01 4.13921088e-01 8.22388113e-01
2.88396567e-01 1.03623152e+00 1.03771138e+00 5.24700403e-01
3.32076520e-01 6.23048067e-01 9.54853594e-01 2.21273676e-01
2.29193732e-01 -9.95643362e-02 3.63240391e-01 2.68850625e-01
-3.69951516e-01 -5.54566979e-01 -7.58089840e-01 5.30382574e-01
-1.72323346e+00 -1.05442750e+00 4.37051952e-02 1.95690799e+00
3.08477044e-01 2.17957154e-01 2.31288761e-01 3.90071608e-02
4.17922109e-01 -2.76379198e-01 -9.31326866e-01 -6.45385623e-01
1.97331667e-01 -4.37798649e-01 5.69681525e-01 4.52926695e-01
-6.53349161e-01 1.07911491e+00 6.17752981e+00 3.15944478e-02
-1.17360628e+00 -4.69701916e-01 -8.07094947e-02 4.91716713e-02
3.17800254e-01 6.31635338e-02 -5.16668081e-01 2.97106504e-01
7.47383654e-01 -1.62753463e-01 5.89610279e-01 9.76778984e-01
4.92464423e-01 -9.51064706e-01 -1.25510347e+00 1.50366738e-01
-3.44478726e-01 -6.18340492e-01 -2.99653113e-01 1.73033908e-01
4.33840394e-01 -1.10834070e-01 1.51752561e-01 4.34602886e-01
7.20813155e-01 -8.79642665e-01 7.87253380e-01 3.92569721e-01
-1.22535415e-01 -7.36987770e-01 3.70254219e-01 8.70751500e-01
-9.78634059e-01 -3.88041168e-01 -4.86176312e-01 -6.51890337e-01
3.09524179e-01 -3.58321488e-01 -1.26912308e+00 4.56428081e-02
7.73627281e-01 2.39141792e-01 -2.33184636e-01 1.21030223e+00
-3.57624024e-01 3.49821478e-01 -5.93501806e-01 -6.58001542e-01
6.62129998e-01 -3.26340675e-01 7.77774751e-01 5.23698449e-01
2.71802843e-01 5.00563800e-01 5.42009622e-02 7.46852160e-01
5.49769223e-01 -1.17163338e-01 -8.79731894e-01 1.73105091e-01
-3.60355601e-02 1.08265889e+00 -1.19089532e+00 -1.50095180e-01
-1.22115292e-01 8.32303882e-01 5.15242755e-01 4.02968019e-01
-7.93486238e-01 -2.55872548e-01 5.47420919e-01 6.87688217e-02
6.16178393e-01 -4.82205570e-01 6.44586161e-02 -8.73105705e-01
-1.92580789e-01 -6.48134768e-01 -3.62870954e-02 -1.17559803e+00
-7.96494782e-01 3.62318635e-01 1.00703038e-01 -1.45454657e+00
-4.28301632e-01 -5.09550095e-01 -3.25213522e-01 2.93279678e-01
-9.99033630e-01 -7.90552557e-01 -1.36690125e-01 3.24879408e-01
4.80971843e-01 3.20372954e-02 4.38387543e-01 -4.59577173e-01
8.87050256e-02 -1.88556835e-01 -2.26255402e-01 -3.75481099e-01
2.50105917e-01 -1.27549088e+00 4.60062213e-02 7.06966579e-01
-1.92953020e-01 4.88492101e-01 1.34017026e+00 -8.13308716e-01
-1.42667329e+00 -5.00765324e-01 1.04004659e-01 -3.49233210e-01
8.99459541e-01 -3.67909074e-01 -9.09757197e-01 6.33111775e-01
1.49803013e-01 -4.61869836e-02 9.57553163e-02 -2.56618023e-01
3.23379457e-01 2.18351454e-01 -1.31112671e+00 1.13114738e+00
9.79694068e-01 1.04286991e-01 -9.58194494e-01 6.24423921e-02
3.19869131e-01 -2.99184591e-01 -5.57261229e-01 6.27702832e-01
7.89987564e-01 -9.50473964e-01 8.04280460e-01 -9.12716866e-01
2.52576053e-01 -5.56856453e-01 -5.33394143e-02 -1.50767171e+00
-4.35047030e-01 -5.10827065e-01 1.67901084e-01 4.84149337e-01
1.74741149e-01 -4.77070332e-01 8.28175128e-01 1.71076711e-02
-6.30868971e-02 -4.22495306e-01 -1.17521000e+00 -1.01623869e+00
4.07278627e-01 6.92154989e-02 5.69901057e-02 5.02478659e-01
3.65811527e-01 4.50803563e-02 -4.81101692e-01 6.40552342e-01
5.31125247e-01 -2.73153675e-03 9.13020134e-01 -9.91299093e-01
-2.76883632e-01 -1.91448241e-01 -5.65304637e-01 -1.04221094e+00
1.78591654e-01 -4.77403343e-01 6.69840336e-01 -1.50888789e+00
-2.64287829e-01 -3.85989159e-01 2.85736602e-02 4.19315964e-01
1.04023278e-01 -2.63789203e-02 1.78699717e-01 -3.62787419e-03
-6.06577277e-01 6.05525076e-01 1.64823830e+00 2.49606103e-01
-4.02388752e-01 2.01382950e-01 -2.42179900e-01 9.50666726e-01
9.60685372e-01 -5.98460853e-01 -4.51546609e-01 -1.29236430e-01
3.75727117e-01 5.40082276e-01 4.42132413e-01 -1.26878834e+00
3.90616417e-01 -5.50685942e-01 2.43684992e-01 -2.66396046e-01
5.11961877e-01 -1.16912889e+00 1.29238605e-01 1.19098651e+00
-3.44516814e-01 1.07265428e-01 1.74476847e-01 6.68417513e-01
2.06257299e-01 -3.27853054e-01 1.02100599e+00 -3.57488215e-01
-8.13454211e-01 -1.59257695e-01 -1.45693648e+00 -7.97712654e-02
1.40487432e+00 -5.70332706e-01 -2.19705179e-01 -2.49066785e-01
-1.03117561e+00 4.93912846e-01 7.42434561e-01 3.26475829e-01
5.36766827e-01 -8.98157418e-01 -3.25006098e-01 -4.28717211e-02
4.55525182e-02 -2.70562738e-01 2.46309623e-01 8.17198277e-01
-1.03007650e+00 1.96321070e-01 -1.07358253e+00 -5.15205801e-01
-1.09792757e+00 7.91451573e-01 4.98266071e-01 2.48945266e-01
-7.58320093e-01 5.44854820e-01 2.69388348e-01 -4.07250285e-01
-2.18010634e-01 -4.14651841e-01 -2.13164955e-01 -1.72129095e-01
-3.64093333e-02 4.82581943e-01 -2.49586925e-01 -6.42540395e-01
-3.43739986e-01 5.70966542e-01 1.33882001e-01 -2.95493394e-01
1.41285777e+00 -2.93792278e-01 3.22904706e-01 5.41594028e-01
5.85700035e-01 -1.73124731e-01 -1.67797494e+00 3.32512856e-01
5.27363829e-02 -2.91559190e-01 -4.98657852e-01 -4.88954335e-01
-4.64742184e-01 8.11247349e-01 3.79693687e-01 4.26768929e-01
7.60118783e-01 -9.08546075e-02 3.26206177e-01 1.09545016e+00
1.05100548e+00 -1.37680268e+00 6.50654674e-01 5.66006541e-01
1.02099431e+00 -1.15338278e+00 1.06364280e-01 -2.07797922e-02
-8.17621946e-01 9.76010501e-01 9.10989881e-01 -6.01603329e-01
1.41966254e-01 3.48139375e-01 2.43650433e-02 2.54099201e-02
-1.02830851e+00 -5.16862929e-01 -3.14760953e-01 9.65275526e-01
-1.51806682e-01 -2.71273702e-02 -5.20492867e-02 -4.40234184e-01
-4.22244519e-01 -3.17820132e-01 1.14403284e+00 1.43546450e+00
-1.09319806e+00 -7.20375657e-01 -3.74361306e-01 -1.02829926e-01
-2.48695880e-01 5.04606068e-01 -1.09491110e-01 1.38685012e+00
5.35115115e-02 8.58333647e-01 4.88209762e-02 2.27450609e-01
3.13741624e-01 -4.32405502e-01 7.60982573e-01 -8.36122334e-01
-2.63367295e-01 -1.59770742e-01 -1.66933626e-01 -5.63042939e-01
-6.21815264e-01 -9.66104269e-01 -1.71720469e+00 1.46742985e-01
1.11950651e-01 1.44123793e-01 4.71963286e-01 8.32841873e-01
-3.03536266e-01 3.82475197e-01 6.27645016e-01 -1.20497549e+00
-4.58854467e-01 -6.37558579e-01 -5.74505806e-01 1.04819886e-01
5.72763085e-01 -1.12167871e+00 -5.79927862e-01 5.26929125e-02] | [4.559063911437988, 1.1442644596099854] |
c302fcb0-ea6e-4aa0-851c-eb82162aa5b8 | beyond-rewards-a-hierarchical-perspective-on | 2206.09046 | null | https://arxiv.org/abs/2206.09046v3 | https://arxiv.org/pdf/2206.09046v3.pdf | Beyond Rewards: a Hierarchical Perspective on Offline Multiagent Behavioral Analysis | Each year, expert-level performance is attained in increasingly-complex multiagent domains, where notable examples include Go, Poker, and StarCraft II. This rapid progression is accompanied by a commensurate need to better understand how such agents attain this performance, to enable their safe deployment, identify limitations, and reveal potential means of improving them. In this paper we take a step back from performance-focused multiagent learning, and instead turn our attention towards agent behavior analysis. We introduce a model-agnostic method for discovery of behavior clusters in multiagent domains, using variational inference to learn a hierarchy of behaviors at the joint and local agent levels. Our framework makes no assumption about agents' underlying learning algorithms, does not require access to their latent states or policies, and is trained using only offline observational data. We illustrate the effectiveness of our method for enabling the coupled understanding of behaviors at the joint and local agent level, detection of behavior changepoints throughout training, discovery of core behavioral concepts, demonstrate the approach's scalability to a high-dimensional multiagent MuJoCo control domain, and also illustrate that the approach can disentangle previously-trained policies in OpenAI's hide-and-seek domain. | ['Been Kim', 'Lucas Dixon', 'Yannick Assogba', 'Andrei Kapishnikov', 'Shayegan Omidshafiei'] | 2022-06-17 | null | null | null | null | ['starcraft-ii'] | ['playing-games'] | [-5.03845036e-01 -4.39154916e-02 -4.71056461e-01 1.52102588e-02
-5.44591963e-01 -8.06228518e-01 9.76024091e-01 2.39410594e-01
-4.14538473e-01 8.77354801e-01 5.87258078e-02 -3.65847170e-01
-5.28988600e-01 -4.32074726e-01 -4.11964297e-01 -7.45688915e-01
-6.79659843e-01 9.00086701e-01 1.01745903e-01 -2.84922004e-01
-1.98224515e-01 3.66019934e-01 -1.07092381e+00 -1.83051601e-01
7.65605092e-01 3.25330973e-01 -2.42505938e-01 1.02008903e+00
6.55334055e-01 1.32269287e+00 -5.73973835e-01 -1.77261174e-01
3.61630350e-01 -2.32450232e-01 -6.35285020e-01 5.57707548e-01
9.21454951e-02 -7.28262007e-01 -4.02134806e-01 9.39507544e-01
-7.17115700e-02 2.54559010e-01 8.35374117e-01 -1.73927927e+00
-5.29839158e-01 5.50649583e-01 -5.24533331e-01 2.84552604e-01
4.35614288e-02 8.42113256e-01 1.09440958e+00 2.96374261e-01
5.08392215e-01 1.13534641e+00 5.46359062e-01 4.91448134e-01
-1.50439405e+00 -2.82738656e-01 5.81012309e-01 6.97706267e-02
-9.02490258e-01 -3.89950663e-01 4.25351739e-01 -8.41890156e-01
1.00639355e+00 -7.48350918e-02 6.94159746e-01 1.34558833e+00
-1.17346860e-01 8.71812761e-01 1.10047317e+00 -5.90071715e-02
5.00006795e-01 1.26769751e-01 1.96015060e-01 1.05084193e+00
2.08712310e-01 3.77984464e-01 -1.38832957e-01 -6.65638328e-01
8.03870916e-01 1.46011218e-01 6.14451021e-02 -6.57945096e-01
-1.16140831e+00 1.02242339e+00 1.93932448e-02 6.29717857e-02
-6.29642665e-01 3.51243854e-01 3.18287790e-01 5.72861910e-01
3.92523468e-01 6.19401038e-01 -5.96447945e-01 -4.29008186e-01
-4.99099433e-01 7.09588110e-01 1.20441544e+00 6.69288993e-01
7.49642432e-01 2.95750707e-01 2.04405606e-01 4.08507824e-01
3.55321914e-01 4.52662498e-01 3.14436227e-01 -1.63364398e+00
2.29916275e-01 5.45456648e-01 6.45897925e-01 -6.23816609e-01
-6.85535192e-01 -2.87025869e-01 -3.25783849e-01 4.17781711e-01
5.45687437e-01 -8.04739892e-01 -5.26245654e-01 1.92828012e+00
6.04811728e-01 1.74855098e-01 3.00962031e-01 8.07549655e-01
-1.17535584e-01 4.65742320e-01 -4.04672176e-02 -2.87851900e-01
1.12007916e+00 -9.90660012e-01 -2.89335459e-01 -3.54745924e-01
7.38090396e-01 -1.30095268e-02 6.99215651e-01 2.63852656e-01
-1.00696337e+00 2.58357413e-02 -8.48804772e-01 5.92525601e-01
-3.25011402e-01 -2.10580543e-01 8.21274042e-01 3.12425077e-01
-1.06536162e+00 3.83202046e-01 -1.52090430e+00 -4.91603464e-01
1.79199994e-01 4.90452379e-01 -8.81337747e-03 4.21444684e-01
-8.55068266e-01 1.11974525e+00 1.70116007e-01 -3.72264802e-01
-1.67145872e+00 -6.28078640e-01 -6.81857109e-01 9.00987536e-02
9.52296257e-01 -6.77291989e-01 1.72254002e+00 -9.75859582e-01
-1.81423903e+00 3.14055145e-01 3.21459562e-01 -5.68299770e-01
3.57635438e-01 2.04478592e-01 -2.70914912e-01 9.03429538e-02
1.63883209e-01 8.31281468e-02 7.77507365e-01 -1.25400579e+00
-8.78113270e-01 -2.36926779e-01 5.17591059e-01 4.57893193e-01
-3.86114597e-01 -1.15787555e-02 1.86915830e-01 -1.76587671e-01
-7.74720788e-01 -1.20416474e+00 -4.67105836e-01 -4.70431209e-01
-1.12939626e-01 -4.40445423e-01 6.58672452e-01 -3.90834510e-01
7.94128418e-01 -2.06429076e+00 5.35090864e-01 1.27402768e-01
6.39218450e-01 -1.33038443e-02 -2.16746598e-01 7.67594397e-01
4.19642657e-01 -1.29843280e-01 -1.35520287e-02 -4.19967711e-01
5.15554845e-01 2.84095615e-01 -7.18916729e-02 7.29732990e-01
-2.77517159e-02 6.04494750e-01 -1.08146966e+00 -1.88776642e-01
3.05481583e-01 -9.87068564e-02 -8.53728473e-01 3.69751990e-01
-7.24970341e-01 4.12369311e-01 -8.21613431e-01 4.09536004e-01
-6.48425743e-02 -5.29408455e-01 7.04330921e-01 4.13163006e-01
-3.31677735e-01 5.93231060e-02 -9.85946298e-01 1.13131142e+00
-1.12968273e-01 3.72813046e-01 6.06990933e-01 -1.14900160e+00
2.59207368e-01 3.96633953e-01 9.37647343e-01 -2.78374314e-01
1.53423622e-01 -1.54144198e-01 4.42832828e-01 -7.38004565e-01
1.49772733e-01 2.03556046e-02 -2.69726783e-01 8.63569081e-01
4.76076007e-02 4.61052991e-02 4.33746040e-01 2.11951464e-01
1.33155942e+00 -5.40949143e-02 2.77429789e-01 -3.27081293e-01
1.30091067e-02 5.09577394e-01 4.65419143e-01 9.49399590e-01
-6.49914086e-01 -3.89121205e-01 7.93305993e-01 -3.69105369e-01
-9.70072269e-01 -1.05673409e+00 3.75682354e-01 1.42355776e+00
8.32776129e-02 -2.52035260e-01 -6.10127211e-01 -7.07490563e-01
4.09610063e-01 5.67011058e-01 -8.07886600e-01 -1.47917539e-01
-5.34858704e-01 -9.93215263e-01 4.35063481e-01 1.48292124e-01
3.75427663e-01 -6.89171433e-01 -8.93535793e-01 2.84004390e-01
8.45190138e-02 -1.00551713e+00 -2.66744018e-01 -9.58963484e-02
-5.96738160e-01 -1.38932228e+00 -3.18589687e-01 -3.71928751e-01
3.52260739e-01 -6.01766445e-03 9.81116176e-01 -1.34342581e-01
-1.37637898e-01 1.15414524e+00 -8.35589170e-02 -2.29543224e-01
-8.20167780e-01 1.36645973e-01 6.11926019e-01 1.16187017e-02
2.11628512e-01 -7.30264127e-01 -5.76009572e-01 1.27830282e-01
-6.02042913e-01 -2.12695017e-01 4.15059984e-01 5.33800900e-01
-3.80668581e-01 -8.68516043e-02 4.99250054e-01 -5.09528637e-01
1.06573510e+00 -9.80541408e-01 -1.13596082e+00 3.97232771e-01
-6.51631355e-01 -2.76497938e-03 7.37923145e-01 -7.13796556e-01
-8.77198517e-01 -2.97268033e-01 7.03808367e-01 -4.26669538e-01
-3.67070258e-01 5.45970082e-01 4.95048016e-01 2.79118776e-01
6.11104608e-01 1.44700304e-01 5.28329790e-01 -2.65562087e-01
4.31357414e-01 3.11501116e-01 1.67807370e-01 -8.76545131e-01
7.96004891e-01 4.42387134e-01 -1.47646978e-01 -7.61447966e-01
-4.19041216e-01 -3.66443656e-02 -6.81152791e-02 -3.44499320e-01
8.28559995e-01 -1.05962479e+00 -1.67715466e+00 4.20641661e-01
-6.59959137e-01 -1.10774899e+00 -2.00111531e-02 5.75853467e-01
-6.57390773e-01 1.49513245e-01 -9.69788194e-01 -1.00641894e+00
3.08532476e-01 -1.18166184e+00 5.21442950e-01 3.17648977e-01
-1.75817296e-01 -1.49388945e+00 7.77673542e-01 3.56574565e-01
3.80244136e-01 3.30466069e-02 1.06329775e+00 -1.03746271e+00
-6.74670875e-01 7.31058866e-02 1.98223710e-01 -5.80649674e-02
2.98410386e-01 1.55448139e-01 -4.81345028e-01 -7.38050342e-01
-1.52765617e-01 -6.56819701e-01 1.31732419e-01 4.89838570e-01
4.04326677e-01 -7.13741124e-01 -3.15616250e-01 2.02802241e-01
1.12425733e+00 3.31255794e-01 -3.96514148e-01 4.42577869e-01
4.76140708e-01 5.62441945e-01 1.00246094e-01 7.30930030e-01
1.08624482e+00 6.24294460e-01 5.99505424e-01 9.82238874e-02
4.43684816e-01 -1.23543032e-02 7.89048672e-01 6.13997281e-01
-2.22242609e-01 -1.71374112e-01 -9.12480354e-01 5.59883177e-01
-2.31094813e+00 -1.16270971e+00 3.59037548e-01 1.84127808e+00
7.42566824e-01 -1.36900455e-01 9.62394655e-01 -7.48865962e-01
3.21454197e-01 2.24088863e-01 -1.08437741e+00 -2.74734885e-01
3.80390137e-01 -4.96879637e-01 5.22037983e-01 7.77238786e-01
-1.02615082e+00 8.77724826e-01 6.84392691e+00 2.03964949e-01
-7.60524035e-01 2.38604620e-01 3.71833503e-01 -4.11032438e-01
1.89489394e-01 4.04423513e-02 -4.41971809e-01 3.11639428e-01
1.00647676e+00 -2.28143975e-01 1.14813733e+00 8.05518687e-01
4.14354980e-01 -1.63930282e-01 -1.46169448e+00 5.58296084e-01
-1.20705374e-01 -1.17090118e+00 -5.57629526e-01 4.46218967e-01
8.38248372e-01 4.09588039e-01 1.00852638e-01 7.51441777e-01
1.57129180e+00 -8.06421518e-01 4.21450317e-01 2.83713132e-01
-1.08516179e-01 -4.81943667e-01 9.99084711e-02 7.62241006e-01
-7.72617042e-01 -5.18734097e-01 1.18601106e-01 -5.29330373e-01
-9.03349295e-02 -2.81310856e-01 -9.91563022e-01 2.45569557e-01
4.34291810e-01 7.99742341e-01 -1.66900367e-01 5.24376214e-01
2.12684035e-01 7.37701178e-01 -3.96510720e-01 -2.32243389e-01
6.18431628e-01 -6.02852941e-01 7.06152618e-01 7.39862263e-01
-2.21926957e-01 1.26196966e-01 9.81645167e-01 8.20720911e-01
2.26313323e-01 -4.40461338e-01 -5.07785439e-01 -4.41165864e-01
4.21658278e-01 1.07595229e+00 -4.35495168e-01 -5.08958697e-01
-5.58290184e-01 6.10896111e-01 6.82060480e-01 7.55053043e-01
-9.59863603e-01 3.06246161e-01 1.31056392e+00 -3.84803146e-01
2.28693321e-01 -5.50818980e-01 2.67494082e-01 -1.49748206e+00
-3.83059949e-01 -1.46024537e+00 5.13586581e-01 -1.92067161e-01
-1.29035926e+00 2.31949612e-02 2.73175538e-01 -8.06552589e-01
-8.89859021e-01 -4.61049825e-01 -5.48999667e-01 4.15574014e-01
-1.27926135e+00 -9.38124418e-01 2.03889340e-01 7.68729568e-01
5.07204950e-01 -4.08691496e-01 6.79918766e-01 2.81319730e-02
-1.01426423e+00 1.42965466e-01 4.34934348e-01 1.56248212e-01
3.30996901e-01 -1.28051960e+00 -9.20019001e-02 6.42497778e-01
5.04251681e-02 3.71286303e-01 9.75088298e-01 -5.37091315e-01
-1.74398458e+00 -6.19559884e-01 -2.30878115e-01 -7.74477661e-01
1.39161909e+00 -3.43287289e-01 -4.95601237e-01 1.25329220e+00
4.68366027e-01 -4.34463948e-01 5.17573059e-01 4.68712866e-01
-2.22351953e-01 -1.69698410e-02 -8.57795179e-01 9.62818742e-01
6.95940971e-01 -3.65651995e-01 -5.27702153e-01 6.03961408e-01
5.17068326e-01 -1.29144490e-01 -8.32207978e-01 1.32904248e-02
3.41405988e-01 -8.43861580e-01 7.51413584e-01 -1.22565269e+00
2.20565766e-01 -5.26727065e-02 -1.17257349e-02 -1.63607144e+00
-3.81685346e-01 -9.77443814e-01 -3.74583334e-01 7.34945357e-01
3.83554101e-01 -8.42049181e-01 7.38878965e-01 9.20138001e-01
8.53775218e-02 -3.84016514e-01 -7.05935717e-01 -7.38359332e-01
1.55518532e-01 1.34042144e-01 3.63314837e-01 1.20136952e+00
4.15647745e-01 3.05250674e-01 -6.02604330e-01 6.64246738e-01
9.58008170e-01 3.56205732e-01 1.10586560e+00 -9.68855798e-01
-8.62269282e-01 -7.55614758e-01 -4.93069477e-02 -1.09038281e+00
4.19151247e-01 -4.80816394e-01 -1.62503392e-01 -1.06254661e+00
3.83862287e-01 -2.79585481e-01 -1.41600162e-01 5.10175765e-01
8.30993801e-02 -3.96550298e-01 3.54453355e-01 4.14925486e-01
-1.03887665e+00 6.60373688e-01 8.94959509e-01 -8.47590491e-02
-3.13466370e-01 -1.16677620e-02 -4.84786838e-01 8.28954458e-01
7.63959646e-01 -3.24642926e-01 -6.23967111e-01 -4.27926123e-01
-1.15454951e-02 3.37929696e-01 7.03200519e-01 -6.70612752e-01
3.56325805e-01 -9.67930019e-01 -2.47190937e-01 2.20001727e-01
4.11445111e-01 -7.77312100e-01 5.45642711e-02 7.93394446e-01
-5.07354379e-01 3.23941678e-01 4.68410589e-02 7.54202902e-01
2.42340595e-01 6.91751912e-02 5.56258321e-01 -2.97239065e-01
-5.41616499e-01 3.50139916e-01 -9.02167499e-01 3.37968946e-01
1.44629729e+00 3.15090299e-01 -5.19777179e-01 -6.95745289e-01
-8.72634709e-01 9.53038752e-01 4.92780477e-01 2.34165728e-01
-1.91029850e-02 -7.98992455e-01 -7.45226502e-01 -3.70262414e-02
-2.17195764e-01 -5.02414525e-01 2.15666488e-01 9.33080018e-01
-1.60939157e-01 1.03837237e-01 -1.63028780e-02 -5.58357477e-01
-1.07010162e+00 6.45731509e-01 7.99204946e-01 -4.46539044e-01
-4.50138032e-01 2.49769807e-01 1.86599925e-01 -5.93540967e-01
2.15846300e-01 -2.00276628e-01 2.75667105e-02 -9.24960971e-02
2.24145547e-01 4.35555130e-01 -6.95160270e-01 -1.57659993e-01
-2.29768127e-01 -2.79605761e-02 -3.29410464e-01 -3.89825016e-01
1.53804362e+00 -1.34422213e-01 -2.77074147e-03 6.46139860e-01
7.31864572e-01 -2.70655453e-01 -1.93304718e+00 -1.69281885e-01
7.19459727e-02 -7.89514184e-02 -1.41523793e-01 -8.06535959e-01
-7.31399775e-01 3.31450760e-01 2.43492246e-01 9.02501881e-01
4.74322140e-01 2.82740623e-01 3.28113258e-01 6.94062948e-01
4.30103362e-01 -1.14691639e+00 3.10455829e-01 4.22985196e-01
3.93846303e-01 -1.44609070e+00 6.39143586e-02 2.97472119e-01
-7.74448872e-01 8.13154995e-01 8.24336469e-01 -3.82564902e-01
5.06150663e-01 2.85840869e-01 -3.24686728e-02 -4.73375618e-01
-1.38057172e+00 -2.35659704e-01 -3.70134205e-01 5.19501209e-01
-2.58578777e-01 3.35082650e-01 4.23695743e-01 7.80537352e-02
1.32753909e-01 -2.95970410e-01 7.42864370e-01 1.02743769e+00
-4.50180352e-01 -8.39537680e-01 -1.56775802e-01 3.88075233e-01
-1.66876137e-01 3.46795022e-01 -2.85451859e-01 1.08415020e+00
-3.75763506e-01 9.97537315e-01 1.43143892e-01 3.12668309e-02
2.33033404e-01 5.81440777e-02 4.88407999e-01 -4.70735103e-01
-5.60800672e-01 4.33885120e-02 2.07353055e-01 -5.68625331e-01
-5.62258601e-01 -9.95173097e-01 -8.22249353e-01 -6.55044734e-01
1.35009110e-01 3.23069274e-01 3.28203559e-01 1.14820516e+00
4.87639219e-01 4.63983119e-01 7.84377754e-01 -8.09753120e-01
-1.26712263e+00 -8.60088468e-01 -4.03650671e-01 4.03811693e-01
9.11314547e-01 -8.11063290e-01 -5.64511359e-01 -1.19018897e-01] | [3.746082067489624, 2.019031286239624] |
14cbe56c-0fb9-4956-a318-7a3a2e4a35f1 | pricing-algorithmic-insurance | 2106.00839 | null | https://arxiv.org/abs/2106.00839v2 | https://arxiv.org/pdf/2106.00839v2.pdf | Algorithmic Insurance | As machine learning algorithms start to get integrated into the decision-making process of companies and organizations, insurance products are being developed to protect their owners from liability risk. Algorithmic liability differs from human liability since it is based on a single model compared to multiple heterogeneous decision-makers and its performance is known a priori for a given set of data. Traditional actuarial tools for human liability do not take these properties into consideration, primarily focusing on the distribution of historical claims. We propose, for the first time, a quantitative framework to estimate the risk exposure of insurance contracts for machine-driven liability, introducing the concept of algorithmic insurance. Specifically, we present an optimization formulation to estimate the risk exposure of a binary classification model given a pre-defined range of premiums. We adjust the formulation to account for uncertainty in the resulting losses using robust optimization. Our approach outlines how properties of the model, such as accuracy, interpretability, and generalizability, can influence the insurance contract evaluation. To showcase a practical implementation of the proposed framework, we present a case study of medical malpractice in the context of breast cancer detection. Our analysis focuses on measuring the effect of the model parameters on the expected financial loss and identifying the aspects of algorithmic performance that predominantly affect the risk of the contract. | ['Agni Orfanoudaki', 'Dimitris Bertsimas'] | 2021-06-01 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 2.82603055e-01 6.39796257e-01 -3.79461676e-01 -4.76034194e-01
-8.92790675e-01 -5.10333002e-01 1.93117514e-01 4.47517365e-01
-4.00589168e-01 6.80369377e-01 8.32312256e-02 -7.79822707e-01
-6.84237480e-01 -8.40654194e-01 -7.07113683e-01 -4.54023600e-01
1.91819891e-01 5.74105561e-01 -3.23528945e-01 -3.25794667e-02
1.55732527e-01 4.38554913e-01 -1.02947307e+00 1.13659911e-01
1.14009416e+00 1.20162463e+00 -2.46873200e-01 1.69866785e-01
2.47392192e-01 7.29175627e-01 -6.16018414e-01 -9.08099174e-01
4.62716788e-01 -2.11279020e-01 -5.06446958e-01 -1.56785939e-02
-9.18583274e-02 -5.19870996e-01 1.88317120e-01 9.14785326e-01
2.38104403e-01 -4.56767589e-01 8.90672028e-01 -1.08495724e+00
-3.22927147e-01 1.09987998e+00 -8.12067371e-03 -1.50733158e-01
1.51593760e-01 -2.99384855e-02 1.05635428e+00 1.81328002e-02
4.24814790e-01 7.86404490e-01 6.31350756e-01 1.81970701e-01
-1.23228085e+00 -9.38011482e-02 1.55553631e-02 -3.34687442e-01
-1.15329635e+00 -1.71933159e-01 6.99068427e-01 -9.48322594e-01
5.22807300e-01 4.34347838e-01 4.04770374e-01 6.62356853e-01
5.66265404e-01 2.90200830e-01 1.04610121e+00 -5.42457759e-01
5.06465137e-01 4.85695690e-01 2.31230289e-01 1.78540930e-01
7.38700449e-01 3.47348422e-01 1.45168751e-01 -7.03046799e-01
4.44407105e-01 8.22525937e-03 -1.50985807e-01 -5.07243216e-01
-7.06604898e-01 1.05838561e+00 2.44855527e-02 1.52258620e-01
-4.96969372e-01 1.42269909e-01 2.29288235e-01 3.46382946e-01
2.47881159e-01 4.68425632e-01 -4.89617854e-01 1.00568473e-01
-9.33494925e-01 3.86444181e-01 9.19034719e-01 5.14311016e-01
1.58877105e-01 -2.62078613e-01 -4.68280733e-01 4.32297349e-01
2.30568528e-01 1.66589558e-01 2.65022129e-01 -9.74398732e-01
7.33448565e-01 5.80317497e-01 4.59480613e-01 -5.76800287e-01
-9.40265432e-02 -5.89772999e-01 -1.86516643e-01 6.72211587e-01
6.13163471e-01 -4.39116240e-01 -4.79002059e-01 1.55860758e+00
-3.58358733e-02 -6.67370856e-01 1.87389910e-01 4.96361524e-01
-1.95170507e-01 -5.41895116e-03 1.31625041e-01 -4.18758661e-01
1.10730350e+00 -3.24454904e-01 -6.15178704e-01 2.64449775e-01
5.90265512e-01 -4.31312293e-01 8.33878696e-01 5.13726830e-01
-1.31098819e+00 -1.01259775e-01 -1.02325392e+00 3.92624140e-01
-2.07641423e-02 7.30949193e-02 5.01582921e-01 1.26986110e+00
-3.92558247e-01 9.33774769e-01 -6.01494670e-01 1.43691748e-01
5.31587124e-01 2.30993450e-01 2.50029027e-01 3.34997535e-01
-1.24553490e+00 1.10901427e+00 -1.86618771e-02 1.37215346e-01
-7.20086515e-01 -7.76247859e-01 -4.32841510e-01 3.00532967e-01
3.63348246e-01 -8.27076674e-01 1.15887678e+00 -1.24478447e+00
-9.26128626e-01 6.14744067e-01 4.66939181e-01 -9.96860266e-01
1.25895107e+00 -1.96325943e-01 -4.09614265e-01 -1.36387587e-01
-4.70153876e-02 -2.53856272e-01 4.55378920e-01 -9.65580702e-01
-4.17940557e-01 -2.37728834e-01 3.63136679e-01 -1.76763132e-01
-9.83545333e-02 1.69392958e-01 3.19863528e-01 -9.14667130e-01
-1.75411686e-01 -8.51320863e-01 -5.10045767e-01 -2.16725305e-01
-3.56016666e-01 3.53578269e-01 -5.09847105e-02 -6.24639034e-01
1.43758476e+00 -2.00450659e+00 -2.24291384e-01 4.73164707e-01
-2.47131467e-01 -2.25013793e-01 6.42013371e-01 4.83622998e-01
-1.11262769e-01 2.19007462e-01 -7.46772826e-01 -1.11383907e-02
2.76597083e-01 -7.21363798e-02 -2.86216110e-01 4.98760045e-01
1.46514967e-01 6.21701717e-01 -3.10984045e-01 -3.66078854e-01
4.80506662e-03 2.01566592e-01 -5.17587125e-01 6.53101206e-02
-2.48802453e-01 -2.69968472e-02 -6.69782221e-01 7.21582413e-01
5.11427879e-01 -8.57375190e-02 2.34516814e-01 1.00330800e-01
-6.30910322e-02 1.53757051e-01 -9.38280821e-01 1.03420675e+00
-4.33595061e-01 -3.44768018e-02 -1.21917546e-01 -9.99404609e-01
6.57600164e-01 4.06827033e-01 7.12335527e-01 -1.03931807e-01
2.43494526e-01 2.81334668e-01 4.09121126e-01 -4.98740494e-01
7.43688047e-02 -4.78360832e-01 -3.49731207e-01 6.22680187e-01
-4.49941248e-01 -5.32262260e-03 -7.26116523e-02 -1.05306186e-01
9.99059081e-01 5.24573680e-03 3.85146260e-01 -4.32800710e-01
2.42543295e-01 1.39809757e-01 6.04001045e-01 4.53528911e-01
-1.66485086e-01 6.29164875e-01 9.89667952e-01 -2.82185107e-01
-8.67684782e-01 -1.06429422e+00 -6.82930410e-01 2.76184022e-01
-2.55146533e-01 4.47602868e-01 -1.05031800e+00 -7.38187671e-01
6.01782858e-01 1.15513480e+00 -6.63162410e-01 -1.14714481e-01
-1.79822221e-01 -1.22662640e+00 2.55629718e-01 1.93971857e-01
-9.68173966e-02 -9.78861094e-01 -1.24155915e+00 2.57679045e-01
3.31234217e-01 -6.53577209e-01 -3.13326120e-01 9.72892493e-02
-9.96508777e-01 -1.20523608e+00 -6.68549955e-01 3.64412647e-03
6.57029569e-01 -6.99775636e-01 1.05400074e+00 6.96568713e-02
-2.69365311e-01 4.55215275e-01 -2.00827599e-01 -7.86533475e-01
-7.81969488e-01 -7.75676519e-02 -3.10838401e-01 2.29043543e-01
5.37465997e-02 -3.67133051e-01 -4.35233444e-01 1.20931767e-01
-9.77889836e-01 -4.17093635e-01 7.31216848e-01 6.84460521e-01
2.60480642e-01 2.96729296e-01 8.78239989e-01 -1.39660311e+00
8.88798296e-01 -5.53730726e-01 -7.26768434e-01 7.43119419e-01
-1.24045968e+00 2.36317933e-01 9.12567824e-02 -2.68052697e-01
-1.34228361e+00 8.57012719e-02 7.68129379e-02 -4.81016487e-02
6.21584356e-02 8.11606228e-01 -4.88926977e-01 4.36266512e-02
4.77954656e-01 -5.39439380e-01 4.66273837e-02 -4.86722738e-01
7.20642582e-02 5.21554172e-01 2.25423738e-01 -6.49047017e-01
4.34835345e-01 4.29059803e-01 -3.60652730e-02 -1.13314539e-02
-7.07311034e-01 2.69380566e-02 -4.24769670e-01 -1.57281369e-01
7.36691952e-01 -5.92544734e-01 -7.25192666e-01 2.01014459e-01
-8.34875107e-01 -1.20922282e-01 -5.87143958e-01 6.74083173e-01
-8.92461300e-01 3.66318375e-02 -3.41940194e-01 -1.24959934e+00
-2.46140361e-01 -1.23635626e+00 4.14873779e-01 -2.44666785e-01
-3.66649568e-01 -1.08527756e+00 2.96449624e-02 6.79497004e-01
3.74559045e-01 8.72072279e-01 1.27771235e+00 -6.95261717e-01
-5.42906821e-01 -4.85425591e-01 2.44255200e-01 6.51225686e-01
5.43152466e-02 -1.21867331e-03 -8.40195835e-01 -1.32345837e-02
6.53260231e-01 4.58626039e-02 6.61644042e-01 7.70182550e-01
9.27074909e-01 -5.02416611e-01 -1.16140239e-01 3.04872870e-01
1.69711697e+00 3.47835004e-01 5.21730185e-01 4.08525348e-01
2.26129770e-01 1.21211922e+00 9.24769878e-01 6.25325859e-01
8.31839740e-02 6.61960244e-01 6.27166033e-01 4.65645045e-02
7.16406226e-01 -3.61505002e-02 2.35003456e-01 -1.63475573e-01
-1.67951822e-01 1.09493313e-03 -1.03249097e+00 4.47520614e-01
-1.68064773e+00 -8.22868288e-01 5.62551245e-02 2.51619506e+00
8.53746831e-01 7.47461140e-01 3.31036210e-01 1.94649249e-01
6.85853124e-01 -2.74552286e-01 -6.50409520e-01 -7.33309507e-01
3.38540785e-02 -7.02950731e-02 1.12410045e+00 5.89379787e-01
-7.69389927e-01 6.41445369e-02 6.79723740e+00 4.61233228e-01
-7.74170399e-01 7.40261301e-02 1.24521589e+00 -2.27551147e-01
-8.02865565e-01 2.79219061e-01 -4.65597421e-01 6.28290415e-01
9.21247423e-01 -4.83052880e-01 4.85635810e-02 9.61739600e-01
4.41852957e-01 -5.52666597e-02 -1.09002495e+00 1.19882256e-01
-3.20830673e-01 -1.23213589e+00 -1.40251189e-01 4.97247964e-01
5.87208867e-01 -4.37165320e-01 3.18286419e-01 -1.34637713e-01
2.57000506e-01 -1.10561180e+00 1.16653430e+00 1.12037039e+00
4.92276639e-01 -8.91319573e-01 1.06127310e+00 3.35056692e-01
-2.11384565e-01 -4.98458356e-01 4.22245786e-02 -3.50222029e-02
4.81371999e-01 9.72709596e-01 -6.80774748e-01 4.76630211e-01
3.39901090e-01 -1.23814099e-01 -3.23851377e-01 7.87233710e-01
2.75981098e-01 4.88774747e-01 5.19760437e-02 2.06005171e-01
-8.60029161e-02 -3.26254517e-01 2.28801355e-01 6.90635085e-01
5.83726346e-01 -4.70400229e-02 -3.27753633e-01 1.20705330e+00
2.65294909e-01 8.68253484e-02 -3.77601057e-01 -5.39321154e-02
2.17118695e-01 7.57643878e-01 -3.29286963e-01 6.21817186e-02
-5.03114343e-01 3.50321680e-01 -1.15379438e-01 1.81894377e-01
-9.43751574e-01 -1.14280939e-01 4.40360039e-01 5.79995871e-01
1.21368706e-01 3.43949854e-01 -8.80895138e-01 -6.60087347e-01
2.95610875e-01 -8.86450410e-01 5.35211682e-01 -2.57761747e-01
-1.18094075e+00 4.34770614e-01 2.69356996e-01 -1.18521559e+00
-3.48525852e-01 -5.78670979e-01 -5.73725045e-01 1.14513230e+00
-1.38285685e+00 -8.49035382e-01 4.32370514e-01 1.07621163e-01
-7.48140812e-02 -4.05553222e-01 4.94297236e-01 7.21332505e-02
-4.32710826e-01 7.51912653e-01 2.22885638e-01 -6.53115287e-02
3.33296478e-01 -1.21353757e+00 1.52541921e-01 6.80159628e-01
-2.23537624e-01 5.80941975e-01 9.37537551e-01 -7.57702887e-01
-6.99309289e-01 -7.40257502e-01 8.48577738e-01 -6.55117035e-01
7.06106365e-01 -3.95637006e-03 -5.75169802e-01 4.82285440e-01
-2.61239171e-01 -4.26370770e-01 6.69289291e-01 8.43553916e-02
7.15243956e-03 -3.37547988e-01 -1.64960885e+00 2.19727576e-01
5.52472532e-01 -3.09130192e-01 -5.57614088e-01 1.97016284e-01
7.62719274e-01 -2.11800858e-02 -1.26018620e+00 5.29531002e-01
7.40050197e-01 -1.34416437e+00 8.85869503e-01 -4.61766630e-01
2.34579310e-01 2.33562842e-01 -1.15530670e-01 -6.49804831e-01
1.11206040e-01 -5.78030109e-01 2.05325291e-01 1.21030700e+00
9.83929455e-01 -8.30654025e-01 7.76764810e-01 1.32953906e+00
1.44493520e-01 -1.13374257e+00 -1.05055594e+00 -7.64185786e-01
2.66469181e-01 -4.50185984e-01 8.91619861e-01 6.46532357e-01
-1.21985272e-01 -5.18894792e-01 -2.70845234e-01 1.19825639e-01
7.05167890e-01 3.16158831e-01 2.28729457e-01 -1.22011793e+00
-7.09679246e-01 -4.08641011e-01 -1.24006018e-01 6.43719658e-02
-1.04131065e-01 -4.09658581e-01 -2.73478836e-01 -8.69405210e-01
1.26603305e-01 -7.44699061e-01 -5.58649421e-01 6.39460534e-02
-8.91058072e-02 -5.58545291e-01 1.72807038e-01 1.84051841e-01
2.60280222e-01 1.04771070e-01 7.42475092e-01 6.05141819e-02
-1.35886431e-01 7.56743312e-01 -9.41459715e-01 7.50377953e-01
7.86711216e-01 -6.29919112e-01 -5.00553250e-01 -7.36627504e-02
3.41958612e-01 3.72285157e-01 4.60118920e-01 -7.24813223e-01
-5.93150146e-02 -5.34810126e-01 -1.12401783e-01 -1.66716710e-01
-7.39800856e-02 -1.07395327e+00 7.08085835e-01 8.17265749e-01
-7.28313029e-01 -1.04596823e-01 -3.40515643e-01 5.49925506e-01
-2.15099066e-01 -7.33825326e-01 5.51507115e-01 -7.82438591e-02
2.14977369e-01 -1.99153312e-02 -1.21364787e-01 -1.80319220e-01
1.23272014e+00 -9.99137461e-02 -1.09308623e-01 -3.31978351e-01
-8.27726722e-01 7.02087581e-02 7.30607331e-01 1.76104307e-02
6.59662560e-02 -1.08832562e+00 -7.82233536e-01 -2.67671406e-01
4.01233695e-02 -4.25998718e-01 1.76937103e-01 6.52371883e-01
-6.04693890e-01 5.29434144e-01 -6.26947284e-02 4.53145988e-02
-8.95967245e-01 8.32455397e-01 6.60808980e-01 -5.29680669e-01
-1.21310242e-01 3.54033917e-01 1.71242550e-01 9.06289369e-02
1.41706005e-01 -6.27460897e-01 -1.93640795e-02 4.13878672e-02
-1.33174807e-02 4.69958931e-01 1.46451682e-01 -3.96457195e-01
-1.23538584e-01 1.99272260e-01 2.37552911e-01 -4.93575811e-01
1.18447101e+00 -8.41177553e-02 2.92798374e-02 4.91019577e-01
6.95018351e-01 1.46249592e-01 -1.28217983e+00 2.55096555e-01
5.12195647e-01 -2.84101933e-01 -5.10739721e-02 -1.03802955e+00
-8.85104775e-01 5.19573510e-01 4.77908999e-01 5.31232715e-01
1.14767826e+00 -1.28876090e-01 4.18425202e-01 1.38419624e-02
3.91062587e-01 -1.30042112e+00 -6.74850404e-01 -4.71382648e-01
1.05259800e+00 -1.12827885e+00 2.35537812e-01 -3.71618181e-01
-8.61329317e-01 7.53418505e-01 -8.22442845e-02 -1.32683739e-01
9.17482972e-01 3.15221459e-01 2.00876668e-01 1.04117412e-02
-5.54900765e-01 2.27069035e-01 3.47901344e-01 3.39596987e-01
3.78456295e-01 5.20528376e-01 -8.69842470e-01 1.23799479e+00
-2.14543357e-01 1.04510270e-01 5.10624290e-01 9.29892719e-01
-2.07710668e-01 -1.31576943e+00 -5.53628147e-01 5.87589681e-01
-1.01754272e+00 5.27306274e-02 -2.41090491e-01 8.44498038e-01
4.21984076e-01 7.96241820e-01 -1.24269709e-01 1.46368206e-01
6.72497928e-01 2.60967109e-02 1.44205242e-01 -5.09693682e-01
-6.97733581e-01 1.55858146e-02 3.98546576e-01 -2.13148624e-01
-4.10339326e-01 -1.17495608e+00 -9.13889110e-01 -5.28401993e-02
-2.83410341e-01 2.32354775e-01 5.73876262e-01 9.96850789e-01
3.79327834e-02 3.78617465e-01 6.32617295e-01 -4.24425378e-02
-1.33345520e+00 -5.70585012e-01 -8.80580664e-01 3.70195210e-01
2.01782331e-01 -5.42255878e-01 -6.37274325e-01 -1.01728849e-01] | [7.642518997192383, 5.01938009262085] |
e8596996-f980-4c7c-b81d-9def68cc71ff | simhaze-game-engine-simulated-data-for-real | 2305.16481 | null | https://arxiv.org/abs/2305.16481v1 | https://arxiv.org/pdf/2305.16481v1.pdf | SimHaze: game engine simulated data for real-world dehazing | Deep models have demonstrated recent success in single-image dehazing. Most prior methods consider fully supervised training and learn from paired clean and hazy images, where a hazy image is synthesized based on a clean image and its estimated depth map. This paradigm, however, can produce low-quality hazy images due to inaccurate depth estimation, resulting in poor generalization of the trained models. In this paper, we explore an alternative approach for generating paired clean-hazy images by leveraging computer graphics. Using a modern game engine, our approach renders crisp clean images and their precise depth maps, based on which high-quality hazy images can be synthesized for training dehazing models. To this end, we present SimHaze: a new synthetic haze dataset. More importantly, we show that training with SimHaze alone allows the latest dehazing models to achieve significantly better performance in comparison to previous dehazing datasets. Our dataset and code will be made publicly available. | ['Yu Hen Hu', 'Yin Li', 'Bochen Guan', 'Jiang Li', 'Liang Shang', 'XiaoYu Zhang', 'Yanli Liu', 'Fangzhou Mu', 'Huan Xu', 'Zhengyang Lou'] | 2023-05-25 | null | null | null | null | ['image-dehazing', 'single-image-dehazing'] | ['computer-vision', 'computer-vision'] | [ 5.41786194e-01 2.34075457e-01 6.30207956e-01 -9.09422617e-03
-7.14386165e-01 -3.15668702e-01 6.43810630e-01 -2.94725478e-01
-3.27593833e-02 7.40828335e-01 -5.22652231e-02 -2.14492753e-01
3.43772054e-01 -1.11316812e+00 -9.78693187e-01 -9.65126395e-01
3.11239183e-01 1.95640236e-01 3.68674070e-01 -5.16891837e-01
2.79936910e-01 7.87218846e-03 -1.96905792e+00 4.10401314e-01
1.19487369e+00 8.23099434e-01 3.73843700e-01 1.22634494e+00
2.07075223e-01 9.08214629e-01 -8.68556678e-01 -4.25436318e-01
5.62129557e-01 -6.49467766e-01 -1.96068436e-01 -3.76685038e-02
9.35303986e-01 -8.98048341e-01 -2.91895717e-01 1.07013512e+00
2.57521123e-01 3.98362540e-02 8.16382766e-01 -1.18225443e+00
-8.83733630e-01 -5.97825758e-02 -3.22203428e-01 -1.58015475e-01
1.65972605e-01 2.33993083e-01 5.79470873e-01 -8.61869752e-01
2.85869420e-01 1.03280067e+00 5.33972681e-01 7.73379385e-01
-1.04751801e+00 -7.41665483e-01 -2.10859671e-01 2.02435076e-01
-1.32174838e+00 -4.91019696e-01 7.14604676e-01 -2.46711209e-01
7.57865191e-01 1.67852700e-01 5.80650568e-01 9.39754248e-01
4.61644799e-01 4.60205466e-01 1.26827216e+00 -5.87398529e-01
5.74393213e-01 -1.62295606e-02 -5.78175902e-01 6.96179628e-01
5.41315496e-01 4.42814112e-01 -5.38665950e-01 3.28431249e-01
8.43809962e-01 2.25196145e-02 -3.32378894e-01 -2.16940641e-01
-8.93527985e-01 6.26337230e-01 5.82322180e-01 -9.32522193e-02
-3.34267974e-01 3.39625150e-01 -5.08301914e-01 3.49240243e-01
6.43196166e-01 6.61893427e-01 2.04468176e-01 9.95136127e-02
-1.12408817e+00 5.31338573e-01 5.33021331e-01 8.92928243e-01
1.15526843e+00 3.60556901e-01 2.62179613e-01 5.73444545e-01
7.90640861e-02 7.36171246e-01 8.24272409e-02 -1.33025575e+00
2.35279709e-01 9.61651187e-03 3.08334321e-01 -1.02444553e+00
1.59637898e-01 -9.03122127e-02 -9.60030854e-01 1.20212483e+00
2.88287342e-01 -1.04013979e-01 -1.52057159e+00 1.16938245e+00
7.33783022e-02 6.01995111e-01 2.29111075e-01 8.33062053e-01
6.60901070e-01 1.04602909e+00 -3.85258973e-01 3.31336796e-01
8.68928969e-01 -1.19627440e+00 -7.94616342e-01 -4.35957104e-01
2.86301583e-01 -6.18890285e-01 1.04208910e+00 1.03210938e+00
-1.27864444e+00 -3.56930077e-01 -1.56198752e+00 -2.55925268e-01
-6.21284366e-01 -4.86954004e-01 2.71960795e-01 7.55739391e-01
-1.66418624e+00 5.56181431e-01 -7.80747473e-01 1.48727866e-02
4.86187816e-01 -3.19784768e-02 -1.73875630e-01 -5.71778357e-01
-9.86567795e-01 8.06766808e-01 3.16631824e-01 -3.36937271e-02
-1.28256130e+00 -9.38105822e-01 -1.11376464e+00 -1.73142217e-02
5.94060980e-02 -8.45631063e-01 1.14113700e+00 -8.14070225e-01
-1.69897842e+00 6.65841281e-01 1.47198737e-01 -5.35516202e-01
3.64991605e-01 -3.12842876e-01 -2.32520044e-01 4.53339934e-01
-7.60472193e-02 8.44033718e-01 1.59527254e+00 -1.83658063e+00
-6.34688199e-01 -5.66674806e-02 2.06525877e-01 2.20315829e-01
2.18813140e-02 -5.46230674e-01 -2.00598285e-01 -7.24064529e-01
-2.08761573e-01 -6.25005424e-01 -2.12806627e-01 2.59018928e-01
-2.88134664e-01 6.31643772e-01 7.70053446e-01 -6.85456574e-01
7.34570503e-01 -2.14251518e+00 3.99849452e-02 -7.10087568e-02
6.33144081e-01 4.50251400e-01 -1.54507324e-01 3.52431178e-01
1.10715702e-01 2.53446996e-01 -5.30536771e-01 -7.74457514e-01
-1.52202502e-01 2.80308783e-01 -4.57336128e-01 3.51527393e-01
3.04011852e-01 8.41788411e-01 -9.12424803e-01 -3.53565931e-01
6.31799877e-01 7.35983074e-01 -7.47682273e-01 5.87415695e-01
-3.58465523e-01 4.74240869e-01 1.16815194e-01 5.45611024e-01
1.04481471e+00 1.11966059e-01 -4.38421041e-01 1.42189980e-01
-6.73114136e-02 -7.82729834e-02 -6.54439807e-01 1.62721694e+00
-7.82108963e-01 7.79959977e-01 2.69881606e-01 -6.54126525e-01
7.53972948e-01 1.79584101e-01 -6.12047091e-02 -9.46898937e-01
3.76095548e-02 2.78720498e-01 -5.62762022e-01 -2.48203903e-01
7.30063796e-01 -5.08113980e-01 3.30480248e-01 3.71678263e-01
-1.07756995e-01 -1.22625792e+00 -1.53747618e-01 1.86238497e-01
8.49697530e-01 1.09042257e-01 -9.55594704e-02 -1.11691155e-01
1.50864378e-01 -7.61243573e-04 2.09868811e-02 9.63858604e-01
3.65085639e-02 1.53555131e+00 1.35843396e-01 -4.18718129e-01
-1.35783720e+00 -1.31694579e+00 8.31275061e-02 5.62678993e-01
4.17535126e-01 -2.38383040e-01 -1.05517256e+00 -1.22781493e-01
-3.47521991e-01 9.73424613e-01 -8.90543878e-01 -3.26460481e-01
-5.19784272e-01 -5.13725817e-01 3.86952728e-01 -4.11222689e-02
7.29164839e-01 -8.97318840e-01 -5.08623123e-01 6.76327348e-02
1.35499761e-02 -1.10150576e+00 -2.03763440e-01 -7.35950395e-02
-7.91289985e-01 -8.29990685e-01 -1.04257143e+00 -8.29282820e-01
5.04783452e-01 8.48511994e-01 1.28062379e+00 5.45953214e-01
-1.24454789e-01 3.63488458e-02 -4.97966439e-01 -8.45398724e-01
-6.37070954e-01 -2.84448236e-01 -1.59050345e-01 2.84177288e-02
-2.77264506e-01 -9.24543738e-01 -9.35512066e-01 6.92477077e-02
-1.59077287e+00 5.62680364e-01 5.26146114e-01 6.63553357e-01
4.21777427e-01 3.15509558e-01 1.38729841e-01 -9.06733513e-01
2.38899350e-01 -4.49244469e-01 -6.67603374e-01 -1.04744665e-01
-7.46933341e-01 -3.13805118e-02 6.62478209e-01 4.11132583e-03
-1.28142011e+00 -2.98033148e-01 -2.57366866e-01 -6.31445050e-01
-2.51647949e-01 1.70348197e-01 -3.55077572e-02 -3.72797370e-01
9.28648889e-01 4.10981148e-01 4.18369099e-02 -4.26132649e-01
5.54565430e-01 4.79404271e-01 1.14574301e+00 -3.93202573e-01
1.37677765e+00 9.30406988e-01 -1.65638357e-01 -8.64136875e-01
-7.86546230e-01 -1.04507804e-01 -3.79888445e-01 -8.96414742e-02
1.14741540e+00 -1.23230469e+00 -8.75693858e-02 9.36873674e-01
-1.06938004e+00 -9.15288091e-01 -4.16846909e-02 2.14683384e-01
-5.64501345e-01 4.05161858e-01 -5.06065547e-01 -8.40947390e-01
-2.14842409e-01 -9.20179367e-01 1.48734212e+00 2.53495783e-01
3.64277989e-01 -1.05896151e+00 1.14267178e-01 3.34745407e-01
6.87984526e-01 5.81903815e-01 7.26022303e-01 3.03587437e-01
-1.04226506e+00 -1.64975282e-02 -2.81816334e-01 5.89111805e-01
2.52064615e-01 1.32334709e-01 -1.25155258e+00 -2.22932667e-01
2.24451348e-01 -3.14956158e-01 1.05954313e+00 3.41748148e-01
1.00400519e+00 -2.38788873e-01 1.03935987e-01 1.25437522e+00
1.67995310e+00 -1.66210011e-02 1.15626812e+00 5.96754611e-01
9.55254018e-01 3.58585447e-01 3.87921512e-01 1.90592900e-01
3.61849755e-01 5.73083341e-01 9.23290253e-01 -4.07155991e-01
-6.23300374e-01 -2.92357355e-01 2.51291424e-01 6.09887838e-01
-1.89708725e-01 -6.01839423e-01 -7.32076287e-01 6.01013541e-01
-1.42894053e+00 -6.87222660e-01 -1.61616113e-02 2.06773162e+00
7.06610262e-01 7.61266276e-02 -3.55787665e-01 6.65991828e-02
3.76173139e-01 3.03445131e-01 -2.29547814e-01 -4.59721625e-01
-2.84097642e-01 6.39483869e-01 4.27378058e-01 9.08252835e-01
-8.64320993e-01 1.02488196e+00 6.52677345e+00 6.97649777e-01
-1.06467366e+00 6.81779161e-02 5.44572592e-01 -2.43929759e-01
-6.76374137e-01 -2.56945670e-01 -2.79333949e-01 6.70797706e-01
9.73351717e-01 -1.27471864e-01 6.94176912e-01 4.59590495e-01
1.04545631e-01 -3.99113327e-01 -8.47192943e-01 1.29886675e+00
3.85517627e-01 -1.62325478e+00 2.99163282e-01 1.95711061e-01
1.29667282e+00 -7.86154941e-02 5.29417574e-01 1.81594774e-01
5.07296503e-01 -1.32801950e+00 9.54239607e-01 4.20842707e-01
9.66512084e-01 -7.02912152e-01 6.17064297e-01 3.73862267e-01
-7.45462537e-01 3.06533813e-01 -6.40978813e-01 -2.29098871e-01
6.08494058e-02 6.37453675e-01 -6.34569585e-01 6.93027616e-01
8.34027827e-01 6.85767710e-01 -6.85888231e-01 8.98229480e-01
-5.27918398e-01 5.88115156e-01 -1.88530385e-01 5.11838853e-01
2.91440487e-01 -3.38577449e-01 1.81473434e-01 1.00304008e+00
6.47579849e-01 3.62472683e-01 -3.15680623e-01 9.78454769e-01
4.07941081e-02 -4.78682250e-01 -9.00201559e-01 2.96053588e-01
1.07771583e-01 9.27768469e-01 -4.08009559e-01 -4.64294672e-01
-3.56063128e-01 1.26959741e+00 1.35637358e-01 5.09832263e-01
-8.57388735e-01 -5.24616241e-01 9.33293223e-01 1.06915645e-01
3.75005603e-01 -3.39894295e-01 -5.91020346e-01 -1.45454741e+00
-1.88110843e-01 -8.70539367e-01 -1.13561466e-01 -1.34871626e+00
-9.86149132e-01 6.77463293e-01 -5.49734235e-02 -1.17402828e+00
-3.09333503e-01 -7.11151063e-01 -8.65370095e-01 1.04257166e+00
-1.99371564e+00 -1.04963696e+00 -9.59348083e-01 5.44568598e-01
6.03891790e-01 2.11619928e-01 7.15932488e-01 8.11270699e-02
-2.23788500e-01 4.28820729e-01 2.01165810e-01 -2.87843913e-01
6.85687780e-01 -1.47024095e+00 9.28984880e-01 1.18240023e+00
1.22334622e-01 1.63034782e-01 1.13326430e+00 -4.73568738e-01
-1.19031084e+00 -1.30477762e+00 2.57097274e-01 -8.07872534e-01
2.37790942e-01 -5.68684638e-01 -1.23187184e+00 4.93419886e-01
5.47000110e-01 2.40607187e-02 2.82558948e-01 -7.05401778e-01
-4.48327631e-01 9.47032273e-02 -1.08005595e+00 9.01307762e-01
8.62579882e-01 -5.54368854e-01 -5.13277531e-01 2.23187581e-01
9.65497494e-01 -6.30781949e-01 -4.66491699e-01 1.96523696e-01
3.65585744e-01 -1.60469663e+00 1.02432263e+00 1.72222801e-03
8.46537232e-01 -5.79637945e-01 -3.78938229e-03 -1.67341590e+00
-2.07833499e-01 -7.17447937e-01 -1.99992061e-01 5.55916250e-01
2.66384631e-01 -4.87926185e-01 1.01579511e+00 3.11717212e-01
-4.11190629e-01 -3.83628011e-01 -5.09248674e-01 -8.16459358e-01
3.56862485e-01 -3.67831826e-01 8.77078593e-01 6.56442642e-01
-5.12921035e-01 -1.30668253e-01 -7.71880925e-01 5.38897514e-01
1.10011804e+00 1.96461697e-04 1.00269592e+00 -7.85917401e-01
-2.92913079e-01 -1.52199164e-01 -4.28504139e-01 -9.21257615e-01
-7.08744675e-02 -4.05474633e-01 5.63131094e-01 -1.55759144e+00
-2.64902472e-01 -1.68006897e-01 -3.40256677e-03 2.73861408e-01
-5.43223381e-01 1.01101542e+00 1.91846937e-01 2.39676744e-01
-3.45283926e-01 7.11177528e-01 1.40122652e+00 -2.79268324e-01
-8.97597671e-02 -3.66649270e-01 -7.62629449e-01 3.87605876e-01
9.39736307e-01 -3.63251418e-01 -5.64579427e-01 -8.52714717e-01
1.42475486e-01 -2.30796054e-01 4.82517421e-01 -1.36732924e+00
1.61973342e-01 -1.04668431e-01 4.45271224e-01 -2.36427769e-01
6.49379849e-01 -5.16444325e-01 2.77974010e-01 3.21727723e-01
5.56741022e-02 -2.13309556e-01 1.94537133e-01 7.16319799e-01
-3.37676466e-01 -1.02595119e-02 9.04005349e-01 -1.80412754e-01
-7.77183652e-01 4.01294112e-01 -4.55258846e-01 -1.67469606e-01
8.01900446e-01 -5.52024662e-01 -5.16028106e-01 -9.02215242e-01
-3.25180143e-01 -2.33673185e-01 1.26472449e+00 2.63374656e-01
9.84688759e-01 -1.13308430e+00 -9.19551373e-01 5.57658494e-01
2.42391974e-01 5.51474988e-01 4.19380635e-01 2.94573694e-01
-1.36373460e+00 1.56822987e-02 -3.20629030e-01 -4.96253133e-01
-8.50601196e-01 8.70359182e-01 3.51713866e-01 2.81072855e-01
-9.25557017e-01 7.77050078e-01 8.75860929e-01 -1.78285856e-02
-1.43079177e-01 -3.95265698e-01 1.90142512e-01 -6.75702453e-01
1.02085114e+00 9.74600911e-02 3.83681178e-01 -3.62526089e-01
1.12132087e-01 5.45944095e-01 1.14259042e-01 -3.18117768e-01
1.30343258e+00 -3.56965363e-01 -3.20542790e-02 6.32346198e-02
1.13445985e+00 1.06210202e-01 -1.78446555e+00 -3.58878821e-02
-5.24630070e-01 -9.11034644e-01 3.07791620e-01 -5.67715228e-01
-1.17105746e+00 1.11024559e+00 4.75455731e-01 1.29212976e-01
1.33085275e+00 -2.42653623e-01 1.01574266e+00 2.28279054e-01
4.60062712e-01 -5.10872841e-01 1.58377901e-01 2.67352521e-01
7.10855901e-01 -1.22531712e+00 -5.93711622e-02 -4.80726808e-01
-3.81025225e-01 9.34430599e-01 5.98823369e-01 -3.71965289e-01
5.79300284e-01 3.36256057e-01 5.09493470e-01 -2.81118572e-01
-7.30551183e-01 -1.13307260e-01 1.24388449e-01 1.12457001e+00
-1.37152284e-01 -6.48083985e-02 5.75228989e-01 7.76388422e-02
-5.19374132e-01 -1.42577142e-01 1.13059390e+00 8.39714170e-01
-6.76714897e-01 -7.55946994e-01 -6.68641686e-01 1.07955940e-01
-2.16535822e-01 -4.25056309e-01 -2.52872884e-01 5.22839248e-01
1.68539882e-01 1.18943489e+00 1.13674544e-01 -4.78618205e-01
1.79070473e-01 -4.02536124e-01 6.69546604e-01 -8.44417870e-01
-7.49248788e-02 -2.48883143e-01 -1.89123794e-01 -6.42514884e-01
-1.56553641e-01 -3.78607959e-02 -7.97933519e-01 -5.85792720e-01
2.04730090e-02 4.01640497e-02 6.68505192e-01 7.70852089e-01
2.61689484e-01 4.78022903e-01 5.20638287e-01 -1.51761794e+00
-5.70581816e-02 -6.40864432e-01 -8.52944553e-01 3.57384086e-01
8.99613321e-01 -5.39837182e-01 -6.71846807e-01 3.28762114e-01] | [10.902064323425293, -3.16693377494812] |
1f97cd94-f106-413b-b9f5-ed0deae4f11e | domain-adaptation-in-lidar-semantic | 2010.12239 | null | https://arxiv.org/abs/2010.12239v3 | https://arxiv.org/pdf/2010.12239v3.pdf | Domain Adaptation in LiDAR Semantic Segmentation by Aligning Class Distributions | LiDAR semantic segmentation provides 3D semantic information about the environment, an essential cue for intelligent systems during their decision making processes. Deep neural networks are achieving state-of-the-art results on large public benchmarks on this task. Unfortunately, finding models that generalize well or adapt to additional domains, where data distribution is different, remains a major challenge. This work addresses the problem of unsupervised domain adaptation for LiDAR semantic segmentation models. Our approach combines novel ideas on top of the current state-of-the-art approaches and yields new state-of-the-art results. We propose simple but effective strategies to reduce the domain shift by aligning the data distribution on the input space. Besides, we propose a learning-based approach that aligns the distribution of the semantic classes of the target domain to the source domain. The presented ablation study shows how each part contributes to the final performance. Our strategy is shown to outperform previous approaches for domain adaptation with comparisons run on three different domains. | ['Ana C. Murillo', 'Luis Montesano', 'Luis Riazuelo', 'Inigo Alonso'] | 2020-10-23 | null | null | null | null | ['lidar-semantic-segmentation'] | ['computer-vision'] | [ 3.22063059e-01 1.52433245e-02 -2.97573775e-01 -7.90796578e-01
-6.03762627e-01 -5.97115278e-01 4.92792368e-01 2.49391958e-01
-7.27783382e-01 7.88233280e-01 -4.93825600e-02 -1.00921139e-01
-1.98809922e-01 -9.07478869e-01 -8.65783215e-01 -4.34773833e-01
3.66290629e-01 1.26170099e+00 9.02161658e-01 -2.84407765e-01
3.20292205e-01 6.93362474e-01 -1.72171748e+00 1.77726552e-01
1.00381553e+00 9.48803842e-01 3.18102121e-01 2.72947907e-01
-6.89855814e-01 5.39799258e-02 -5.28760791e-01 -4.49309349e-02
4.29915845e-01 -1.41725212e-01 -9.93397236e-01 -2.98806634e-02
5.75129211e-01 1.95068374e-01 1.13622040e-01 1.06767452e+00
4.78318185e-01 1.95565149e-01 8.94524574e-01 -1.25655162e+00
-3.80823851e-01 2.73640782e-01 -3.62415195e-01 5.87059669e-02
-6.71859160e-02 -1.43352777e-01 6.71593666e-01 -4.47734952e-01
9.83703017e-01 1.22585094e+00 7.79973209e-01 6.63662016e-01
-1.29759097e+00 -7.33798862e-01 4.03221875e-01 4.97627258e-01
-1.17735064e+00 -5.59446551e-02 8.13293457e-01 -5.27430892e-01
8.66219938e-01 -2.38774046e-01 4.48248386e-01 9.79138315e-01
-2.21892476e-01 6.41314447e-01 1.15956056e+00 -2.66871929e-01
6.31662548e-01 3.66435826e-01 1.91375256e-01 1.47307351e-01
3.99248064e-01 -1.35243848e-01 -5.35741270e-01 1.40629485e-01
3.70528162e-01 -3.40827823e-01 2.11903647e-01 -1.00582004e+00
-7.68324256e-01 8.38469207e-01 4.72723693e-01 4.18774396e-01
-2.44309127e-01 -1.52478084e-01 3.43584299e-01 2.82194111e-02
7.11541772e-01 4.53014374e-01 -9.03979838e-01 6.98669767e-03
-1.00676191e+00 5.41986108e-01 7.98419237e-01 1.01451886e+00
1.01954329e+00 -3.29683334e-01 -6.23253733e-03 9.65819776e-01
2.31525362e-01 5.13558924e-01 3.18975121e-01 -9.52815652e-01
4.23040450e-01 8.31230164e-01 6.13329336e-02 -5.01147926e-01
-6.54192090e-01 -4.33512747e-01 -2.79388309e-01 2.80907691e-01
5.60724914e-01 -5.60998134e-02 -1.26687372e+00 1.81460404e+00
5.21884799e-01 9.19483826e-02 1.98628128e-01 6.90770030e-01
7.32629120e-01 3.08017492e-01 3.54966164e-01 4.85742897e-01
1.06591737e+00 -6.86580360e-01 -1.84130818e-01 -7.48023450e-01
2.65677333e-01 -5.39769351e-01 1.00962102e+00 2.19587266e-01
-5.89092851e-01 -6.85732126e-01 -1.02819586e+00 -1.04418881e-01
-9.06245470e-01 -1.16455480e-01 4.64251816e-01 7.31905103e-01
-8.61968219e-01 6.11943007e-01 -6.89686179e-01 -1.01843357e+00
7.10730672e-01 4.67458218e-01 -3.07564586e-01 -2.65633166e-02
-1.10239911e+00 1.06617939e+00 7.22559273e-01 -5.26178062e-01
-6.26956522e-01 -8.34478259e-01 -7.41709173e-01 -2.61284769e-01
4.59140807e-01 -7.61240184e-01 1.37189305e+00 -9.19651091e-01
-1.30956507e+00 1.15142035e+00 -2.92776138e-01 -6.77883506e-01
5.61546266e-01 -3.04040045e-01 -7.50100464e-02 -3.30051444e-02
3.75834525e-01 1.18840396e+00 6.82762504e-01 -1.47194815e+00
-9.48555350e-01 -5.22279084e-01 -1.74039796e-01 2.14854687e-01
-1.76491290e-01 -3.45708996e-01 -4.78291601e-01 -2.62761176e-01
1.87304944e-01 -9.88096237e-01 -4.62667972e-01 -6.45925999e-02
-9.61162373e-02 -3.52359802e-01 8.40882301e-01 -3.24651480e-01
7.25916326e-01 -2.31579924e+00 1.89070225e-01 1.11073636e-01
-2.14972615e-01 3.51198286e-01 -7.96458349e-02 1.53691038e-01
-1.10488608e-02 1.18039390e-02 -7.19639659e-01 -4.50393379e-01
2.49294788e-01 4.95908797e-01 -2.20629141e-01 2.09821716e-01
2.54528672e-01 5.90990067e-01 -7.40948856e-01 -4.00260359e-01
2.36455634e-01 2.50813007e-01 -6.39030814e-01 1.63983196e-01
-6.32789850e-01 7.36500502e-01 -4.99015540e-01 2.99507618e-01
9.19867933e-01 1.26300812e-01 7.77857602e-02 -2.11043749e-02
-8.67099129e-03 3.89831632e-01 -1.18443251e+00 2.13167119e+00
-3.77919137e-01 4.48052913e-01 -8.59808996e-02 -1.30481863e+00
1.24085188e+00 -2.45037541e-01 5.91095924e-01 -8.22601795e-01
2.86447089e-02 4.28551108e-01 -1.61611304e-01 -4.31134433e-01
5.94847620e-01 -3.06794971e-01 -2.24837750e-01 2.19689295e-01
2.95892447e-01 -5.06150961e-01 1.77492231e-01 -2.25577965e-01
7.54770458e-01 5.88124812e-01 3.17793816e-01 -4.59446579e-01
4.82974052e-01 4.12839711e-01 5.01379550e-01 6.61008298e-01
-4.22489733e-01 6.75302505e-01 4.06404525e-01 -4.98955160e-01
-1.06288421e+00 -1.15376925e+00 -1.31778851e-01 1.19035566e+00
4.07229424e-01 6.88275695e-02 -1.06976330e+00 -8.90169799e-01
3.08838934e-01 9.81534421e-01 -6.63248777e-01 -9.01127458e-02
-5.06734431e-01 -5.64084351e-01 4.73209262e-01 7.40997612e-01
7.41011858e-01 -1.12312746e+00 -8.69007528e-01 1.58299014e-01
-1.59051523e-01 -1.43240607e+00 1.28360957e-01 4.09183502e-01
-9.78684247e-01 -1.00386417e+00 -5.16135991e-01 -6.40734017e-01
3.87767673e-01 1.46748066e-01 1.31849110e+00 -4.71867621e-01
-4.63806204e-02 3.79681736e-01 -2.94881672e-01 -9.22386348e-01
-3.55322331e-01 7.26806402e-01 -1.51903033e-01 -1.41271189e-01
8.49387348e-01 -7.20750213e-01 -3.90815824e-01 3.26205730e-01
-9.68482614e-01 -2.92421162e-01 4.02474791e-01 2.91470051e-01
8.45116615e-01 -4.89484426e-03 5.20757496e-01 -1.19867563e+00
4.00293440e-01 -4.15108532e-01 -6.28690660e-01 5.31312861e-02
-5.85476756e-01 4.17701304e-01 4.02394146e-01 -2.25900784e-01
-1.13166380e+00 5.30770957e-01 -2.90439487e-01 -1.82536080e-01
-9.61133003e-01 7.10934997e-02 -4.78769094e-01 -1.52065139e-02
9.05176699e-01 -3.33553106e-02 -1.18509397e-01 -7.33089328e-01
4.85581636e-01 6.34588778e-01 5.55852592e-01 -7.93649435e-01
7.39235520e-01 7.40037084e-01 1.76816031e-01 -8.65100265e-01
-9.04457808e-01 -7.62087524e-01 -1.17315185e+00 1.21417284e-01
9.80304003e-01 -7.40060210e-01 -8.04755613e-02 5.98765671e-01
-1.13937652e+00 -4.89911318e-01 -5.94881713e-01 8.96907300e-02
-7.09967315e-01 1.08225666e-01 3.40129048e-01 -4.51296866e-01
4.52583581e-02 -1.11617315e+00 1.29129028e+00 3.24812084e-01
-2.31445059e-01 -1.04325008e+00 3.45372975e-01 3.05072367e-01
4.57101852e-01 1.58380434e-01 9.45466638e-01 -1.15789104e+00
-4.69298840e-01 1.95376500e-01 -3.90266299e-01 4.01877284e-01
-4.82651778e-02 -3.41865003e-01 -1.15335298e+00 1.46255523e-01
-1.86224073e-01 6.09802490e-04 1.13028610e+00 5.50924122e-01
1.10068119e+00 4.10436779e-01 -6.10946119e-01 6.73549592e-01
1.42106187e+00 1.67905688e-01 6.01523340e-01 7.74293780e-01
4.82114464e-01 8.74438524e-01 8.89071763e-01 1.09209068e-01
7.21108496e-01 8.05227280e-01 5.86511314e-01 -9.72815827e-02
-2.79794544e-01 -1.78238958e-01 2.36916132e-02 -6.62227720e-02
1.11681998e-01 -1.11146234e-01 -1.16271400e+00 9.54015076e-01
-1.93965220e+00 -5.56593955e-01 4.28066365e-02 2.12854552e+00
5.58365703e-01 3.63222599e-01 3.48947167e-01 -2.75258906e-02
6.41477704e-01 3.39110605e-02 -8.25106740e-01 -4.56602126e-01
2.51611974e-02 6.37003601e-01 8.15940619e-01 3.86658609e-01
-1.44937134e+00 1.35507822e+00 6.33004618e+00 7.25621700e-01
-1.14571130e+00 2.72279292e-01 1.25439569e-01 1.93777636e-01
-2.11908743e-01 -1.30006611e-01 -9.80491281e-01 2.45012805e-01
7.98751652e-01 8.79693180e-02 2.33246133e-01 1.10337675e+00
-4.41823453e-02 -2.88959801e-01 -1.10837877e+00 6.70542717e-01
4.00073826e-02 -1.00754118e+00 -2.17393436e-03 -7.02189505e-02
7.31260538e-01 4.12136316e-01 -7.09605813e-02 2.13815317e-01
3.72791201e-01 -7.38511205e-01 7.62530088e-01 3.77253532e-01
6.19088888e-01 -7.52305984e-01 7.71907330e-01 3.51102650e-01
-1.06886148e+00 -5.61679043e-02 -4.46280122e-01 6.10327674e-03
-1.33865578e-02 5.14124215e-01 -1.09114528e+00 6.71517909e-01
9.68451619e-01 6.89741433e-01 -8.17252636e-01 1.00619411e+00
-3.99617910e-01 4.09352243e-01 -4.87256587e-01 3.22334394e-02
3.27951372e-01 -1.00993358e-01 6.26350343e-01 1.40787542e+00
3.49471658e-01 -3.18810016e-01 2.40922064e-01 9.21878397e-01
5.98824620e-02 5.56316860e-02 -8.37811887e-01 1.17231771e-01
4.93251413e-01 8.09506178e-01 -9.15633559e-01 -2.10576549e-01
-1.43855855e-01 7.23577797e-01 2.76807576e-01 2.45486364e-01
-6.07726336e-01 -3.58033657e-01 9.16520774e-01 3.28694075e-01
5.64755142e-01 -2.74847686e-01 -7.79858530e-01 -6.67638600e-01
-4.70627956e-02 -5.97045958e-01 5.48437417e-01 -5.88396311e-01
-1.31040478e+00 3.80426049e-01 4.05063629e-01 -1.14479232e+00
-1.82173908e-01 -7.68239319e-01 -1.58754900e-01 8.09476733e-01
-1.92383373e+00 -1.24010801e+00 -4.86908138e-01 5.14672995e-01
6.86274886e-01 -2.80618399e-01 8.03171158e-01 2.28160247e-01
-1.93504784e-02 2.77742416e-01 -1.19019402e-02 -8.02575946e-02
8.90009463e-01 -1.33153832e+00 7.27635264e-01 7.68518448e-01
1.14895731e-01 2.57850289e-01 7.29890287e-01 -5.97648859e-01
-6.48739159e-01 -1.23841417e+00 8.19108188e-01 -5.87327838e-01
4.87280339e-01 -2.68655509e-01 -9.91222680e-01 5.11649489e-01
5.25071882e-02 -2.58435518e-01 6.03293777e-01 2.83290565e-01
-4.11434144e-01 -3.27665746e-01 -1.36459720e+00 3.25764358e-01
1.36643457e+00 -2.90099442e-01 -8.38406742e-01 7.03783482e-02
6.03338540e-01 -5.81057131e-01 -4.31556046e-01 6.16588533e-01
4.21820909e-01 -1.23582149e+00 1.03668165e+00 -6.30188286e-01
3.47057313e-01 -3.46761793e-01 -3.55832249e-01 -1.52683949e+00
-1.11905143e-01 -9.00843665e-02 3.00861299e-01 1.07006538e+00
4.33106303e-01 -7.01907992e-01 1.04918921e+00 2.40340114e-01
-2.08280563e-01 -2.49859244e-01 -1.12832427e+00 -8.90669465e-01
4.63979006e-01 -6.29943013e-01 8.31845701e-01 6.64407194e-01
-6.51498020e-01 3.80135626e-01 2.98068821e-01 2.06404179e-01
5.72393835e-01 1.89186290e-01 1.07958794e+00 -1.75580239e+00
9.91133079e-02 -4.52588707e-01 -6.70042932e-01 -8.98193538e-01
5.11913180e-01 -9.08014715e-01 1.87427595e-01 -1.78885376e+00
2.72449534e-02 -5.30975580e-01 -3.40516895e-01 5.45266509e-01
7.49138743e-02 1.07025802e-01 3.32588673e-01 -1.90916255e-01
-5.81436515e-01 4.20972973e-01 9.72524166e-01 -2.29785576e-01
-3.46340537e-01 1.20312758e-01 -7.79228270e-01 8.68699968e-01
1.06721354e+00 -7.49397337e-01 -5.27747273e-01 -5.69546402e-01
5.32687292e-04 -8.23112547e-01 3.33954096e-01 -1.40755379e+00
9.44049209e-02 -2.71320164e-01 1.14262581e-01 -7.50617623e-01
3.89334857e-01 -1.10766637e+00 6.44492730e-02 3.11430663e-01
-1.51183799e-01 -3.41239929e-01 6.10474586e-01 5.13188720e-01
-1.24073148e-01 -3.32056910e-01 1.06330311e+00 -2.33899206e-01
-1.23366666e+00 1.13726325e-01 -5.29825911e-02 3.16386163e-01
9.88840580e-01 -4.58423048e-01 -8.28207731e-02 -9.59361568e-02
-7.74263024e-01 1.54768109e-01 6.72407150e-01 6.40731573e-01
2.47797266e-01 -9.38170314e-01 -5.97872078e-01 2.80327708e-01
3.67291003e-01 3.49156499e-01 1.76654686e-03 3.42829466e-01
-4.51294363e-01 4.40194726e-01 -4.72398251e-01 -9.29493368e-01
-1.18277335e+00 3.08852255e-01 4.80365783e-01 -3.27411413e-01
-3.34985673e-01 8.80727351e-01 1.76308736e-01 -1.01312566e+00
1.65497467e-01 -3.89942288e-01 -2.82391876e-01 2.29101628e-01
-9.78190638e-03 3.20350617e-01 3.31913650e-01 -5.50603151e-01
-5.58967054e-01 1.00607443e+00 9.08662155e-02 -1.53671071e-01
1.50659287e+00 -1.17942125e-01 3.79859321e-02 3.96914423e-01
7.58349657e-01 -2.42409408e-01 -1.34048891e+00 -3.61394018e-01
2.79149383e-01 -3.78984928e-01 -2.58360088e-01 -1.03179049e+00
-8.67841780e-01 1.18206632e+00 8.58205497e-01 -2.26391200e-03
1.19512498e+00 2.09132031e-01 6.07000768e-01 5.19741714e-01
5.16688824e-01 -1.42495823e+00 -1.34677842e-01 7.84948766e-01
5.22366285e-01 -1.38921261e+00 6.53259382e-02 -5.09790182e-01
-6.88330770e-01 8.57484818e-01 6.30054593e-01 -1.34280056e-01
7.01286316e-01 1.20826423e-01 2.35062420e-01 -7.16783702e-02
-1.35542035e-01 -7.07242429e-01 2.14683026e-01 1.28263199e+00
6.06700256e-02 1.66338757e-01 -1.91550016e-01 6.21192157e-01
-2.31186494e-01 6.99828640e-02 8.01699460e-02 7.99409509e-01
-7.22072423e-01 -1.53331769e+00 -3.59196752e-01 4.50848676e-02
-1.48255199e-01 2.48281345e-01 -8.27774644e-01 1.02468920e+00
5.70362151e-01 7.11257339e-01 1.52951842e-02 -1.36740938e-01
7.68086433e-01 4.40257639e-01 5.55363059e-01 -8.84247899e-01
-3.12731206e-01 -2.16959089e-01 7.46649280e-02 -5.96213937e-01
-7.19930768e-01 -8.68814766e-01 -1.37253642e+00 1.74230561e-01
1.84869304e-01 -2.52960734e-02 9.39067006e-01 1.09641564e+00
5.38248658e-01 5.17034531e-01 1.19178794e-01 -1.00359011e+00
-3.64848614e-01 -8.64161253e-01 -3.69750500e-01 6.03154957e-01
1.50890946e-01 -1.06403553e+00 9.22102705e-02 9.03971493e-02] | [8.20639419555664, -2.5773825645446777] |
14a54db4-1ea6-4c05-ae6e-d6be3cbf01d7 | a-reinforcement-learning-based-energy | 2104.04443 | null | https://arxiv.org/abs/2104.04443v2 | https://arxiv.org/pdf/2104.04443v2.pdf | A Reinforcement-Learning-Based Energy-Efficient Framework for Multi-Task Video Analytics Pipeline | Deep-learning-based video processing has yielded transformative results in recent years. However, the video analytics pipeline is energy-intensive due to high data rates and reliance on complex inference algorithms, which limits its adoption in energy-constrained applications. Motivated by the observation of high and variable spatial redundancy and temporal dynamics in video data streams, we design and evaluate an adaptive-resolution optimization framework to minimize the energy use of multi-task video analytics pipelines. Instead of heuristically tuning the input data resolution of individual tasks, our framework utilizes deep reinforcement learning to dynamically govern the input resolution and computation of the entire video analytics pipeline. By monitoring the impact of varying resolution on the quality of high-dimensional video analytics features, hence the accuracy of video analytics results, the proposed end-to-end optimization framework learns the best non-myopic policy for dynamically controlling the resolution of input video streams to globally optimize energy efficiency. Governed by reinforcement learning, optical flow is incorporated into the framework to minimize unnecessary spatio-temporal redundancy that leads to re-computation, while preserving accuracy. The proposed framework is applied to video instance segmentation which is one of the most challenging computer vision tasks, and achieves better energy efficiency than all baseline methods of similar accuracy on the YouTube-VIS dataset. | ['Robert P. Dick', 'Li Shang', 'Ning Gu', 'Tun Lu', 'Dongsheng Li', 'Qin Lv', 'Da Feng', 'Yujiang Wang', 'Mingzhi Dong', 'Yingying Zhao'] | 2021-04-09 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.35428244e-02 -3.64156306e-01 -2.77411044e-01 -2.02191528e-02
-4.36408073e-01 -4.85778064e-01 2.40234584e-01 1.86751671e-02
-7.04868734e-01 3.01163346e-01 -5.41598573e-02 -7.76010081e-02
-1.93200603e-01 -5.88102818e-01 -7.30624735e-01 -6.61361516e-01
-2.28439152e-01 -3.18998173e-02 3.06300074e-01 3.79709005e-01
2.92536348e-01 4.71161366e-01 -1.59603202e+00 8.89391527e-02
7.77908742e-01 1.33526587e+00 4.40630108e-01 9.81258333e-01
1.88900679e-01 1.01757884e+00 -2.51982272e-01 -1.49020091e-01
4.73995715e-01 -3.08223575e-01 -5.58187664e-01 3.80880147e-01
3.87116909e-01 -7.67393589e-01 -7.20121920e-01 8.63348246e-01
4.78766859e-01 4.66863930e-01 3.06468695e-01 -1.39560068e+00
-2.13062853e-01 1.47071749e-01 -5.90421915e-01 8.50240409e-01
-3.65320481e-02 6.18498683e-01 8.04548025e-01 -5.42694211e-01
5.34695387e-01 9.28235531e-01 1.76281214e-01 3.03909808e-01
-1.04521155e+00 -5.11725485e-01 3.07319462e-01 6.65812731e-01
-1.14366150e+00 -7.41691589e-01 7.26114333e-01 -3.97568643e-01
1.02198410e+00 -7.85528570e-02 9.25260663e-01 5.81276357e-01
4.19723511e-01 7.38824785e-01 6.10893011e-01 -1.12613859e-02
6.51775956e-01 -4.14679110e-01 -3.82774174e-01 9.92145061e-01
1.68685094e-02 -2.45970972e-02 -1.07612216e+00 1.53345078e-01
8.05524349e-01 1.22908555e-01 -1.46891713e-01 -2.84287989e-01
-1.05657971e+00 4.41652566e-01 3.61352682e-01 -8.38957578e-02
-5.81248581e-01 5.26502252e-01 5.67158282e-01 2.34135464e-02
3.58147264e-01 2.02127174e-01 -5.07624209e-01 -4.86818433e-01
-1.28454137e+00 3.28162044e-01 3.14347625e-01 7.44583488e-01
6.72703803e-01 2.53727943e-01 -3.54147911e-01 4.83763635e-01
2.23112330e-01 3.16269755e-01 4.82113570e-01 -1.63827109e+00
4.99487132e-01 6.22386873e-01 7.15263560e-02 -1.02127802e+00
-3.13559681e-01 -1.43708184e-01 -8.31375480e-01 3.30465913e-01
2.62489498e-01 -1.87606677e-01 -8.16366494e-01 1.46539772e+00
5.43084562e-01 4.17332977e-01 -2.36263439e-01 1.35422790e+00
3.14298034e-01 6.92587554e-01 3.21591049e-01 -5.50245345e-01
1.28429461e+00 -1.03097153e+00 -7.70401001e-01 1.98503714e-02
2.14424103e-01 -4.54122722e-01 9.94999349e-01 4.57336843e-01
-1.47645628e+00 -5.08298814e-01 -1.08309877e+00 -2.71413207e-01
3.35669592e-02 -7.28870332e-02 3.75785351e-01 3.85954261e-01
-9.20460522e-01 7.60113239e-01 -1.30211699e+00 4.13256809e-02
8.57998133e-01 3.55188191e-01 2.61823773e-01 -1.21326409e-01
-7.50708938e-01 3.84739667e-01 2.77530938e-01 2.85198130e-02
-1.12754989e+00 -1.01474547e+00 -4.71456498e-01 2.56862730e-01
7.38268018e-01 -8.73296976e-01 1.14564049e+00 -1.05964696e+00
-1.58444464e+00 2.88488448e-01 -9.38424841e-02 -5.56017816e-01
7.03056276e-01 -3.04824829e-01 1.65616777e-02 4.65522349e-01
-1.72094494e-01 7.40044355e-01 1.27498519e+00 -6.23336434e-01
-9.71072495e-01 -4.46851492e-01 1.01714142e-01 5.55532992e-01
-6.10222042e-01 -1.69411987e-01 -7.74886370e-01 -4.76202786e-01
-2.31872946e-01 -8.77849281e-01 -2.39137903e-01 2.30460957e-01
1.17103890e-01 -1.20839894e-01 1.04355109e+00 -4.74405468e-01
1.38511932e+00 -2.10132432e+00 3.33261758e-01 -2.25680470e-01
3.90679628e-01 1.15463033e-01 1.67337265e-02 -2.75690556e-01
6.25388563e-01 7.28465617e-02 -1.69547066e-01 -2.87005037e-01
-2.77978003e-01 2.60190696e-01 -2.96044629e-02 5.64441502e-01
2.66208678e-01 7.63112724e-01 -9.19433355e-01 -7.16098607e-01
5.00179529e-01 5.15023947e-01 -7.51682341e-01 5.75421095e-01
-4.62860435e-01 3.43088597e-01 -5.12079298e-01 8.09832096e-01
3.80523622e-01 -4.12484705e-01 9.04553458e-02 -4.17495966e-01
-1.57699645e-01 4.85024136e-03 -1.06024122e+00 1.83042872e+00
-4.58988935e-01 8.48776281e-01 3.60757977e-01 -9.31910694e-01
2.24109858e-01 9.31905359e-02 1.09969759e+00 -9.58212912e-01
1.55698493e-01 -1.29922807e-01 -1.89188287e-01 -7.25760937e-01
6.21689260e-01 5.47829807e-01 1.93963021e-01 3.20842177e-01
-9.04319212e-02 1.84125885e-01 4.96473193e-01 1.84521809e-01
1.35782945e+00 3.64739895e-01 9.99793038e-02 -2.63283495e-02
2.30768844e-01 7.18210042e-02 7.62569904e-01 3.44015479e-01
-5.56156039e-01 6.87871575e-02 3.69474709e-01 -5.94939649e-01
-1.27792549e+00 -9.58404541e-01 1.24389529e-01 1.45018458e+00
4.09800202e-01 -3.86872053e-01 -8.81744266e-01 -3.29106361e-01
-1.05921440e-01 2.52994269e-01 -2.43836701e-01 -1.09595492e-01
-7.76175737e-01 -6.62820280e-01 1.85172871e-01 3.26588959e-01
6.90557420e-01 -1.01994288e+00 -1.67545056e+00 3.90420616e-01
-2.34718800e-01 -1.64160490e+00 -7.86642373e-01 -6.30776137e-02
-1.12437189e+00 -1.20103920e+00 -2.88247913e-01 -2.45285690e-01
5.04420936e-01 3.73110503e-01 1.18368435e+00 1.13380060e-01
-5.50579429e-01 5.23926139e-01 -2.42755651e-01 -8.85247719e-03
-1.26192316e-01 1.38130322e-01 -9.95693058e-02 4.79900502e-02
-8.41018856e-02 -4.70516711e-01 -1.18578136e+00 1.27650827e-01
-1.08240032e+00 1.38693959e-01 4.46494430e-01 4.87212598e-01
1.00750685e+00 3.00635219e-01 2.18577191e-01 -4.06861126e-01
3.96045893e-01 -5.18241465e-01 -9.33160007e-01 3.86196934e-02
-8.25418055e-01 2.16282219e-01 9.55625713e-01 -4.50582027e-01
-7.58331478e-01 2.93421298e-01 4.81606275e-01 -9.70008254e-01
2.12058604e-01 -5.11574373e-03 -3.09372041e-02 4.28205281e-02
1.38382047e-01 2.40128100e-01 -1.93492606e-01 9.29474086e-02
1.92273974e-01 1.82783335e-01 6.62637234e-01 -3.99972975e-01
3.59874099e-01 4.63255882e-01 2.68045604e-01 -7.70441830e-01
-4.46266294e-01 -2.10499018e-01 -4.78305608e-01 -6.98834002e-01
1.09121656e+00 -1.03866971e+00 -1.07844996e+00 3.59132290e-01
-9.77051556e-01 -5.18469989e-01 -1.27946883e-01 3.13642681e-01
-7.75229633e-01 1.90480083e-01 -4.74609256e-01 -7.17047274e-01
-6.61982834e-01 -1.37489831e+00 1.02717316e+00 2.90571392e-01
8.24678391e-02 -6.18472099e-01 -1.55863002e-01 4.98179823e-01
2.23897338e-01 2.94182748e-01 7.89499342e-01 1.80585340e-01
-1.23613131e+00 3.88367355e-01 -2.78785616e-01 1.15507483e-01
-1.42396376e-01 2.89619267e-01 -7.77281284e-01 -3.22056413e-01
-1.66083366e-01 -2.36199051e-01 7.80995369e-01 6.86438024e-01
1.70821106e+00 -4.92756605e-01 5.40800244e-02 1.03090417e+00
1.56734395e+00 1.08070478e-01 3.14624578e-01 3.11029106e-01
8.08624208e-01 2.87911981e-01 6.74989760e-01 8.38806033e-01
5.61490893e-01 6.95673764e-01 8.84338379e-01 2.33490363e-01
-3.44443880e-02 7.98470899e-03 5.05161166e-01 5.47073543e-01
-1.62112996e-01 -2.57983059e-01 -6.19909465e-01 4.93790686e-01
-2.01706886e+00 -9.11644995e-01 3.00937355e-01 2.12204790e+00
5.56116045e-01 1.33079335e-01 3.72071356e-01 6.61567375e-02
3.69772047e-01 4.32206213e-01 -1.16835809e+00 -3.81696820e-01
2.70760477e-01 -1.28385708e-01 7.17898309e-01 9.90352780e-02
-1.02700901e+00 7.74177253e-01 5.52000523e+00 5.70409536e-01
-1.14648485e+00 1.93708032e-01 8.37800324e-01 -9.70676839e-01
-6.88349921e-03 -4.58254606e-01 -3.55283439e-01 8.66470277e-01
1.20224440e+00 -2.63383836e-01 1.11842561e+00 7.24833310e-01
8.50114465e-01 -3.36825103e-01 -1.36682284e+00 1.23680222e+00
-3.52073550e-01 -1.59692645e+00 -1.18710421e-01 1.94631405e-02
6.76134348e-01 1.54617578e-01 1.19702600e-01 -6.19201660e-02
9.59646106e-02 -7.81599283e-01 8.47849667e-01 3.69683743e-01
7.46090651e-01 -9.87227321e-01 1.96120664e-01 2.38098174e-01
-1.40027070e+00 -5.48882306e-01 -2.97745496e-01 1.09025188e-01
1.53882563e-01 6.99629188e-01 -4.43718553e-01 -9.24163908e-02
1.13561857e+00 6.14687443e-01 -2.18587980e-01 7.49458373e-01
2.20572561e-01 3.25118482e-01 -2.24846542e-01 8.21323842e-02
2.20894665e-01 -2.38021120e-01 4.94432122e-01 1.11277747e+00
1.67217419e-01 3.97116423e-01 3.80306453e-01 6.60020471e-01
-6.08653605e-01 -1.27390280e-01 -1.96616232e-01 -1.20622516e-01
6.66502535e-01 1.28575385e+00 -7.99287319e-01 -2.40474269e-01
-3.13416958e-01 9.73275065e-01 4.20728117e-01 3.12718302e-01
-1.07456911e+00 2.56013095e-01 1.02311027e+00 2.20975652e-01
3.43735904e-01 -5.56357265e-01 -4.51823950e-01 -8.49692225e-01
2.17492342e-01 -5.25456309e-01 4.51861411e-01 -5.10459244e-01
-6.57040775e-01 2.81810910e-01 -2.56088495e-01 -1.01690280e+00
-2.60423690e-01 -1.32224545e-01 -4.00496453e-01 3.20389897e-01
-1.78380609e+00 -5.14921308e-01 -6.26189351e-01 7.22468972e-01
1.23658526e+00 -1.87094912e-01 2.14942411e-01 4.10968035e-01
-9.49216008e-01 4.58075434e-01 -3.20840441e-02 -1.68569967e-01
4.16680396e-01 -9.42101538e-01 2.18556136e-01 1.01054859e+00
-9.11505297e-02 -2.09786460e-01 6.11158967e-01 -3.72794956e-01
-2.05142045e+00 -1.37566233e+00 1.61138121e-02 -8.11311826e-02
6.17470384e-01 -3.46869230e-02 -7.41235137e-01 3.26278731e-02
1.72576740e-01 4.81653094e-01 3.79135668e-01 -6.84934855e-01
6.90144533e-03 -3.70576650e-01 -1.09797895e+00 5.42160749e-01
1.11331236e+00 -3.21328104e-01 1.77915350e-01 2.52959639e-01
7.53544927e-01 -3.43654126e-01 -9.37779963e-01 2.51518369e-01
5.16123712e-01 -8.87836337e-01 8.33838701e-01 -4.78757977e-01
7.50444055e-01 -2.69375652e-01 -3.09076607e-02 -9.75443065e-01
-2.98936188e-01 -7.05809057e-01 -9.64896142e-01 7.66366184e-01
-6.69516325e-02 2.09227711e-01 9.82657373e-01 8.20380151e-01
-1.90375354e-02 -1.06073141e+00 -1.09160471e+00 -2.64481723e-01
-6.13961458e-01 -3.78690392e-01 1.63611755e-01 5.42483807e-01
-4.73260492e-01 -1.51643166e-02 -2.47433245e-01 3.07776093e-01
9.73904490e-01 1.02526084e-01 5.83680570e-01 -7.38446712e-01
-3.09952378e-01 -6.17867410e-01 -2.83324063e-01 -9.90486681e-01
1.10166721e-01 -4.39171761e-01 8.54246095e-02 -1.13855720e+00
1.14725791e-01 -2.15025082e-01 -3.36158037e-01 1.19938679e-01
-2.84561396e-01 1.79304466e-01 4.59024012e-01 3.61082554e-01
-1.11497104e+00 4.98011172e-01 1.11486983e+00 -6.38499781e-02
-4.79791105e-01 -2.46926427e-01 -2.09485427e-01 5.49046516e-01
5.60525417e-01 -3.90085727e-01 -6.18844688e-01 -7.41242409e-01
1.90514475e-01 2.90479511e-01 2.66401500e-01 -1.00916779e+00
5.07600546e-01 -5.25309741e-01 5.90990067e-01 -4.08448875e-01
1.61107913e-01 -1.07279718e+00 -1.27370909e-01 4.71380085e-01
-4.08797622e-01 3.74662876e-01 -1.85703114e-02 7.61639237e-01
8.60384777e-02 2.71113008e-01 9.53806520e-01 1.52428774e-02
-1.08934712e+00 8.18084955e-01 -5.12887418e-01 2.43503347e-01
1.26523447e+00 -3.00789058e-01 -9.87725705e-02 2.13681050e-02
-4.06443745e-01 4.90364790e-01 4.10245448e-01 4.50042367e-01
7.20987260e-01 -1.07620764e+00 -5.16364157e-01 6.53923675e-02
-3.07500362e-01 3.44443858e-01 4.79853660e-01 6.39726460e-01
-6.55470729e-01 -8.28451663e-03 -4.55244303e-01 -8.11821282e-01
-1.21308374e+00 5.81799626e-01 4.23008174e-01 -3.28019351e-01
-5.13150096e-01 4.14948076e-01 -1.91129550e-01 5.68676412e-01
4.21931148e-01 -2.69837797e-01 -2.39448342e-02 3.34248580e-02
4.50543404e-01 1.01995075e+00 1.31880328e-01 -2.50063717e-01
-2.64542192e-01 3.03911090e-01 5.58073148e-02 1.54326677e-01
1.31872129e+00 -4.69347984e-01 2.49105856e-01 1.50616705e-01
1.16052246e+00 -6.04622126e-01 -2.12718940e+00 -8.74197036e-02
-1.26172006e-01 -6.13157153e-01 6.85606241e-01 -3.71340722e-01
-1.57025576e+00 6.68519497e-01 7.90460348e-01 -1.94016546e-02
1.64857614e+00 -3.60082567e-01 1.15724194e+00 1.57730341e-01
2.78679132e-01 -1.61321568e+00 3.63813192e-01 1.40616566e-01
3.72655511e-01 -1.26549900e+00 2.13317126e-01 -8.08078721e-02
-5.14947295e-01 1.07871664e+00 8.40983450e-01 -1.56141728e-01
3.69115710e-01 5.59227347e-01 -3.92844528e-01 7.78602511e-02
-1.14049757e+00 1.90881148e-01 3.51989232e-02 2.12194607e-01
1.13136679e-01 -1.27755955e-01 -1.01508461e-01 7.35295489e-02
1.96348950e-01 2.48310104e-01 4.90106404e-01 7.42219090e-01
-4.27776963e-01 -4.14821178e-01 -1.62283063e-01 5.63601851e-01
-5.35344779e-01 2.33106449e-01 1.83234423e-01 2.58786947e-01
1.44718677e-01 9.34844077e-01 5.32715440e-01 -2.77018636e-01
1.83557168e-01 -2.63737589e-01 3.64813775e-01 -3.64425057e-03
-5.38268626e-01 6.45961910e-02 -2.53120601e-01 -1.10897136e+00
-6.93902493e-01 -5.03874183e-01 -1.52134764e+00 -6.04469657e-01
3.05227101e-01 -3.12931657e-01 8.67518961e-01 1.02988029e+00
6.65711582e-01 9.17284012e-01 8.69313061e-01 -1.03473723e+00
-3.51699978e-01 -3.02473396e-01 -9.80813056e-02 3.37174535e-01
5.87332189e-01 -5.07465005e-01 -1.26379654e-01 3.46919954e-01] | [8.978630065917969, 0.1571565419435501] |
c666bf34-ab03-4fe7-94c4-b7a298ba641e | backbone-is-all-your-need-a-simplified | 2203.05328 | null | https://arxiv.org/abs/2203.05328v2 | https://arxiv.org/pdf/2203.05328v2.pdf | Backbone is All Your Need: A Simplified Architecture for Visual Object Tracking | Exploiting a general-purpose neural architecture to replace hand-wired designs or inductive biases has recently drawn extensive interest. However, existing tracking approaches rely on customized sub-modules and need prior knowledge for architecture selection, hindering the tracking development in a more general system. This paper presents a Simplified Tracking architecture (SimTrack) by leveraging a transformer backbone for joint feature extraction and interaction. Unlike existing Siamese trackers, we serialize the input images and concatenate them directly before the one-branch backbone. Feature interaction in the backbone helps to remove well-designed interaction modules and produce a more efficient and effective framework. To reduce the information loss from down-sampling in vision transformers, we further propose a foveal window strategy, providing more diverse input patches with acceptable computational costs. Our SimTrack improves the baseline with 2.5%/2.6% AUC gains on LaSOT/TNL2K and gets results competitive with other specialized tracking algorithms without bells and whistles. | ['Wanli Ouyang', 'Wei Wu', 'Weihao Gan', 'Bo Li', 'Qiuhong Shen', 'Lei Qiao', 'Lei Bai', 'Peixia Li', 'BoYu Chen'] | 2022-03-10 | null | null | null | null | ['visual-object-tracking'] | ['computer-vision'] | [ 2.49693412e-02 -7.58431926e-02 -1.96345374e-01 -5.49014509e-02
-4.65368569e-01 -8.34761679e-01 4.05680299e-01 -5.55657387e-01
-6.23509705e-01 3.91812891e-01 6.22051656e-02 -8.97690579e-02
1.72971580e-02 -2.36810729e-01 -8.49966645e-01 -7.01520681e-01
2.94858694e-01 1.31645754e-01 8.58259916e-01 8.24484322e-03
8.91535208e-02 3.06876689e-01 -1.64952230e+00 1.99268639e-01
9.83293891e-01 1.01536775e+00 4.68969733e-01 3.82175744e-01
-9.43531394e-02 3.04781497e-01 -5.70952356e-01 -6.52397931e-01
6.62665844e-01 -1.06374741e-01 -3.01234633e-01 -2.70653367e-01
1.26359618e+00 -1.77971184e-01 -2.31648296e-01 1.03458178e+00
8.95575881e-01 -1.45562440e-01 3.83569330e-01 -1.22750247e+00
-4.88613665e-01 7.08872139e-01 -8.54427934e-01 1.64678186e-01
-1.18975565e-01 3.78651112e-01 9.80479121e-01 -9.59496081e-01
6.67568564e-01 1.12138057e+00 1.02889824e+00 7.22715378e-01
-1.39819670e+00 -8.72115076e-01 6.27439082e-01 -4.14694436e-02
-1.14036930e+00 -4.35329169e-01 5.90102494e-01 -4.00266916e-01
8.63094389e-01 5.26481830e-02 8.83468926e-01 1.23358500e+00
6.30724505e-02 1.09802866e+00 7.65898883e-01 -2.59326100e-01
-3.86756100e-02 1.36589199e-01 9.41236913e-02 7.24912405e-01
6.30234778e-01 1.99141681e-01 -6.51762366e-01 1.19156949e-01
9.38052595e-01 9.56538394e-02 -4.15516734e-01 -1.04000139e+00
-1.29600561e+00 5.20597637e-01 8.62221599e-01 1.76210254e-01
-2.43220195e-01 3.95910621e-01 4.10928905e-01 1.20654851e-01
7.88607821e-03 6.77540958e-01 -4.21200335e-01 1.21688582e-02
-1.28629577e+00 1.49924904e-01 3.61520916e-01 1.19161510e+00
2.66946435e-01 5.52710779e-02 -5.60950994e-01 7.13804424e-01
2.78144777e-01 6.89552009e-01 3.52549434e-01 -9.76093888e-01
3.53900403e-01 7.48351157e-01 2.46655464e-01 -6.03518128e-01
-5.38006902e-01 -1.12821805e+00 -3.04795355e-01 2.78219134e-01
7.23105609e-01 -2.43768170e-01 -1.15964937e+00 1.84135401e+00
3.98761451e-01 -8.67913663e-02 -4.54354435e-01 1.10902631e+00
6.04349732e-01 -2.31869631e-02 -2.35146135e-02 3.38489950e-01
1.53917897e+00 -1.48910308e+00 -4.36514676e-01 -2.74744213e-01
4.41673279e-01 -7.50125110e-01 1.20273316e+00 4.63861644e-01
-9.63507056e-01 -5.62403560e-01 -1.09455538e+00 -1.42847419e-01
-2.25489914e-01 6.93329215e-01 7.67775416e-01 6.22178316e-01
-1.23095632e+00 4.21020091e-01 -1.09993780e+00 -3.94690096e-01
7.16128707e-01 8.09692860e-01 -1.57914519e-01 3.22348058e-01
-5.86817265e-01 6.55844927e-01 7.78919086e-02 8.77318010e-02
-5.86829364e-01 -9.14421976e-01 -4.61137086e-01 -4.04723249e-02
6.95910394e-01 -9.35960531e-01 1.27849483e+00 -7.75514364e-01
-1.54749358e+00 5.12352586e-01 -4.40441221e-02 -5.68262577e-01
5.98113358e-01 -7.86535561e-01 -7.72445574e-02 -9.27401856e-02
-2.91910935e-02 1.13717198e+00 1.20396304e+00 -8.25151503e-01
-7.82909095e-01 -1.64627537e-01 1.88654229e-01 2.27909014e-02
-5.99160254e-01 -1.29385561e-01 -1.02983201e+00 -7.88000166e-01
-2.67719720e-02 -1.18881238e+00 -3.99303548e-02 3.56074512e-01
-4.05784696e-01 -5.58165349e-02 6.51300728e-01 -2.66491383e-01
1.34708464e+00 -2.04837823e+00 2.51905620e-01 -1.10577308e-01
5.02992868e-01 5.63819051e-01 -4.41010594e-01 -2.46805530e-02
2.00294778e-01 -3.86119306e-01 2.37950280e-01 -4.17240232e-01
9.91893560e-02 -2.20819250e-01 -2.32778847e-01 3.10808450e-01
1.53281335e-02 9.54933167e-01 -7.75973082e-01 -3.90299290e-01
9.11757872e-02 5.37823081e-01 -9.85427320e-01 -2.01587319e-01
-2.61332899e-01 2.56084353e-01 -3.96687090e-01 8.27438414e-01
5.78811944e-01 -5.21894813e-01 -1.49034753e-01 -6.28461480e-01
-3.82098347e-01 2.26615861e-01 -8.39845300e-01 2.01459527e+00
-4.12856907e-01 8.65356445e-01 1.42473370e-01 -3.32186252e-01
7.85187066e-01 -1.22021675e-01 4.09430385e-01 -7.45647728e-01
2.87758231e-01 1.62709042e-01 2.16722533e-01 -9.07948390e-02
3.25692356e-01 5.15602469e-01 1.42748907e-01 1.67603403e-01
1.30753770e-01 5.12613654e-01 1.32919922e-01 6.19368255e-02
1.08792865e+00 7.43190408e-01 -1.40003800e-01 -2.06320345e-01
2.47877166e-01 1.38577446e-01 6.47937596e-01 8.59585583e-01
-3.84853750e-01 6.26791894e-01 9.45213214e-02 -4.32594180e-01
-7.67661035e-01 -9.73814249e-01 -3.19780260e-02 1.33812296e+00
1.85725629e-01 -5.89251876e-01 -8.23485553e-01 -9.08748627e-01
1.01010561e-01 3.23248833e-01 -6.05370998e-01 -1.29113883e-01
-6.30903602e-01 -4.43427175e-01 6.42986834e-01 8.21242511e-01
5.57925403e-01 -7.98465908e-01 -1.11917305e+00 1.06943071e-01
6.17461093e-02 -1.04075801e+00 -9.09739077e-01 3.04521292e-01
-8.93876433e-01 -9.32686210e-01 -9.73148525e-01 -5.78754365e-01
6.41646981e-01 4.79387373e-01 9.73227024e-01 -2.42784798e-01
-4.22478706e-01 9.11116004e-02 -9.94211882e-02 -3.09354812e-01
4.15576786e-01 5.48295796e-01 1.27642483e-01 -6.66163936e-02
4.77547914e-01 -3.62377763e-01 -9.67002571e-01 5.70534468e-01
-3.30740064e-01 7.84004703e-02 1.02266002e+00 8.97835970e-01
3.17716330e-01 -4.80033427e-01 1.74956024e-01 -6.31367624e-01
1.86347127e-01 1.32379770e-01 -1.08541930e+00 2.86022961e-01
-7.20379770e-01 4.28082019e-01 5.70434809e-01 -8.59275818e-01
-9.88227367e-01 4.88453865e-01 3.73154491e-01 -8.75817418e-01
2.78694928e-01 -1.64904475e-01 -7.34299794e-02 -4.54902738e-01
8.83577585e-01 -9.83478967e-03 -7.50794038e-02 -6.54812038e-01
3.78438175e-01 1.95828557e-01 6.01141512e-01 -3.86551410e-01
9.43356752e-01 4.70176101e-01 -2.12712303e-01 -3.59658211e-01
-9.43593204e-01 -3.98105025e-01 -6.30076408e-01 -5.37511781e-02
6.18404448e-01 -1.01232898e+00 -8.72749805e-01 3.35763007e-01
-9.54683542e-01 -2.90807575e-01 -1.43180460e-01 7.04303443e-01
-8.06984007e-02 3.42227258e-02 -2.28854880e-01 -6.14145517e-01
-5.53828835e-01 -1.31313753e+00 1.18454123e+00 5.44841945e-01
-2.15725884e-01 -4.89119262e-01 7.34298229e-02 2.32702136e-01
7.56179333e-01 -1.76266253e-01 2.51224160e-01 -4.96281654e-01
-1.03686702e+00 -1.74331665e-01 -3.58453482e-01 4.06723842e-02
-8.77711847e-02 -2.84814071e-02 -1.10550833e+00 -5.57631493e-01
-4.90589499e-01 -7.21043795e-02 1.15388119e+00 5.81097007e-01
8.49263012e-01 -5.15456544e-03 -7.99373806e-01 1.04749072e+00
1.11540020e+00 2.88334545e-02 1.88607082e-01 4.84099746e-01
7.88270593e-01 3.61674696e-01 5.52785158e-01 1.63295329e-01
2.04002634e-01 1.05397093e+00 4.09218788e-01 -1.59622580e-01
-4.97207493e-01 -2.04616770e-01 7.88285077e-01 3.16651225e-01
-1.77178293e-01 -5.39810322e-02 -7.38324106e-01 5.54715633e-01
-1.94461536e+00 -6.65453672e-01 1.07601374e-01 2.25182509e+00
7.52226889e-01 2.91662186e-01 5.47389448e-01 -5.65692186e-01
7.99985528e-01 -3.13006938e-02 -9.30630207e-01 2.45183110e-01
7.79902935e-02 1.45341411e-01 8.89867604e-01 1.27272189e-01
-1.20422268e+00 1.21025944e+00 6.32228279e+00 7.60994136e-01
-1.33737707e+00 5.66253774e-02 -2.14758739e-01 -7.88831174e-01
2.75123548e-02 5.92897460e-02 -1.48861730e+00 4.58263963e-01
5.42671740e-01 3.26617360e-01 3.06515396e-01 9.10934627e-01
-1.73234791e-01 1.17208473e-01 -1.11199427e+00 1.15872180e+00
-1.04887530e-01 -1.24183691e+00 -7.01940358e-02 1.92707136e-01
5.04306793e-01 4.29161876e-01 2.93667972e-01 3.23975325e-01
3.99450153e-01 -5.97675145e-01 8.12845051e-01 5.29993355e-01
6.60906315e-01 -2.98943937e-01 4.51643914e-01 -7.04176351e-02
-1.30566776e+00 -1.18455172e-01 -3.69828254e-01 3.35936427e-01
-1.19510144e-01 3.31778347e-01 -6.97615385e-01 3.66177708e-01
9.21636820e-01 5.27401686e-01 -9.38304365e-01 1.70223951e+00
-1.57721803e-01 3.94382089e-01 -6.92045450e-01 -1.78243175e-01
1.92267686e-01 8.96744952e-02 7.62149096e-01 1.06024587e+00
2.41978794e-01 -7.32622445e-01 5.91058433e-02 8.37073088e-01
-1.00705042e-01 -1.55690312e-01 -4.70430106e-01 1.01508470e-02
7.43986070e-01 1.46980524e+00 -6.75469041e-01 -3.78302149e-02
-6.82803333e-01 9.11534905e-01 3.60901147e-01 2.97069579e-01
-1.24708557e+00 -5.00492990e-01 5.92400551e-01 2.46562541e-01
8.03688288e-01 -4.03170697e-02 -1.76946178e-01 -1.23636508e+00
1.71247602e-01 -8.41391683e-01 2.04298019e-01 -4.29823637e-01
-9.33572292e-01 7.65043259e-01 -1.37731507e-01 -1.51983905e+00
-1.02428786e-01 -7.41340637e-01 -4.50578243e-01 6.12627864e-01
-1.31689978e+00 -1.39826810e+00 -3.92215610e-01 5.73669255e-01
3.41780961e-01 -3.22233588e-01 3.21750790e-01 4.72435445e-01
-8.50323737e-01 1.32401514e+00 -9.14927572e-02 1.13567941e-01
1.17751360e+00 -1.29688156e+00 5.43414831e-01 1.02007413e+00
2.76867419e-01 1.06610036e+00 5.14200568e-01 -5.24691224e-01
-1.51326680e+00 -1.08459032e+00 3.84611845e-01 -6.77226424e-01
5.97242832e-01 -6.49023294e-01 -5.89922965e-01 5.79760313e-01
2.99858242e-01 1.05625495e-01 3.97312731e-01 3.97787094e-01
-5.97546995e-01 -4.18509245e-01 -7.77695179e-01 1.01373231e+00
1.57192767e+00 -2.24781305e-01 -4.19272691e-01 -8.70668292e-02
5.98089457e-01 -3.84813875e-01 -6.11579955e-01 3.29106301e-01
1.03229845e+00 -7.91134477e-01 1.10670483e+00 -2.51317948e-01
-2.56782234e-01 -6.36911988e-01 2.37317428e-01 -9.99578238e-01
-6.46524370e-01 -9.87559259e-01 -3.82590950e-01 1.03684676e+00
5.21517694e-01 -5.17880619e-01 1.06396687e+00 4.44003016e-01
-2.14701846e-01 -7.16070354e-01 -6.79598689e-01 -1.02734721e+00
-2.60259032e-01 -9.53212474e-03 3.70891035e-01 4.09935594e-01
-3.66160423e-01 5.66420376e-01 -3.11563671e-01 4.48640063e-02
9.34845150e-01 3.07688326e-01 1.08326817e+00 -1.41619718e+00
-4.81393397e-01 -7.88058877e-01 -1.79589957e-01 -1.51426733e+00
-3.05866450e-01 -7.66719222e-01 2.81623937e-02 -1.05088878e+00
-8.18959251e-03 -3.65028083e-01 -3.00595790e-01 8.07496190e-01
-7.52486214e-02 4.10303652e-01 3.72918487e-01 3.44161212e-01
-8.15027714e-01 5.83671808e-01 1.39291418e+00 -2.50619147e-02
-4.26721334e-01 -4.83258404e-02 -7.25900292e-01 7.64386892e-01
5.13958037e-01 -3.66502672e-01 -5.74217379e-01 -6.29476190e-01
1.02802522e-01 -5.92716575e-01 4.07737911e-01 -1.32935727e+00
6.41672671e-01 3.17169845e-01 5.48228145e-01 -8.11228931e-01
2.93656647e-01 -8.63542795e-01 1.40716255e-01 4.83662188e-01
-2.13363528e-01 2.05684751e-02 4.41747963e-01 4.91658717e-01
3.94418724e-02 -4.84526902e-02 8.07197034e-01 1.63326532e-01
-5.85733473e-01 2.92866528e-01 8.49707797e-02 -4.65506837e-02
9.18209553e-01 -5.02622664e-01 -5.36949456e-01 1.66827053e-01
-6.05532467e-01 3.99198681e-01 7.14429140e-01 5.69045246e-01
1.91196963e-01 -1.22796392e+00 -2.30779305e-01 2.85297215e-01
1.29486233e-01 -1.04096368e-01 -7.08775297e-02 1.20507312e+00
-2.08148763e-01 6.56169415e-01 -4.22780037e-01 -7.87415981e-01
-1.20492291e+00 5.02585828e-01 4.31790769e-01 -1.56371400e-01
-8.94940972e-01 1.14112437e+00 5.57967961e-01 -9.27288234e-02
6.66566491e-01 -4.90449220e-01 -1.51421964e-01 2.03551143e-01
4.34148610e-01 2.36772612e-01 -1.12518212e-02 -2.62695432e-01
-5.58416963e-01 8.11138093e-01 -6.01846159e-01 -2.31107734e-02
1.09483290e+00 1.30474782e-02 4.11338568e-01 2.96985097e-02
7.09014297e-01 1.79014131e-01 -1.81256664e+00 -4.20771569e-01
8.38429928e-02 -4.28908795e-01 2.94964820e-01 -9.91040945e-01
-1.41200984e+00 7.89819777e-01 8.40728164e-01 -1.39046535e-01
1.12431991e+00 -1.12992972e-01 8.15330029e-01 4.91471559e-01
4.14529651e-01 -1.03557515e+00 2.30497364e-02 2.82503337e-01
8.26363504e-01 -8.75148773e-01 -7.95190930e-02 -3.22817355e-01
-4.40548867e-01 8.48464429e-01 1.07333434e+00 -2.63955623e-01
2.17385069e-01 4.02758569e-01 1.77782886e-02 -3.84207219e-01
-8.19383860e-01 -5.09327710e-01 6.43303394e-01 5.60262680e-01
4.17456806e-01 -3.27336431e-01 -5.08331805e-02 6.09481096e-01
-1.08039558e-01 5.82784303e-02 -1.01068661e-01 8.08466673e-01
-2.78331727e-01 -1.22069383e+00 -3.32327664e-01 4.11540151e-01
-3.54988873e-01 -2.01705039e-01 -5.06115913e-01 9.46388841e-01
1.72939971e-01 3.76033098e-01 -3.57650332e-02 -5.07843375e-01
4.60834414e-01 1.00127757e-01 7.45027304e-01 -4.58940536e-01
-9.61458027e-01 5.17891407e-01 -6.23482093e-02 -9.15843070e-01
-3.75115275e-01 -6.51561558e-01 -9.32733715e-01 5.34105673e-02
-7.73593187e-01 -2.02919200e-01 4.36838746e-01 4.73003358e-01
9.07894611e-01 6.57321155e-01 -3.63883143e-03 -9.27004099e-01
-5.79245925e-01 -8.40593040e-01 -5.86658604e-02 -8.98571238e-02
4.87999797e-01 -1.36673009e+00 -3.63153033e-02 -1.32663384e-01] | [6.2879414558410645, -2.116497039794922] |
ebf774ad-7273-4900-9f77-7bed09a1b64a | synthetic-data-for-face-recognition-current | 2305.01021 | null | https://arxiv.org/abs/2305.01021v1 | https://arxiv.org/pdf/2305.01021v1.pdf | Synthetic Data for Face Recognition: Current State and Future Prospects | Over the past years, deep learning capabilities and the availability of large-scale training datasets advanced rapidly, leading to breakthroughs in face recognition accuracy. However, these technologies are foreseen to face a major challenge in the next years due to the legal and ethical concerns about using authentic biometric data in AI model training and evaluation along with increasingly utilizing data-hungry state-of-the-art deep learning models. With the recent advances in deep generative models and their success in generating realistic and high-resolution synthetic image data, privacy-friendly synthetic data has been recently proposed as an alternative to privacy-sensitive authentic data to overcome the challenges of using authentic data in face recognition development. This work aims at providing a clear and structured picture of the use-cases taxonomy of synthetic face data in face recognition along with the recent emerging advances of face recognition models developed on the bases of synthetic data. We also discuss the challenges facing the use of synthetic data in face recognition development and several future prospects of synthetic data in the domain of face recognition. | ['Naser Damer', 'Julian Fierrez', 'Vitomir Struc', 'Fadi Boutros'] | 2023-05-01 | null | null | null | null | ['face-recognition'] | ['computer-vision'] | [ 3.39898556e-01 2.02225819e-01 3.31071764e-01 -9.86466587e-01
-5.42185068e-01 -1.50696516e-01 7.15019166e-01 -6.94223881e-01
-4.87892628e-01 7.39953816e-01 -1.38151824e-01 1.91399232e-01
-4.21700962e-02 -8.44263256e-01 -3.93619448e-01 -5.34565806e-01
1.68854982e-01 6.40466511e-01 -8.21072757e-01 -1.84888884e-01
-4.96676452e-02 9.46385205e-01 -1.91070020e+00 4.13173020e-01
6.91580415e-01 1.09715283e+00 -4.89070773e-01 2.09177688e-01
4.55888622e-02 1.56626880e-01 -7.44457126e-01 -1.12539268e+00
5.57110250e-01 -4.79434460e-01 -4.48766708e-01 2.03357980e-01
6.36693478e-01 -6.32142127e-01 -4.05987948e-01 7.66065359e-01
9.91906226e-01 -2.08394185e-01 3.81431520e-01 -1.67258978e+00
-9.64019835e-01 4.96457927e-02 -3.04318219e-01 -2.45662898e-01
3.19555670e-01 2.42774516e-01 2.28888661e-01 -1.19929147e+00
6.59059167e-01 1.21052969e+00 6.57512009e-01 1.15815246e+00
-1.14276457e+00 -9.72714245e-01 -6.01619542e-01 9.90661234e-02
-1.55893493e+00 -9.97771323e-01 6.04065120e-01 -2.93562025e-01
6.95424676e-01 1.76957503e-01 5.13002932e-01 1.71383274e+00
-1.22520171e-01 5.34421563e-01 1.08664238e+00 -4.52673793e-01
-1.01421280e-02 4.64466840e-01 -4.37698185e-01 5.19715011e-01
5.53871870e-01 4.65378374e-01 -7.36971915e-01 -2.15083987e-01
6.21077359e-01 -1.52924165e-01 -1.20343968e-01 -2.93564141e-01
-6.33119583e-01 7.95945168e-01 1.91292129e-02 2.25364238e-01
-4.63751704e-01 -1.23024985e-01 3.34278703e-01 3.32549393e-01
4.44166631e-01 3.46631825e-01 -2.21257433e-01 5.86166941e-02
-1.23949611e+00 3.64893526e-01 7.51937509e-01 7.55411863e-01
4.84858930e-01 7.74113297e-01 1.40839502e-01 8.84853363e-01
3.66197020e-01 6.05412781e-01 3.39795023e-01 -6.88909054e-01
1.63054317e-01 5.33667207e-01 4.55731787e-02 -9.40488935e-01
-2.49755513e-02 -4.54662234e-01 -9.40299511e-01 2.27580041e-01
5.59601128e-01 -1.21117212e-01 -1.10713375e+00 1.63358033e+00
2.00679928e-01 3.42879817e-02 2.55765736e-01 4.38238025e-01
8.11308265e-01 2.77399927e-01 6.53016642e-02 -4.86402251e-02
1.18390775e+00 -6.28740266e-02 -7.29225218e-01 1.56895556e-02
2.30137214e-01 -5.73456168e-01 6.75301075e-01 3.05055439e-01
-8.84101093e-01 -6.43593967e-01 -8.44630063e-01 1.04864329e-01
-6.78009629e-01 1.92933232e-01 6.21739268e-01 1.56800795e+00
-1.19295549e+00 2.91613162e-01 -5.77674389e-01 -4.36691940e-01
1.30166852e+00 7.47121036e-01 -8.52679133e-01 -5.79702199e-01
-1.21560824e+00 9.33669686e-01 1.67957395e-01 5.48710644e-01
-1.06329322e+00 -7.54686415e-01 -8.95948350e-01 3.60003002e-02
1.19494103e-01 -3.96914631e-01 7.03099668e-01 -1.05902517e+00
-1.28770113e+00 1.20269513e+00 1.38669834e-01 -3.99159819e-01
6.62292600e-01 -2.74244063e-02 -7.90096700e-01 -2.42738992e-01
-2.81408459e-01 7.67242432e-01 7.94822454e-01 -1.13119912e+00
-1.64640456e-01 -8.50606561e-01 -4.70101476e-01 -2.84262538e-01
-4.60496366e-01 2.19546586e-01 -8.15558583e-02 -3.40434790e-01
-4.14417237e-01 -9.84710574e-01 2.81800758e-02 2.29757279e-01
-4.21132287e-03 1.81498110e-01 1.15960121e+00 -7.00089872e-01
5.95797896e-01 -2.08356500e+00 -3.32650065e-01 2.57479191e-01
-6.97804540e-02 9.93766427e-01 -2.55996585e-01 3.94263536e-01
-3.16813380e-01 2.36644670e-01 -1.78599745e-01 -4.10809577e-01
-1.18433043e-01 1.97609246e-01 -3.02373052e-01 3.62625808e-01
5.93445837e-01 8.83897126e-01 -5.65980196e-01 -1.58611059e-01
1.80025548e-01 1.07872176e+00 -3.11320364e-01 2.86447555e-01
1.41985893e-01 6.48633659e-01 -3.90814453e-01 8.08633029e-01
1.16809011e+00 3.39183778e-01 7.54205063e-02 7.12457299e-02
2.37271965e-01 -3.33328873e-01 -9.23378170e-01 1.26642072e+00
-2.77676433e-01 6.69601679e-01 8.83057192e-02 -7.41643071e-01
1.18676531e+00 6.50322378e-01 4.87595826e-01 -6.63337767e-01
2.20334336e-01 2.89819658e-01 1.78792551e-01 -4.31012422e-01
2.35827252e-01 -1.01156853e-01 3.23806286e-01 3.45871300e-01
3.17234933e-01 -1.91542119e-01 -1.23683669e-01 -3.89866710e-01
5.20942807e-01 4.33455288e-01 -3.99850681e-02 -5.08418307e-02
7.34126270e-01 -2.93827981e-01 5.92475832e-01 2.95448035e-01
-2.92220294e-01 6.47335172e-01 1.90057576e-01 -8.49321544e-01
-9.61978793e-01 -6.90569878e-01 -2.64444619e-01 3.78304154e-01
-7.30082631e-01 -1.14583075e-01 -1.06593347e+00 -6.57591760e-01
1.13353118e-01 7.16310263e-01 -8.14190090e-01 -3.16540718e-01
-4.46223885e-01 -9.22976851e-01 1.12799239e+00 3.03817570e-01
9.86869037e-01 -1.07105255e+00 -4.26627606e-01 -4.23222743e-02
2.30161563e-01 -1.05925202e+00 1.94770470e-01 -3.66399586e-01
-4.46519643e-01 -1.00296021e+00 -9.03007090e-01 -4.44439262e-01
8.20853412e-01 -1.85081378e-01 8.89177859e-01 -6.16833800e-04
-7.57438838e-01 4.42723900e-01 -3.04010771e-02 -7.49099493e-01
-7.30360568e-01 -2.43286535e-01 3.21028888e-01 4.17876363e-01
6.00637913e-01 -3.39548677e-01 -5.81425667e-01 4.16704625e-01
-1.06746411e+00 -3.06713313e-01 3.95936519e-01 1.04068995e+00
2.55274445e-01 -1.04597837e-01 1.04904902e+00 -9.48659241e-01
7.12219238e-01 -3.62538576e-01 -5.37776113e-01 4.48462486e-01
-8.72846663e-01 -1.56853572e-01 5.28015018e-01 -1.81121573e-01
-1.42136741e+00 7.88508430e-02 -3.99948478e-01 -4.44693267e-01
-3.61986250e-01 2.37813279e-01 -5.10549486e-01 -5.04474401e-01
7.65338719e-01 3.26101065e-01 5.11935353e-01 -2.32905298e-01
1.90822527e-01 8.63837540e-01 3.33691925e-01 -5.41373372e-01
6.67085052e-01 4.04385924e-01 2.85424650e-01 -1.00515234e+00
-3.37194264e-01 3.67724866e-01 -6.41247451e-01 -2.69121081e-01
5.09202242e-01 -6.84193790e-01 -6.98296905e-01 9.98916268e-01
-8.28107774e-01 3.68477285e-01 -3.47559869e-01 2.01280028e-01
-4.07750010e-01 2.21766785e-01 -2.31108367e-01 -1.09242070e+00
-5.05559683e-01 -1.38426673e+00 9.17277098e-01 3.90576869e-01
-1.33516612e-02 -5.99589467e-01 -2.66889542e-01 6.91938460e-01
7.97728598e-01 6.42921984e-01 4.99612808e-01 -7.66002536e-01
-4.58984435e-01 -8.42301250e-01 -5.34723774e-02 6.97133660e-01
3.50537091e-01 1.17410064e-01 -1.53174388e+00 -4.09570128e-01
1.96842358e-01 -6.38207197e-01 3.67843926e-01 5.71586490e-02
1.20541358e+00 -2.52763689e-01 -1.08126268e-01 6.73940301e-01
1.07697260e+00 4.72124547e-01 9.66326058e-01 -2.00535581e-01
3.46757263e-01 9.89699483e-01 5.57326674e-01 3.45958740e-01
8.38415176e-02 5.20780683e-01 3.28883708e-01 5.44740744e-02
-5.36104664e-02 -3.39476794e-01 -7.07136914e-02 -1.75621007e-02
-1.08098142e-01 -2.85474330e-01 -1.03254735e+00 2.54306287e-01
-1.20860505e+00 -1.07185030e+00 3.08992952e-01 2.39371490e+00
5.08003175e-01 -3.33990961e-01 -8.88078734e-02 -2.57194545e-02
5.80543935e-01 -1.27972424e-01 -7.32449293e-01 -4.11921024e-01
-2.87650555e-01 4.70207483e-01 1.18301548e-01 8.42435926e-04
-8.87752831e-01 8.96266580e-01 6.69796610e+00 5.03555298e-01
-1.35351682e+00 -1.97944015e-01 1.23565269e+00 1.56248823e-01
-1.60245523e-01 -4.12125289e-01 -9.25173700e-01 4.19010341e-01
1.21537352e+00 -1.04763217e-01 3.42771113e-01 1.07674444e+00
-7.72041082e-02 1.76879093e-01 -1.20249796e+00 1.14345694e+00
2.95461595e-01 -1.32682669e+00 3.91076237e-01 4.69222963e-01
5.61178505e-01 -2.07085460e-01 7.87204146e-01 1.12099126e-01
-8.13843161e-02 -1.51551640e+00 1.96206510e-01 5.47917783e-01
1.28690207e+00 -8.94543469e-01 8.92318189e-01 1.31098837e-01
-5.32574713e-01 5.09377867e-02 -4.40435499e-01 3.78854841e-01
-3.35663259e-01 3.75461727e-02 -1.22747445e+00 5.66062808e-01
5.45243979e-01 2.58646965e-01 -4.82157707e-01 6.60776198e-01
2.41667002e-01 3.59460533e-01 -3.22704911e-01 3.63882124e-01
-1.36215061e-01 -2.20783919e-01 6.81406036e-02 7.77694404e-01
5.45897782e-01 -6.65120482e-02 -5.91137946e-01 1.11957800e+00
-2.80782998e-01 -4.20278832e-02 -9.06328380e-01 -5.56850731e-01
4.83112097e-01 1.12657762e+00 -2.71574557e-01 1.71476007e-01
-4.16084915e-01 8.36773992e-01 -4.22393531e-03 2.88503736e-01
-4.38179135e-01 1.29302740e-01 8.56094420e-01 1.54055074e-01
-1.11228049e-01 -2.86212638e-02 -2.42387727e-01 -9.57774282e-01
-1.68544978e-01 -1.31495285e+00 3.21547955e-01 -5.29920697e-01
-1.12254345e+00 1.08157384e+00 -8.05120170e-02 -9.90583360e-01
-6.43322647e-01 -8.03695261e-01 -2.87591398e-01 1.24684548e+00
-1.19636464e+00 -1.61464822e+00 -2.61070490e-01 5.68917334e-01
3.00308257e-01 -9.21978951e-01 1.51793492e+00 3.25330287e-01
-5.26560366e-01 1.15585768e+00 2.11793453e-01 3.49283993e-01
7.61486053e-01 -4.15533006e-01 7.91863263e-01 6.94272518e-01
1.89065099e-01 8.46564174e-01 3.80046964e-01 -4.71282065e-01
-1.45906246e+00 -8.61632586e-01 5.43259799e-01 -5.82987368e-01
-1.02825269e-01 -6.08116210e-01 -8.31034660e-01 5.68736851e-01
-1.27385020e-01 1.18201032e-01 1.27836156e+00 -2.26180509e-01
-4.21556562e-01 -5.41095734e-01 -1.99017537e+00 3.51379693e-01
6.35605216e-01 -6.23918772e-01 -1.66470766e-01 1.90346166e-01
-2.26310808e-02 -1.50312603e-01 -9.59738076e-01 5.24171710e-01
8.96401882e-01 -9.43013608e-01 9.09892142e-01 -6.90509379e-01
3.93897444e-02 3.28860432e-02 -9.62002948e-02 -1.12419236e+00
1.87553659e-01 -6.86279178e-01 2.33872697e-01 1.42376912e+00
3.42391074e-01 -7.61738181e-01 1.29562140e+00 1.40510285e+00
3.83817732e-01 -7.08424449e-01 -1.16407919e+00 -3.89510006e-01
2.46216590e-03 -1.10675886e-01 9.31873500e-01 9.40226734e-01
-6.72182620e-01 -3.10704142e-01 -6.39980018e-01 -1.96385831e-01
8.79694045e-01 -2.86895573e-01 8.40716183e-01 -1.45739567e+00
3.16118419e-01 4.03713286e-02 -8.20344508e-01 -1.11595869e-01
2.75784671e-01 -6.14213049e-01 -5.26513875e-01 -1.03485370e+00
-1.09415710e-01 -3.02548915e-01 -9.92976725e-02 4.09850240e-01
2.17968002e-01 6.14958584e-01 9.42427590e-02 -2.55895585e-01
3.94166440e-01 5.74467838e-01 1.25506699e+00 -1.20650632e-02
2.63467342e-01 2.19977781e-01 -7.96765625e-01 3.27119142e-01
6.97100699e-01 -1.77083760e-01 -6.31129622e-01 -2.39428431e-01
-1.46612033e-01 5.12850471e-02 8.41685236e-02 -9.46025312e-01
3.92199531e-02 6.09450676e-02 1.06115091e+00 -4.02148962e-01
9.03097749e-01 -1.18935657e+00 6.89750314e-01 4.58159804e-01
-1.06577866e-01 -1.78869352e-01 4.94409174e-01 2.34655589e-01
-1.69196233e-01 4.48106676e-02 1.00148618e+00 -1.65593192e-01
-6.51113391e-01 6.34198964e-01 -2.81260908e-02 -3.83336753e-01
1.20956230e+00 -6.54168785e-01 -2.64334053e-01 -6.15325212e-01
-7.41780698e-01 -7.92707577e-02 4.83225882e-01 7.62125194e-01
1.01047862e+00 -1.11052084e+00 -1.05005169e+00 1.00220549e+00
1.31608069e-01 -3.50240082e-01 4.28804517e-01 1.24942504e-01
-5.48997641e-01 3.88133794e-01 -9.59746957e-01 -3.36242527e-01
-1.39650142e+00 4.52017397e-01 5.96382558e-01 2.19442815e-01
-9.77080315e-02 9.80595231e-01 1.27111971e-01 -4.35094208e-01
7.55464211e-02 5.54266751e-01 6.00619838e-02 -1.44536020e-02
6.09703302e-01 2.66904414e-01 1.67462230e-01 -9.63964760e-01
-4.81543422e-01 8.74904618e-02 -2.83824295e-01 -9.46910307e-02
1.52377391e+00 2.45233312e-01 -1.49165422e-01 -1.23723134e-01
1.00831461e+00 -3.68576497e-01 -1.03516722e+00 5.52409813e-02
-6.61783889e-02 -9.19178903e-01 -2.19861329e-01 -1.09139872e+00
-1.35767138e+00 9.27099764e-01 9.44919527e-01 -2.71312118e-01
1.08000791e+00 -6.04881644e-01 4.08836782e-01 2.74281830e-01
6.38139904e-01 -8.68215859e-01 -2.57850826e-01 1.09628096e-01
1.08946276e+00 -1.41128647e+00 3.45754884e-02 -3.55815679e-01
-6.85714483e-01 1.14827752e+00 6.81877375e-01 3.26024950e-01
7.24875748e-01 2.63605028e-01 2.00792044e-01 -5.77504486e-02
-2.18973383e-01 2.22931594e-01 2.62642503e-01 1.15068138e+00
5.62435210e-01 -2.24518195e-01 -1.63946986e-01 2.65281379e-01
-1.83835506e-01 2.29854420e-01 2.82672673e-01 8.08938146e-01
2.28421599e-01 -1.84161210e+00 -3.26807141e-01 5.86591065e-01
-6.39572382e-01 1.37718871e-01 -6.49103880e-01 9.71162975e-01
2.09173843e-01 8.77598584e-01 -1.67882323e-01 -4.07450438e-01
1.97387367e-01 4.73425537e-01 5.32490015e-01 -3.31339300e-01
-6.13848329e-01 -5.02558410e-01 8.24384466e-02 -3.65868688e-01
-1.66763917e-01 -7.73175299e-01 -5.83554447e-01 -5.23621321e-01
-2.25855619e-01 1.23428345e-01 1.20926666e+00 6.61261201e-01
7.91648686e-01 1.37527380e-02 5.99041283e-01 -6.86484516e-01
-6.22535646e-01 -9.61920679e-01 -4.95938331e-01 2.25191787e-01
3.22273858e-02 -4.76116240e-01 2.92389542e-02 -8.87572244e-02] | [12.919297218322754, 0.7312828302383423] |
714a14f0-0012-44a1-80da-a69e56267251 | panoptic-segmentation-with-an-end-to-end-cell | null | null | https://doi.org/10.1007/978-3-030-00934-2_27 | https://www.semanticscholar.org/paper/Panoptic-Segmentation-with-an-End-to-End-Cell-R-CNN-Zhang-Song/be403b1a9fa44c860044e79b8d9704cc0b217417 | Panoptic Segmentation with an End-to-End Cell R-CNN for Pathology Image Analysis | The morphological clues of various cancer cells are essential for pathologists to determine the stages of cancers. In order to obtain the quantitative morphological information, we present an end-to-end network for panoptic segmentation of pathology images. Recently, many methods have been proposed, focusing on the semantic-level or instance-level cell segmentation. Unlike existing cell segmentation methods, the proposed network unifies detecting, localizing objects and assigning pixel-level class information to regions with large overlaps such as the background. This unifier is obtained by optimizing the novel semantic loss, the bounding box loss of Region Proposal Network (RPN), the classifier loss of RPN, the background-foreground classifier loss of segmentation Head instead of class-specific loss, the bounding box loss of proposed cell object, and the mask loss of cell object. The results demonstrate that the proposed method not only outperforms state-of-the-art approaches to the 2017 MICCAI Digital Pathology Challenge dataset, but also proposes an effective and end-to-end solution for the panoptic segmentation challenge. | ['Si-Qi Liu', 'Yang song', 'Weidong Cai', 'Heng Huang', 'Haozhe Jia', 'Donghao Zhang', 'Dongnan Liu', 'Yong Xia'] | 2018-09-28 | null | null | null | miccai-2018-2018-9 | ['nuclear-segmentation'] | ['medical'] | [ 3.51651371e-01 2.93875873e-01 -2.40546897e-01 -2.81980932e-01
-9.68500078e-01 -2.65352786e-01 2.56593198e-01 6.29827976e-01
-7.42295623e-01 5.63156962e-01 -2.99026459e-01 -1.16895281e-01
1.23403724e-02 -6.62764370e-01 -3.33798856e-01 -1.25723028e+00
3.01590841e-02 5.57059526e-01 6.46983087e-01 2.89332092e-01
2.51532018e-01 8.25788498e-01 -1.01790893e+00 2.22077966e-01
8.40001523e-01 1.19430280e+00 3.02029103e-01 5.83076119e-01
-3.99253190e-01 4.55424845e-01 -3.05253863e-01 -2.20644906e-01
2.31603861e-01 -2.33357757e-01 -9.61530268e-01 1.63884804e-01
5.19455671e-01 9.31887142e-03 -4.32004593e-02 1.53721273e+00
4.33261067e-01 -2.61457175e-01 9.04322386e-01 -9.41879749e-01
9.67561454e-02 3.68908942e-01 -1.00584483e+00 4.77368265e-01
-5.47616601e-01 2.82528430e-01 8.66178751e-01 -4.79290307e-01
8.64144385e-01 1.09630108e+00 6.85783446e-01 5.31784892e-01
-1.05259919e+00 -5.00392556e-01 2.08893605e-02 1.03302091e-01
-1.68000269e+00 -1.46551393e-02 5.04429698e-01 -4.71052885e-01
4.78433698e-01 3.97791505e-01 5.11445463e-01 2.87683487e-01
4.51490045e-01 1.10436416e+00 9.45301950e-01 -2.88416177e-01
6.63425848e-02 2.01116294e-01 6.32838428e-01 8.17942679e-01
3.38712335e-01 -3.51605117e-01 1.06593944e-01 -1.70911234e-02
7.14221716e-01 3.44257206e-01 -2.90719271e-01 -4.21577841e-01
-1.08609009e+00 4.88358706e-01 5.90509117e-01 3.99647295e-01
-1.26353741e-01 1.10766917e-01 6.10877812e-01 -3.13053429e-01
6.15263700e-01 5.22616878e-02 -3.95265222e-01 4.86144900e-01
-1.32471013e+00 -8.56714919e-02 5.89027405e-01 5.18373430e-01
4.87563729e-01 -5.65783739e-01 -5.31967700e-01 7.28023052e-01
4.41754431e-01 2.32784953e-02 5.30238152e-01 -5.50180972e-01
6.30487427e-02 1.17036629e+00 -2.58092582e-01 -5.95806897e-01
-8.34103048e-01 -6.09358430e-01 -1.08790731e+00 8.84223953e-02
8.73946130e-01 6.19150624e-02 -1.36949778e+00 1.34869552e+00
7.44699955e-01 2.14927673e-01 -2.64142871e-01 9.07347798e-01
8.89913738e-01 4.53658551e-01 1.84242815e-01 -3.47683042e-01
1.66217661e+00 -1.03392828e+00 -6.59641385e-01 -1.90333615e-03
8.29142213e-01 -6.13739729e-01 7.25686371e-01 2.01371402e-01
-1.14567530e+00 -1.86988547e-01 -8.19780648e-01 -9.64007154e-02
-2.96574295e-01 2.90274471e-01 5.32611549e-01 5.32808661e-01
-1.18384600e+00 3.73847246e-01 -1.04727674e+00 -5.75479031e-01
9.83450353e-01 4.80202287e-01 -1.50253206e-01 2.29545590e-02
-4.70556349e-01 6.12069607e-01 6.07945025e-01 6.82684174e-03
-8.65397096e-01 -1.06501257e+00 -5.40634453e-01 9.57785100e-02
4.73360837e-01 -5.78950703e-01 9.02812779e-01 -7.63313949e-01
-1.10691369e+00 1.40994811e+00 -6.17619753e-02 -5.34104645e-01
6.73427761e-01 4.79141653e-01 5.53971231e-02 4.64414328e-01
4.64252494e-02 1.14013922e+00 1.97329864e-01 -1.11148930e+00
-1.17891812e+00 -5.99583209e-01 -4.40823108e-01 1.11935116e-01
-6.30102456e-02 -1.27592161e-01 -7.34840751e-01 -5.95096707e-01
2.92883039e-01 -6.93270922e-01 -6.48442447e-01 5.39788902e-01
-6.87786400e-01 -2.34196767e-01 1.04992855e+00 -7.01631248e-01
9.65778887e-01 -2.32402921e+00 -1.09335639e-01 4.13868695e-01
2.50808477e-01 1.47490397e-01 1.54809490e-01 -6.58137739e-01
2.31682599e-01 2.34677702e-01 -6.12570882e-01 -4.74864006e-01
-2.72484452e-01 -3.30507359e-03 2.61416674e-01 8.64300430e-01
-7.94626772e-02 6.92893326e-01 -7.35979974e-01 -1.22634816e+00
8.63295719e-02 3.40746790e-01 -5.33681095e-01 1.33120596e-01
-1.38450623e-01 4.60056365e-01 -2.82464683e-01 1.02607620e+00
1.00718665e+00 -4.00571257e-01 -1.46251656e-02 -2.80460805e-01
1.29072472e-01 -2.21913010e-01 -1.07368731e+00 1.47444403e+00
1.16932787e-01 3.48709226e-01 7.02740550e-01 -8.51680040e-01
5.83991885e-01 1.10688314e-01 7.83292592e-01 -2.43501738e-01
3.16512942e-01 3.42693418e-01 1.42879918e-01 -2.69633949e-01
1.55300796e-01 -2.91010737e-01 5.68968542e-02 -5.64538725e-02
1.65004078e-02 -1.41336992e-01 3.61217052e-01 1.22725576e-01
9.87123966e-01 -3.48925889e-01 2.18224585e-01 -6.67607069e-01
8.31529915e-01 6.52753562e-02 9.06996191e-01 3.85370076e-01
-6.34416938e-01 8.83685708e-01 7.92023420e-01 -2.70255417e-01
-6.47538304e-01 -1.08564162e+00 -4.94387835e-01 6.69872880e-01
5.32618225e-01 2.73063928e-01 -9.87816036e-01 -1.06103015e+00
-3.33183147e-02 3.24644417e-01 -6.62322581e-01 6.46818280e-02
-6.96424365e-01 -1.24557734e+00 5.90845704e-01 2.84299731e-01
5.13137281e-01 -8.88209283e-01 -5.13718903e-01 2.52656132e-01
-2.29944006e-01 -1.13769197e+00 -6.53184593e-01 3.02538216e-01
-1.19726801e+00 -1.24655926e+00 -1.01621914e+00 -1.17888486e+00
1.28433049e+00 3.84403467e-02 9.81035292e-01 3.50798070e-01
-9.58185434e-01 -1.10710621e-01 -2.45719645e-02 -3.04925323e-01
-3.86954010e-01 8.47941786e-02 -6.18745089e-01 1.25280797e-01
2.25922793e-01 -1.11985818e-01 -9.43218887e-01 4.38701749e-01
-1.01632559e+00 5.31353280e-02 8.58604968e-01 7.73041427e-01
1.35743117e+00 7.12226406e-02 1.89948872e-01 -1.07386744e+00
8.33390504e-02 -2.09319785e-01 -6.78593934e-01 4.59396333e-01
-4.04181182e-01 -3.66107792e-01 3.03655237e-01 -2.37280399e-01
-8.56500030e-01 3.77225876e-01 -2.51551241e-01 -1.34185523e-01
-2.34292135e-01 1.21656321e-01 -5.45268953e-02 -2.92678714e-01
2.51691997e-01 2.84438550e-01 9.34734270e-02 -2.52092332e-01
1.56173423e-01 3.77882153e-01 5.60902774e-01 -6.35423809e-02
2.56158262e-01 1.10635293e+00 3.73699129e-01 -6.03882849e-01
-7.53585160e-01 -9.15152133e-01 -6.55268669e-01 -2.01300308e-01
1.12931871e+00 -6.53574586e-01 -6.00477636e-01 6.26037836e-01
-1.02343786e+00 -1.50251269e-01 -5.03931761e-01 2.70052522e-01
-4.28829521e-01 2.88702250e-01 -9.14825737e-01 -4.83141810e-01
-6.70053780e-01 -1.41891170e+00 1.19311190e+00 5.04279733e-01
-7.41191804e-02 -9.02878284e-01 -1.17838733e-01 4.86949235e-01
1.05896205e-01 5.13199687e-01 1.13227260e+00 -7.26986945e-01
-5.77650368e-01 -2.92930126e-01 -4.61647779e-01 3.09939116e-01
-1.32023846e-03 1.95438981e-01 -6.89435840e-01 -4.04553771e-01
-2.37251654e-01 -1.17616160e-02 1.29976666e+00 1.08856177e+00
1.50226426e+00 -2.01463923e-01 -9.96944845e-01 1.07972455e+00
1.65101171e+00 2.36178562e-01 6.38159573e-01 2.49916464e-02
6.70135081e-01 8.00864816e-01 6.99856937e-01 8.45601857e-02
1.73598170e-01 3.43697160e-01 5.70757389e-01 -7.30401039e-01
-3.70174646e-01 3.26925308e-01 -1.82424903e-01 3.17548513e-01
4.27247822e-01 -2.94049650e-01 -1.05924940e+00 1.09912312e+00
-1.61332107e+00 -4.63007689e-01 -1.51741773e-01 1.76615882e+00
8.98984969e-01 2.95309454e-01 7.30440244e-02 1.73735805e-02
1.07408071e+00 3.44867222e-02 -7.01153159e-01 -8.12322740e-03
-7.48634711e-02 -7.57094612e-03 7.03137636e-01 4.31317717e-01
-1.29610634e+00 8.95728767e-01 5.94382811e+00 1.52130103e+00
-1.07766092e+00 2.97299922e-01 1.41523862e+00 -1.99221462e-01
1.39270201e-01 -2.46129721e-01 -1.28000569e+00 4.38542098e-01
3.26117098e-01 3.69147211e-01 -2.86625206e-01 5.99824548e-01
8.64167958e-02 -5.04226267e-01 -1.08542860e+00 6.41881406e-01
-8.35389793e-02 -1.51560926e+00 2.13529840e-01 2.85014957e-01
7.64353096e-01 -9.33942124e-02 -9.74527448e-02 -5.54958694e-02
1.04497019e-02 -9.09947991e-01 3.93906802e-01 4.60166305e-01
7.64591813e-01 -8.08997750e-01 1.00805438e+00 2.26833284e-01
-1.10493410e+00 5.27348444e-02 -3.21247131e-01 6.65113807e-01
7.15707093e-02 7.81112611e-01 -1.03670716e+00 1.26874700e-01
5.34554601e-01 3.91063601e-01 -5.91038704e-01 1.48189735e+00
2.21331075e-01 5.85063338e-01 -3.88727099e-01 -9.35751200e-02
3.63645464e-01 -1.17980771e-01 6.25841618e-01 1.58998525e+00
1.63789287e-01 2.78449692e-02 2.36360326e-01 8.68340433e-01
-2.58905053e-01 1.58222884e-01 3.44705552e-01 2.78784096e-01
2.45942935e-01 1.80447459e+00 -1.69445181e+00 -3.35087866e-01
1.08548492e-01 5.92158794e-01 5.54762296e-02 1.31106898e-01
-6.72753692e-01 -3.25704694e-01 4.16021526e-01 3.93547118e-01
1.26358896e-01 3.17143112e-01 -7.31124401e-01 -4.89172846e-01
-1.52882516e-01 -4.36368257e-01 7.79963493e-01 -9.14201513e-02
-1.20285332e+00 1.41542956e-01 -2.00812742e-01 -1.07590437e+00
5.66271305e-01 -6.15492046e-01 -7.97589481e-01 6.62236214e-01
-1.66476667e+00 -1.03913689e+00 -3.58653724e-01 3.01822513e-01
5.35141528e-01 6.54041916e-02 1.66251391e-01 3.89987081e-01
-7.06781268e-01 8.50535095e-01 8.45763758e-02 2.84037054e-01
5.76481104e-01 -1.35099411e+00 -2.95529962e-01 7.15590179e-01
-5.44177294e-01 1.27079204e-01 4.71701980e-01 -5.88741183e-01
-7.64303327e-01 -1.34735978e+00 6.06042683e-01 -1.07128382e-01
3.57223332e-01 -1.51873857e-01 -7.95479178e-01 1.40616462e-01
-3.78500521e-02 4.84767616e-01 6.31452739e-01 -4.52242970e-01
3.94802541e-01 -4.38572228e-01 -1.81040740e+00 6.53142929e-01
4.46975946e-01 9.65304226e-02 -3.62021923e-02 4.88139361e-01
7.37049043e-01 -5.64081192e-01 -8.13827574e-01 6.23663485e-01
2.30917037e-01 -6.55407250e-01 9.88236368e-01 -1.06213599e-01
2.56726921e-01 -4.23251808e-01 2.45722517e-01 -8.25559437e-01
-3.89252424e-01 -2.32253149e-01 2.40640968e-01 1.11955988e+00
5.36212087e-01 -2.38649130e-01 1.25751066e+00 1.67987049e-01
-6.49104953e-01 -1.22316670e+00 -1.40166247e+00 -4.85149145e-01
2.59978741e-01 9.55036953e-02 3.34123820e-01 6.19175911e-01
-3.56275439e-01 -1.38811678e-01 2.48770580e-01 1.56949714e-01
7.57458329e-01 -1.53643489e-01 3.53718728e-01 -1.25078070e+00
6.59056902e-02 -1.13270724e+00 -5.37582815e-01 -8.04844141e-01
-1.58613548e-01 -1.14603889e+00 1.39506742e-01 -1.68907177e+00
6.79299712e-01 -3.63775611e-01 -6.77793860e-01 2.96623468e-01
-3.69228393e-01 3.85929793e-01 -3.98988537e-02 2.53116786e-01
-8.55224371e-01 2.04697922e-02 1.74348164e+00 -4.44182634e-01
-9.64298621e-02 3.12588625e-02 -6.31066680e-01 9.98277366e-01
5.68153203e-01 -6.74263537e-01 7.28965476e-02 1.50786728e-01
-1.12261251e-01 -4.47606519e-02 4.81332600e-01 -1.16812360e+00
5.06562173e-01 -2.18214348e-01 5.56867957e-01 -1.23872364e+00
1.76238269e-01 -7.34615862e-01 -2.38956511e-01 8.08044314e-01
-2.57312030e-01 -7.06766844e-01 1.43289313e-01 6.06705487e-01
-2.25167260e-01 -2.92531848e-01 1.48604012e+00 -3.22954595e-01
-2.32580751e-01 4.12031293e-01 -3.02849472e-01 7.34057724e-02
1.60694206e+00 -7.49349713e-01 -4.55693007e-01 4.45110053e-01
-6.56545758e-01 6.47979617e-01 2.21143171e-01 -2.53933877e-01
3.57856125e-01 -8.37052703e-01 -9.41883147e-01 -5.30659147e-02
8.64443853e-02 5.98972380e-01 4.87485975e-01 1.31452823e+00
-9.99675572e-01 4.21936065e-01 -1.39396815e-02 -9.15566087e-01
-1.45153213e+00 1.98485389e-01 8.34104180e-01 -9.65655208e-01
-4.89753157e-01 1.54146504e+00 7.97216177e-01 -4.11352128e-01
4.45730150e-01 -7.29162395e-01 -3.59613985e-01 3.02788466e-02
3.28804255e-01 4.90553975e-01 7.58055523e-02 -5.26328564e-01
-5.72147071e-01 5.23943126e-01 -5.09111345e-01 3.30405742e-01
9.12508607e-01 2.97856834e-02 -5.01349092e-01 6.99142218e-02
1.15923369e+00 -4.45650160e-01 -1.40788722e+00 -3.44440728e-01
2.30840072e-02 -4.09698002e-02 4.29705441e-01 -1.09061539e+00
-1.42373872e+00 9.25715566e-01 1.08732927e+00 -1.55436888e-01
1.19024765e+00 1.65365860e-01 1.14390182e+00 -1.68856397e-01
7.88897835e-03 -1.35955799e+00 -2.09738672e-01 4.20057029e-01
3.57220620e-01 -1.12836802e+00 2.66072303e-01 -7.14402556e-01
-3.20144653e-01 9.61351812e-01 6.45017028e-01 4.44888100e-02
7.72382379e-01 3.60754937e-01 3.59657779e-02 -2.51890212e-01
-4.20774460e-01 -1.67617857e-01 4.25175458e-01 3.67188275e-01
2.70687610e-01 2.57516295e-01 -4.71227735e-01 7.23203182e-01
1.46428779e-01 -2.82083333e-01 2.79176742e-01 5.58607757e-01
-7.41741836e-01 -6.67648375e-01 -3.89212787e-01 7.56847620e-01
-8.74334335e-01 -3.64010036e-02 -4.70730186e-01 8.43353033e-01
5.22213280e-01 4.48578596e-01 4.05441552e-01 2.45056674e-01
5.40716201e-03 -4.40840423e-01 2.46806413e-01 -5.22536814e-01
-8.48174930e-01 3.62386614e-01 -3.30036283e-01 -2.05005154e-01
-2.61059552e-01 -5.72312474e-01 -1.75190008e+00 1.19295500e-01
-4.25295979e-01 -1.10189468e-02 5.39953053e-01 7.68614471e-01
-9.13251638e-02 5.81630707e-01 4.66063946e-01 -6.12240255e-01
-1.81328923e-01 -8.80203724e-01 -7.51096010e-01 2.51543134e-01
3.29642087e-01 -2.60166585e-01 -3.33189607e-01 6.71175420e-02] | [14.960859298706055, -3.0541627407073975] |
32fa9865-4561-4fec-b78f-4fa5dc7140bd | autonomous-golf-putting-with-data-driven-and | 2211.08081 | null | https://arxiv.org/abs/2211.08081v1 | https://arxiv.org/pdf/2211.08081v1.pdf | Autonomous Golf Putting with Data-Driven and Physics-Based Methods | We are developing a self-learning mechatronic golf robot using combined data-driven and physics-based methods, to have the robot autonomously learn to putt the ball from an arbitrary point on the green. Apart from the mechatronic control design of the robot, this task is accomplished by a camera system with image recognition and a neural network for predicting the stroke velocity vector required for a successful hole-in-one. To minimize the number of time-consuming interactions with the real system, the neural network is pretrained by evaluating basic physical laws on a model, which approximates the golf ball dynamics on the green surface in a data-driven manner. Thus, we demonstrate the synergetic combination of data-driven and physics-based methods on the golf robot as a mechatronic example system. | ['Ansgar Trächtler', 'Julia Timmermann', 'Niklas Fittkau', 'Annika Junker'] | 2022-11-15 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [-2.15143621e-01 3.43955129e-01 -5.48842400e-02 1.12310641e-01
-1.11406237e-01 -2.53868967e-01 2.15203598e-01 -3.46809059e-01
-4.05439079e-01 3.18484098e-01 -5.73232710e-01 -2.86832482e-01
-6.52465940e-01 -8.48939180e-01 -1.16675067e+00 -5.72660148e-01
1.03248641e-01 9.37090874e-01 9.83150229e-02 -5.67609310e-01
3.11811686e-01 9.02838528e-01 -1.54654264e+00 -2.08354533e-01
5.40074527e-01 7.02802837e-01 2.56515443e-01 7.89731503e-01
6.57377481e-01 7.94257939e-01 -6.01192154e-02 4.35557812e-02
5.56730092e-01 -3.44792068e-01 -4.37508672e-01 9.96710733e-02
2.09658042e-01 -2.14024350e-01 -9.32015300e-01 7.45556176e-01
-3.34439687e-02 3.90257865e-01 1.05564225e+00 -1.34153783e+00
-1.73045605e-01 1.08991452e-01 -6.91557629e-03 -5.08942008e-01
1.21543497e-01 5.94733834e-01 7.79826164e-01 -7.69594669e-01
6.72654927e-01 9.75204587e-01 7.70943105e-01 5.36535561e-01
-7.79249609e-01 -2.92253852e-01 -4.35300797e-01 3.13671052e-01
-1.28510725e+00 8.34401846e-02 8.63885820e-01 -7.55739808e-01
8.09111536e-01 -2.49380335e-01 1.06007493e+00 7.31013656e-01
1.07690251e+00 3.73317719e-01 3.71903062e-01 -4.77607071e-01
5.36158085e-01 -3.44734818e-01 -1.38807327e-01 1.05804443e+00
3.34337771e-01 6.21150911e-01 -3.26232553e-01 2.97809541e-01
1.27647626e+00 3.08943223e-02 9.72279906e-02 -1.12012649e+00
-6.30701661e-01 6.42800748e-01 5.12816370e-01 -1.85718194e-01
-4.93111670e-01 6.94660306e-01 -3.53267305e-02 1.05385490e-01
-2.99395651e-01 9.38606143e-01 -4.57510889e-01 -3.33435982e-01
-4.41742867e-01 5.92323422e-01 9.88574028e-01 1.06535506e+00
7.74710059e-01 2.44648263e-01 5.09947419e-01 2.74996847e-01
5.49544871e-01 4.86484349e-01 1.90265000e-01 -1.33758008e+00
2.44971514e-01 8.94963205e-01 2.96155810e-01 -6.08364165e-01
-6.26961172e-01 1.02665313e-01 -3.64416361e-01 1.05516422e+00
3.97773147e-01 -6.26686811e-01 -7.28844881e-01 1.32412720e+00
3.71598303e-01 -5.50444499e-02 4.50084060e-02 1.27685606e+00
8.34997967e-02 7.97472298e-01 -3.48613083e-01 4.79437411e-01
9.46516037e-01 -7.53033280e-01 -1.30107328e-01 -2.76720107e-01
8.09454501e-01 -1.26575008e-01 1.14302874e+00 6.89548075e-01
-1.07321239e+00 -5.63390732e-01 -1.60142100e+00 -9.61790234e-02
-1.60038441e-01 3.70936990e-01 2.35588402e-01 9.03771296e-02
-4.27249908e-01 1.10452449e+00 -1.17479849e+00 -1.40636101e-01
1.20048061e-01 5.19764483e-01 -3.57221395e-01 3.61959189e-01
-7.97340214e-01 1.62108362e+00 3.09929907e-01 -5.20956554e-02
-1.14403415e+00 -7.99791753e-01 -8.06149483e-01 7.08338842e-02
2.91035384e-01 -8.52207661e-01 1.38354635e+00 -4.03671950e-01
-2.09848523e+00 4.01345670e-01 5.17025232e-01 -4.88810241e-01
6.04577422e-01 -5.17122924e-01 3.49445105e-01 1.73339069e-01
-4.05456811e-01 5.12123227e-01 8.47683072e-01 -1.39640343e+00
-4.34964985e-01 -1.10362180e-01 6.60680756e-02 4.09519285e-01
-8.13491195e-02 -9.55375493e-01 -1.58220574e-01 -9.81642976e-02
4.65447903e-02 -1.41335964e+00 -2.75701076e-01 2.25386500e-01
-1.69882506e-01 -2.11814702e-01 1.05158389e+00 -1.95987865e-01
3.94562453e-01 -1.94681776e+00 4.32747900e-01 3.24657261e-01
-1.42658532e-01 8.61752331e-02 3.66711020e-02 8.24575603e-01
1.59966320e-01 -5.87797165e-01 6.56865016e-02 9.12059471e-03
-2.15950772e-01 2.00768873e-01 -2.00094953e-01 4.79876369e-01
8.57024938e-02 7.99409807e-01 -8.09471011e-01 2.07722448e-02
4.11115170e-01 2.54214168e-01 -5.47771513e-01 3.52689892e-01
-4.89548147e-01 3.22399080e-01 -6.03238881e-01 2.02776477e-01
1.64881647e-01 2.59496689e-01 3.00935153e-02 -1.05373427e-01
-2.23744705e-01 -5.20780459e-02 -9.20620680e-01 1.78548384e+00
-8.56814384e-01 5.31738997e-01 2.23326400e-01 -1.01625574e+00
1.13735509e+00 -2.86464207e-02 6.83848619e-01 -4.19413179e-01
4.60697681e-01 1.45846128e-01 -6.76251426e-02 -7.26765394e-01
5.81461132e-01 -3.86219531e-01 -1.69282541e-01 2.00675592e-01
2.90821325e-02 -9.78540838e-01 -1.65930763e-01 -2.64356077e-01
1.17336917e+00 8.34479630e-01 -2.65469819e-01 -2.35278860e-01
2.64333487e-01 1.05241942e+00 1.89940110e-01 4.43439543e-01
3.65163654e-01 5.67362309e-01 8.55172426e-02 -3.85724813e-01
-1.48580515e+00 -1.05259454e+00 3.17432314e-01 4.69750375e-01
7.19671249e-01 -6.93478733e-02 -1.08550799e+00 -2.19056606e-01
4.12216872e-01 6.97784603e-01 -2.62629926e-01 -9.36501443e-01
-6.28669083e-01 -1.24403276e-01 1.79543883e-01 5.61615705e-01
1.78983420e-01 -8.81413460e-01 -1.43569446e+00 3.73011351e-01
6.24212086e-01 -7.78106511e-01 -1.14592358e-01 2.97543645e-01
-8.24376047e-01 -1.63913512e+00 -2.37586141e-01 -8.90679955e-01
4.95819837e-01 -7.37275509e-03 5.02525210e-01 -2.11121023e-01
-6.57031596e-01 7.82712817e-01 1.24538802e-01 -5.15499234e-01
-7.26282895e-01 4.16957550e-02 2.42457151e-01 -4.49785680e-01
-2.06327468e-01 -4.30928886e-01 -5.96548855e-01 3.90423924e-01
-4.21896785e-01 4.35978770e-01 5.81138730e-01 6.73536301e-01
4.66808826e-01 5.40774584e-01 1.86073646e-01 -1.03613533e-01
4.93614018e-01 -1.04593866e-01 -9.39322114e-01 -1.34230718e-01
-4.88197595e-01 -7.37350807e-02 8.87575448e-01 -6.49030566e-01
-8.70601714e-01 8.68986905e-01 2.74407268e-01 -7.49952734e-01
-3.91324647e-02 4.20431376e-01 -1.44535154e-01 -2.27610737e-01
6.71888709e-01 7.60023370e-02 6.13329768e-01 -1.78707212e-01
4.33198571e-01 3.79082978e-01 7.70383060e-01 -6.94908321e-01
9.30468738e-01 3.30289185e-01 6.92932248e-01 -8.90096545e-01
-2.66043872e-01 -1.37608573e-01 -8.35956633e-01 -7.07568824e-01
7.26030529e-01 -6.51760340e-01 -1.36807215e+00 7.71445930e-01
-1.11427689e+00 -8.98778737e-01 -6.78931892e-01 6.57691836e-01
-1.37254405e+00 -1.32531047e-01 -5.57410657e-01 -9.51106548e-01
-1.37310818e-01 -1.08197296e+00 8.53038907e-01 1.85282126e-01
-2.39857703e-01 -5.97155750e-01 2.93503255e-01 1.68883353e-01
-4.56457064e-02 3.35386127e-01 1.01628792e+00 -1.96583509e-01
-5.88533878e-01 -7.20330000e-01 1.60859942e-01 2.05029503e-01
1.07450727e-02 9.79594290e-02 -5.19927204e-01 -3.46645087e-01
2.50160486e-01 -3.84683996e-01 1.06254116e-01 3.46926183e-01
7.74663806e-01 -1.52058229e-01 -4.63004559e-01 2.60910392e-01
1.54885960e+00 5.21869421e-01 3.31026226e-01 4.73331690e-01
8.64084661e-01 7.08815157e-01 7.70580590e-01 2.42164940e-01
3.68453026e-01 6.09990954e-01 8.71090651e-01 2.25079253e-01
-5.66288345e-02 -5.30461848e-01 3.18752557e-01 2.90377259e-01
-9.62510332e-03 9.22692865e-02 -1.16825640e+00 2.32936904e-01
-2.21429420e+00 -4.10122693e-01 -2.35250697e-01 2.15523195e+00
2.38854468e-01 1.23833977e-01 6.32304549e-02 1.09890193e-01
5.17015934e-01 -7.32553244e-01 -9.81012464e-01 -2.80058205e-01
5.09700060e-01 -2.28147298e-01 6.60077095e-01 4.84021097e-01
-6.76852465e-01 8.19120944e-01 6.32715368e+00 4.64518815e-01
-1.28244901e+00 -3.19839388e-01 9.89355147e-02 -1.11555792e-01
3.96220803e-01 1.36001438e-01 -6.84899628e-01 2.37184823e-01
1.07681537e+00 -3.29014868e-01 6.10252202e-01 1.19472694e+00
4.05368626e-01 -2.52204448e-01 -1.44723737e+00 8.19852173e-01
-2.12150797e-01 -1.55085754e+00 -3.51092331e-02 1.35372490e-01
4.67254490e-01 1.37411384e-02 -2.42424771e-01 1.00194767e-01
3.57961267e-01 -6.96080565e-01 1.08681893e+00 8.57413411e-01
3.71106565e-01 -6.45799696e-01 1.81386352e-01 8.25374544e-01
-8.32570374e-01 -4.84827071e-01 -1.41851887e-01 -3.51083279e-01
1.54770494e-01 2.29987249e-01 -8.17088604e-01 2.65638411e-01
3.72668117e-01 5.20421505e-01 4.85041849e-02 1.10086775e+00
-1.71718419e-01 1.38618037e-01 -4.76444691e-01 -5.43533087e-01
1.90074399e-01 -5.41453242e-01 6.35413468e-01 3.94044906e-01
4.26329643e-01 1.00504890e-01 -1.60197578e-02 9.75743592e-01
2.71320999e-01 -2.57700264e-01 -8.82905066e-01 2.37227291e-01
-2.17894055e-02 1.24408662e+00 -4.92582977e-01 4.40721586e-02
1.30822450e-01 4.73555475e-01 9.00493786e-02 1.66399598e-01
-8.27863932e-01 -5.81604004e-01 5.20151794e-01 5.55736840e-01
5.97451925e-02 -7.93957531e-01 -4.94961739e-01 -7.03426898e-01
-1.10480733e-01 -2.73584306e-01 -2.56486326e-01 -1.39655828e+00
-8.25354517e-01 -4.26627360e-02 2.43917748e-01 -1.48863912e+00
-3.35516542e-01 -1.18191230e+00 -7.41616964e-01 5.50815344e-01
-9.01223242e-01 -7.73456812e-01 -3.09763610e-01 2.70618141e-01
4.31022376e-01 -3.19607437e-01 4.86212015e-01 -1.83929622e-01
-4.01369661e-01 -1.74853772e-01 4.92282547e-02 -3.05842131e-01
1.92052588e-01 -1.15673411e+00 2.26347014e-01 4.29132015e-01
-2.02020824e-01 1.90927476e-01 7.09672928e-01 -8.05881739e-01
-2.29933071e+00 -7.74766326e-01 5.88099696e-02 -3.85788530e-01
9.74830747e-01 -2.80965954e-01 -7.30404735e-01 3.42094749e-01
-2.75424391e-01 -3.44952136e-01 -1.46757856e-01 -4.07009274e-01
4.42519486e-01 -2.40119502e-01 -1.00475407e+00 9.94439304e-01
8.70113611e-01 -2.41215080e-01 -8.03399861e-01 5.52562475e-01
1.28418252e-01 -7.11253107e-01 -7.97323048e-01 4.19351608e-01
7.59821594e-01 -1.67161658e-01 8.47696781e-01 -5.69100320e-01
7.11931527e-01 -3.24036300e-01 2.74773300e-01 -1.37954795e+00
-1.75169975e-01 -8.35905433e-01 -2.24707276e-01 5.74016631e-01
1.34747133e-01 -4.57666427e-01 1.39559507e+00 7.36023843e-01
-3.93029928e-01 -1.03349650e+00 -1.24965763e+00 -9.97012794e-01
4.31187838e-01 -4.67062712e-01 2.92706251e-01 1.65989339e-01
4.90725785e-01 1.52212128e-01 -1.87453046e-01 6.00825012e-01
3.29373181e-01 -7.15192268e-03 1.06583667e+00 -1.19291425e+00
-2.54496008e-01 -2.61385620e-01 -4.63772595e-01 -9.77193236e-01
2.84057766e-01 -4.93136317e-01 6.42731249e-01 -1.48495829e+00
-3.25865090e-01 -4.85738397e-01 2.45246798e-01 3.77565831e-01
5.63987494e-01 -2.50926137e-01 1.67694017e-01 4.11382198e-01
-2.24505551e-03 5.82535923e-01 1.37664807e+00 -8.80432352e-02
-5.15137196e-01 1.05727091e-01 1.08868338e-01 9.09051120e-01
7.48037159e-01 -4.03440177e-01 -5.50405562e-01 -3.17562848e-01
2.92451084e-01 5.23673534e-01 5.14868557e-01 -1.39132965e+00
7.44808197e-01 -3.25311005e-01 2.51560599e-01 -4.89601225e-01
6.46241069e-01 -1.19335842e+00 2.20780388e-01 1.09781063e+00
-1.40192196e-01 1.38837650e-01 2.46612713e-01 4.63139266e-01
1.20734707e-01 -4.17774439e-01 8.02231014e-01 -1.81343928e-02
-4.98438507e-01 2.63264328e-01 -7.34227657e-01 -4.17116910e-01
1.32324588e+00 -4.21110004e-01 -1.42886996e-01 -2.54324079e-01
-6.96680844e-01 2.78802484e-01 6.23184204e-01 4.85748202e-01
5.82859755e-01 -9.64079916e-01 -4.85620275e-02 4.47687626e-01
-8.78318995e-02 3.71509105e-01 7.99454972e-02 2.11108834e-01
-1.11979842e+00 1.49706721e-01 -4.88838851e-01 -7.18875587e-01
-7.01595426e-01 2.41565838e-01 8.36798072e-01 1.03669025e-01
-9.65290666e-01 2.07990810e-01 -1.85714945e-01 -5.67267716e-01
-1.57047257e-01 -4.20226276e-01 1.98045343e-01 -5.94527841e-01
-3.66132081e-01 5.85367501e-01 1.13357924e-01 -3.23558182e-01
1.84145123e-02 8.48808169e-01 4.07199174e-01 -3.36272746e-01
1.42964745e+00 2.93173432e-01 4.77065980e-01 4.25822556e-01
6.40742362e-01 -5.93071282e-01 -1.98004699e+00 4.64420646e-01
-1.33124530e-01 9.26876515e-02 1.99218109e-01 -6.71966910e-01
-8.77841651e-01 7.43924081e-01 4.91017491e-01 4.62821275e-02
4.64603692e-01 -3.03760260e-01 8.02134395e-01 1.01651692e+00
6.04610920e-01 -1.50570619e+00 2.74457633e-01 7.08782971e-01
1.07857823e+00 -8.14154685e-01 5.01677347e-03 -2.05458313e-01
-3.82369637e-01 1.50293314e+00 9.42598999e-01 -9.77984667e-01
7.79000223e-01 5.09409249e-01 7.68063664e-02 -3.96160632e-01
-4.42317307e-01 5.00289440e-01 2.94549614e-01 4.46366549e-01
-3.98754984e-01 -1.77276477e-01 1.93028346e-01 4.32371110e-01
-4.50096548e-01 3.52221102e-01 7.06579149e-01 1.34592009e+00
-6.12917364e-01 -8.47135067e-01 -3.39070916e-01 -3.65613960e-02
4.02396142e-01 6.46730363e-01 -1.82697281e-01 1.26102400e+00
-5.26333079e-02 4.67473030e-01 2.09558532e-01 -4.04331893e-01
9.23535526e-01 -3.43053676e-02 7.99370825e-01 -7.41415560e-01
-3.61429930e-01 -3.96069497e-01 -9.73471925e-02 -6.84552372e-01
1.70155376e-01 -5.56228161e-01 -1.71206439e+00 1.84796363e-01
-1.33792549e-01 5.13614854e-03 1.05077624e+00 1.07134759e+00
4.20426875e-01 3.66790533e-01 5.96135080e-01 -1.51644552e+00
-7.15055585e-01 -5.75188398e-01 -5.45223534e-01 2.96208173e-01
1.81423664e-01 -9.65642273e-01 -1.70808569e-01 -1.72293112e-02] | [4.840149879455566, 1.29497230052948] |
ec1c8aa4-8aa3-402f-93a5-abdb2bbd244b | carbon-efficient-neural-architecture-search | 2307.04131 | null | https://arxiv.org/abs/2307.04131v1 | https://arxiv.org/pdf/2307.04131v1.pdf | Carbon-Efficient Neural Architecture Search | This work presents a novel approach to neural architecture search (NAS) that aims to reduce energy costs and increase carbon efficiency during the model design process. The proposed framework, called carbon-efficient NAS (CE-NAS), consists of NAS evaluation algorithms with different energy requirements, a multi-objective optimizer, and a heuristic GPU allocation strategy. CE-NAS dynamically balances energy-efficient sampling and energy-consuming evaluation tasks based on current carbon emissions. Using a recent NAS benchmark dataset and two carbon traces, our trace-driven simulations demonstrate that CE-NAS achieves better carbon and search efficiency than the three baselines. | ['Tian Guo', 'Yiyang Zhao'] | 2023-07-09 | null | null | null | null | ['architecture-search'] | ['methodology'] | [-2.46448033e-02 -7.13947058e-01 -3.30681264e-01 -2.50358731e-01
-5.94281673e-01 -5.30772328e-01 4.41775858e-01 -1.76135227e-01
-7.15543270e-01 6.09591067e-01 -8.06012005e-02 -5.34868956e-01
-1.27894625e-01 -1.02775288e+00 -8.04200232e-01 -6.11924231e-01
3.19414884e-01 6.42815769e-01 7.39298910e-02 2.62078077e-01
4.57342774e-01 6.05172694e-01 -1.58155131e+00 -1.77846640e-01
1.01330698e+00 1.08992052e+00 3.98242384e-01 6.93324566e-01
-5.78778014e-02 2.74156302e-01 -5.69785237e-01 -2.09119678e-01
6.06640220e-01 -3.89161795e-01 -6.72430456e-01 -7.98690796e-01
6.95063770e-02 -2.93877631e-01 -2.91161593e-02 1.13174236e+00
7.59158254e-01 3.82265449e-01 5.93104184e-01 -1.14908397e+00
-3.17459732e-01 7.41389632e-01 -4.34838742e-01 4.78453785e-01
-6.91833556e-01 2.94289440e-01 7.52246976e-01 -3.47950041e-01
1.70038179e-01 1.09606266e+00 4.83654022e-01 7.17595160e-01
-1.12259281e+00 -9.43096101e-01 1.70234635e-01 4.49789166e-01
-1.40516901e+00 -4.49214399e-01 7.06380308e-01 -1.16469055e-01
1.76574802e+00 7.50208795e-01 9.18663502e-01 9.60855067e-01
2.20475346e-01 2.21927658e-01 8.95557761e-01 -4.20501888e-01
9.33945954e-01 -1.41640410e-01 2.55858123e-01 8.01639259e-01
7.77949750e-01 5.93494773e-01 -4.47417527e-01 -4.70869124e-01
1.77833527e-01 -3.33980381e-01 -5.64605184e-03 -2.89281290e-02
-7.57931232e-01 5.80378294e-01 3.86979640e-01 -1.05486639e-01
-6.11991942e-01 9.10041451e-01 4.97791052e-01 -2.66293854e-01
4.21011806e-01 7.31529474e-01 -6.03071034e-01 -2.76558489e-01
-8.34865987e-01 2.63433158e-01 6.22411370e-01 8.01192343e-01
3.21893275e-01 5.34492850e-01 -4.09913659e-01 7.47271121e-01
4.29497510e-01 6.17704809e-01 5.99185944e-01 -1.22264338e+00
2.91395336e-01 5.06727934e-01 -8.74778628e-03 -3.81829798e-01
-4.49150205e-01 -7.45446265e-01 -6.70862257e-01 1.70079872e-01
-4.08323288e-01 -4.18685168e-01 -8.38716626e-01 1.97487903e+00
3.82143945e-01 1.36874899e-01 -1.52438000e-01 8.31407964e-01
1.65744081e-01 9.29075658e-01 2.74933904e-01 -2.11049244e-01
1.22897625e+00 -1.44265640e+00 -4.30152893e-01 -7.74703473e-02
3.05304259e-01 5.69628030e-02 1.09770131e+00 1.94824338e-01
-1.40018046e+00 1.55138643e-02 -1.30672741e+00 1.73702419e-01
-5.24608016e-01 1.17704809e-01 7.66260207e-01 1.19442046e+00
-1.08666241e+00 6.26467407e-01 -1.23896837e+00 -2.02936426e-01
3.49411875e-01 4.18632865e-01 7.63652325e-01 3.53892237e-01
-6.65630519e-01 7.99139738e-01 6.97830260e-01 -3.95877451e-01
-1.23024130e+00 -1.04429185e+00 -5.95478863e-02 4.14333850e-01
4.42724884e-01 -1.26256454e+00 1.36620152e+00 -4.99431282e-01
-1.83984375e+00 3.08647066e-01 3.14212888e-02 -5.65152705e-01
1.77921355e-01 -2.33161330e-01 -2.35447079e-01 -2.44927749e-01
-6.58575892e-01 6.78231537e-01 4.71244991e-01 -1.08281910e+00
-4.94787425e-01 -4.39239711e-01 -2.01109678e-01 3.82467389e-01
-7.72697866e-01 1.41719207e-01 -6.68479741e-01 -4.63279665e-01
-5.32390118e-01 -1.10671055e+00 -4.51145291e-01 -4.26521987e-01
-4.96127933e-01 -6.71588480e-02 3.81295979e-01 -4.71674711e-01
1.66070580e+00 -1.48662841e+00 3.79745185e-01 3.37135017e-01
-1.82397902e-01 2.13133246e-01 -2.71712750e-01 -1.99021682e-01
2.58877158e-01 3.05688739e-01 -2.64568835e-01 -2.59097695e-01
3.15786809e-01 9.39055234e-02 1.97091162e-01 2.56278981e-02
-2.76548326e-01 6.84495270e-01 -5.88750780e-01 -2.16897383e-01
-7.63233602e-02 4.40650374e-01 -8.35813701e-01 2.11781606e-01
-8.09543610e-01 -2.84092158e-01 -4.35428888e-01 1.02668619e+00
6.31569624e-01 -4.13065583e-01 3.33099723e-01 -3.03449750e-01
-4.97711867e-01 2.90837497e-01 -7.65213490e-01 1.58088505e+00
-6.34964824e-01 1.36244655e-01 6.61714301e-02 -5.66945434e-01
5.07774770e-01 -2.89425761e-01 4.75400716e-01 -1.34497190e+00
3.29755276e-01 3.62375915e-01 -1.97637841e-01 -4.07436006e-02
5.99551022e-01 6.63480699e-01 3.27750057e-01 7.03142822e-01
-4.83405083e-01 4.15583886e-02 2.26329163e-01 -3.27547014e-01
1.33669043e+00 1.37965515e-01 1.70983344e-01 -8.91287446e-01
3.33767474e-01 2.75054067e-01 7.04452038e-01 4.12980169e-01
-1.90494075e-01 -1.34563670e-01 -1.33934498e-01 -4.28265423e-01
-1.26723254e+00 -8.42085123e-01 5.36558330e-02 1.46138167e+00
8.66256207e-02 -5.32760978e-01 -1.33513713e+00 -4.77401972e-01
-3.34264725e-01 1.22488034e+00 -2.53420204e-01 -3.97059411e-01
-6.63770854e-01 -1.29523802e+00 5.26091814e-01 4.44585383e-01
7.67650962e-01 -6.13017261e-01 -1.62130177e+00 2.82698367e-02
5.90000562e-02 -3.93181413e-01 -7.72993863e-01 7.16812015e-01
-1.01984394e+00 -7.67329693e-01 -2.48368576e-01 -2.80616045e-01
4.93839920e-01 1.97357371e-01 1.56030500e+00 1.41920149e-01
-5.08695662e-01 7.78887868e-02 2.10069001e-01 -7.54306793e-01
-1.72622845e-01 3.64574879e-01 -1.17610306e-01 -7.37641990e-01
2.60024905e-01 -3.19253087e-01 -1.00237143e+00 3.20666358e-02
-7.66593337e-01 2.06528693e-01 4.31415856e-01 6.70098066e-01
1.20106161e+00 3.26413125e-01 1.20920055e-01 -6.39434874e-01
9.28403437e-01 -6.02728426e-01 -1.43487692e+00 7.23796785e-01
-1.93353522e+00 6.04383707e-01 6.76064134e-01 -1.37067467e-01
-1.02204490e+00 1.34192631e-01 6.77682236e-02 -5.89423001e-01
3.17067206e-01 2.98773199e-01 -1.73518211e-01 -6.57500699e-02
7.42753327e-01 2.37442940e-01 -7.30169535e-01 -6.54745758e-01
6.10579923e-02 2.66192645e-01 6.01911426e-01 -9.42919075e-01
4.47222590e-01 -1.69772238e-01 3.01483840e-01 -2.35888287e-01
-4.55586582e-01 -1.40193235e-02 -2.11792868e-02 -1.28742561e-01
8.51555109e-01 -8.56371701e-01 -6.34784698e-01 4.42970395e-01
-1.09037471e+00 -5.29420197e-01 -2.14420155e-01 3.36941570e-01
-2.14491785e-01 -2.59873152e-01 -3.37567508e-01 -9.25939679e-01
-1.38183689e+00 -1.27200353e+00 7.19343007e-01 4.44137543e-01
2.10288420e-01 -7.07475483e-01 4.75635737e-01 2.76346833e-01
9.60566461e-01 1.11570857e-01 1.43172848e+00 -1.72018394e-01
-8.29462051e-01 5.48718512e-01 -2.71043837e-01 1.38370261e-01
-3.11381906e-01 1.48295283e-01 -8.48707139e-01 -4.33540285e-01
-2.73622274e-01 -2.19109789e-01 8.00281167e-01 6.18653238e-01
1.93511915e+00 -6.97428524e-01 -3.21428031e-01 1.11282349e+00
1.99574113e+00 7.83794165e-01 1.57828107e-01 7.47730732e-01
5.12929797e-01 -1.02044761e-01 2.73147881e-01 6.39164209e-01
4.42455649e-01 8.03367913e-01 9.74501729e-01 2.37205178e-01
-2.40141735e-03 -1.49780896e-03 3.38797659e-01 1.19723344e+00
-3.00396115e-01 -6.25680864e-01 -1.06885874e+00 3.72875631e-01
-1.65281236e+00 -5.37554204e-01 2.15568319e-01 2.15741158e+00
1.00349629e+00 -1.28372639e-01 2.90784329e-01 -2.21116409e-01
5.24075627e-01 6.96305260e-02 -1.50953841e+00 -8.87483418e-01
2.39923820e-01 4.88245875e-01 1.03640437e+00 3.29054706e-02
-6.52911127e-01 4.99543160e-01 7.39332438e+00 8.17309856e-01
-8.72926056e-01 4.58472520e-01 5.75589061e-01 -8.18227708e-01
-7.26881206e-01 -2.76955634e-01 -6.89031959e-01 6.86145782e-01
1.87653351e+00 -6.75683856e-01 1.32516158e+00 1.09782481e+00
1.20111309e-01 1.62802398e-01 -1.03712678e+00 1.05602825e+00
-8.26031044e-02 -1.72489822e+00 -2.64656311e-03 9.52849761e-02
8.23890924e-01 6.15008593e-01 6.12091012e-02 1.69289947e-01
6.35716021e-01 -9.54819798e-01 9.93325412e-01 3.76044005e-01
7.55338967e-01 -1.18710721e+00 3.35553825e-01 1.27884641e-01
-1.39129484e+00 -4.33041304e-01 -2.99223930e-01 5.17427027e-01
-1.46019951e-01 4.39096749e-01 -4.27515984e-01 4.54589725e-01
1.01404214e+00 -1.19705118e-01 -5.37985265e-01 1.19556797e+00
2.44475111e-01 6.41942263e-01 -4.90549654e-01 -5.20414889e-01
2.32277915e-01 -5.08224666e-01 4.55633879e-01 1.31543267e+00
6.02323711e-01 1.53332710e-01 -1.98155805e-01 1.34958720e+00
-4.55148757e-01 -1.22907571e-01 6.44527003e-02 -2.67330676e-01
1.17218864e+00 1.03108358e+00 -5.47054946e-01 -1.02533065e-01
9.69915539e-02 6.14506245e-01 1.68653578e-01 2.83482037e-02
-1.29908919e+00 -2.46107459e-01 8.68300617e-01 -5.32734036e-01
-1.64100677e-01 -1.23640612e-01 -6.42026603e-01 -5.29873490e-01
-2.96150297e-01 -8.92150939e-01 4.72253591e-01 -6.98656380e-01
-6.74717188e-01 6.71966672e-01 5.95313348e-02 -2.83446312e-01
4.72029895e-02 -3.10323566e-01 -6.96034193e-01 7.85405576e-01
-1.68145776e+00 -6.27319157e-01 -6.18468046e-01 3.33637625e-01
8.70953143e-01 -3.23224783e-01 8.30518782e-01 5.44167161e-01
-1.41175771e+00 7.72091389e-01 3.27487737e-01 -1.01423848e+00
-6.83371425e-02 -1.20361495e+00 8.16096842e-01 8.95847201e-01
-2.18516648e-01 4.24514890e-01 5.66805243e-01 -5.99642932e-01
-1.90817463e+00 -1.53709900e+00 1.50026381e-01 1.36480540e-01
2.77867109e-01 -1.68764085e-01 -4.90605503e-01 1.00586616e-01
4.25036877e-01 -5.17868042e-01 5.96042871e-01 -1.23319469e-01
-2.37547904e-01 -2.03787386e-01 -1.29783404e+00 7.18703449e-01
1.41127539e+00 -1.95901781e-01 1.86460033e-01 5.26094496e-01
1.03218353e+00 -3.98533940e-01 -7.38131642e-01 5.01587927e-01
6.09745443e-01 -6.47810578e-01 9.47847366e-01 -5.57179213e-01
8.41504782e-02 -1.06929392e-01 -5.21414936e-01 -1.59113193e+00
-5.30574739e-01 -3.36192131e-01 -7.73658335e-01 8.73387933e-01
4.55827773e-01 -2.93507934e-01 9.22450423e-01 8.40296447e-01
-3.46076339e-01 -9.76127744e-01 -9.31393802e-01 -1.02056670e+00
1.05572538e-02 -2.95075029e-01 1.30076551e+00 5.84280193e-01
-8.06088626e-01 8.35330039e-02 -4.07569623e-03 2.19057664e-01
9.33486700e-01 1.18143454e-01 8.28649178e-02 -1.10341358e+00
-4.53329444e-01 -9.04199302e-01 5.63278079e-01 -3.39361697e-01
1.40756860e-01 -7.92915106e-01 -1.74805392e-02 -1.27367449e+00
5.86789370e-01 -4.87332880e-01 -7.86790550e-01 6.08147264e-01
3.48474495e-02 -2.85626441e-01 8.87761116e-02 1.56822816e-01
-5.33263028e-01 8.00660431e-01 4.84176874e-01 -3.91192317e-01
-2.27038413e-01 -3.09411645e-01 -4.83450174e-01 4.21659410e-01
8.47779632e-01 -6.42204702e-01 -7.90733337e-01 -9.51033831e-01
3.02987456e-01 -4.26103383e-01 -8.92368630e-02 -1.11745679e+00
4.24232841e-01 -6.39925539e-01 5.24694249e-02 -8.23250353e-01
1.14272550e-01 -1.06906176e+00 9.57489491e-01 1.13222992e+00
-4.61642474e-01 8.74518931e-01 2.88973629e-01 5.78746140e-01
6.21696949e-01 -2.97408432e-01 1.10401177e+00 -9.65870544e-02
-6.77179158e-01 2.71614939e-01 -1.16931289e-01 -2.82910228e-01
1.12940967e+00 1.38350338e-01 -9.77597892e-01 6.20161593e-01
1.75076976e-01 4.89397764e-01 4.22228783e-01 2.47311160e-01
3.97128016e-01 -1.38546014e+00 -2.71276534e-01 -1.58730850e-01
-8.37437809e-02 -1.70626000e-01 7.57900998e-02 3.57323080e-01
-8.81444514e-01 8.40820074e-01 -3.32414627e-01 -2.86145478e-01
-1.22546947e+00 3.75844777e-01 7.49892712e-01 -4.71100807e-01
-2.37253368e-01 9.43341911e-01 -4.32944506e-01 -3.83242637e-01
6.59691334e-01 -1.67199969e-01 6.49944618e-02 -4.39292938e-01
3.03915203e-01 1.25673628e+00 4.55035090e-01 1.24771088e-01
-4.78091896e-01 3.41995299e-01 3.34477812e-01 2.82586575e-03
1.43780744e+00 2.86835171e-02 -7.51326308e-02 6.86883852e-02
8.19414735e-01 -5.25788367e-01 -1.08178484e+00 1.21914834e-01
1.56001925e-01 -3.24269533e-01 9.44029510e-01 -1.44967782e+00
-1.54637027e+00 3.32414627e-01 1.24772596e+00 -2.36877263e-01
1.70618987e+00 -6.99720562e-01 9.93795812e-01 7.65586734e-01
5.07248759e-01 -1.71031809e+00 -1.57987162e-01 4.05972421e-01
5.06849468e-01 -6.19615138e-01 6.38317391e-02 2.07632959e-01
-1.38021991e-01 6.87519968e-01 1.17846024e+00 3.59615147e-01
3.75248462e-01 5.51413894e-01 -6.39445961e-01 -1.88848466e-01
-1.44944787e+00 1.27828285e-01 7.51084536e-02 2.82383591e-01
-1.85662389e-01 2.76176274e-01 -3.16221327e-01 6.09086156e-01
2.77527496e-02 -2.24847980e-02 -5.39748035e-02 8.67894769e-01
-5.16704082e-01 -9.26021814e-01 -2.24168375e-01 7.90196538e-01
-3.43675971e-01 -9.85586047e-02 -5.17317414e-01 2.77512491e-01
1.36510387e-01 6.25878811e-01 7.52421692e-02 -7.46191919e-01
2.90844053e-01 1.46923497e-01 3.68385613e-01 -1.41244709e-01
-1.05328918e+00 -1.17647953e-01 3.85825276e-01 -1.01437080e+00
-1.13137886e-01 -4.29639101e-01 -1.24887919e+00 -6.16753876e-01
-3.71343464e-01 1.56457812e-01 1.70091105e+00 4.38561261e-01
7.95839250e-01 1.12767160e+00 7.13789880e-01 -6.38600588e-01
-8.06990862e-01 -3.68638635e-01 -3.20464373e-01 -3.70340347e-01
-1.30057111e-01 -5.14681458e-01 -1.25617996e-01 -4.09185290e-01] | [8.387685775756836, 3.244055986404419] |
6338210b-b6d9-4b07-a021-1be08e749473 | amplitude-spectrum-transformation-for-open | 2202.04287 | null | https://arxiv.org/abs/2202.04287v1 | https://arxiv.org/pdf/2202.04287v1.pdf | Amplitude Spectrum Transformation for Open Compound Domain Adaptive Semantic Segmentation | Open compound domain adaptation (OCDA) has emerged as a practical adaptation setting which considers a single labeled source domain against a compound of multi-modal unlabeled target data in order to generalize better on novel unseen domains. We hypothesize that an improved disentanglement of domain-related and task-related factors of dense intermediate layer features can greatly aid OCDA. Prior-arts attempt this indirectly by employing adversarial domain discriminators on the spatial CNN output. However, we find that latent features derived from the Fourier-based amplitude spectrum of deep CNN features hold a more tractable mapping with domain discrimination. Motivated by this, we propose a novel feature space Amplitude Spectrum Transformation (AST). During adaptation, we employ the AST auto-encoder for two purposes. First, carefully mined source-target instance pairs undergo a simulation of cross-domain feature stylization (AST-Sim) at a particular layer by altering the AST-latent. Second, AST operating at a later layer is tasked to normalize (AST-Norm) the domain content by fixing its latent to a mean prototype. Our simplified adaptation technique is not only clustering-free but also free from complex adversarial alignment. We achieve leading performance against the prior arts on the OCDA scene segmentation benchmarks. | ['R. Venkatesh Babu', 'Varun Jampani', 'Suvaansh Bhambri', 'Akshay Kulkarni', 'Jogendra Nath Kundu'] | 2022-02-09 | null | null | null | null | ['scene-segmentation'] | ['computer-vision'] | [ 5.40866852e-01 6.44240826e-02 -6.59907833e-02 -5.11343300e-01
-1.07432151e+00 -1.15165544e+00 8.50790858e-01 -2.57092237e-01
-3.52124423e-01 4.82301831e-01 3.39140519e-02 -1.49026886e-01
-5.87168820e-02 -5.11452317e-01 -9.33804989e-01 -7.68367589e-01
2.26594314e-01 4.59432125e-01 2.32264008e-02 -2.20429227e-01
-5.86198792e-02 6.08229578e-01 -9.81514692e-01 3.50954056e-01
9.54714835e-01 1.03383636e+00 -2.21861769e-02 3.89152795e-01
1.57636717e-01 2.48372361e-01 -6.99703634e-01 -5.25956213e-01
8.03496838e-01 -6.29082799e-01 -8.02451491e-01 1.93283960e-01
6.32550895e-01 -1.63501173e-01 -3.56963247e-01 1.20349681e+00
3.31468701e-01 2.63254523e-01 1.02902329e+00 -1.39942372e+00
-8.08949471e-01 3.97277206e-01 -5.89146197e-01 1.75385088e-01
2.19844077e-02 5.33172607e-01 9.50768828e-01 -7.61878252e-01
6.81527793e-01 1.21871269e+00 6.82908118e-01 6.50388360e-01
-1.69353914e+00 -9.63912189e-01 1.99975267e-01 -8.70215073e-02
-1.28272462e+00 -3.46052587e-01 1.06250966e+00 -5.24194241e-01
6.80698216e-01 2.75078882e-03 2.95334369e-01 1.56387520e+00
7.20219910e-02 5.41251302e-01 1.18413925e+00 -2.33071640e-01
3.10792416e-01 -1.81400459e-02 -1.76565453e-01 3.12603861e-01
-5.29931765e-03 1.27241090e-01 -3.89367253e-01 -1.45662621e-01
8.76648664e-01 -2.74942577e-01 -1.39354557e-01 -6.62270486e-01
-1.18404102e+00 8.99132371e-01 5.54302692e-01 3.30202729e-02
-2.06331789e-01 -5.36938012e-02 4.26179856e-01 5.35872698e-01
4.18388516e-01 8.69425654e-01 -6.01541758e-01 7.77111277e-02
-9.04004753e-01 3.72304767e-01 4.76359129e-01 9.71229672e-01
9.13342595e-01 2.54889697e-01 -2.73855090e-01 8.13870788e-01
-1.04633057e-02 4.59794343e-01 6.89664304e-01 -1.13642263e+00
4.83865827e-01 3.55210066e-01 -2.04660222e-01 -6.79549396e-01
-1.91936806e-01 -5.05883515e-01 -7.83517241e-01 4.04651880e-01
8.94224346e-01 -2.61214912e-01 -1.29874098e+00 2.15378022e+00
3.28565747e-01 2.73629606e-01 1.62882596e-01 8.79109740e-01
1.10395394e-01 3.97043824e-01 1.82371587e-01 1.97747797e-01
1.18816721e+00 -7.86152720e-01 -2.21451834e-01 -3.83562237e-01
4.22389418e-01 -7.52113283e-01 1.36168337e+00 2.45930403e-01
-8.57771695e-01 -7.45615304e-01 -1.11778343e+00 -2.26336449e-01
-6.04029477e-01 -6.06476367e-02 3.13847601e-01 5.90067208e-01
-7.08971322e-01 6.49988234e-01 -8.53834093e-01 -3.70312124e-01
7.71413386e-01 3.65078479e-01 -5.58751047e-01 9.41294245e-03
-1.16157353e+00 6.87822878e-01 3.18793803e-01 -3.59725356e-01
-9.04247880e-01 -1.12889469e+00 -9.65017438e-01 -1.78012043e-01
1.01939805e-01 -8.95612538e-01 1.06541288e+00 -1.41615760e+00
-1.76617765e+00 9.57985044e-01 6.52723610e-02 -6.61157906e-01
6.30108356e-01 -1.79159269e-01 -3.51101786e-01 1.53303176e-01
3.13164026e-01 8.04255664e-01 1.53311431e+00 -1.12321866e+00
-3.41453940e-01 -4.54971373e-01 4.91305366e-02 2.85609424e-01
-2.88648963e-01 -2.94248432e-01 -1.31010965e-01 -1.14217055e+00
7.34718069e-02 -1.09524298e+00 -9.59614664e-02 2.51639813e-01
-4.97715473e-01 1.40383929e-01 8.70934308e-01 -5.87850451e-01
6.83711231e-01 -2.53681064e+00 1.19069353e-01 3.40094477e-01
1.61399782e-01 2.47264117e-01 -4.63381439e-01 6.89818189e-02
-4.47765291e-01 2.26599723e-02 -4.58822429e-01 -2.28077993e-01
4.62839454e-02 6.33229092e-02 -6.27214253e-01 5.89077413e-01
8.58875155e-01 9.31688786e-01 -8.48920405e-01 -2.00370029e-01
6.31051287e-02 2.13542327e-01 -8.64350855e-01 1.02075405e-01
-3.45359117e-01 7.96738446e-01 -4.02643770e-01 4.50718045e-01
7.61064827e-01 -2.16179997e-01 -5.46672046e-02 -2.94304401e-01
2.69134104e-01 2.28369519e-01 -9.00160134e-01 2.14384484e+00
-3.97013962e-01 7.08126009e-01 -5.47618465e-03 -1.15746033e+00
9.17725503e-01 -4.16095220e-02 4.70157593e-01 -6.26155317e-01
1.05343752e-01 1.58814922e-01 2.81430751e-01 -2.35293992e-02
3.13493341e-01 -2.47099742e-01 -3.26116562e-01 2.60309309e-01
4.33172524e-01 -2.41091534e-01 -2.72853583e-01 8.93938914e-02
1.21695840e+00 4.17604864e-01 2.48419091e-01 -4.11790043e-01
3.06883186e-01 2.17410296e-01 5.66042721e-01 6.52390897e-01
-4.66375798e-01 9.71702516e-01 4.72652644e-01 -1.33550078e-01
-1.27256918e+00 -1.64612377e+00 -1.66994989e-01 1.05458581e+00
1.02130368e-01 1.66586503e-01 -8.31818521e-01 -9.70772028e-01
3.08508784e-01 5.46001911e-01 -8.57666075e-01 -5.27477860e-01
-5.39417803e-01 -4.16573137e-01 9.03595209e-01 6.77433014e-01
6.35385990e-01 -7.12501764e-01 -2.96636760e-01 7.45930895e-02
4.06391285e-02 -1.31138217e+00 -7.25166917e-01 6.11128986e-01
-6.30596995e-01 -8.62100661e-01 -9.82674479e-01 -7.01745868e-01
5.25163531e-01 -2.44426914e-02 9.09858167e-01 -6.96336210e-01
-4.73808832e-02 2.45856017e-01 -2.51702696e-01 -1.92162231e-01
-4.37511563e-01 3.04209739e-01 1.96814597e-01 2.10310176e-01
4.29711401e-01 -9.43758190e-01 -5.12403905e-01 2.91853130e-01
-1.02933896e+00 -2.15657800e-01 5.86237013e-01 1.00178707e+00
5.97024083e-01 -1.53910577e-01 5.98648131e-01 -9.08708870e-01
5.73595643e-01 -5.08636653e-01 -4.29364473e-01 3.57060954e-02
-1.79127082e-01 2.96905100e-01 9.61659431e-01 -9.48401332e-01
-1.02426398e+00 2.44057000e-01 1.70271993e-02 -9.34655368e-01
-4.75442082e-01 3.46199200e-02 -6.00687325e-01 2.48255003e-02
1.16859126e+00 6.42872155e-02 4.88851778e-02 -2.62310594e-01
5.70489168e-01 4.03494060e-01 9.67517853e-01 -8.37193072e-01
1.37144375e+00 5.98708272e-01 -1.69859231e-01 -6.68766677e-01
-7.33816385e-01 -3.16631556e-01 -8.47479999e-01 3.02345783e-01
1.03670967e+00 -1.03655899e+00 -1.52433068e-01 6.04576886e-01
-1.04958868e+00 -6.28196359e-01 -6.30624890e-01 2.41983801e-01
-6.70592368e-01 9.09277648e-02 -2.87291557e-01 -1.26703456e-01
5.23271300e-02 -1.12008548e+00 1.09842050e+00 1.46308079e-01
-4.36635137e-01 -9.67835903e-01 3.70302126e-02 2.26932213e-01
2.60297924e-01 5.64143002e-01 9.81928885e-01 -1.15858448e+00
-2.88853645e-01 9.80781391e-02 -2.32393041e-01 6.07329965e-01
3.78509700e-01 -1.85505718e-01 -1.35415876e+00 -3.83962333e-01
-2.17838362e-02 -3.01133722e-01 7.85896003e-01 2.85719723e-01
1.17975771e+00 -6.04080036e-02 -1.71208173e-01 1.14155626e+00
1.12968957e+00 1.25650361e-01 4.61831212e-01 4.33699101e-01
7.72224724e-01 3.61953855e-01 4.95587498e-01 1.45535678e-01
-5.88640198e-02 6.11166120e-01 1.69710815e-01 -2.15666860e-01
-3.10733765e-01 -3.41046423e-01 4.29359943e-01 3.43282223e-01
3.77328753e-01 -5.21649867e-02 -9.06899333e-01 6.29268706e-01
-1.35488319e+00 -6.60375297e-01 3.40266258e-01 1.92155361e+00
9.73206580e-01 3.35612208e-01 3.71602744e-01 -4.35243659e-02
5.91086626e-01 1.25932217e-01 -1.13949215e+00 -2.81483769e-01
-2.45866343e-01 6.14396632e-01 7.16312528e-01 2.77451485e-01
-1.32901764e+00 9.56079662e-01 5.27465534e+00 9.66041267e-01
-1.21856844e+00 5.92903160e-02 5.31078756e-01 1.58368461e-02
-3.33331376e-01 -1.37890205e-01 -4.98779953e-01 4.81522560e-01
6.98280931e-01 -1.67840615e-01 5.35960555e-01 9.77272987e-01
-1.91795483e-01 3.19508851e-01 -1.46220207e+00 7.95193553e-01
-2.47229204e-01 -1.16891873e+00 3.19328845e-01 2.39128500e-01
9.03028548e-01 1.65005829e-02 6.76031947e-01 3.96149695e-01
4.73513991e-01 -9.49668229e-01 7.76363194e-01 2.19272628e-01
1.16828251e+00 -6.79347694e-01 2.87434936e-01 5.54422997e-02
-1.04610479e+00 -5.99763580e-02 -2.24775285e-01 1.97017998e-01
-1.71704084e-01 2.86761791e-01 -8.92051578e-01 4.91271496e-01
5.23526788e-01 6.59887731e-01 -5.03789723e-01 5.34365952e-01
6.10219948e-02 4.69351292e-01 -2.81356514e-01 6.88305020e-01
3.69155824e-01 -1.61936671e-01 6.68767393e-01 1.08150589e+00
1.68730631e-01 -1.57902688e-01 -7.20081571e-03 1.23130703e+00
-2.60075808e-01 -2.61278778e-01 -7.97216892e-01 -1.31688863e-01
5.11999130e-01 9.13509071e-01 -6.62291050e-01 -1.56329088e-02
-3.52916807e-01 1.50753784e+00 2.75285780e-01 6.77383721e-01
-9.77373362e-01 -3.69996518e-01 1.32617819e+00 1.17195718e-01
4.12119627e-01 -2.42590621e-01 -5.64660549e-01 -1.10521984e+00
-7.43052587e-02 -1.08299577e+00 2.04120040e-01 -5.01653552e-01
-1.72807968e+00 5.50938845e-01 6.83343783e-02 -1.45596528e+00
-3.47745180e-01 -6.65653884e-01 -5.59667110e-01 1.21158135e+00
-1.38696527e+00 -1.42113268e+00 -6.06117956e-02 1.03764606e+00
6.50046587e-01 -3.58366042e-01 7.04548359e-01 2.34463960e-01
-3.82287294e-01 1.08709359e+00 1.27139986e-01 4.45755541e-01
1.02436054e+00 -1.33188903e+00 8.76597524e-01 9.84448075e-01
1.95064440e-01 6.74807012e-01 6.39581919e-01 -5.66060901e-01
-9.64598656e-01 -1.32197738e+00 1.44495487e-01 -7.50368953e-01
8.35844398e-01 -5.61974406e-01 -1.12358129e+00 8.03876281e-01
8.86780967e-04 3.00904155e-01 4.84053910e-01 -1.64281040e-01
-7.26923406e-01 -1.16672941e-01 -1.27658212e+00 6.22598767e-01
1.09378278e+00 -1.00673687e+00 -6.70710921e-01 -4.10744660e-02
9.31936502e-01 -4.24449116e-01 -8.42607081e-01 4.15901124e-01
4.39740568e-01 -7.43625343e-01 1.14074802e+00 -7.44963825e-01
4.57227081e-01 -2.70673573e-01 -2.37735659e-01 -1.54079974e+00
-4.09401894e-01 -7.35488236e-01 1.79988906e-01 1.24295413e+00
2.93632865e-01 -6.98725760e-01 8.01197886e-01 5.00181913e-01
-1.97034940e-01 -4.43129510e-01 -9.71432030e-01 -1.15721262e+00
8.09212327e-01 -4.13200885e-01 6.51009440e-01 1.21455729e+00
-4.07032430e-01 3.41352880e-01 1.03718303e-01 3.62895548e-01
5.24847209e-01 -1.47489116e-01 9.34423983e-01 -1.16816497e+00
-4.23845023e-01 -5.45945585e-01 -3.58699828e-01 -1.10998583e+00
3.58162910e-01 -9.37831104e-01 -7.66988248e-02 -7.45876372e-01
-2.14021966e-01 -4.50599343e-01 -3.69603276e-01 3.70133072e-01
-8.99055377e-02 2.55935967e-01 2.38684937e-01 1.38048381e-01
-1.36732459e-01 4.84085202e-01 1.42010117e+00 -2.14155525e-01
-2.88550109e-01 -8.40849727e-02 -6.99128211e-01 7.70129800e-01
7.90351808e-01 -5.15030801e-01 -6.32761121e-01 -3.88610959e-01
-1.93295687e-01 -3.12912166e-01 5.85777402e-01 -1.10610116e+00
1.44877229e-02 -1.83581233e-01 5.59258163e-01 -1.67077065e-01
3.67765814e-01 -8.63463640e-01 -3.37912440e-02 1.21234618e-01
-3.49570453e-01 -1.81033731e-01 4.90454763e-01 6.10826492e-01
-1.52839765e-01 6.55064881e-02 1.11353278e+00 1.39056683e-01
-7.08114028e-01 2.24608928e-01 -3.63406911e-02 5.23048103e-01
9.20076668e-01 -5.34291208e-01 -1.80402637e-01 -1.69502422e-01
-7.63229311e-01 -1.75243035e-01 7.09665179e-01 4.30888891e-01
2.25730717e-01 -1.32896674e+00 -4.95582104e-01 6.79469883e-01
1.31245032e-01 2.02583477e-01 1.47711471e-01 5.75627387e-01
-2.08264917e-01 3.01896147e-02 -5.53050756e-01 -7.19211698e-01
-7.15076447e-01 4.32833225e-01 4.96401936e-01 -2.41881102e-01
-4.83787388e-01 1.02288806e+00 7.10063934e-01 -4.84861135e-01
1.01321852e-02 -4.61451769e-01 2.97339916e-01 -9.56648402e-03
6.95597306e-02 4.83957026e-03 2.76433285e-02 -6.75251484e-01
-2.83093601e-01 5.09198010e-01 -1.57835484e-01 -2.50887305e-01
1.10689855e+00 -3.36407535e-02 5.82491279e-01 2.01494306e-01
1.56237233e+00 2.78358757e-02 -1.92089653e+00 -4.01831627e-01
-8.54141340e-02 -3.32113266e-01 -3.16370755e-01 -8.59132409e-01
-9.30424929e-01 8.85508955e-01 6.55777872e-01 4.18392755e-02
1.33380723e+00 -5.09035736e-02 7.75306940e-01 1.27489984e-01
4.94676121e-02 -1.02522969e+00 3.47588688e-01 5.72707772e-01
8.38044763e-01 -1.19835711e+00 -4.06525075e-01 -1.74354434e-01
-8.38963211e-01 8.32127750e-01 7.17141032e-01 -4.63516980e-01
6.26677513e-01 1.78169191e-01 2.33473495e-01 3.28426063e-02
-2.18705073e-01 -2.27351099e-01 5.25486946e-01 1.07802367e+00
-3.32296863e-02 -4.80202734e-02 4.10718471e-01 7.88930237e-01
-4.17374849e-01 -3.84206057e-01 9.21640024e-02 5.93980610e-01
4.63075377e-02 -1.16708076e+00 -4.21485305e-01 3.03993791e-01
-4.17800009e-01 -1.20954879e-01 -4.02628779e-01 1.03308225e+00
3.01026225e-01 5.01964629e-01 2.46870086e-01 -4.35100943e-01
3.91649395e-01 3.17340344e-01 4.41034764e-01 -7.39042997e-01
-4.79422957e-01 -1.39204673e-02 -4.08018321e-01 -6.19408906e-01
-1.76556155e-01 -8.68728280e-01 -1.13795078e+00 -2.92711542e-03
8.55594054e-02 -2.06783161e-01 4.99947101e-01 1.01936805e+00
4.44890261e-01 6.40739322e-01 6.68080449e-01 -8.81773233e-01
-7.64774859e-01 -8.36834311e-01 -5.99859059e-01 8.06616962e-01
6.56292319e-01 -7.86921203e-01 -3.31960410e-01 3.64662468e-01] | [9.809328079223633, 2.005810260772705] |
1a1ab5da-4070-486e-80e2-7b123c08f3b4 | avis-autonomous-visual-information-seeking | 2306.08129 | null | https://arxiv.org/abs/2306.08129v1 | https://arxiv.org/pdf/2306.08129v1.pdf | AVIS: Autonomous Visual Information Seeking with Large Language Models | In this paper, we propose an autonomous information seeking visual question answering framework, AVIS. Our method leverages a Large Language Model (LLM) to dynamically strategize the utilization of external tools and to investigate their outputs, thereby acquiring the indispensable knowledge needed to provide answers to the posed questions. Responding to visual questions that necessitate external knowledge, such as "What event is commemorated by the building depicted in this image?", is a complex task. This task presents a combinatorial search space that demands a sequence of actions, including invoking APIs, analyzing their responses, and making informed decisions. We conduct a user study to collect a variety of instances of human decision-making when faced with this task. This data is then used to design a system comprised of three components: an LLM-powered planner that dynamically determines which tool to use next, an LLM-powered reasoner that analyzes and extracts key information from the tool outputs, and a working memory component that retains the acquired information throughout the process. The collected user behavior serves as a guide for our system in two key ways. First, we create a transition graph by analyzing the sequence of decisions made by users. This graph delineates distinct states and confines the set of actions available at each state. Second, we use examples of user decision-making to provide our LLM-powered planner and reasoner with relevant contextual instances, enhancing their capacity to make informed decisions. We show that AVIS achieves state-of-the-art results on knowledge-intensive visual question answering benchmarks such as Infoseek and OK-VQA. | ['Alireza Fathi', 'Cordelia Schmid', 'David A Ross', 'Yizhou Sun', 'Kai-Wei Chang', 'Chen Sun', 'Ahmet Iscen', 'Ziniu Hu'] | 2023-06-13 | null | null | null | null | ['visual-question-answering-1', 'question-answering'] | ['computer-vision', 'natural-language-processing'] | [ 1.16703644e-01 1.79175600e-01 -1.45398071e-02 -4.17176336e-01
-6.77395403e-01 -1.31303668e+00 7.12237537e-01 3.25847477e-01
-1.88857779e-01 -4.96608876e-02 4.54806656e-01 -8.48708153e-01
-1.72535479e-01 -7.36830890e-01 -4.25047636e-01 7.93442056e-02
3.42030793e-01 5.57673395e-01 4.23931539e-01 -1.96076259e-01
6.30229235e-01 5.03521025e-01 -1.83758128e+00 6.94309950e-01
8.05428803e-01 9.09777522e-01 3.84629101e-01 8.72245908e-01
-6.45032942e-01 1.44738245e+00 -7.12253332e-01 -6.56292439e-02
3.34526226e-02 -2.62997955e-01 -1.51323938e+00 6.86415359e-02
1.76296994e-01 -5.04899859e-01 3.26793224e-01 6.85476661e-01
7.98362717e-02 3.29902321e-01 5.15993714e-01 -1.41041088e+00
-5.76666951e-01 5.87310314e-01 -1.11823892e-02 1.39420331e-01
1.09084582e+00 9.57089245e-01 9.40205336e-01 -6.26636326e-01
1.02577794e+00 1.42060816e+00 -1.33288473e-01 4.38261241e-01
-9.09444571e-01 -1.93036020e-01 3.57574910e-01 3.96829873e-01
-9.54721153e-01 -5.54316282e-01 7.47816443e-01 -6.73130989e-01
1.19899917e+00 6.94231391e-01 7.04124272e-01 8.54269326e-01
-1.56379640e-01 9.94043231e-01 9.78981197e-01 -6.07044935e-01
6.55740976e-01 2.66126513e-01 4.82685089e-01 1.01120913e+00
-2.98966348e-01 -3.28081727e-01 -9.01454806e-01 -3.23076636e-01
3.92455369e-01 -1.47686645e-01 -8.02697465e-02 -3.72906804e-01
-8.81829917e-01 3.93735617e-01 2.51988590e-01 3.23429972e-01
-4.89166975e-01 8.92571509e-02 2.95559093e-02 5.24540782e-01
-3.40239584e-01 1.06631577e+00 -2.57645726e-01 -5.24314225e-01
-4.54377919e-01 1.23674154e-01 1.29301822e+00 1.07749057e+00
9.94463503e-01 -4.75232631e-01 -4.62975770e-01 4.12627697e-01
5.34817994e-01 4.49569225e-01 1.04448386e-01 -1.49347186e+00
3.76445651e-01 1.56517339e+00 4.74386334e-01 -9.56970394e-01
-2.92505115e-01 6.41077086e-02 2.00626820e-01 3.03139538e-01
5.26574135e-01 1.67391390e-01 -9.15935755e-01 1.36433387e+00
5.20781219e-01 -5.06736457e-01 -1.70765460e-01 1.03957176e+00
7.15788305e-01 5.87002873e-01 3.44137698e-01 1.59173653e-01
1.78401434e+00 -7.33936727e-01 -5.18967330e-01 -4.85629082e-01
8.05441797e-01 -4.77609932e-01 1.72661698e+00 3.56227547e-01
-1.04888892e+00 -2.80914038e-01 -7.81887591e-01 -3.45919997e-01
-4.94922876e-01 -1.50930211e-01 4.97731149e-01 1.78792253e-01
-9.62654412e-01 4.86577526e-02 -6.32748306e-01 -4.52839702e-01
3.00973475e-01 1.83221754e-02 8.59540328e-03 -2.50702083e-01
-8.50459516e-01 8.83755326e-01 -6.98745549e-02 -1.02422960e-01
-1.25705969e+00 -6.92108154e-01 -7.95717239e-01 3.80126715e-01
1.01823497e+00 -8.74041498e-01 1.77438295e+00 -6.89640164e-01
-1.34718132e+00 8.53399038e-01 -2.51643419e-01 1.46808594e-01
4.72146481e-01 -1.53526723e-01 -1.42866075e-01 1.96001768e-01
2.88097084e-01 4.17555153e-01 7.68976152e-01 -1.24605310e+00
-6.59747660e-01 -6.84025049e-01 6.88282430e-01 1.78581446e-01
7.91340880e-03 1.82343721e-01 -9.24652278e-01 2.70123512e-01
-2.76419401e-01 -7.73752570e-01 -2.29943302e-02 -3.91709879e-02
-5.23669600e-01 -3.66784275e-01 6.16586685e-01 -4.77932602e-01
1.52094269e+00 -1.98810232e+00 1.97108462e-01 4.45431143e-01
4.13465708e-01 6.24038558e-03 -3.09018582e-01 8.10368836e-01
3.95181447e-01 4.33504879e-01 8.18606019e-02 -1.44016847e-01
3.15226674e-01 -1.58692617e-02 -2.90542543e-01 -2.01631427e-01
8.64367411e-02 1.14145756e+00 -9.13001716e-01 -7.26424813e-01
2.89609611e-01 -1.19340718e-02 -4.27734882e-01 7.43092656e-01
-9.50840533e-01 2.53213853e-01 -8.73556137e-01 1.02195084e+00
1.67702153e-01 -7.69095600e-01 3.91534328e-01 -1.03580877e-01
-3.20254058e-01 2.92827606e-01 -1.10095894e+00 1.69122827e+00
-7.69784510e-01 3.60976487e-01 2.22436875e-01 -2.38324136e-01
5.38513720e-01 -1.64756700e-01 -1.06536232e-01 -9.92201746e-01
7.49124363e-02 -1.87720269e-01 -2.97237217e-01 -1.16278040e+00
3.43899965e-01 5.90008557e-01 -2.84116358e-01 9.10529077e-01
-1.71770796e-01 -1.66841254e-01 3.09783995e-01 6.70418382e-01
1.73315728e+00 1.95121348e-01 4.56545830e-01 2.05087587e-02
4.73711938e-01 5.59451282e-01 2.99485680e-02 9.71738815e-01
-7.28967879e-03 -2.46423230e-01 9.39313829e-01 -6.16303802e-01
-5.26800632e-01 -8.79458845e-01 6.62353814e-01 1.39276314e+00
3.26076239e-01 -7.20047832e-01 -6.11391842e-01 -7.86253273e-01
-8.82564709e-02 1.17878234e+00 -6.50942028e-01 -1.60361856e-01
-3.25078130e-01 4.39748138e-01 2.18155071e-01 4.28126305e-01
4.82898980e-01 -1.55785942e+00 -1.76973212e+00 -1.66326433e-01
-3.55284899e-01 -8.42361093e-01 -4.51939434e-01 -1.01525262e-01
-3.99825245e-01 -1.53437030e+00 1.77634090e-01 -2.38136247e-01
7.21082449e-01 2.24210992e-01 1.52564836e+00 4.48362470e-01
-3.73237461e-01 1.15722072e+00 -4.48201358e-01 -3.61293226e-01
-5.56511939e-01 2.80560218e-02 -7.34468341e-01 -8.84359237e-03
3.95199627e-01 -3.37562382e-01 -7.12212503e-01 5.00230610e-01
-9.92135108e-01 4.04632837e-01 5.59290946e-01 6.49656057e-02
3.79520208e-01 -2.95948178e-01 -1.79019049e-02 -9.70246613e-01
1.09554338e+00 -5.54860234e-01 -8.60109210e-01 8.98555219e-01
-5.56826830e-01 5.89111865e-01 4.72137839e-01 -3.00668627e-01
-1.27971435e+00 1.63741976e-01 3.96261156e-01 -2.40111649e-01
-3.17894489e-01 7.07862079e-01 -3.49459529e-01 3.98057252e-01
8.98953259e-01 7.92808980e-02 -1.48787066e-01 -4.59572732e-01
7.78261960e-01 6.83253467e-01 5.89784384e-01 -9.21105683e-01
6.39472783e-01 3.26389700e-01 -4.28510487e-01 -5.24466932e-01
-5.71322203e-01 -6.97447419e-01 -6.27083406e-02 -6.48079455e-01
8.07169497e-01 -1.66440129e-01 -1.65568209e+00 6.51406646e-02
-1.22114682e+00 -7.19958544e-01 -2.93703049e-01 -4.18353379e-01
-6.04495764e-01 -7.42660537e-02 -1.48734391e-01 -1.22552824e+00
-2.69619823e-01 -1.35793388e+00 1.08527863e+00 3.65248293e-01
-7.04453170e-01 -5.01215339e-01 3.83589752e-02 7.40008891e-01
5.93778849e-01 1.91185772e-02 1.30452704e+00 -6.65095210e-01
-1.13449752e+00 9.55099985e-02 -1.59315690e-01 -4.23501462e-01
-7.24101812e-02 1.93254557e-02 -8.52966130e-01 2.50133634e-01
-2.69354165e-01 -4.19544935e-01 1.40922129e-01 -2.87521362e-01
1.24258423e+00 -3.71218592e-01 -4.25994039e-01 1.75587520e-01
1.14003766e+00 5.24234414e-01 5.20674467e-01 4.37932342e-01
5.15350759e-01 8.31385672e-01 5.94035387e-01 4.20123816e-01
7.65821099e-01 6.41461492e-01 6.81307733e-01 2.27725089e-01
7.70949051e-02 -4.35795009e-01 1.70045003e-01 -6.10621180e-03
1.98448956e-01 -2.94470698e-01 -1.39835310e+00 6.26665294e-01
-2.14545417e+00 -7.56909013e-01 2.66310364e-01 1.95288157e+00
7.03778982e-01 -1.40700936e-01 -3.01328227e-02 -2.27469921e-01
9.99193341e-02 2.23543823e-01 -1.02974653e+00 -4.67662632e-01
6.78637862e-01 1.94989499e-02 -1.51009813e-01 5.74042559e-01
-4.16083783e-01 1.08221960e+00 5.56254625e+00 2.11059645e-01
-7.03858912e-01 -1.90583110e-01 1.69571564e-01 -2.61757690e-02
-6.85851276e-01 4.47459370e-01 -3.40118915e-01 2.38433912e-01
8.80602241e-01 -6.35209978e-02 7.99548864e-01 9.44932818e-01
2.10848391e-01 -6.59813762e-01 -1.41822171e+00 8.21663737e-01
-1.14057273e-01 -1.37856722e+00 5.21569960e-02 -1.42059013e-01
-1.21385783e-01 -3.52126479e-01 -8.21130648e-02 2.97116846e-01
7.58095622e-01 -9.28858697e-01 8.15408468e-01 9.47757483e-01
5.49539745e-01 -3.84466380e-01 -3.33941318e-02 6.95285261e-01
-1.00780201e+00 -3.37245941e-01 4.98914450e-01 -7.60563537e-02
1.59870684e-01 1.95406765e-01 -1.12351882e+00 3.52404922e-01
7.28755593e-01 1.56163648e-01 -7.75764167e-01 5.39395630e-01
-6.01161838e-01 4.16159272e-01 -2.99854606e-01 -3.84486198e-01
-8.13132599e-02 8.22549537e-02 5.22416413e-01 8.26812446e-01
-1.77705109e-01 2.57618427e-01 2.57559121e-01 1.29861712e+00
2.41685156e-02 -7.18341321e-02 -6.14330828e-01 -3.09562117e-01
6.79505289e-01 1.30673289e+00 -6.36960745e-01 -5.86563885e-01
-2.41655275e-01 8.21148694e-01 5.59642911e-01 6.43425047e-01
-5.03888249e-01 -2.96722800e-01 5.71791828e-01 4.04845953e-01
-4.54040393e-02 -2.35654414e-01 2.04978907e-03 -9.34230447e-01
2.25916967e-01 -1.28967214e+00 6.46806777e-01 -1.49416816e+00
-6.50581062e-01 5.75626433e-01 2.03869194e-01 -6.10755503e-01
-4.76675481e-01 -3.08029562e-01 -5.34271836e-01 9.17635143e-01
-1.20681500e+00 -1.02207720e+00 -8.14059436e-01 5.56664169e-01
5.09257674e-01 1.89910606e-01 7.14832842e-01 -2.12049171e-01
-4.89564210e-01 -2.97343768e-02 -1.03116000e+00 -1.55403897e-01
4.15239125e-01 -1.14628708e+00 4.14254636e-01 7.92942166e-01
1.16633147e-01 9.70511377e-01 5.21308720e-01 -8.84222150e-01
-2.01605535e+00 -5.09370923e-01 6.21566951e-01 -1.05928564e+00
6.16723835e-01 -3.38716567e-01 -9.28277075e-01 7.59449542e-01
1.24501735e-01 -3.06221783e-01 6.55431867e-01 1.27327383e-01
-6.05521023e-01 2.19058618e-01 -9.75724041e-01 8.64125550e-01
1.34737110e+00 -8.38305235e-01 -9.01954651e-01 7.43060857e-02
7.64260530e-01 -3.52034330e-01 -4.10689920e-01 -1.22544043e-01
6.05649054e-01 -9.76285219e-01 6.22346222e-01 -1.15867817e+00
5.01198769e-01 -6.15880191e-01 -7.23267645e-02 -1.06880081e+00
-9.25366208e-02 -8.69160056e-01 -3.52093458e-01 9.86646771e-01
4.80109125e-01 -5.22580504e-01 4.60769176e-01 1.50594366e+00
8.16046819e-02 -6.84247792e-01 -4.34004277e-01 -2.57246882e-01
-9.28691983e-01 -8.12467694e-01 8.15248013e-01 6.15895331e-01
2.36853421e-01 3.90789539e-01 1.78485766e-01 1.83850542e-01
1.91972762e-01 5.15728474e-01 1.02238595e+00 -1.03004086e+00
-3.38865191e-01 -3.28882068e-01 2.05301970e-01 -1.17704964e+00
-1.48435785e-02 -8.10040534e-01 6.69719651e-02 -2.02375388e+00
8.51576552e-02 -3.43647003e-01 1.52572906e-02 9.79080021e-01
5.87533740e-03 -7.30997980e-01 4.21799451e-01 4.11996394e-01
-1.10330939e+00 -8.66506919e-02 1.02396595e+00 -2.34805897e-01
-6.12543166e-01 -2.24890396e-01 -9.52984452e-01 7.59556532e-01
3.93985689e-01 -1.90051228e-01 -5.74939549e-01 -5.80816090e-01
8.27520669e-01 3.46627980e-01 5.40946841e-01 -7.71994293e-01
6.54889762e-01 -6.93996608e-01 1.35153234e-01 -4.97254640e-01
1.82974905e-01 -1.07297206e+00 2.28933692e-01 3.68991852e-01
-7.50213802e-01 3.13922256e-01 9.67245363e-03 4.10481125e-01
6.68735504e-02 -1.06775522e-01 5.31000458e-02 -3.62123728e-01
-1.11502016e+00 -4.68168855e-02 -7.15064824e-01 1.49126202e-01
1.07757783e+00 -7.73989111e-02 -7.40251005e-01 -5.35645843e-01
-4.95132178e-01 8.61157656e-01 5.78962266e-01 5.98971307e-01
5.80875635e-01 -7.60148883e-01 -2.90992931e-02 8.32716599e-02
6.72869205e-01 -2.96111703e-02 1.53137147e-01 4.72007245e-01
-4.02905136e-01 1.49178728e-01 -1.09058142e-01 -4.48901474e-01
-1.17628586e+00 8.07398498e-01 2.98086733e-01 -2.57957280e-01
-3.90550166e-01 4.83971775e-01 -2.21662093e-02 -2.65636235e-01
2.17386648e-01 -5.19574285e-01 -3.50447655e-01 1.55779809e-01
7.85006106e-01 3.55930954e-01 1.96205135e-02 3.10670439e-04
-6.70780659e-01 3.83563042e-01 9.26012248e-02 -3.75795633e-01
1.02973092e+00 1.08802058e-02 -4.96011972e-02 3.61440212e-01
6.71133697e-01 5.78729473e-02 -1.01656020e+00 -1.85437754e-01
2.32340068e-01 -5.06225109e-01 -3.64929557e-01 -1.52789009e+00
-4.91171956e-01 6.58222258e-01 4.77010868e-02 5.65599263e-01
1.11119926e+00 4.75672275e-01 3.02521467e-01 7.35255420e-01
5.44299722e-01 -1.09774482e+00 3.56757164e-01 3.04187834e-01
1.21421754e+00 -1.03766823e+00 -3.16671342e-01 -3.38702351e-01
-7.38082349e-01 9.13463831e-01 9.20586288e-01 6.20479941e-01
1.55390203e-01 3.52672279e-01 3.96494508e-01 -8.64725590e-01
-1.28354335e+00 -2.77450711e-01 2.09234312e-01 4.31595236e-01
-7.91406631e-02 4.51390538e-03 1.76354662e-01 5.39650798e-01
6.24408387e-02 2.62149423e-01 3.49053383e-01 1.36079514e+00
-5.12774825e-01 -9.52372134e-01 -3.57933044e-01 4.42293793e-01
2.48955578e-01 1.98038593e-01 -1.08062041e+00 6.44438922e-01
-3.67753506e-01 1.12983227e+00 -1.23809882e-01 -3.32532167e-01
8.21817875e-01 3.77784222e-01 2.39509225e-01 -6.78177774e-01
-9.18974340e-01 -6.39114618e-01 3.85886967e-01 -1.26965308e+00
6.12050295e-03 -4.30450052e-01 -1.43301356e+00 1.59237478e-02
1.94965556e-01 1.45914126e-02 8.28071713e-01 1.04712439e+00
8.99912059e-01 5.79798877e-01 1.90679997e-01 -3.13423216e-01
-5.29234290e-01 -5.45367658e-01 3.57113361e-01 6.72056258e-01
1.59978718e-01 -5.31197727e-01 -2.01395929e-01 2.27681804e-03] | [9.02309799194336, 7.105389595031738] |
0f84f319-3e92-4eb6-b5c2-ce85ee19c609 | decoupled-novel-object-captioner | 1804.03803 | null | http://arxiv.org/abs/1804.03803v2 | http://arxiv.org/pdf/1804.03803v2.pdf | Decoupled Novel Object Captioner | Image captioning is a challenging task where the machine automatically
describes an image by sentences or phrases. It often requires a large number of
paired image-sentence annotations for training. However, a pre-trained
captioning model can hardly be applied to a new domain in which some novel
object categories exist, i.e., the objects and their description words are
unseen during model training. To correctly caption the novel object, it
requires professional human workers to annotate the images by sentences with
the novel words. It is labor expensive and thus limits its usage in real-world
applications.
In this paper, we introduce the zero-shot novel object captioning task where
the machine generates descriptions without extra sentences about the novel
object. To tackle the challenging problem, we propose a Decoupled Novel Object
Captioner (DNOC) framework that can fully decouple the language sequence model
from the object descriptions. DNOC has two components. 1) A Sequence Model with
the Placeholder (SM-P) generates a sentence containing placeholders. The
placeholder represents an unseen novel object. Thus, the sequence model can be
decoupled from the novel object descriptions. 2) A key-value object memory
built upon the freely available detection model, contains the visual
information and the corresponding word for each object. The SM-P will generate
a query to retrieve the words from the object memory. The placeholder will then
be filled with the correct word, resulting in a caption with novel object
descriptions. The experimental results on the held-out MSCOCO dataset
demonstrate the ability of DNOC in describing novel concepts in the zero-shot
novel object captioning task. | ['Linchao Zhu', 'Yu Wu', 'Yi Yang', 'Lu Jiang'] | 2018-04-11 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 6.90129876e-01 2.27118745e-01 -2.57473327e-02 -2.54887879e-01
-8.56589317e-01 -4.79048580e-01 5.69510877e-01 2.15483442e-01
-4.28134024e-01 7.56659150e-01 -5.74119911e-02 -1.25492434e-03
5.94594538e-01 -6.52581573e-01 -1.13061750e+00 -7.46434093e-01
4.61960346e-01 6.08718932e-01 5.10693371e-01 -2.39487141e-02
2.71589100e-01 1.40692458e-01 -1.63156998e+00 5.38573563e-01
5.29689670e-01 1.05074155e+00 1.15592837e+00 5.34923196e-01
-4.86284137e-01 5.68734586e-01 -6.01592720e-01 -4.35986966e-01
5.85426949e-02 -4.67511356e-01 -6.56771004e-01 2.79397607e-01
3.49093944e-01 -5.29884338e-01 -1.30583823e-01 1.11441755e+00
5.25872469e-01 -1.84890609e-02 4.71046746e-01 -1.39314604e+00
-1.01095712e+00 3.59077662e-01 -4.40574706e-01 9.51371342e-02
3.44672978e-01 4.52179641e-01 8.27730894e-01 -1.25428104e+00
8.72297347e-01 1.15090334e+00 7.90168568e-02 9.22664642e-01
-1.15043473e+00 -6.34392679e-01 2.54924446e-01 2.95637459e-01
-1.55682993e+00 -2.99943268e-01 7.10646272e-01 -4.90433455e-01
6.86887264e-01 1.55672505e-01 5.15929461e-01 1.26843476e+00
-2.98354384e-02 9.68492925e-01 5.62206328e-01 -2.75529653e-01
2.79299706e-01 5.67383587e-01 -4.13768888e-02 3.71740073e-01
7.24531859e-02 -1.64443672e-01 -3.24483335e-01 1.07168719e-01
5.68110228e-01 1.82643190e-01 -2.99483329e-01 -4.59934533e-01
-1.42368984e+00 7.19901145e-01 5.71474433e-01 2.71846652e-01
-4.62012619e-01 3.26523751e-01 4.77752864e-01 -2.04198897e-01
2.38560334e-01 2.87850708e-01 -2.39644200e-01 1.42544284e-01
-5.96441925e-01 2.64131188e-01 4.00065124e-01 1.37621117e+00
8.85997832e-01 -3.83129328e-01 -6.09763086e-01 6.97880983e-01
2.14905739e-01 8.13496411e-01 5.33061624e-01 -5.39127886e-01
5.14936149e-01 4.58267480e-01 3.50835353e-01 -9.62396860e-01
1.50695294e-01 -2.56774068e-01 -5.42974591e-01 -2.88995415e-01
-1.77372098e-01 2.12952346e-01 -1.04542959e+00 1.69199622e+00
3.25993896e-01 2.65585274e-01 4.62438911e-01 1.06752706e+00
1.19451869e+00 1.19337463e+00 3.53231728e-01 -3.36277872e-01
1.70820129e+00 -1.27823567e+00 -8.24737072e-01 -7.84068465e-01
3.35950971e-01 -7.53067911e-01 1.10458350e+00 -2.65543431e-01
-9.70089316e-01 -8.12636375e-01 -8.84812057e-01 -2.22268224e-01
-3.68730426e-01 3.77162956e-02 8.91289189e-02 -6.51334822e-02
-8.74205530e-01 -6.99543878e-02 -4.01990205e-01 -5.20651400e-01
4.03865755e-01 1.68055609e-01 -4.01215434e-01 -3.23865652e-01
-1.13952363e+00 9.17413533e-01 1.00240588e+00 9.91630927e-02
-1.27749014e+00 -3.87143910e-01 -1.24662590e+00 2.65080422e-01
4.53481287e-01 -7.50594437e-01 1.48019004e+00 -1.36781335e+00
-8.36279690e-01 1.03226662e+00 -4.27636981e-01 -3.75110924e-01
1.72719717e-01 8.04635882e-02 -2.93956637e-01 4.00479525e-01
5.89036286e-01 1.37360859e+00 1.05981410e+00 -1.65515649e+00
-6.77503645e-01 -5.15951067e-02 5.50965704e-02 2.45231554e-01
-2.51554310e-01 -1.18606888e-01 -6.86143696e-01 -5.85680544e-01
-1.66037858e-01 -8.16168964e-01 4.22899127e-02 2.89261252e-01
-2.27979183e-01 -3.61180216e-01 1.02156591e+00 -4.60687339e-01
8.41615438e-01 -2.49821258e+00 -1.89682007e-01 -4.18587387e-01
-1.92339029e-02 4.02688682e-01 -5.09191990e-01 4.91070479e-01
-1.27986103e-01 4.06828821e-02 -3.98232251e-01 -3.23917150e-01
-1.96359232e-01 4.02538598e-01 -7.52436101e-01 -2.68367236e-03
6.96281791e-01 1.37135017e+00 -1.25560009e+00 -8.28726590e-01
9.86494273e-02 3.26756626e-01 -1.80118978e-01 6.49468601e-01
-7.01484442e-01 2.10760087e-01 -4.57649618e-01 4.50791806e-01
7.69264519e-01 -5.46985567e-01 -2.29186788e-01 -2.95292288e-01
-3.90239037e-03 -2.30291158e-01 -7.31614053e-01 1.76745427e+00
-2.37586901e-01 5.03091753e-01 -3.01965594e-01 -7.31315851e-01
1.02386510e+00 4.00066227e-01 1.97707396e-02 -7.11206436e-01
-3.18867564e-02 3.12204927e-01 -3.47200960e-01 -1.05770433e+00
4.91177499e-01 -4.57127184e-01 -2.08866790e-01 3.09367567e-01
1.07840739e-01 -9.08640772e-02 3.21545243e-01 3.72045845e-01
8.66116762e-01 3.52671668e-02 2.78878957e-01 1.27457350e-01
7.03552485e-01 1.58338472e-01 4.47212696e-01 8.82195413e-01
-1.77002117e-01 9.99840736e-01 2.28893995e-01 -3.73106688e-01
-1.24483085e+00 -1.33389163e+00 2.77707223e-02 9.42711234e-01
6.57355011e-01 1.25607420e-02 -6.03636026e-01 -6.70563698e-01
-1.82823002e-01 8.39300454e-01 -5.70342720e-01 -3.97614986e-01
-5.04482627e-01 -1.08825915e-01 5.93501665e-02 4.58753169e-01
5.83123147e-01 -1.77330375e+00 -6.52737498e-01 2.45966852e-01
-5.66251934e-01 -1.39182460e+00 -7.76178956e-01 -1.54596061e-01
-5.22247314e-01 -8.03211570e-01 -1.08869231e+00 -1.47671449e+00
1.11729622e+00 4.81202006e-01 8.97227347e-01 1.19627804e-01
-1.80724531e-01 3.70269984e-01 -5.52642941e-01 -4.40205127e-01
-4.74246532e-01 -2.17886329e-01 -2.11503401e-01 3.33172053e-01
3.78528953e-01 -5.87978251e-02 -5.28378785e-01 2.62665600e-02
-1.31899190e+00 6.79302335e-01 9.19730961e-01 8.95062625e-01
7.49523997e-01 -4.79807198e-01 7.11463034e-01 -4.78809327e-01
2.79984087e-01 -6.44082189e-01 -3.25914294e-01 4.34616446e-01
-7.86905885e-02 1.43541053e-01 5.05381405e-01 -8.36208403e-01
-8.75005484e-01 5.46599269e-01 -3.40808630e-02 -4.95786160e-01
-1.24352956e-02 4.57235754e-01 -2.12635919e-01 3.30275476e-01
4.22264814e-01 8.89683068e-01 5.02052829e-02 -3.22600484e-01
3.66130769e-01 9.39649880e-01 7.70897806e-01 -3.06955785e-01
9.15320396e-01 4.52898979e-01 -4.41633910e-01 -6.41796172e-01
-1.10382187e+00 -5.21563947e-01 -4.10375208e-01 -2.14977726e-01
1.15329504e+00 -1.04778957e+00 -5.54123819e-01 2.15259999e-01
-1.93434894e+00 1.89298600e-01 -4.54352617e-01 2.38583699e-01
-7.63589501e-01 2.21220762e-01 -1.35154426e-01 -7.75913358e-01
-5.09409606e-01 -1.15484703e+00 1.44245648e+00 4.69687700e-01
4.48523723e-02 -4.81698394e-01 -1.99775815e-01 4.14899141e-01
1.80445060e-01 6.67748824e-02 9.33059931e-01 -7.19103396e-01
-9.17655587e-01 -1.99928239e-01 -6.07432425e-01 4.34737116e-01
-1.21952429e-01 -6.18901670e-01 -9.08564806e-01 -2.63570398e-01
8.32184497e-03 -4.56780732e-01 6.37035429e-01 -9.74970087e-02
1.09762180e+00 -4.36261982e-01 -5.55630624e-01 4.88278363e-03
1.49675834e+00 4.55824584e-01 6.12692118e-01 1.99672118e-01
4.80125755e-01 5.73881865e-01 8.20982814e-01 1.74317226e-01
3.19569230e-01 5.62890649e-01 5.40458441e-01 -1.07661381e-01
-9.73766074e-02 -5.90610862e-01 3.07113707e-01 5.78394175e-01
6.33395374e-01 -6.27726376e-01 -9.27599490e-01 1.01862109e+00
-2.04443645e+00 -8.20262551e-01 1.04678638e-01 1.86338401e+00
9.54728782e-01 1.02457190e-02 -4.76166189e-01 -3.91765267e-01
1.12012303e+00 -9.25825629e-03 -7.88034379e-01 -1.79909497e-01
-1.24696597e-01 -1.85759887e-01 3.71950753e-02 3.18654746e-01
-8.07091832e-01 1.08744156e+00 4.93640041e+00 7.39982486e-01
-1.02419055e+00 2.54464418e-01 3.58490437e-01 1.34292161e-02
-2.48407900e-01 8.40508863e-02 -9.58342314e-01 6.22430265e-01
7.06975579e-01 -6.09426260e-01 5.84932864e-02 9.54872787e-01
1.96999282e-01 -1.79460555e-01 -1.31864583e+00 1.32021797e+00
7.35260069e-01 -1.27847910e+00 7.02395022e-01 -3.13632160e-01
5.29880643e-01 -1.79360703e-01 1.16635427e-01 3.94428849e-01
-4.00274605e-01 -7.58159280e-01 9.95240510e-01 6.31582677e-01
8.83937001e-01 -2.18436673e-01 8.40861917e-01 6.22442782e-01
-1.07283568e+00 -1.57491013e-01 -4.50166613e-01 1.82642490e-02
5.31079352e-01 2.35441998e-01 -1.13664985e+00 1.88895077e-01
4.78619486e-01 6.13662660e-01 -5.75801551e-01 1.07151854e+00
-3.09526771e-01 1.07867531e-01 6.17396683e-02 -1.89575076e-01
3.47536832e-01 2.53545433e-01 5.83965778e-01 9.17894602e-01
3.92209768e-01 2.16879666e-01 1.14870422e-01 1.12595236e+00
-2.34571174e-01 1.76057979e-01 -6.17435932e-01 -3.46758932e-01
2.65517503e-01 1.13443804e+00 -6.12986207e-01 -6.24142587e-01
-4.18894470e-01 1.41862607e+00 2.22388655e-01 4.86275107e-01
-8.64967108e-01 -3.89472574e-01 2.74944723e-01 2.29495257e-01
7.03299046e-01 5.78634217e-02 1.77651659e-01 -9.59741950e-01
4.92073894e-01 -4.61759657e-01 1.81009382e-01 -1.60327387e+00
-1.45057487e+00 7.96600223e-01 3.40365283e-02 -1.33925736e+00
-2.37610638e-01 -5.83366394e-01 -4.08137798e-01 8.63346279e-01
-1.50079858e+00 -1.31551945e+00 -5.75609505e-01 2.71367133e-01
8.98960233e-01 1.60361335e-01 7.83477426e-01 2.81923175e-01
-3.07117850e-01 2.26429597e-01 -1.04891673e-01 1.49719253e-01
6.28213704e-01 -6.86145604e-01 3.86830121e-01 6.57858491e-01
4.58348952e-02 3.92355204e-01 8.56581867e-01 -8.62040699e-01
-9.82169926e-01 -1.45074594e+00 1.27583468e+00 -4.49193478e-01
4.02915746e-01 -8.53986442e-01 -1.04045343e+00 6.48406327e-01
6.39512017e-02 2.60931849e-01 4.19090807e-01 -8.81846130e-01
-1.98398322e-01 -1.14170015e-01 -9.80714142e-01 5.92739940e-01
8.06877673e-01 -5.56296289e-01 -9.07624543e-01 6.41883075e-01
1.39823020e+00 -1.95770532e-01 -1.57121390e-01 3.03284883e-01
2.66642243e-01 -2.49450177e-01 8.65324557e-01 -5.76095998e-01
7.58732021e-01 -7.37757802e-01 -6.38576820e-02 -7.89320469e-01
-3.09995189e-02 -1.22200273e-01 -1.54128999e-01 1.26216590e+00
4.13529217e-01 -2.22154006e-01 4.38704044e-01 5.73385000e-01
-1.22014090e-01 -6.20317698e-01 -9.52880502e-01 -9.53581214e-01
-4.47335303e-01 -1.94222018e-01 5.80347776e-01 5.09433687e-01
-3.00136477e-01 6.47832632e-01 -3.88865054e-01 9.98868346e-02
3.31462473e-01 2.60088265e-01 6.79638326e-01 -7.86885977e-01
6.06036745e-02 1.87595338e-01 -5.88013172e-01 -1.31236458e+00
2.31528923e-01 -1.00239766e+00 6.78008020e-01 -1.78278065e+00
8.15078259e-01 -1.69008255e-01 -8.92620161e-02 5.85615158e-01
-3.56503516e-01 4.19361979e-01 3.60399872e-01 3.85452777e-01
-9.42750275e-01 8.63295436e-01 1.37405121e+00 -5.95358074e-01
-5.68134934e-02 -4.15104926e-01 -6.87683940e-01 4.24946338e-01
5.00779867e-01 -7.56403804e-01 -3.99498880e-01 -5.99471807e-01
3.78486365e-02 5.11552058e-02 7.43191183e-01 -8.75245214e-01
3.53052109e-01 -2.35544413e-01 2.90255219e-01 -7.31816471e-01
4.87563580e-01 -9.61862028e-01 1.13695174e-01 5.42226017e-01
-3.60913664e-01 -5.95975295e-02 2.68369526e-01 6.37274861e-01
-1.90936983e-01 -6.05138123e-01 6.93001628e-01 -2.93400288e-01
-1.05584621e+00 4.75517213e-01 -4.44582961e-02 -1.15814880e-01
1.56404793e+00 -2.70133585e-01 -3.41562718e-01 -1.95347682e-01
-7.91221678e-01 4.50447381e-01 4.15362775e-01 9.47639644e-01
1.01977956e+00 -1.47178614e+00 -8.10005844e-01 2.06001207e-01
9.00232494e-01 2.59273827e-01 4.37966317e-01 2.91930348e-01
-3.71729702e-01 4.21235323e-01 -1.34112164e-01 -7.13948011e-01
-1.14408541e+00 1.30518448e+00 9.84451845e-02 2.02789485e-01
-5.74133933e-01 7.13042498e-01 7.90015399e-01 2.54371874e-02
2.57398002e-02 2.32934188e-02 -1.91974998e-01 -6.75713818e-04
7.81405985e-01 -4.91793513e-01 -4.74499047e-01 -9.66339469e-01
-3.04882050e-01 3.94373745e-01 -3.72425497e-01 -1.22944370e-01
1.13341188e+00 -3.78201693e-01 -2.80833989e-01 7.73106992e-01
1.46245623e+00 -7.92569220e-01 -1.19741666e+00 -3.71303141e-01
-1.20062709e-01 -3.95698607e-01 -3.75178218e-01 -7.63908088e-01
-5.52715421e-01 8.55505109e-01 7.46877730e-01 -2.71331847e-01
1.04314876e+00 5.32851040e-01 1.21758485e+00 4.45311099e-01
3.08874875e-01 -8.71066034e-01 4.67292905e-01 3.46873492e-01
1.24335778e+00 -1.39750099e+00 -4.46340173e-01 -4.75790530e-01
-7.72941589e-01 8.68985951e-01 9.45939600e-01 1.85614690e-01
1.83711216e-01 -2.64490873e-01 4.48367707e-02 -1.42637044e-01
-7.17079937e-01 -3.09035987e-01 1.36630177e-01 5.89865267e-01
-2.12961361e-01 -2.05553517e-01 -3.18527162e-01 8.94222975e-01
7.96393231e-02 7.03796744e-02 5.20819068e-01 9.56491053e-01
-5.72493732e-01 -7.23601043e-01 -1.84679464e-01 1.70687869e-01
2.19844710e-02 -3.89331639e-01 -1.70337453e-01 3.79720300e-01
4.21100676e-01 7.62006164e-01 4.54248875e-01 -1.79162577e-01
3.36435646e-01 1.15925960e-01 -1.41765596e-02 -1.11608875e+00
-2.64949739e-01 -4.58481878e-01 -4.88714159e-01 -1.89607039e-01
-2.61084765e-01 -2.53298640e-01 -1.47119725e+00 3.78714681e-01
-3.19809139e-01 3.89621466e-01 7.45955944e-01 9.85652447e-01
4.81484532e-01 3.08552682e-01 5.65770090e-01 -4.92743731e-01
-2.87112623e-01 -8.80175233e-01 -2.75277704e-01 6.91511929e-01
5.34993291e-01 -5.19420087e-01 -3.38614494e-01 6.41597271e-01] | [10.84164047241211, 1.0852768421173096] |
58e796c6-3205-4fff-a32c-5a0dbe16a0cc | improving-noise-robustness-of-contrastive | 2110.15430 | null | https://arxiv.org/abs/2110.15430v1 | https://arxiv.org/pdf/2110.15430v1.pdf | Improving Noise Robustness of Contrastive Speech Representation Learning with Speech Reconstruction | Noise robustness is essential for deploying automatic speech recognition (ASR) systems in real-world environments. One way to reduce the effect of noise interference is to employ a preprocessing module that conducts speech enhancement, and then feed the enhanced speech to an ASR backend. In this work, instead of suppressing background noise with a conventional cascaded pipeline, we employ a noise-robust representation learned by a refined self-supervised framework for noisy speech recognition. We propose to combine a reconstruction module with contrastive learning and perform multi-task continual pre-training on noisy data. The reconstruction module is used for auxiliary learning to improve the noise robustness of the learned representation and thus is not required during inference. Experiments demonstrate the effectiveness of our proposed method. Our model substantially reduces the word error rate (WER) for the synthesized noisy LibriSpeech test sets, and yields around 4.1/7.5% WER reduction on noisy clean/other test sets compared to data augmentation. For the real-world noisy speech from the CHiME-4 challenge (1-channel track), we have obtained the state of the art ASR performance without any denoising front-end. Moreover, we achieve comparable performance to the best supervised approach reported with only 16% of labeled data. | ['DeLiang Wang', 'Jinyu Li', 'Takuya Yoshioka', 'Shujie Liu', 'Chengyi Wang', 'Yiming Wang', 'Xiaofei Wang', 'Yao Qian', 'Heming Wang'] | 2021-10-28 | null | null | null | null | ['auxiliary-learning', 'noisy-speech-recognition'] | ['methodology', 'speech'] | [ 5.20390332e-01 6.73222542e-02 5.17829895e-01 -4.77180839e-01
-1.52534580e+00 -2.87521601e-01 4.68131900e-01 -9.43317339e-02
-8.63836348e-01 4.16062683e-01 4.63896215e-01 -5.44481456e-01
2.06737071e-01 -3.94689828e-01 -6.56049371e-01 -9.01866019e-01
4.56438392e-01 -5.89208007e-02 -2.20081490e-02 -5.09984314e-01
-3.19130719e-01 3.15227568e-01 -1.52063763e+00 4.59733993e-01
7.93064296e-01 9.86913562e-01 6.09034479e-01 9.18749571e-01
2.40165591e-02 7.24747598e-01 -1.08488357e+00 -3.20158333e-01
1.93560243e-01 -2.89565921e-01 -4.77472067e-01 3.37896705e-01
2.43601128e-01 -1.60218537e-01 -8.22132230e-01 1.09975815e+00
1.00836921e+00 5.91147482e-01 6.81007802e-02 -4.17931736e-01
-3.91115129e-01 9.46933806e-01 -2.09153026e-01 3.71532440e-01
7.55793676e-02 2.76544958e-01 8.10693443e-01 -1.15320694e+00
4.32665460e-02 1.30733049e+00 4.12045985e-01 8.17824423e-01
-1.23123455e+00 -6.56845689e-01 2.07120076e-01 1.80651665e-01
-1.34620488e+00 -1.43248570e+00 6.47068918e-01 3.29371780e-01
1.10653949e+00 4.33510333e-01 1.17035560e-01 1.58418131e+00
-5.73078156e-01 8.76623571e-01 1.13080001e+00 -5.30689240e-01
3.26870292e-01 -1.52477950e-01 1.85742781e-01 2.03922272e-01
-1.35742351e-01 1.44746840e-01 -7.37412632e-01 1.19872957e-01
1.99155524e-01 -4.82891411e-01 -4.63016272e-01 6.57768190e-01
-9.04394686e-01 3.40757459e-01 2.55335063e-01 3.27664256e-01
-4.41411525e-01 1.96487337e-01 5.51112890e-01 5.33784151e-01
9.11735713e-01 7.89576694e-02 -6.66817248e-01 -3.09908420e-01
-1.18426180e+00 -1.85865924e-01 5.28068364e-01 7.31402397e-01
3.52081358e-01 6.75134242e-01 -3.70950699e-01 1.57227540e+00
5.28947234e-01 7.60904312e-01 5.18713534e-01 -5.57596207e-01
7.88438439e-01 -1.05169371e-01 -1.56368509e-01 -2.63062775e-01
-1.14100941e-01 -9.31709468e-01 -1.05402827e+00 1.57402437e-02
1.16436616e-01 -2.25278094e-01 -1.56620431e+00 1.59108770e+00
2.86154859e-02 3.38155180e-01 5.15232861e-01 7.48433888e-01
9.05053616e-01 8.44192326e-01 5.73580824e-02 -2.80364305e-01
1.10660124e+00 -1.04964519e+00 -1.18323815e+00 -4.92001325e-01
4.55351233e-01 -1.00710511e+00 1.17605340e+00 6.41335726e-01
-1.06640053e+00 -6.35093868e-01 -1.08092201e+00 -6.49475530e-02
-1.54302374e-01 4.75138217e-01 -8.86047408e-02 1.06096137e+00
-1.09490526e+00 4.65292990e-01 -9.50595379e-01 6.35560900e-02
2.93164283e-01 2.46873692e-01 -3.96034688e-01 -3.66688222e-01
-1.19074667e+00 7.85867870e-01 2.42150769e-01 6.25788271e-01
-1.09801233e+00 -5.66738844e-01 -1.04872942e+00 1.02449380e-01
5.48698664e-01 -2.04133078e-01 1.38379180e+00 -6.29304826e-01
-2.04283142e+00 5.04826903e-01 -4.77646351e-01 -6.46897256e-01
3.61947596e-01 -5.56131721e-01 -8.40929806e-01 -9.67506096e-02
-2.83638835e-01 1.10737763e-01 1.07209718e+00 -1.24303794e+00
-3.57503712e-01 -1.69094488e-01 -4.14482713e-01 1.96186438e-01
-3.00174862e-01 2.52376318e-01 -7.41360903e-01 -1.03229427e+00
4.02661175e-01 -6.51187599e-01 -4.51454639e-01 -6.67323828e-01
-3.10007721e-01 -3.93662602e-02 1.01109135e+00 -1.21863568e+00
1.21130168e+00 -2.47125673e+00 -1.25094831e-01 3.53466064e-01
-1.69786736e-01 8.66491377e-01 -5.25207758e-01 9.86318067e-02
-2.76737660e-01 2.52970327e-02 -4.26400095e-01 -9.69284892e-01
-1.35128200e-01 2.41805628e-01 -3.46585959e-01 4.50865895e-01
4.82713670e-01 4.89048213e-01 -7.74363935e-01 1.65952787e-01
4.83728707e-01 8.81787658e-01 -3.36051941e-01 4.76997405e-01
-1.37465820e-02 4.71971780e-01 1.18581191e-01 3.82701188e-01
7.85194218e-01 4.43957895e-01 8.25134963e-02 -1.71096712e-01
2.40084808e-02 1.12420774e+00 -1.47197807e+00 1.72949624e+00
-8.99424374e-01 5.38856685e-01 7.27827787e-01 -9.08999801e-01
1.17105079e+00 6.34719491e-01 -1.89252868e-01 -8.11321735e-01
-1.84883438e-02 4.30038810e-01 8.68009403e-02 -2.84168452e-01
2.47873560e-01 -4.20226380e-02 2.17414171e-01 -9.29274559e-02
3.08922023e-01 -2.68738627e-01 -1.99033059e-02 3.33108455e-02
1.52230012e+00 -1.54040694e-01 -2.23826859e-02 1.35396523e-02
7.83774853e-01 -6.63372397e-01 5.52211344e-01 8.15075338e-01
-5.94186001e-02 9.51348782e-01 2.57251821e-02 2.56693989e-01
-9.16766405e-01 -1.02406132e+00 -1.03658929e-01 1.10132337e+00
-4.85234529e-01 -4.99917924e-01 -8.65424871e-01 -4.78344530e-01
-4.17642295e-01 8.77456665e-01 -7.11406320e-02 -1.22195862e-01
-7.81773150e-01 -8.62979114e-01 9.62993860e-01 4.94975239e-01
5.60494602e-01 -9.08025444e-01 4.76073354e-01 3.12689990e-01
-2.97258675e-01 -1.40695381e+00 -4.88748938e-01 7.29037046e-01
-4.51651454e-01 -3.72384608e-01 -5.90816975e-01 -6.56044185e-01
4.35036272e-01 4.56191599e-01 7.20137298e-01 5.21524698e-02
1.75223500e-01 2.45693088e-01 -5.86843908e-01 -1.03362933e-01
-9.37027633e-01 4.41778824e-03 3.06792140e-01 2.94612348e-01
-5.50078042e-03 -4.88399714e-01 -3.85100901e-01 3.29573631e-01
-9.11210120e-01 -4.22688931e-01 6.47354364e-01 1.17867541e+00
6.69686735e-01 1.52056724e-01 9.89130199e-01 -6.15924895e-01
5.78890860e-01 -2.47861415e-01 -6.35493159e-01 8.35425556e-02
-2.75587797e-01 7.36671835e-02 7.21874475e-01 -3.52615535e-01
-1.50519693e+00 1.73148721e-01 -1.06512547e+00 -3.57091099e-01
-2.86970824e-01 4.56717938e-01 -7.36380994e-01 2.56859243e-01
6.98313892e-01 2.55142570e-01 -2.48996913e-01 -8.83018553e-01
4.80938405e-01 1.43086803e+00 6.79918408e-01 -3.22810590e-01
9.77944672e-01 3.68530899e-02 -5.22795439e-01 -1.37419784e+00
-8.10048580e-01 -7.15040922e-01 -2.33713239e-01 1.46096915e-01
4.11742479e-01 -1.41381562e+00 -4.31349695e-01 5.67323804e-01
-1.11975932e+00 -4.23602968e-01 -2.90881604e-01 6.00871027e-01
-2.33130097e-01 5.19031584e-01 -8.30838859e-01 -1.21399665e+00
-6.87197208e-01 -1.31674206e+00 1.06343734e+00 -1.88115701e-01
1.07417285e-01 -4.47150439e-01 -2.57066965e-01 7.42232859e-01
7.55218089e-01 -7.09050119e-01 3.11677068e-01 -8.20373297e-01
-2.71508753e-01 -7.80472234e-02 -1.61908865e-02 1.16649830e+00
2.83455282e-01 -3.73400599e-01 -1.76209509e+00 -4.40855682e-01
3.08903962e-01 -2.76615351e-01 1.30940604e+00 2.28702456e-01
1.32589698e+00 -1.31041244e-01 2.47042835e-01 5.65267265e-01
9.02961314e-01 1.29488617e-01 9.43351805e-01 1.78252012e-02
5.73231280e-01 3.31548661e-01 5.10829210e-01 1.77006990e-01
-2.20012367e-01 5.39220333e-01 1.22087345e-01 -1.97768927e-01
-8.02499950e-01 -1.62816986e-01 7.46066511e-01 1.35042036e+00
2.05495775e-01 -4.73057747e-01 -7.79778123e-01 5.16433120e-01
-1.46737146e+00 -7.90213823e-01 -6.17656950e-03 2.13430071e+00
1.04634762e+00 2.17034817e-01 -2.25584298e-01 6.70965374e-01
7.76588321e-01 2.43664563e-01 -2.21311286e-01 -2.28319958e-01
-4.85689610e-01 7.16051936e-01 4.79092181e-01 7.50389695e-01
-1.15742254e+00 1.14429820e+00 5.68838978e+00 1.32907212e+00
-9.62905049e-01 3.16862971e-01 7.36138940e-01 -1.78463429e-01
-1.32999286e-01 -3.93186063e-01 -7.33628273e-01 2.35952839e-01
1.52274764e+00 2.97500134e-01 7.36431599e-01 7.00578570e-01
7.96618700e-01 9.57893133e-02 -7.64390230e-01 1.02406538e+00
2.42088307e-02 -9.91301894e-01 -2.99894631e-01 -3.20077658e-01
5.16260326e-01 3.80981565e-01 9.12896320e-02 5.17805457e-01
2.13049680e-01 -1.03766775e+00 6.98563814e-01 1.49682894e-01
8.12059045e-01 -8.08417499e-01 8.84935617e-01 2.57383585e-01
-1.06553650e+00 -3.04508675e-02 -2.60508090e-01 2.30300874e-01
1.20981060e-01 1.02402163e+00 -9.89356101e-01 4.58894789e-01
6.90204501e-01 2.27679312e-01 -2.97987401e-01 9.96213555e-01
-6.36687338e-01 1.32568979e+00 -4.47906196e-01 3.30450177e-01
-3.68795805e-02 -1.36455968e-02 6.77331150e-01 1.61455202e+00
1.62933528e-01 2.63100415e-02 -1.33486822e-01 3.87107134e-01
-5.74931860e-01 5.21163642e-03 -2.73520380e-01 1.39236078e-03
4.97261882e-01 1.13673270e+00 -2.96588212e-01 -2.50473887e-01
-3.71429026e-01 1.16614294e+00 1.76023930e-01 6.22036099e-01
-6.34773433e-01 -4.90621269e-01 8.16239953e-01 -1.68969378e-01
4.36966240e-01 -3.31739873e-01 -3.32907140e-01 -1.06865919e+00
2.61918783e-01 -1.33792591e+00 -6.98179379e-02 -5.34851074e-01
-1.27233624e+00 1.01468289e+00 -6.68541610e-01 -7.82913983e-01
-6.12832233e-03 -5.18762231e-01 -4.69171345e-01 1.19548631e+00
-1.76336861e+00 -9.60100770e-01 2.65040211e-02 4.50072408e-01
1.01665390e+00 -1.98453665e-01 9.12097991e-01 7.37464249e-01
-8.22101414e-01 8.29632878e-01 1.65745988e-01 2.11480454e-01
7.07427859e-01 -1.22333241e+00 9.58287179e-01 1.43666315e+00
2.33506784e-01 5.40514350e-01 6.33443594e-01 -6.12176716e-01
-1.42602265e+00 -1.37118411e+00 6.05986595e-01 -4.07137312e-02
6.72051489e-01 -7.46030927e-01 -1.16643989e+00 4.01448399e-01
1.94830850e-01 3.29132080e-01 4.14514691e-01 1.10550284e-01
-4.93005216e-01 -3.84779602e-01 -9.57974434e-01 5.28212726e-01
1.10729158e+00 -9.00308192e-01 -5.72224975e-01 9.26166773e-02
1.16252780e+00 -4.70153898e-01 -5.74869931e-01 4.38147366e-01
-8.06188360e-02 -3.91520292e-01 8.46910834e-01 -4.12598312e-01
-2.11062849e-01 -5.43127894e-01 -5.99925578e-01 -1.74582648e+00
2.51067523e-03 -1.01543450e+00 -6.61742091e-02 1.63613284e+00
8.37348163e-01 -3.95094693e-01 5.62890232e-01 2.79935271e-01
-7.62382925e-01 -2.70790696e-01 -1.18698192e+00 -7.63198733e-01
-3.96237463e-01 -1.07216227e+00 1.85119063e-01 4.84326273e-01
-3.92162919e-01 4.18304771e-01 -4.73009765e-01 6.34641051e-01
5.32145679e-01 -8.56697381e-01 5.72261631e-01 -5.45757890e-01
-4.84995663e-01 3.00781671e-02 -1.14144310e-01 -1.17924130e+00
2.57951349e-01 -8.21130097e-01 6.15146399e-01 -1.37954378e+00
-4.19958204e-01 -2.15164453e-01 -4.95678216e-01 5.05916715e-01
-4.26828861e-01 1.35004982e-01 5.58936484e-02 -4.05555129e-01
-4.64946985e-01 9.38795447e-01 7.94788659e-01 -2.38684863e-01
-3.05134714e-01 2.99878448e-01 -6.11223876e-01 5.62088966e-01
8.13962936e-01 -4.58754271e-01 -1.92621276e-01 -6.19000316e-01
-3.26761693e-01 -2.07383335e-01 9.00532901e-02 -1.02748764e+00
1.53318420e-01 4.26438510e-01 2.25530013e-01 -4.98221278e-01
6.37739420e-01 -6.88553631e-01 -3.01304549e-01 1.95160806e-01
-4.00905669e-01 -4.61877704e-01 3.86019915e-01 5.43758631e-01
-3.77646714e-01 -2.18202606e-01 1.03717446e+00 2.19900548e-01
-4.33118433e-01 -1.33015305e-01 -6.45124018e-01 -1.07427701e-01
1.75590932e-01 3.61757368e-01 -3.45313042e-01 -3.88131469e-01
-9.56112027e-01 -7.79256299e-02 -2.58515388e-01 3.67916465e-01
8.23485374e-01 -9.24733996e-01 -1.02836072e+00 3.80442500e-01
-4.05803835e-03 -1.21486532e-02 3.47568959e-01 5.55455923e-01
-7.73517862e-02 -9.33962595e-03 6.64927781e-01 -3.86596799e-01
-1.31881928e+00 1.17915310e-01 4.94934887e-01 -7.86354616e-02
-6.37934923e-01 9.71909881e-01 -3.08898419e-01 -5.09833634e-01
7.95319200e-01 -3.46801639e-01 9.25124288e-02 -1.83398545e-01
8.09167564e-01 5.35652995e-01 1.00758815e+00 -5.88197708e-01
-2.90721476e-01 -7.68748373e-02 -1.58779964e-01 -4.86064702e-01
1.58911788e+00 -2.24507049e-01 2.06483737e-01 2.26189196e-01
1.08706522e+00 4.10001993e-01 -1.02925646e+00 -5.14659882e-01
9.63287614e-03 -2.33597770e-01 6.62491083e-01 -1.02990770e+00
-1.04992318e+00 7.89857566e-01 7.47915208e-01 4.24983092e-02
1.34419072e+00 -1.31429419e-01 8.33481312e-01 7.21772194e-01
-1.15372315e-01 -1.23272908e+00 2.66652629e-02 7.71689832e-01
1.11660528e+00 -1.28680301e+00 -4.70952272e-01 -5.61204851e-01
-4.69348818e-01 8.12199116e-01 3.58246595e-01 1.75123155e-01
5.21734536e-01 7.59578943e-01 4.64242339e-01 3.04695904e-01
-7.61586428e-01 -7.03511059e-01 2.50913233e-01 5.90516925e-01
4.69109535e-01 -7.91949481e-02 -2.09775735e-02 8.40211987e-01
-1.99915484e-01 -5.07123530e-01 2.84071654e-01 6.13260090e-01
-6.15620792e-01 -1.24413276e+00 -6.62895739e-01 1.67298049e-01
-6.33648455e-01 -4.88628477e-01 -1.10428900e-01 8.49330053e-02
-1.90415949e-01 1.78957629e+00 -2.09783688e-01 -4.04950887e-01
8.56345177e-01 2.67560691e-01 9.37072858e-02 -7.80681908e-01
-8.56022894e-01 6.59053624e-01 4.83930230e-01 -4.69196469e-01
-9.49330702e-02 -3.57589841e-01 -1.10143614e+00 1.11757042e-02
-6.29519701e-01 2.90004890e-02 1.11865520e+00 9.74510729e-01
2.78783172e-01 1.08009481e+00 8.71642232e-01 -6.34386778e-01
-7.82092214e-01 -1.38933289e+00 -2.98960894e-01 3.01635325e-01
5.66108942e-01 -1.43080652e-01 -6.39310658e-01 -8.88413563e-02] | [14.765353202819824, 6.235950469970703] |
52d5e71c-488d-4067-b189-de1c448a282c | non-backtracking-walks-reveal-compartments-in | 2003.09949 | null | https://arxiv.org/abs/2003.09949v2 | https://arxiv.org/pdf/2003.09949v2.pdf | Non-backtracking walks reveal compartments in sparse chromatin interaction networks | Chromatin communities stabilized by protein machinery play essential role in gene regulation and refine global polymeric folding of the chromatin fiber. However, treatment of these communities in the framework of the classical network theory (stochastic block model, SBM) does not take into account intrinsic linear connectivity of the chromatin loci. Here we propose the "polymer" block model, paving the way for community detection in polymer networks. On the basis of this new model we modify the non-backtracking flow operator and suggest the first protocol for annotation of compartmental domains in sparse single cell Hi-C matrices. In particular, we prove that our approach corresponds to the maximum entropy principle. The benchmark analyses demonstrates that the spectrum of the polymer non-backtracking operator resolves the true compartmental structure up to the theoretical detectability threshold, while all commonly used operators fail above it. We test various operators on real data and conclude that the sizes of the non-backtracking single cell domains are most close to the sizes of compartments from the population data. Moreover, the found domains clearly segregate in the gene density and correlate with the population compartmental mask, corroborating biological significance of our annotation of the chromatin compartmental domains in single cells Hi-C matrices. | ['S. Ulianov', 'S. V. Razin', 'S. Nechaev', 'A. Gorsky', 'K. Polovnikov'] | 2020-03-22 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 3.61142844e-01 2.88867921e-01 -6.98923916e-02 3.59029591e-01
1.51523873e-01 -9.97845054e-01 7.05696046e-01 3.73707592e-01
-3.50440413e-01 1.02047002e+00 4.76327762e-02 -3.38903219e-01
-8.03244948e-01 -9.39111352e-01 -7.27506042e-01 -1.32033777e+00
-1.07179224e-01 1.00187874e+00 6.99973226e-01 -1.57648534e-01
1.32904395e-01 5.18439412e-01 -1.41996872e+00 2.55142182e-01
7.72853971e-01 2.20248163e-01 2.51796413e-02 8.26643646e-01
-2.44995818e-01 2.50383615e-01 -4.26618546e-01 -1.15357138e-01
8.66465047e-02 -6.82128668e-01 -5.22941411e-01 7.24307150e-02
1.86077785e-02 4.65699345e-01 -1.15000084e-01 1.27317595e+00
2.39435807e-01 -3.51363033e-01 1.14816141e+00 -1.06060064e+00
-2.39461407e-01 7.47906446e-01 -4.78499055e-01 2.41806343e-01
1.95841193e-01 -4.42376845e-02 1.11123633e+00 -6.79894030e-01
1.34653068e+00 9.51640546e-01 7.19121277e-01 1.42108053e-01
-1.69751847e+00 -5.64159527e-02 -2.42968544e-01 -7.00518340e-02
-1.53861856e+00 -1.10845994e-02 6.49230063e-01 -8.56054306e-01
4.45657820e-01 4.26201016e-01 1.01235807e+00 8.14253151e-01
3.75009626e-01 9.07776505e-02 1.32885695e+00 -3.48826081e-01
4.83078361e-01 -1.26935542e-01 3.07632685e-01 6.76070988e-01
9.87199128e-01 -2.93165118e-01 -3.50889444e-01 -3.59265774e-01
6.91394925e-01 -1.21222429e-01 -2.93899775e-01 -6.45934045e-01
-1.29453349e+00 5.73905051e-01 -1.53817400e-01 8.25732768e-01
-4.02542688e-02 -3.59376594e-02 4.34782147e-01 4.92326133e-02
2.74629682e-01 1.73696339e-01 -1.48137629e-01 6.43695071e-02
-1.01991534e+00 1.05758622e-01 9.13340092e-01 7.44099677e-01
7.18145251e-01 -5.12350261e-01 1.16954058e-01 2.14504182e-01
3.04629028e-01 4.01098311e-01 2.38415539e-01 -7.57830143e-01
8.64493623e-02 8.88698161e-01 -1.88172311e-02 -1.08655775e+00
-5.69128275e-01 -8.06314468e-01 -1.30411434e+00 -2.15954706e-02
1.06353343e+00 4.62532043e-03 -3.44255656e-01 1.49743319e+00
3.87474865e-01 6.11827821e-02 -8.59106109e-02 5.67192256e-01
2.33623788e-01 3.38997364e-01 -2.57588714e-01 -6.01129413e-01
1.51293278e+00 -2.44572937e-01 -7.17819810e-01 5.02914667e-01
4.42875981e-01 -3.25307727e-01 5.51008880e-01 4.29574579e-01
-7.95138955e-01 -1.66431829e-01 -9.30157006e-01 5.90854108e-01
-3.40974271e-01 -1.16704062e-01 3.10428292e-01 7.26512253e-01
-1.01604450e+00 8.13980579e-01 -6.82307839e-01 -6.71773195e-01
1.59437656e-01 2.37060934e-01 -5.03613949e-01 2.11028636e-01
-7.79864550e-01 4.85651940e-01 4.81724948e-01 4.70045567e-01
-5.94817758e-01 -2.13226229e-01 -2.17953876e-01 2.23716363e-01
2.78946966e-01 -4.89999712e-01 1.27224982e-01 -6.37202919e-01
-9.88143921e-01 9.32407439e-01 5.88687100e-02 -4.27192807e-01
6.65204823e-01 5.56543589e-01 -1.05679281e-01 4.28386241e-01
-1.85285524e-01 1.30914986e-01 4.12539423e-01 -1.31753147e+00
-3.12313717e-02 -6.87709153e-02 -3.84739816e-01 -3.06143969e-01
-8.44894275e-02 -3.17842066e-01 1.66246384e-01 -1.84651256e-01
4.28204060e-01 -1.02759600e+00 -5.75498529e-02 -1.77114472e-01
-6.15972877e-01 6.21811002e-02 4.67132628e-01 -2.31489107e-01
1.13474631e+00 -2.04191256e+00 4.72604603e-01 8.35876822e-01
5.96286297e-01 -1.57194585e-01 2.35235170e-01 8.36483717e-01
-3.40684354e-01 3.04059058e-01 -3.86027396e-01 1.52364686e-01
1.87474698e-01 2.93365717e-01 -3.26590575e-02 1.08639753e+00
-3.57017219e-02 4.25415277e-01 -9.06064510e-01 -6.84312284e-01
-2.00675264e-01 2.59117842e-01 -7.35370815e-01 -3.50992411e-01
-3.09941351e-01 5.02325952e-01 -2.95804814e-02 2.38939524e-01
9.29566801e-01 -4.81968731e-01 9.71515596e-01 6.14295602e-02
-2.72733659e-01 -2.06230059e-01 -1.43697846e+00 1.05817354e+00
5.92036545e-01 5.96934736e-01 2.71838874e-01 -9.21985090e-01
7.74043977e-01 2.45771989e-01 5.79408348e-01 1.89065501e-01
-5.33013837e-04 3.24084371e-01 4.33824629e-01 -1.84705064e-01
2.02647313e-01 -7.66723454e-02 2.65479982e-01 3.09365630e-01
1.06604643e-01 4.04265016e-01 5.67833006e-01 2.51938492e-01
1.46795416e+00 -8.74607265e-02 3.95830959e-01 -1.00392640e+00
7.49415219e-01 -1.31043449e-01 4.14491981e-01 5.59594810e-01
-2.58370936e-01 5.88098288e-01 1.59376347e+00 -1.33918803e-02
-1.32310545e+00 -9.81590688e-01 -2.12250218e-01 5.41558385e-01
1.09759830e-01 -4.66774762e-01 -9.95373249e-01 -3.20165932e-01
8.02334920e-02 -1.40995815e-01 -8.06664705e-01 1.63436949e-01
-4.18570668e-01 -1.22906637e+00 6.02009475e-01 -1.80069298e-01
-3.65210744e-03 -7.10751772e-01 -3.56697768e-01 6.99857366e-04
-2.23850742e-01 -8.61050844e-01 -2.35336676e-01 4.35415685e-01
-8.44120085e-01 -1.36400723e+00 -6.09205782e-01 -5.39364696e-01
9.61980939e-01 -2.13954955e-01 7.11357772e-01 3.93203229e-01
-2.81205863e-01 5.74155934e-02 -8.31285641e-02 3.51662904e-01
-7.80048788e-01 1.65855438e-01 1.42130414e-02 7.93132558e-02
-2.71779411e-02 -8.09162021e-01 -5.98355174e-01 4.28220361e-01
-8.60195994e-01 -1.40542626e-01 3.36372823e-01 7.55049348e-01
5.76020896e-01 4.30936426e-01 2.18831092e-01 -9.37046468e-01
3.96609664e-01 -7.29603767e-01 -8.06857944e-01 1.93348490e-02
-7.36961544e-01 3.75048250e-01 5.19608557e-01 -2.67949224e-01
-5.89806497e-01 6.41037896e-02 5.32807857e-02 -2.16923933e-02
-7.34826848e-02 4.20747101e-01 -3.48753661e-01 1.18384948e-02
4.18668032e-01 4.59955245e-01 9.27962959e-02 -4.36696917e-01
8.30180123e-02 2.91501880e-01 2.03514874e-01 -6.39317095e-01
7.82437205e-01 8.85415971e-01 8.17241311e-01 -8.96863282e-01
3.28297392e-02 -3.85566413e-01 -8.20243180e-01 -5.30579448e-01
7.85791039e-01 -4.99210715e-01 -1.16880143e+00 2.38255590e-01
-1.06377435e+00 -1.37749270e-01 -3.86645943e-01 2.07841501e-01
-5.88475287e-01 9.40331578e-01 -9.02550042e-01 -1.16759574e+00
1.91810980e-01 -8.69656920e-01 7.78361499e-01 -1.69612080e-01
-9.71221253e-02 -1.06363058e+00 6.44812584e-01 -1.48759037e-01
9.16637182e-02 5.01485348e-01 1.35436285e+00 -6.40549183e-01
-6.63906991e-01 5.64155839e-02 5.51037937e-02 -1.89254895e-01
-4.28019673e-01 5.94503284e-01 -6.96562290e-01 -2.10778922e-01
-3.05367351e-01 7.72536516e-01 8.53890598e-01 3.92235279e-01
4.50929105e-01 -1.29920200e-01 -5.31204581e-01 3.53843182e-01
1.76867282e+00 -2.38760084e-01 7.25485861e-01 1.43561721e-01
2.30096132e-01 8.96193266e-01 5.50029948e-02 3.97169769e-01
-3.08260530e-01 5.27789533e-01 4.13160771e-01 2.58121520e-01
-5.35253026e-02 -2.72957534e-01 4.09621119e-01 9.78583455e-01
-2.81196594e-01 -4.12750602e-01 -1.06729925e+00 5.24596453e-01
-1.93505764e+00 -1.13792419e+00 -9.09911931e-01 2.19299507e+00
7.56409049e-01 5.49420297e-01 4.80283469e-01 2.68381536e-01
1.12071753e+00 -1.35861278e-01 8.12537521e-02 -8.05360153e-02
-6.81606472e-01 -7.46393576e-02 7.49859929e-01 7.56727040e-01
-7.40536213e-01 2.45686755e-01 6.78213263e+00 9.02114868e-01
-5.28036356e-01 2.27878958e-01 1.01793915e-01 2.05745816e-01
-4.40190732e-01 3.85124385e-01 -7.90347397e-01 8.41136754e-01
8.80170882e-01 -4.96800765e-02 2.51588911e-01 4.12505418e-01
2.68639952e-01 -4.03019905e-01 -7.30876267e-01 4.51470137e-01
-3.80680233e-01 -1.37376928e+00 -2.74201602e-01 8.56487870e-01
6.74523056e-01 -2.88844168e-01 -3.56101215e-01 -3.82259101e-01
2.38245167e-02 -6.91306174e-01 5.56082606e-01 7.94065893e-01
6.71671152e-01 -4.52655047e-01 8.48589897e-01 4.56351012e-01
-1.23407257e+00 4.23579849e-02 -3.50171536e-01 -5.44295497e-02
1.90524163e-03 1.03177011e+00 -8.88670325e-01 4.63850230e-01
4.47832756e-02 4.58118170e-01 -5.90368211e-01 1.05639160e+00
2.15178296e-01 7.11526871e-01 -4.95608509e-01 -1.93910390e-01
-2.03930736e-01 -8.82497787e-01 8.88409913e-01 1.30990422e+00
3.19036275e-01 -4.26526129e-01 -4.04880702e-01 9.49906290e-01
1.21179000e-01 2.84443218e-02 -6.46551132e-01 -3.45049471e-01
2.63328195e-01 1.17491663e+00 -1.63711047e+00 -1.79267362e-01
-2.33337469e-02 6.86142027e-01 1.90254435e-01 1.98501885e-01
-5.92193365e-01 -2.09250346e-01 4.65870112e-01 6.47233367e-01
5.35901248e-01 -3.31188112e-01 -2.24133804e-01 -9.57265973e-01
-1.64312258e-01 -5.09763658e-01 2.31247526e-02 -4.22256470e-01
-1.23099160e+00 1.09047405e-01 -9.11685899e-02 -1.11331868e+00
-6.32306607e-03 -6.08788013e-01 -3.82113934e-01 5.24185181e-01
-9.53718126e-01 -7.33016133e-01 -7.99061432e-02 2.36937851e-01
-2.97742814e-01 1.55809745e-01 7.54259408e-01 3.73087406e-01
-5.74244082e-01 -1.28607258e-01 7.55980074e-01 -1.04538910e-02
3.23231936e-01 -1.52037013e+00 -9.79463290e-03 9.03715611e-01
-2.09616005e-01 9.33000267e-01 1.20228076e+00 -1.10861754e+00
-1.15882397e+00 -5.87935507e-01 9.23504889e-01 -4.22372222e-01
1.03630972e+00 -9.65804160e-01 -8.11740756e-01 2.64531195e-01
3.21451575e-01 -1.22228712e-01 5.60420334e-01 -2.67314469e-03
-6.54389858e-02 5.97574972e-02 -9.07430887e-01 3.92928898e-01
1.06853008e+00 -3.68362278e-01 -2.74802595e-01 5.33597767e-01
3.40814769e-01 1.52231589e-01 -1.07526755e+00 2.51427712e-03
5.94524205e-01 -1.20055366e+00 7.36340642e-01 -4.21699256e-01
3.47855836e-01 -6.57960594e-01 1.02072366e-01 -7.04203427e-01
-3.72388899e-01 -7.82820642e-01 -1.03211127e-01 1.27023435e+00
2.52367973e-01 -5.82993448e-01 9.48309779e-01 -2.80434728e-01
1.74150631e-01 -4.31038618e-01 -1.37952459e+00 -7.39605308e-01
8.00539777e-02 2.57206649e-01 2.10294887e-01 9.40970540e-01
2.50734448e-01 2.86793709e-01 -2.14746036e-02 -9.98562127e-02
8.84706676e-01 -1.07481793e-01 4.36659098e-01 -1.56155872e+00
-6.17353320e-01 -6.55166805e-01 -5.51276922e-01 -5.22093236e-01
-4.60798405e-02 -8.72289121e-01 -1.53113790e-02 -1.16868901e+00
5.74607313e-01 -3.18294048e-01 -4.07147296e-02 -1.28268167e-01
3.34363610e-01 1.28048867e-01 -5.34353890e-02 4.70048487e-01
-6.63952053e-01 2.06512609e-03 1.10084724e+00 7.98134953e-02
1.76275253e-01 -2.64076978e-01 -3.55146289e-01 6.53280258e-01
4.26882505e-01 -6.42404556e-01 -7.69410431e-02 5.83939612e-01
1.04533613e+00 -6.87040016e-02 3.16793621e-01 -1.06012058e+00
3.37073088e-01 -4.91757132e-02 1.50612056e-01 -7.10236490e-01
6.67093024e-02 -8.68951082e-01 8.38277340e-01 8.17451894e-01
-2.34408289e-01 -1.02238385e-02 -1.98151693e-01 1.11636448e+00
2.44208395e-01 -6.04340434e-01 6.93073630e-01 -1.58023030e-01
1.62890390e-01 -3.85419041e-01 -1.04883552e+00 -9.95906144e-02
1.02316439e+00 -3.36800128e-01 -6.60812020e-01 2.39124298e-02
-1.00082374e+00 -3.00977200e-01 7.96858072e-01 -5.19337833e-01
1.74730733e-01 -1.05226266e+00 -5.59353769e-01 1.57309279e-01
-8.93648043e-02 -4.86135215e-01 3.21976960e-01 1.38549745e+00
-9.72453535e-01 4.75423455e-01 -1.89191744e-01 -9.31285918e-01
-1.25367737e+00 6.85994506e-01 4.80995268e-01 -4.91837382e-01
-3.57353538e-01 3.17786515e-01 1.94311030e-02 -1.23928703e-01
-1.53476894e-01 -3.93649578e-01 -3.09831709e-01 3.01203519e-01
3.11238319e-02 5.11329174e-01 -1.72833264e-01 -6.05998755e-01
-3.60942245e-01 6.07369006e-01 3.82636130e-01 2.77829356e-02
1.21926749e+00 -2.00317755e-01 -8.93582463e-01 6.25443339e-01
8.16513181e-01 5.66007674e-01 -9.51765597e-01 2.94431806e-01
3.57505500e-01 -1.13267906e-01 -6.32349551e-01 -3.27515662e-01
-5.19414842e-01 6.62441194e-01 4.98032629e-01 8.68829727e-01
6.65526032e-01 8.86997506e-02 1.00740388e-01 1.35256514e-01
2.26327136e-01 -9.22427297e-01 -2.13025540e-01 3.11411113e-01
3.70500892e-01 -3.30489486e-01 8.93287063e-02 -7.50315964e-01
-4.52840421e-03 1.17589819e+00 7.03912899e-02 -4.12848324e-01
6.67726576e-01 4.90084201e-01 -7.60615051e-01 -3.68688554e-01
-8.32452059e-01 -2.47089341e-01 -2.58156210e-01 5.86550713e-01
2.61945784e-01 1.55777752e-01 -8.64830196e-01 3.58802974e-01
1.80537656e-01 -2.31708765e-01 9.37298834e-01 5.71099758e-01
-6.24175549e-01 -1.14733338e+00 -5.80709457e-01 2.27901220e-01
-2.28547066e-01 2.27089785e-02 -5.94060481e-01 8.86742055e-01
4.54850614e-01 5.00818431e-01 1.02007814e-01 -1.19137324e-01
-1.53191306e-03 3.30679595e-01 6.01963818e-01 -3.15707892e-01
-4.68291730e-01 5.48613012e-01 1.53421899e-02 3.64590250e-02
-4.97541100e-01 -8.20945442e-01 -1.10269332e+00 -5.18064439e-01
-4.65030015e-01 5.48922062e-01 3.69709909e-01 7.86329746e-01
3.26134980e-01 4.81175750e-01 1.34766355e-01 -4.06459749e-01
-2.81870842e-01 -7.82183588e-01 -1.21206319e+00 2.53275603e-01
2.45774344e-01 -6.49466991e-01 -7.06669033e-01 1.36399686e-01] | [6.869426250457764, 5.152101516723633] |
9f1ed521-02b0-4cc9-8854-3a9c29de8c17 | efficient-learning-of-pinball-twsvm-using | 2107.06744 | null | https://arxiv.org/abs/2107.06744v1 | https://arxiv.org/pdf/2107.06744v1.pdf | Efficient Learning of Pinball TWSVM using Privileged Information and its applications | In any learning framework, an expert knowledge always plays a crucial role. But, in the field of machine learning, the knowledge offered by an expert is rarely used. Moreover, machine learning algorithms (SVM based) generally use hinge loss function which is sensitive towards the noise. Thus, in order to get the advantage from an expert knowledge and to reduce the sensitivity towards the noise, in this paper, we propose privileged information based Twin Pinball Support Vector Machine classifier (Pin-TWSVMPI) where expert's knowledge is in the form of privileged information. The proposed Pin-TWSVMPI incorporates privileged information by using correcting function so as to obtain two nonparallel decision hyperplanes. Further, in order to make computations more efficient and fast, we use Sequential Minimal Optimization (SMO) technique for obtaining the classifier and have also shown its application for Pedestrian detection and Handwritten digit recognition. Further, for UCI datasets, we first implement a procedure which extracts privileged information from the features of the dataset which are then further utilized by Pin-TWSVMPI that leads to enhancement in classification accuracy with comparatively lesser computational time. | ['Aman Pal', 'Reshma Rastogi'] | 2021-07-14 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 2.20014751e-01 -2.16286570e-01 -3.12928528e-01 -3.23025405e-01
-3.14072967e-02 -3.51829320e-01 1.70869783e-01 2.20331162e-01
-5.53140819e-01 1.12689829e+00 -2.93267936e-01 -5.13087809e-01
-4.69543517e-01 -8.77186954e-01 -4.24430668e-01 -8.69033873e-01
3.75305206e-01 3.03382259e-02 4.38149393e-01 -1.82466224e-01
5.35496771e-01 6.50020540e-01 -1.58720493e+00 6.87732100e-02
1.15917170e+00 1.00160718e+00 1.09070122e-01 4.12406832e-01
-3.37018468e-03 6.87214792e-01 -4.83137161e-01 -3.79839003e-01
4.34452266e-01 -2.85492390e-01 -5.79337299e-01 9.34413597e-02
-1.19975254e-01 -1.67694688e-01 -2.91469097e-01 9.26302850e-01
3.00107986e-01 2.31485531e-01 9.15087223e-01 -1.22653925e+00
-3.43185633e-01 1.01641268e-01 -7.14776933e-01 1.56636819e-01
1.28154650e-01 -3.36929597e-02 5.96291244e-01 -5.24017572e-01
3.99302393e-01 8.66568983e-01 3.65316421e-01 2.02308774e-01
-9.30209219e-01 -4.61866349e-01 -5.99199831e-02 6.81425691e-01
-1.30254889e+00 -5.32642156e-02 9.30580795e-01 -4.42176759e-01
4.45397884e-01 5.19812882e-01 5.41638911e-01 6.98252618e-01
2.66265959e-01 9.29431021e-01 1.24711156e+00 -6.71237826e-01
2.38943964e-01 7.69634485e-01 6.24752641e-01 5.61824203e-01
3.95364583e-01 -2.35712947e-03 -2.74262190e-01 -5.79509437e-02
3.61794174e-01 1.05467312e-01 -3.46490800e-01 -3.93927991e-01
-7.10973263e-01 7.82215595e-01 2.95949548e-01 3.47855985e-01
-8.74902830e-02 -6.89911306e-01 4.59256560e-01 1.62599400e-01
-5.20321280e-02 1.73651949e-01 -3.25106442e-01 2.24574748e-02
-8.13990474e-01 -2.08288617e-02 8.51598501e-01 4.13213223e-01
6.10162973e-01 4.29224037e-02 1.64000630e-01 6.62777662e-01
2.71149039e-01 3.40346038e-01 7.03169465e-01 -2.72503108e-01
4.93622005e-01 9.43799078e-01 4.02367935e-02 -1.27875638e+00
-1.88327834e-01 -3.86664540e-01 -8.10089886e-01 6.64161444e-01
6.32043123e-01 -7.46660531e-02 -7.72617161e-01 1.21878803e+00
3.29275429e-01 -6.75471649e-02 3.39947045e-01 9.42126870e-01
4.28265423e-01 5.49050331e-01 -4.95165102e-02 -1.62088752e-01
1.33093238e+00 -6.35797262e-01 -3.43568534e-01 -6.08074255e-02
4.48527902e-01 -7.43291020e-01 8.71505320e-01 8.33587527e-01
-2.17849046e-01 -5.41650891e-01 -1.38315594e+00 2.19776422e-01
-6.64030135e-01 4.60415930e-01 5.15708208e-01 9.73742425e-01
-3.42090666e-01 5.98046839e-01 -5.98897159e-01 -3.16304594e-01
4.31529462e-01 5.19585550e-01 -6.75056875e-01 -5.59212873e-04
-1.06060326e+00 1.23226345e+00 8.34938586e-01 1.32694751e-01
-1.99188724e-01 -1.19105138e-01 -7.37736881e-01 3.07886279e-03
3.45194310e-01 6.19371980e-02 5.31814396e-01 -1.13552511e+00
-1.54694140e+00 4.97348398e-01 4.69722366e-03 -4.74954844e-01
6.75744712e-01 5.65268919e-02 -3.44193071e-01 1.40706211e-01
-2.45649114e-01 6.59638420e-02 8.34560812e-01 -8.17938209e-01
-6.18723691e-01 -6.70920908e-01 -3.57890390e-02 2.24443853e-01
-4.79182482e-01 -3.82333137e-02 7.49713331e-02 -3.68503839e-01
1.72038138e-01 -6.56243324e-01 5.45126311e-02 -1.88452765e-01
-5.96535683e-01 -1.55841276e-01 1.24637187e+00 -9.19985533e-01
1.16052592e+00 -2.25792384e+00 1.02663105e-02 6.35524392e-01
-1.29920304e-01 6.75388277e-01 5.77100396e-01 9.85594764e-02
-2.06260029e-02 -3.84195119e-01 -3.09736758e-01 3.84406745e-01
-2.87906915e-01 2.67041326e-01 -7.78787062e-02 2.86708176e-01
7.65813366e-02 1.50604427e-01 -3.39867175e-01 -6.87553763e-01
5.24482906e-01 3.53863478e-01 -2.36836791e-01 -5.50989695e-02
2.15352103e-01 2.66721547e-01 -6.62039638e-01 5.97916186e-01
8.46003592e-01 1.60384372e-01 8.85558948e-02 -2.98864335e-01
-2.50364631e-01 -5.24419963e-01 -1.53032887e+00 7.33556390e-01
-1.93686336e-01 5.47025383e-01 -1.65433049e-01 -1.48411524e+00
1.28274643e+00 2.19306111e-01 1.78378627e-01 -2.99948007e-01
4.95093763e-01 3.31930697e-01 2.13286411e-02 -5.37746012e-01
3.17001134e-01 1.70999691e-01 3.75127494e-01 2.18890402e-02
-9.76528898e-02 4.63880897e-01 9.22928452e-02 -2.28258073e-01
4.75802839e-01 1.34698242e-01 7.22248971e-01 -2.07759470e-01
1.10110521e+00 1.38533130e-01 5.11159360e-01 3.42954844e-01
-3.42697680e-01 8.99321213e-02 5.85325539e-01 -4.16695207e-01
-8.55253160e-01 -5.86536527e-01 -4.94422406e-01 5.56841016e-01
1.29041508e-01 2.29143441e-01 -6.34840965e-01 -7.25631177e-01
1.71958387e-01 5.72720885e-01 -3.18033278e-01 -1.60410494e-01
-5.31415820e-01 -9.18092668e-01 4.46549624e-01 5.00522435e-01
9.93975639e-01 -7.16728091e-01 -8.11539948e-01 1.40346035e-01
1.76671267e-01 -6.19518638e-01 -9.95200872e-03 2.91124880e-01
-7.72292912e-01 -1.43996859e+00 -7.82029390e-01 -6.05340123e-01
6.87696874e-01 8.07132274e-02 8.14110637e-02 -1.50843889e-01
-6.01785779e-01 2.99201929e-03 -4.65981662e-01 -4.51964438e-01
-1.08554043e-01 1.07277468e-01 2.59880107e-02 2.55894512e-01
6.29061520e-01 -2.82842964e-01 -3.52944076e-01 3.35977644e-01
-7.53151059e-01 -2.14063510e-01 6.82620645e-01 9.45570409e-01
1.05484009e-01 4.16080594e-01 4.90136653e-01 -6.10642135e-01
3.43625307e-01 -2.80707181e-01 -8.37080717e-01 3.70883763e-01
-5.98400176e-01 2.30312943e-01 8.36652815e-01 -3.86456609e-01
-1.38173258e+00 1.50757536e-01 -4.99978624e-02 1.30168656e-02
-2.14024827e-01 3.47198695e-01 -5.18143415e-01 -4.67065483e-01
8.02148879e-01 5.40961742e-01 1.67590082e-01 -5.39056420e-01
-2.12168336e-01 1.12636554e+00 2.70807356e-01 -4.28694874e-01
4.99273121e-01 2.02168778e-01 4.48416799e-01 -9.94562805e-01
-5.90481043e-01 -2.80622929e-01 -6.91030383e-01 -1.26497597e-01
8.50897253e-01 -3.05200517e-01 -1.01938832e+00 6.83056295e-01
-7.84200311e-01 4.83778924e-01 2.99507260e-01 7.60943592e-01
-1.90096408e-01 7.18647659e-01 -2.52454013e-01 -1.23010290e+00
-2.92619407e-01 -1.18593884e+00 1.44770667e-01 6.81704342e-01
4.52051759e-02 -8.64154398e-01 -4.20032889e-01 4.66760039e-01
2.37892181e-01 4.07452404e-01 9.00554240e-01 -9.93368447e-01
-4.81870323e-01 -5.91271520e-01 -3.26440603e-01 7.83165872e-01
1.80792719e-01 8.37233514e-02 -7.42240906e-01 -1.61109924e-01
-2.43241135e-02 -1.91017762e-02 6.80313349e-01 2.39613712e-01
9.63419259e-01 -3.08886558e-01 -3.28597099e-01 3.91817153e-01
1.54043138e+00 7.00323284e-01 7.09447265e-01 8.27594519e-01
4.79429066e-01 5.24816692e-01 8.21511805e-01 4.30284142e-01
8.22348818e-02 6.19602621e-01 9.87073332e-02 -3.94992940e-02
4.85639542e-01 -4.51317877e-02 1.16228193e-01 2.04387337e-01
-2.44507983e-01 1.36152819e-01 -6.16612732e-01 2.48406723e-01
-1.81454611e+00 -8.35835695e-01 -1.01728454e-01 2.54240108e+00
7.57808328e-01 2.42137447e-01 1.14278369e-01 9.64875817e-01
8.80319297e-01 -1.60559759e-01 -5.14182568e-01 -5.34733534e-01
-1.45752653e-01 2.38630801e-01 6.64475262e-01 4.37952220e-01
-1.25209773e+00 6.02800488e-01 4.75086308e+00 1.03530240e+00
-1.30467486e+00 -3.00542891e-01 5.29726505e-01 3.39037478e-01
4.06640291e-01 1.07380167e-01 -9.10488546e-01 6.40074551e-01
3.14413577e-01 -1.50004417e-01 2.16030687e-01 1.17988563e+00
4.23009433e-02 -6.56072915e-01 -7.39712119e-01 9.68195677e-01
-7.21762031e-02 -5.64723372e-01 -1.09202191e-01 2.73663756e-02
1.66537300e-01 -8.80218744e-01 4.36021648e-02 1.00420825e-01
-1.90163657e-01 -6.85797155e-01 2.81374037e-01 5.56981325e-01
3.02163810e-01 -1.10349250e+00 1.07152772e+00 5.94528139e-01
-7.87855864e-01 -1.93478629e-01 -5.81067026e-01 -9.52252597e-02
-3.03034425e-01 5.94900012e-01 -8.97554636e-01 7.93874383e-01
4.10383016e-01 3.11088830e-01 -4.52781022e-01 1.23401761e+00
-1.35757446e-01 3.64825666e-01 -3.60001743e-01 -3.18821877e-01
1.36413410e-01 -2.73624033e-01 5.66347063e-01 9.32641029e-01
1.43026382e-01 1.37502536e-01 9.34868678e-02 3.97882015e-01
4.78693873e-01 5.57150066e-01 -4.75548118e-01 1.17012627e-01
3.73622030e-01 9.81811285e-01 -7.53828764e-01 -2.59358883e-01
-8.86256620e-02 1.06214881e+00 7.37599656e-02 1.97351068e-01
-5.78104436e-01 -1.01063240e+00 3.81488472e-01 -5.45101182e-04
3.04791749e-01 -5.22655398e-02 -4.06928986e-01 -1.10698080e+00
1.16737731e-01 -7.44771481e-01 4.83458757e-01 -3.87803197e-01
-9.35234845e-01 5.77483058e-01 4.04065251e-02 -1.23850048e+00
-4.39268947e-02 -1.07118452e+00 -4.75114822e-01 1.08343554e+00
-1.18131936e+00 -1.06623602e+00 -2.11657241e-01 7.31847465e-01
1.40837520e-01 -4.76267636e-01 5.58157563e-01 2.50839978e-01
-8.23989689e-01 7.51796901e-01 2.57764488e-01 3.65489155e-01
5.89419603e-01 -8.97524238e-01 -5.01411200e-01 9.47643340e-01
-2.67422885e-01 6.11476183e-01 8.00737441e-01 -6.14946187e-01
-9.89852667e-01 -6.63556695e-01 7.24943399e-01 -4.37635444e-02
2.34919295e-01 1.29612982e-01 -7.91454256e-01 4.19719785e-01
-1.38980031e-01 -3.26158166e-01 7.26209402e-01 -5.82773015e-02
-2.64714241e-01 -4.04822230e-01 -1.67719936e+00 4.38261479e-01
1.74436942e-01 -2.10432425e-01 -7.25549340e-01 2.30625109e-03
-7.33367197e-05 -3.32187116e-01 -6.29148722e-01 3.35750699e-01
7.15632498e-01 -9.63068783e-01 7.75534630e-01 -6.27901435e-01
1.56555146e-01 -5.67469656e-01 -1.59737766e-01 -1.04015064e+00
-9.94271114e-02 7.74023086e-02 1.29878506e-01 1.07723665e+00
2.67002404e-01 -1.03471851e+00 9.66072559e-01 8.03763568e-01
2.95070380e-01 -8.22847962e-01 -9.32685971e-01 -8.38605404e-01
-2.48920411e-01 -9.67437103e-02 4.92296547e-01 8.73753190e-01
3.31240147e-01 -1.52308106e-01 -6.16391182e-01 1.05698861e-01
8.32994401e-01 -8.88998210e-02 5.62126637e-01 -1.30043590e+00
-3.69318902e-01 -1.14277728e-01 -1.02481747e+00 -4.19125617e-01
-2.73387637e-02 -6.45863533e-01 -3.77829760e-01 -1.00013006e+00
9.67993811e-02 -4.41930652e-01 -3.16570252e-01 5.42278886e-01
-2.11754084e-01 1.91368148e-01 1.48811206e-01 1.79404557e-01
-5.64683266e-02 2.77066618e-01 9.57277119e-01 4.86340262e-02
-4.15292561e-01 4.75878537e-01 -6.16275549e-01 7.93469429e-01
8.86980772e-01 -2.89130419e-01 -3.30721647e-01 7.03682303e-02
-2.93493807e-01 -3.32600139e-02 4.07442808e-01 -1.14625311e+00
1.89885587e-01 -2.25522667e-01 8.75248909e-01 -4.95232701e-01
3.53694826e-01 -1.00251913e+00 -1.01442948e-01 6.58527970e-01
5.90856373e-02 -4.79509294e-01 7.31711984e-02 3.84440571e-01
-2.29818746e-01 -6.28471375e-01 9.07043099e-01 -4.80937026e-02
-6.32059157e-01 -1.28606811e-01 -3.93456221e-01 -5.97951710e-01
1.49124289e+00 -7.30805397e-01 -1.61956355e-01 6.61033615e-02
-7.25337386e-01 8.66370425e-02 2.14947343e-01 6.97235242e-02
5.82073927e-01 -1.04795325e+00 -3.53483826e-01 2.39398524e-01
-4.89592291e-02 -5.15837610e-01 2.98942208e-01 7.74100900e-01
-7.30633140e-01 7.00629234e-01 -6.83613718e-01 -3.04363817e-01
-1.69260776e+00 5.38508475e-01 1.68583080e-01 3.86925437e-03
-3.61476988e-01 6.52958572e-01 -4.76217091e-01 -1.33458227e-01
1.33653238e-01 -2.00331919e-02 -7.09776878e-01 4.09464240e-02
5.38830161e-01 7.97537148e-01 6.09445944e-02 -6.47312224e-01
-5.06480694e-01 5.35824418e-01 -2.97938228e-01 -2.54717749e-02
9.59507525e-01 3.33760500e-01 -9.61572602e-02 7.71922767e-02
1.04506063e+00 2.96640769e-02 -9.69357073e-01 -1.58736691e-01
2.23369658e-01 -8.67691457e-01 -2.84165866e-03 -7.59147704e-01
-7.37779856e-01 7.81112909e-01 7.75323570e-01 1.82550419e-02
1.23711324e+00 -5.47705770e-01 5.03688812e-01 5.71439266e-01
5.01245141e-01 -1.24850595e+00 -5.26041150e-01 1.54226333e-01
4.32576418e-01 -1.25627220e+00 7.58654326e-02 -5.84769011e-01
-7.90434480e-01 1.50893819e+00 6.74615026e-01 -1.83564778e-02
7.73852885e-01 9.66914594e-02 -9.85628739e-02 4.21554029e-01
-1.25488520e-01 -8.94111320e-02 3.54747504e-01 6.20567083e-01
8.67891535e-02 1.54264048e-01 -7.70241559e-01 8.90176713e-01
-1.33889049e-01 2.58753479e-01 3.71056527e-01 1.09200358e+00
-8.24013889e-01 -1.34500802e+00 -7.12097585e-01 5.59987068e-01
-3.76625866e-01 1.47044837e-01 -1.99095443e-01 6.79098308e-01
2.95048356e-01 9.25528824e-01 -4.83270317e-01 -4.13791418e-01
2.69385815e-01 2.45272487e-01 4.84919906e-01 -1.67933449e-01
-2.64398783e-01 -5.54641962e-01 -1.28211960e-01 -2.98518017e-02
-1.00453019e-01 -5.55875897e-01 -1.05369055e+00 -3.81723106e-01
-5.53116620e-01 3.49971503e-01 7.74810255e-01 1.21765208e+00
-5.08199558e-02 2.65766591e-01 6.21183634e-01 -4.19736296e-01
-7.18133867e-01 -8.20313394e-01 -7.35948682e-01 1.40460739e-02
-7.08610343e-04 -8.35147738e-01 -3.80946994e-01 -1.95279002e-01] | [8.221360206604004, 4.045428276062012] |
7adc6d81-aacd-41b6-b5be-0b9857472362 | concept-aware-clustering-for-decentralized | 2306.12768 | null | https://arxiv.org/abs/2306.12768v1 | https://arxiv.org/pdf/2306.12768v1.pdf | Concept-aware clustering for decentralized deep learning under temporal shift | Decentralized deep learning requires dealing with non-iid data across clients, which may also change over time due to temporal shifts. While non-iid data has been extensively studied in distributed settings, temporal shifts have received no attention. To the best of our knowledge, we are first with tackling the novel and challenging problem of decentralized learning with non-iid and dynamic data. We propose a novel algorithm that can automatically discover and adapt to the evolving concepts in the network, without any prior knowledge or estimation of the number of concepts. We evaluate our algorithm on standard benchmark datasets and demonstrate that it outperforms previous methods for decentralized learning. | ['Olof Mogren', 'Martin Willbo', 'Edvin Listo Zec', 'Emilie Klefbom', 'Marcus Toftås'] | 2023-06-22 | null | null | null | null | ['clustering'] | ['methodology'] | [-4.79634404e-01 -3.47603530e-01 -1.03710622e-01 -4.66687530e-01
-3.14921498e-01 -6.37299895e-01 4.76581484e-01 1.26331210e-01
-5.23866713e-01 9.77015138e-01 5.85151836e-03 -7.50548989e-02
-4.97460455e-01 -4.46595252e-01 -6.38291001e-01 -7.99383163e-01
-6.54087067e-01 1.18649602e+00 4.03162479e-01 2.21551582e-02
-2.50670873e-02 4.11801010e-01 -1.36582494e+00 8.95473137e-02
2.99593568e-01 9.67861354e-01 -2.00393004e-03 4.40778702e-01
-5.63965514e-02 7.79190302e-01 -7.57686675e-01 -1.17147110e-01
5.10880232e-01 -1.83644831e-01 -7.60764718e-01 1.45702288e-01
4.44038153e-01 -4.23149198e-01 -6.91589892e-01 7.95213044e-01
6.26130223e-01 2.32548967e-01 6.84543788e-01 -1.37762916e+00
-3.68874520e-01 8.83734703e-01 -8.37955296e-01 7.50845671e-01
-2.21244007e-01 -5.92802987e-02 9.41878557e-01 -5.23051679e-01
7.52440631e-01 1.41156006e+00 5.98335385e-01 5.24015188e-01
-1.18701684e+00 -7.27185309e-01 7.89682984e-01 6.50810301e-01
-1.23405302e+00 -2.46598691e-01 8.09849024e-01 -3.33128244e-01
7.07922816e-01 -4.78555650e-01 5.62192619e-01 1.19972515e+00
-5.38888499e-02 9.08717334e-01 8.46409917e-01 -2.04516873e-01
6.20501339e-01 8.65792297e-03 -2.75540650e-01 3.04906696e-01
1.57358468e-01 -2.82962561e-01 -7.78138876e-01 -2.34075874e-01
5.48221588e-01 2.54285216e-01 -6.55124187e-02 -6.90454364e-01
-7.91923046e-01 8.01588535e-01 2.29885966e-01 4.80618507e-01
-7.55836010e-01 3.51854324e-01 5.30283630e-01 1.02222693e+00
5.95521092e-01 -3.60283591e-02 -1.09078765e+00 -4.08758044e-01
-7.14037538e-01 1.66271627e-01 1.30379426e+00 9.46918130e-01
9.16750431e-01 -1.03178523e-01 2.07692355e-01 6.94168329e-01
-9.18465108e-02 3.22243333e-01 5.19533157e-01 -9.55044925e-01
6.00358069e-01 3.30151796e-01 1.14822701e-01 -8.88624310e-01
-4.38005388e-01 -5.16976953e-01 -9.75961387e-01 1.69793963e-02
4.55789864e-01 -7.14723468e-01 -3.61561418e-01 1.83993340e+00
5.60411692e-01 2.80169725e-01 -5.84438220e-02 6.60203040e-01
-7.05121905e-02 4.54194367e-01 -3.45906109e-01 -5.61783612e-01
6.02883220e-01 -8.78718138e-01 -5.72245598e-01 1.15252063e-01
3.10137957e-01 -3.40720475e-01 3.91352952e-01 6.44242942e-01
-7.52714276e-01 3.47179472e-02 -4.88984644e-01 3.30332458e-01
-1.09583229e-01 -5.45394778e-01 6.38440967e-01 2.63078451e-01
-1.42701924e+00 3.64241064e-01 -8.95970464e-01 -6.89158738e-01
6.77497506e-01 4.09023106e-01 -6.93480298e-02 -1.76417381e-01
-9.54555988e-01 4.46329236e-01 2.19616309e-01 -8.50662738e-02
-1.13509941e+00 -6.98051095e-01 9.71982330e-02 1.40978813e-01
1.02470040e+00 -6.23474419e-01 1.76027751e+00 -1.09606552e+00
-1.46282959e+00 2.04084992e-01 4.58658189e-02 -6.31267369e-01
8.73816311e-01 -7.01638311e-02 -9.67992619e-02 1.49814487e-01
-9.34120864e-02 1.56606525e-01 9.36658978e-01 -1.25547147e+00
-9.19591367e-01 -5.00901937e-01 1.14528835e-01 4.72461969e-01
-9.48124826e-01 -2.72208631e-01 -5.06725430e-01 -4.60924596e-01
-1.10332102e-01 -1.00969052e+00 -1.59741685e-01 3.73240232e-01
-1.69324242e-02 -7.49047935e-01 1.42088103e+00 1.66558884e-02
8.64575505e-01 -2.00045204e+00 1.05683610e-01 2.56764531e-01
5.24669230e-01 3.65601555e-02 -1.41548902e-01 8.02355289e-01
3.77521545e-01 -5.36297187e-02 1.71306625e-01 -5.44070661e-01
7.90726542e-02 5.72687149e-01 -2.55492270e-01 5.41703224e-01
-3.80294323e-01 5.79455912e-01 -1.02525866e+00 -5.08882046e-01
-3.54068160e-01 2.24321693e-01 -6.98327661e-01 -1.45926327e-03
-4.33826596e-01 3.74794722e-01 -7.16232777e-01 5.41260719e-01
4.56804752e-01 -6.02405429e-01 7.59829044e-01 1.39974713e-01
1.76831692e-01 -1.82067424e-01 -1.25399756e+00 1.41425169e+00
-3.55987430e-01 8.30790222e-01 4.26687539e-01 -1.57886004e+00
7.03619063e-01 6.28264904e-01 1.11622405e+00 -7.22204924e-01
-2.19041005e-01 1.03001483e-01 7.42631629e-02 -5.02942801e-01
2.45974101e-02 -1.34783790e-01 2.37056129e-02 9.28136826e-01
1.63666576e-01 2.15454221e-01 1.42484546e-01 3.57606739e-01
1.46065140e+00 -5.99585295e-01 -9.65125952e-03 -1.63303465e-01
-5.48702180e-02 -3.06755662e-01 7.30321050e-01 1.33828187e+00
-3.48150492e-01 1.00872308e-01 7.45867670e-01 -7.80867577e-01
-9.09967542e-01 -9.71621990e-01 1.07091755e-01 1.33787465e+00
1.16476163e-01 -5.82563616e-02 -3.39779139e-01 -9.08279955e-01
4.47606951e-01 -4.18395288e-02 -6.90190315e-01 2.38171950e-01
-3.94849777e-01 -5.10607660e-01 5.28280884e-02 3.25026184e-01
3.46347958e-01 -9.58330274e-01 -4.17626262e-01 6.75186753e-01
8.05087611e-02 -9.79376853e-01 -4.75994319e-01 1.92084581e-01
-1.00170529e+00 -1.14071262e+00 -4.84779596e-01 -8.16631794e-01
5.04271626e-01 4.12381947e-01 1.19763494e+00 -6.57779202e-02
-4.54540610e-01 1.12490773e+00 -2.84156948e-01 -5.89533210e-01
-2.57565305e-02 5.55152953e-01 2.42071956e-01 2.11981237e-01
2.70245850e-01 -1.05433309e+00 -8.86369467e-01 3.68226618e-01
-1.05049360e+00 -4.86459196e-01 5.17628610e-01 7.45374084e-01
3.32199872e-01 4.49716270e-01 8.24365854e-01 -1.05407596e+00
8.14441919e-01 -8.94995689e-01 -5.79415321e-01 3.15456688e-01
-7.77795196e-01 1.23477979e-02 8.00316751e-01 -8.62775445e-01
-1.00679088e+00 -4.53347117e-02 7.72311687e-01 -5.69784939e-01
-1.52879030e-01 6.41799569e-01 2.90731728e-01 1.13384187e-01
4.96535480e-01 1.61704961e-02 2.83231307e-03 -6.96371078e-01
2.86859065e-01 6.96844935e-01 1.15312263e-01 -8.65277886e-01
8.69806647e-01 8.34544718e-01 3.35531309e-04 -6.06795907e-01
-8.03633809e-01 -4.82588410e-01 -6.91963553e-01 -2.36608773e-01
5.54491915e-02 -9.49252307e-01 -9.83418286e-01 7.17453480e-01
-8.70038629e-01 -7.70166278e-01 -3.47121894e-01 3.30896020e-01
-5.86512864e-01 2.93303132e-01 -7.01529920e-01 -6.96936548e-01
-3.09150130e-01 -5.28132379e-01 4.59793150e-01 1.77793577e-01
1.90252721e-01 -1.34357953e+00 4.24203634e-01 -3.11190821e-02
8.30200076e-01 -8.63480121e-02 6.38055384e-01 -8.10716391e-01
-7.62108326e-01 -1.46595210e-01 8.58373716e-02 1.12515911e-01
4.55658019e-01 -2.74828881e-01 -4.98773217e-01 -8.16727281e-01
4.14834544e-02 -6.51038468e-01 6.86094761e-01 1.90909266e-01
1.28853476e+00 -8.24103653e-01 -3.20851177e-01 1.93629146e-01
1.53375721e+00 2.79864073e-01 -1.12831406e-01 3.22560251e-01
3.45619977e-01 4.36430573e-01 3.39467287e-01 1.28547394e+00
6.69875979e-01 6.97937965e-01 5.33064127e-01 2.08269656e-01
1.23079166e-01 -4.11531106e-02 -3.84464487e-02 8.47052574e-01
2.67301500e-01 -5.27683377e-01 -1.08413553e+00 9.32404995e-01
-2.23409939e+00 -1.08781743e+00 4.80124295e-01 1.68842018e+00
1.10789108e+00 -1.34032637e-01 3.93830180e-01 -4.09841388e-01
8.46648574e-01 -5.79615682e-02 -1.36503732e+00 -7.75397383e-03
-1.07100584e-01 -8.70245229e-03 4.74801749e-01 1.44923091e-01
-8.74159098e-01 8.55053246e-01 6.51149893e+00 3.85243773e-01
-1.20169103e+00 2.68751532e-01 4.50541317e-01 -6.26612902e-01
-3.03294323e-02 -1.24304160e-01 -6.44674540e-01 4.78510261e-01
6.30692363e-01 -4.84934211e-01 7.76739657e-01 7.88541257e-01
1.06711291e-01 1.23621419e-01 -1.18729591e+00 9.79779363e-01
-9.58713219e-02 -1.05854511e+00 -2.01540589e-01 7.52571821e-02
1.36567044e+00 6.23635530e-01 1.43807948e-01 3.41216147e-01
7.83888221e-01 -3.81242275e-01 4.46650475e-01 7.86624491e-01
1.53055593e-01 -7.34130263e-01 5.54056764e-01 5.28024554e-01
-9.35429454e-01 -6.28561795e-01 -3.02586585e-01 4.35919315e-02
-6.19797669e-02 5.09622097e-01 -8.80012512e-01 1.62133694e-01
1.15617979e+00 9.76403415e-01 -4.10973221e-01 1.31959713e+00
2.37317368e-01 7.98035502e-01 -6.20451093e-01 -1.33281529e-01
3.41571748e-01 1.06979132e-01 4.61972654e-01 8.08453381e-01
2.31473312e-01 -8.50019306e-02 4.76207912e-01 2.74780452e-01
-3.87064695e-01 6.14352748e-02 -5.50032079e-01 4.20197695e-02
8.65227282e-01 8.22999477e-01 -5.98419189e-01 -3.89069289e-01
-6.33673072e-01 7.56222546e-01 7.29521036e-01 8.06318402e-01
-2.61269689e-01 -7.64986798e-02 8.28650117e-01 -3.44252884e-02
8.26316595e-01 -3.07174802e-01 3.43436837e-01 -1.21855831e+00
1.85408160e-01 -8.58102798e-01 8.73652637e-01 -5.77850603e-02
-2.06930304e+00 3.16956401e-01 -1.35221973e-01 -9.17498469e-01
-3.51029962e-01 -2.04922289e-01 -3.83125842e-01 1.24505788e-01
-1.51835465e+00 -4.92919028e-01 -1.85403496e-01 1.12285507e+00
5.04083991e-01 -5.22047460e-01 4.24618363e-01 3.06428254e-01
-5.23007154e-01 6.44272804e-01 1.05413008e+00 -3.23534459e-02
1.04406738e+00 -1.32905328e+00 -5.00121787e-02 2.62408763e-01
-3.63580696e-03 3.75473678e-01 7.18537331e-01 -3.58804882e-01
-1.51459503e+00 -1.06586444e+00 3.78458500e-01 -5.22709787e-02
1.07864118e+00 -3.71571720e-01 -1.00419521e+00 7.28355944e-01
1.04020290e-01 2.80438095e-01 5.14927506e-01 3.37819695e-01
-5.00887871e-01 -8.40443194e-01 -1.08945966e+00 2.17475623e-01
1.04470539e+00 -3.83320212e-01 -9.79872942e-02 5.64953923e-01
6.23993158e-01 -5.68308383e-02 -7.57652521e-01 4.45336103e-02
2.80107975e-01 -6.95227504e-01 6.02328062e-01 -7.06241488e-01
-2.05872998e-01 5.48781939e-02 5.91125898e-03 -1.23940587e+00
-2.46126145e-01 -8.25359344e-01 -5.31956196e-01 1.10499406e+00
1.60036504e-01 -9.27037776e-01 1.11283612e+00 4.44094300e-01
2.88826019e-01 -5.68584919e-01 -1.45781314e+00 -9.84925032e-01
1.35443792e-01 5.16603813e-02 4.14116621e-01 1.00266480e+00
-1.80107951e-02 8.74067098e-02 -6.92174077e-01 -2.97552720e-03
7.49032438e-01 2.60375291e-01 7.67246008e-01 -1.37109029e+00
-4.84116733e-01 -4.64371920e-01 -3.85884553e-01 -1.03026915e+00
3.16745222e-01 -4.83487308e-01 8.65861252e-02 -1.29713166e+00
3.66982877e-01 -4.50562537e-01 -4.86485302e-01 7.23780572e-01
2.60365397e-01 -1.88425601e-01 3.32719982e-02 5.71146011e-01
-1.42038977e+00 8.90694380e-01 8.20794582e-01 -1.26610979e-01
-3.37477803e-01 3.97966802e-01 -7.17017353e-01 1.66324586e-01
1.10539377e+00 -9.28179979e-01 -7.15796292e-01 -6.42278314e-01
2.22943142e-01 -2.40741745e-02 4.00823392e-02 -9.03493643e-01
6.76648617e-01 -3.69966090e-01 -6.14927933e-02 -5.31637430e-01
-7.27227330e-02 -1.02271390e+00 -3.09011880e-02 5.29214799e-01
-3.53368819e-01 2.08381072e-01 -5.74570484e-02 1.29267716e+00
-8.46915096e-02 7.79778063e-02 5.54122210e-01 -1.40984848e-01
-4.98226851e-01 9.34120953e-01 -6.18435919e-01 3.90552789e-01
1.09013629e+00 2.22785980e-01 -2.44004890e-01 -9.33444202e-01
-7.22218215e-01 7.87549198e-01 3.13426673e-01 5.65551519e-01
2.97117651e-01 -1.09868503e+00 -5.51650763e-01 -4.13705409e-01
-1.77759975e-01 1.20867997e-01 4.08181012e-01 6.18346274e-01
-1.28273129e-01 2.72628278e-01 5.63530996e-02 -6.64453149e-01
-9.72560763e-01 8.26833129e-01 3.61359566e-01 -1.64181560e-01
-8.31100941e-01 4.90883797e-01 -3.09614856e-02 -3.43169034e-01
8.32915783e-01 3.87772322e-01 3.30265202e-02 5.63840687e-01
2.62235731e-01 4.05209005e-01 -1.63863018e-01 3.86312008e-02
-3.24250162e-01 1.86460137e-01 -6.52678967e-01 -8.49670693e-02
1.99579370e+00 -5.32782316e-01 -1.54581293e-01 8.48830044e-01
1.31411386e+00 -6.33114994e-01 -1.73256683e+00 -1.14339685e+00
2.11016804e-01 -2.24793345e-01 -1.62173390e-01 -5.62473118e-01
-1.40789592e+00 6.36065066e-01 6.99851990e-01 3.72571409e-01
1.08612800e+00 3.59606773e-01 5.99437356e-01 9.44252372e-01
6.11664176e-01 -1.39787924e+00 8.41180921e-01 7.10426152e-01
6.14388108e-01 -1.14100516e+00 -2.41007775e-01 4.64420855e-01
-3.61124873e-01 1.27800083e+00 7.80367732e-01 -4.37951274e-03
1.11432600e+00 2.33162463e-01 1.16517097e-01 -1.88496426e-01
-1.56789708e+00 -8.30630362e-02 -3.26511681e-01 5.52800536e-01
-1.35648936e-01 -2.45995104e-01 1.89726315e-02 -5.63691705e-02
1.53636217e-01 -2.05833279e-03 4.14815187e-01 1.22931266e+00
-4.34509695e-01 -1.13500977e+00 -3.55486684e-02 4.68482822e-01
-3.62196177e-01 3.63112807e-01 -3.65552902e-01 5.55777907e-01
-1.07475065e-01 7.62520790e-01 2.50042677e-01 6.87094107e-02
1.09374017e-01 -9.58467871e-02 3.03106308e-01 -5.63970625e-01
-3.12173367e-01 2.92949863e-02 -4.23410594e-01 -2.83800423e-01
-4.50150430e-01 -1.08341396e+00 -7.85085678e-01 -3.97766590e-01
4.97917049e-02 2.56241679e-01 5.78431070e-01 9.22926068e-01
5.82240701e-01 1.00861840e-01 1.20211589e+00 -6.74175143e-01
-1.15174103e+00 -8.41830134e-01 -7.02105522e-01 3.26040506e-01
5.70521474e-01 -5.83605528e-01 -5.92749596e-01 -1.48723990e-01] | [5.871539115905762, 6.113372325897217] |
99163416-6b5b-4aca-bd81-075827c11b7d | variance-reduced-proxskip-algorithm-theory | 2207.04338 | null | https://arxiv.org/abs/2207.04338v1 | https://arxiv.org/pdf/2207.04338v1.pdf | Variance Reduced ProxSkip: Algorithm, Theory and Application to Federated Learning | We study distributed optimization methods based on the {\em local training (LT)} paradigm: achieving communication efficiency by performing richer local gradient-based training on the clients before parameter averaging. Looking back at the progress of the field, we {\em identify 5 generations of LT methods}: 1) heuristic, 2) homogeneous, 3) sublinear, 4) linear, and 5) accelerated. The 5${}^{\rm th}$ generation, initiated by the ProxSkip method of Mishchenko, Malinovsky, Stich and Richt\'{a}rik (2022) and its analysis, is characterized by the first theoretical confirmation that LT is a communication acceleration mechanism. Inspired by this recent progress, we contribute to the 5${}^{\rm th}$ generation of LT methods by showing that it is possible to enhance them further using {\em variance reduction}. While all previous theoretical results for LT methods ignore the cost of local work altogether, and are framed purely in terms of the number of communication rounds, we show that our methods can be substantially faster in terms of the {\em total training cost} than the state-of-the-art method ProxSkip in theory and practice in the regime when local computation is sufficiently expensive. We characterize this threshold theoretically, and confirm our theoretical predictions with empirical results. | ['Peter Richtárik', 'Kai Yi', 'Grigory Malinovsky'] | 2022-07-09 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-1.53219386e-03 2.19339520e-01 -2.10303336e-01 -5.85346967e-02
-1.19784701e+00 -4.72667515e-01 5.51169634e-01 3.36087979e-02
-8.19765925e-01 1.09559178e+00 -1.33907601e-01 -5.81039488e-01
-6.89665675e-01 -5.54496765e-01 -9.44733739e-01 -1.05969918e+00
-3.75854909e-01 7.62890160e-01 2.13643983e-02 -3.68445888e-02
3.62668186e-01 5.04079521e-01 -1.22369373e+00 -8.29100981e-02
7.09903240e-01 1.26811421e+00 4.07816395e-02 1.01985168e+00
-1.31954581e-01 9.15180564e-01 -7.32889891e-01 -4.64784175e-01
3.77934724e-01 -8.03498566e-01 -1.27834296e+00 -7.67171755e-02
-9.78645980e-02 -2.82262146e-01 -2.41409466e-01 9.48468983e-01
8.34171176e-01 1.10830314e-01 3.23441684e-01 -1.21837115e+00
-8.25410709e-02 1.14370108e+00 -6.23506963e-01 1.76299021e-01
-9.86179411e-02 2.91390233e-02 8.40273798e-01 -4.21717644e-01
7.00677633e-01 8.63309264e-01 6.81298375e-01 4.61384803e-01
-1.33885729e+00 -6.58772707e-01 1.06786020e-01 -2.77895704e-02
-1.34273958e+00 -5.31010628e-01 4.04072672e-01 -6.14519306e-02
8.84537041e-01 4.83225137e-01 4.23153132e-01 6.18999243e-01
-2.50006169e-01 7.43472695e-01 1.31322122e+00 -6.87636137e-01
3.74200404e-01 1.78932115e-01 6.06602319e-02 8.85534883e-01
1.23533852e-01 9.18964073e-02 -6.53148234e-01 -5.42546868e-01
5.22441149e-01 -2.94854909e-01 -3.37220281e-01 -4.92893606e-02
-1.00519812e+00 7.91676342e-01 8.88925642e-02 3.26631695e-01
-3.30873996e-01 7.99945951e-01 3.26065451e-01 5.90285897e-01
7.13784814e-01 2.22348511e-01 -6.46939218e-01 -7.20569491e-01
-1.11699736e+00 1.92780972e-01 1.14488804e+00 1.02215719e+00
9.23759520e-01 -1.56514719e-01 1.52243944e-02 7.49756396e-01
-6.67102560e-02 7.20385432e-01 2.17669263e-01 -1.60209775e+00
6.99488759e-01 2.06990212e-01 9.43176597e-02 -2.76397824e-01
-5.48506856e-01 -6.27764225e-01 -7.90516019e-01 -2.18550651e-03
4.40884560e-01 -7.23777115e-01 -1.02013238e-01 1.81051707e+00
1.78100243e-01 -8.48422348e-02 -2.82117873e-01 4.24646318e-01
2.09257364e-01 5.98590791e-01 -5.10985494e-01 -7.50650227e-01
6.82195067e-01 -1.02962387e+00 -2.04370588e-01 4.39090163e-01
1.31925571e+00 -7.66596138e-01 6.71769619e-01 5.18272161e-01
-1.59757543e+00 -7.88482465e-03 -7.46248841e-01 4.34094399e-01
-8.28828886e-02 -7.08793029e-02 8.24595571e-01 1.06405652e+00
-1.70578325e+00 9.57477927e-01 -7.62071908e-01 -4.10808146e-01
2.89665908e-01 6.89303577e-01 1.18924022e-01 1.38803467e-01
-6.15475595e-01 4.47353750e-01 -1.06284030e-01 -1.06034167e-01
-7.91134596e-01 -4.89631474e-01 -1.11508407e-01 5.41786812e-02
5.38132548e-01 -6.72499299e-01 1.25444567e+00 -7.88175344e-01
-1.93464541e+00 5.90446472e-01 -2.78976440e-01 -5.54931343e-01
7.76651800e-01 1.07086278e-01 2.60321885e-01 7.63980746e-02
-4.02735084e-01 1.49924308e-01 4.43709970e-01 -1.14553607e+00
-6.66572809e-01 -4.19255286e-01 1.59329563e-01 1.50043160e-01
-5.81130624e-01 2.37968713e-01 -3.50926161e-01 -2.34977424e-01
-2.42778286e-01 -1.23463702e+00 -4.23071146e-01 -5.02272487e-01
-4.15630460e-01 -2.84421712e-01 4.35429782e-01 -2.35799536e-01
1.25902069e+00 -1.73728466e+00 1.96017131e-01 7.42472529e-01
5.79722345e-01 3.37313861e-02 3.01175565e-03 8.14867854e-01
3.15136909e-01 3.34517062e-01 1.01226345e-01 -7.11952090e-01
3.34265590e-01 3.78441773e-02 5.65816052e-02 7.53170073e-01
-8.19971621e-01 7.06480324e-01 -6.45118535e-01 -5.02086997e-01
1.07341642e-02 3.28979909e-01 -6.39809310e-01 -6.10352568e-02
-1.33089796e-01 4.47391480e-01 -5.96618116e-01 2.96050578e-01
5.57321310e-01 -7.28901088e-01 3.37721169e-01 3.05813104e-01
-1.98692217e-01 3.15720409e-01 -1.33500862e+00 1.35544872e+00
-6.54595256e-01 5.26057601e-01 6.00155652e-01 -1.32194686e+00
2.40813106e-01 2.31383979e-01 8.86825264e-01 -3.68089795e-01
3.31553459e-01 4.93329704e-01 -4.35756952e-01 -1.21702038e-01
2.47129306e-01 3.87227416e-01 1.78883731e-01 1.09912348e+00
-1.01006113e-01 6.34503067e-02 2.45268807e-01 3.70740533e-01
1.52455616e+00 -9.37080532e-02 -4.13567930e-01 -4.55280215e-01
1.88435286e-01 -1.81583136e-01 5.90121113e-02 1.26399326e+00
-9.62534919e-02 1.03321448e-01 6.37614489e-01 -9.23231915e-02
-1.02901936e+00 -8.57317924e-01 -2.14197487e-02 1.59135234e+00
6.21512197e-02 -6.89163625e-01 -1.19275653e+00 -4.00355309e-01
-2.18601674e-01 4.68394905e-01 -5.93364716e-01 2.45477900e-01
-6.44250631e-01 -1.14066207e+00 6.90285861e-01 2.42634207e-01
7.22568989e-01 -6.79700553e-01 -3.18224460e-01 1.36101395e-01
-1.81699425e-01 -1.01530683e+00 -4.34409946e-01 4.17014629e-01
-9.98212159e-01 -6.05508983e-01 -8.34614158e-01 -3.01225215e-01
5.75345755e-01 -6.05114102e-02 1.03216290e+00 2.44695872e-01
-2.62563467e-01 6.86362088e-01 -2.33752415e-01 -2.21799448e-01
-3.32815170e-01 3.87600750e-01 -6.65302901e-03 -2.83487380e-01
-1.88930213e-01 -8.74418855e-01 -9.84704971e-01 3.14584732e-01
-4.79905307e-01 -2.53292024e-01 7.15494871e-01 7.79662371e-01
3.08087140e-01 1.77314058e-01 9.99285281e-02 -9.94753778e-01
6.50043070e-01 -3.23206544e-01 -7.38117874e-01 1.06306128e-01
-1.07833159e+00 3.60805780e-01 6.10656798e-01 -2.59035468e-01
-8.79366040e-01 -2.52505004e-01 -2.16569334e-01 3.11827920e-02
3.70473117e-01 1.59813419e-01 4.58600491e-01 -6.45378649e-01
6.64544404e-01 4.09186959e-01 -4.25632000e-02 -3.67212921e-01
5.06905615e-01 6.00111008e-01 -4.13330346e-02 -1.02937317e+00
5.03232181e-01 6.06065810e-01 3.99066299e-01 -7.82563925e-01
-5.54572642e-01 -2.32376367e-01 5.09984232e-02 -1.58608750e-01
4.11245912e-01 -5.33751786e-01 -1.65769362e+00 1.74491018e-01
-9.58737433e-01 -7.77759433e-01 -4.30881560e-01 4.94263977e-01
-8.45235348e-01 4.77792740e-01 -9.37999308e-01 -1.24318159e+00
-4.56746459e-01 -1.19820082e+00 9.48319733e-01 -5.29954731e-02
3.28438468e-02 -1.03175485e+00 -1.33308142e-01 4.30883199e-01
9.12016749e-01 -1.75160065e-01 6.61255717e-01 -5.86997747e-01
-7.62385547e-01 -3.05723906e-01 -3.77469063e-01 3.32819968e-01
-5.99025309e-01 -1.11905359e-01 -7.12127686e-01 -5.96992254e-01
-1.50075974e-02 -2.63364375e-01 6.29690409e-01 5.81232488e-01
1.42085290e+00 -5.52531183e-01 -5.26058316e-01 7.19703078e-01
1.57819664e+00 -2.47693717e-01 4.59458947e-01 8.50326195e-02
3.33286911e-01 2.96523184e-01 -1.44480944e-01 9.46737647e-01
3.65081847e-01 7.47665226e-01 3.62847805e-01 -1.07928485e-01
1.16453260e-01 1.21653013e-01 4.18572634e-01 9.73252356e-01
-7.17690825e-01 -2.26541504e-01 -5.74956000e-01 4.23983961e-01
-1.97210026e+00 -9.15524125e-01 6.14936762e-02 2.59347439e+00
8.33978534e-01 2.09322453e-01 5.14159679e-01 4.23739143e-02
6.86983109e-01 -3.17851491e-02 -1.77659661e-01 -5.46230078e-01
-2.95607392e-02 5.62155545e-01 1.27827561e+00 6.13614798e-01
-4.59406942e-01 7.07103372e-01 6.50989866e+00 1.35549009e+00
-8.64680231e-01 6.19022429e-01 7.93898404e-01 -5.66157460e-01
-1.05414845e-01 1.40205413e-01 -8.51528227e-01 6.63829327e-01
1.40962029e+00 -1.19553149e-01 1.09347677e+00 8.71379316e-01
7.02245981e-02 -2.86367089e-01 -1.05271113e+00 1.04900801e+00
-2.10188493e-01 -1.47347844e+00 -4.29429471e-01 5.51638544e-01
9.73704100e-01 3.76481563e-01 9.37064961e-02 2.33622491e-01
9.70125556e-01 -9.73017395e-01 1.82819813e-01 2.59386420e-01
5.94859779e-01 -8.10413480e-01 7.00059414e-01 5.25973141e-01
-1.08708358e+00 -2.31580272e-01 -1.64868563e-01 -5.58294766e-02
1.69455454e-01 8.54236066e-01 -4.01929200e-01 5.53870022e-01
8.67875278e-01 -2.81855971e-01 -7.16469809e-02 7.75025725e-01
2.69529730e-01 9.10544217e-01 -1.00813878e+00 -5.26093602e-01
3.28077674e-01 -2.87742734e-01 4.59412694e-01 1.03093088e+00
3.53977680e-01 7.31309056e-02 8.25928822e-02 4.58068520e-01
-3.61875772e-01 2.64892429e-01 -1.94653615e-01 4.08472540e-03
5.04809499e-01 1.04687381e+00 -6.79848254e-01 -4.33038980e-01
-3.95631827e-02 9.94996071e-01 4.62832838e-01 4.19792354e-01
-8.79993021e-01 -5.18148065e-01 2.42868006e-01 4.20387685e-02
3.48252475e-01 -3.28703642e-01 -4.15316373e-01 -7.47075021e-01
3.03323656e-01 -4.20077294e-01 2.35549659e-01 -2.10134715e-01
-8.85314882e-01 5.57585239e-01 -8.57962668e-02 -5.60493886e-01
-2.27600738e-01 -4.17136550e-01 -3.21813524e-01 6.69545114e-01
-1.16856277e+00 -7.68971384e-01 -1.27576869e-02 7.50388026e-01
8.92581865e-02 1.27773970e-01 6.35325372e-01 3.85741055e-01
-2.72503316e-01 1.14266360e+00 8.61532271e-01 -2.45608374e-01
3.89370501e-01 -1.10779095e+00 -1.19730808e-01 3.72694731e-01
-1.20107889e-01 6.57878876e-01 7.44210958e-01 -5.37634864e-02
-1.73608077e+00 -5.51804125e-01 9.85954165e-01 -5.45244396e-01
6.73332214e-01 -3.32753211e-01 -5.46286292e-02 5.23483753e-01
2.86615580e-01 -5.94556779e-02 5.16115665e-01 3.73039007e-01
4.37794030e-02 -4.59878653e-01 -1.07719111e+00 6.01926446e-01
1.42229080e+00 -3.48475486e-01 3.32526714e-01 9.77552414e-01
5.60235143e-01 -2.20720276e-01 -9.43743289e-01 1.21812448e-01
3.14068973e-01 -1.19898617e+00 6.80225015e-01 -2.13986218e-01
5.98070249e-02 3.50288302e-01 -1.19789854e-01 -9.77062106e-01
2.44713724e-01 -1.54155910e+00 -1.32103384e-01 9.01803911e-01
6.44783974e-01 -9.62131262e-01 9.79397535e-01 4.40727413e-01
-6.87460378e-02 -1.04360902e+00 -1.27441871e+00 -9.78770971e-01
2.47091994e-01 -5.14187753e-01 3.90501946e-01 5.75636566e-01
4.09494132e-01 1.52133808e-01 -4.00670022e-01 -1.75478280e-01
7.19997823e-01 -5.33839166e-02 9.75261986e-01 -8.62595081e-01
-9.72666800e-01 -9.04724658e-01 6.05983008e-03 -1.59272325e+00
-1.80765495e-01 -5.20845830e-01 -1.58002466e-01 -1.20124030e+00
3.91507030e-01 -7.38910556e-01 -2.78966397e-01 3.16593885e-01
3.76441151e-01 1.36144236e-01 6.65676519e-02 3.73377353e-01
-1.15034664e+00 3.09037000e-01 7.89470673e-01 2.84279048e-01
-2.82973945e-01 3.86582226e-01 -4.40795153e-01 3.26495111e-01
5.56123376e-01 -3.73571754e-01 -1.55111983e-01 -3.48754734e-01
7.13525891e-01 2.44548157e-01 5.09753004e-02 -8.24001431e-01
5.28412640e-01 -8.95281509e-03 -1.89399093e-01 -2.53436506e-01
4.76351887e-01 -6.41395271e-01 -9.74247158e-02 5.74309468e-01
-3.49413723e-01 1.06406599e-01 -2.51086622e-01 3.69341522e-01
3.95700395e-01 -8.13387409e-02 7.86681950e-01 -9.15213972e-02
1.49331808e-01 3.92165333e-01 -4.16714847e-01 1.20472118e-01
8.88789415e-01 3.69278677e-02 -6.81280792e-01 -7.95047998e-01
-6.84288859e-01 2.94960439e-01 1.76748216e-01 -6.34732068e-01
-7.15404451e-02 -7.80664802e-01 -5.79122782e-01 -3.62504989e-01
-3.98252189e-01 -1.99308023e-01 2.37128183e-01 1.64559042e+00
-5.71642101e-01 3.27241987e-01 5.05023420e-01 -5.12448847e-01
-1.04468739e+00 4.24712330e-01 3.80240977e-01 -6.26561701e-01
-3.21671665e-01 1.25906086e+00 -3.23134959e-01 -2.55119413e-01
5.59674561e-01 1.05896793e-01 7.42207110e-01 -4.23572779e-01
2.60396421e-01 1.04705536e+00 9.86970812e-02 -2.12731268e-02
-2.44975433e-01 4.87605482e-01 -2.31154524e-02 -5.31998396e-01
1.35462129e+00 -4.62831736e-01 -4.29098219e-01 1.92485139e-01
1.42425454e+00 1.80765331e-01 -8.99167478e-01 -5.33748746e-01
-2.19599888e-01 -1.34705827e-01 -2.60180309e-02 -6.38896048e-01
-1.12802553e+00 6.93323553e-01 6.14055276e-01 5.36539495e-01
1.05767429e+00 1.83987707e-01 9.48851883e-01 6.77437484e-01
7.65436292e-01 -1.58957303e+00 -7.15602338e-02 3.91477764e-01
2.99790144e-01 -9.43942904e-01 1.90659225e-01 -2.64333367e-01
-1.85471922e-01 9.85671699e-01 1.28214825e-02 -5.70055805e-02
9.13963318e-01 4.32012111e-01 -4.43192810e-01 -1.53801367e-01
-7.74127841e-01 -2.42113739e-01 -4.51899707e-01 8.12722463e-03
1.95774078e-01 -6.05318062e-02 -6.97663426e-01 4.12805200e-01
-3.29365820e-01 -2.82511525e-02 1.67257994e-01 1.07290137e+00
-3.48481923e-01 -1.24284685e+00 -1.46804348e-01 4.84116852e-01
-6.98099613e-01 -2.17341453e-01 -2.76487857e-01 7.39199698e-01
-9.57973078e-02 1.08385336e+00 -3.20765048e-01 -3.44606817e-01
-8.85269791e-02 -5.96114658e-02 7.53897965e-01 -2.10560877e-02
-8.49325240e-01 1.97025642e-01 2.62967825e-01 -6.96947098e-01
-2.02725202e-01 -3.57968152e-01 -1.10800540e+00 -1.21679354e+00
-5.41086495e-01 6.11364603e-01 9.16826546e-01 1.19589674e+00
6.16056323e-01 2.17861250e-01 9.74944770e-01 -9.35985982e-01
-9.37291861e-01 -8.50970566e-01 -7.20631719e-01 1.00652659e-02
4.71492559e-02 -2.05776095e-01 -8.37504923e-01 -2.74234235e-01] | [6.304009437561035, 4.8940510749816895] |
82aca780-8264-4635-a6e2-fe4c40335a2c | normalized-multivariate-time-series-causality | 2104.11360 | null | https://arxiv.org/abs/2104.11360v1 | https://arxiv.org/pdf/2104.11360v1.pdf | Normalized multivariate time series causality analysis and causal graph reconstruction | Causality analysis is an important problem lying at the heart of science, and is of particular importance in data science and machine learning. An endeavor during the past 16 years viewing causality as real physical notion so as to formulate it from first principles, however, seems to go unnoticed. This study introduces to the community this line of work, with a long-due generalization of the information flow-based bivariate time series causal inference to multivariate series, based on the recent advance in theoretical development. The resulting formula is transparent, and can be implemented as a computationally very efficient algorithm for application. It can be normalized, and tested for statistical significance. Different from the previous work along this line where only information flows are estimated, here an algorithm is also implemented to quantify the influence of a unit to itself. While this forms a challenge in some causal inferences, here it comes naturally, and hence the identification of self-loops in a causal graph is fulfilled automatically as the causalities along edges are inferred. To demonstrate the power of the approach, presented here are two applications in extreme situations. The first is a network of multivariate processes buried in heavy noises (with the noise-to-signal ratio exceeding 100), and the second a network with nearly synchronized chaotic oscillators. In both graphs, confounding processes exist. While it seems to be a huge challenge to reconstruct from given series these causal graphs, an easy application of the algorithm immediately reveals the desideratum. Particularly, the confounding processes have been accurately differentiated. Considering the surge of interest in the community, this study is very timely. | ['X. San Liang'] | 2021-04-23 | null | null | null | null | ['graph-reconstruction'] | ['graphs'] | [ 4.64981467e-01 3.31568420e-01 7.03405812e-02 1.78058669e-01
1.62668273e-01 -6.63046777e-01 9.96509969e-01 3.98497432e-01
-2.40394115e-01 9.94770706e-01 5.12412935e-02 -5.67369580e-01
-7.87742615e-01 -9.68792081e-01 -5.21342039e-01 -1.18673015e+00
-7.48642564e-01 2.60233879e-01 1.62815645e-01 -2.94171929e-01
3.37402612e-01 5.94847560e-01 -1.33247077e+00 -3.72992724e-01
6.37773931e-01 5.77946067e-01 -3.21192257e-02 6.32714272e-01
8.63053426e-02 8.24063063e-01 -5.68585336e-01 -5.26871502e-01
-8.03967938e-02 -6.63284659e-01 -6.97148502e-01 -2.47807756e-01
-2.30603382e-01 2.20730558e-01 -2.41373330e-01 8.69463861e-01
2.94550866e-01 -2.72301555e-01 6.27457619e-01 -1.40807009e+00
-3.10599059e-01 7.31820583e-01 -4.31069195e-01 3.43787700e-01
2.77306139e-01 -2.21924596e-02 1.02636492e+00 -4.81772900e-01
4.61569577e-01 1.08336341e+00 5.81777573e-01 -2.00971793e-02
-1.52260494e+00 -4.54995751e-01 -5.84930703e-02 1.13446191e-01
-1.06395662e+00 1.16983438e-02 8.57169628e-01 -6.98093653e-01
3.52427781e-01 4.03027803e-01 7.99628615e-01 1.17093074e+00
4.56958145e-01 2.06191003e-01 1.23773193e+00 -4.83502924e-01
4.50952321e-01 -1.16280593e-01 1.51142567e-01 3.16844761e-01
7.35223651e-01 4.23990995e-01 -5.50438285e-01 -2.28584841e-01
7.71941006e-01 -1.66348800e-01 -3.88242126e-01 -2.59583473e-01
-1.46463287e+00 7.85938144e-01 1.83383703e-01 8.40635777e-01
-2.80287534e-01 3.02945644e-01 3.85325253e-01 6.02780640e-01
5.19076288e-01 4.03265685e-01 -1.94820017e-01 -1.31124765e-01
-8.22697222e-01 3.46934587e-01 1.16772294e+00 3.24404180e-01
4.20252204e-01 2.68692374e-02 2.70405561e-01 1.32571599e-02
1.28900021e-01 5.23274481e-01 -6.93317950e-02 -5.91585755e-01
5.69287427e-02 4.99930084e-01 4.89597917e-02 -1.63156903e+00
-4.13293421e-01 -5.90487838e-01 -1.35355151e+00 2.42452234e-01
9.73785877e-01 -3.00349861e-01 2.47866809e-02 1.73766935e+00
2.68043846e-01 3.55235577e-01 -2.07373336e-01 7.81266689e-01
1.41180709e-01 4.85189766e-01 -5.92184365e-02 -8.42407823e-01
1.12438071e+00 6.72146864e-03 -1.00313115e+00 3.42841476e-01
2.03835964e-01 -6.85850859e-01 5.37909567e-01 5.68057597e-01
-6.70959830e-01 -1.60151988e-01 -8.89503181e-01 5.11996806e-01
-4.31455404e-01 -5.32060504e-01 8.98392498e-01 7.25839436e-01
-1.05936921e+00 1.02047610e+00 -6.47281766e-01 -5.22200525e-01
-1.21989881e-03 2.45564356e-01 -2.27591455e-01 2.73854882e-01
-1.50932562e+00 6.96438313e-01 1.60729125e-01 4.13055241e-01
-4.45951492e-01 -7.14691997e-01 -4.01165575e-01 -4.18056548e-02
4.93977606e-01 -7.38167465e-01 6.37127280e-01 -9.33625698e-01
-1.20832884e+00 4.46733683e-01 -1.73154905e-01 -4.39510107e-01
9.25238490e-01 2.82757729e-01 -5.36169231e-01 9.13502425e-02
2.60457154e-02 -2.93081373e-01 7.83608556e-01 -1.33702886e+00
-2.72922546e-01 -4.36570227e-01 -1.40734330e-01 -4.30922121e-01
-1.62370205e-01 4.02403437e-02 1.81680188e-01 -8.11302066e-01
1.13057174e-01 -6.25157416e-01 -2.44953588e-01 -2.51494288e-01
-4.84800786e-01 -2.00782523e-01 5.41662693e-01 -4.53063101e-01
1.38229144e+00 -1.82909191e+00 1.83721885e-01 5.30265808e-01
4.95077819e-01 -2.97793776e-01 3.08938295e-01 9.31195915e-01
-5.06764829e-01 3.62262666e-01 -6.33145571e-01 5.49074747e-02
-1.18698075e-01 2.24180341e-01 -6.16698086e-01 8.29729676e-01
4.03980196e-01 6.08186364e-01 -1.31402719e+00 -2.52993226e-01
1.71770826e-01 4.11540270e-01 -2.63191879e-01 6.01117872e-02
8.16977397e-02 6.56763852e-01 -4.86226618e-01 1.30387425e-01
3.76160204e-01 -2.78613955e-01 4.29402232e-01 8.15719292e-02
-5.50334692e-01 -8.29906315e-02 -1.44969869e+00 9.46078122e-01
-6.42697290e-02 8.25409710e-01 -5.38077056e-02 -1.54185867e+00
8.43394697e-01 7.05877542e-01 8.45761001e-01 -4.99456763e-01
1.79455042e-01 2.98450530e-01 2.05508173e-01 -5.79903305e-01
2.36782029e-01 -3.13586503e-01 -4.43763360e-02 4.61623460e-01
-1.76847994e-01 2.80681811e-02 5.09891748e-01 2.04892799e-01
1.43371511e+00 -8.92402455e-02 4.74721491e-01 -6.57005489e-01
4.59236443e-01 1.65700540e-02 3.10952783e-01 6.62203729e-01
3.36943828e-02 2.13299260e-01 1.24638748e+00 -3.89969498e-01
-9.98358607e-01 -1.04870760e+00 -3.35027903e-01 1.77208140e-01
7.52127916e-03 -3.39407623e-01 -5.12807846e-01 -2.68783122e-01
8.03921372e-02 3.33846033e-01 -9.48237538e-01 1.94029391e-01
-5.77261984e-01 -1.08564389e+00 4.21984613e-01 -9.65133905e-02
2.38517642e-01 -1.02377045e+00 -6.06093705e-01 3.50812256e-01
-1.08877063e-01 -8.55160296e-01 3.28633219e-01 2.18313575e-01
-9.32606757e-01 -1.35537255e+00 -6.19885385e-01 -1.99374601e-01
4.39957619e-01 7.69585744e-03 1.22405171e+00 -2.34790612e-02
-3.21854174e-01 2.59277314e-01 -1.97069257e-01 -4.59381938e-01
-6.33934677e-01 -3.32438231e-01 1.52008399e-01 2.63680756e-01
7.99745023e-02 -1.14607573e+00 -3.79044801e-01 2.59400278e-01
-9.52218056e-01 -4.09993619e-01 3.56485009e-01 7.48124540e-01
2.88481768e-02 4.82506871e-01 7.73779750e-01 -7.75400400e-01
7.11888194e-01 -7.56160915e-01 -7.08683252e-01 4.15661000e-02
-6.75683916e-01 4.03761044e-02 8.34925592e-01 -5.29351793e-02
-7.41907954e-01 -3.18364739e-01 2.97626317e-01 -8.44050839e-04
-1.80541068e-01 7.04502404e-01 -3.74845304e-02 2.86447853e-01
5.80618918e-01 -1.60481930e-02 1.13555558e-01 -3.18990648e-01
3.43839139e-01 1.37281030e-01 4.45402175e-01 -5.07424414e-01
9.85383630e-01 7.54757583e-01 6.09312952e-01 -1.12290621e+00
-2.22726092e-01 -2.68087834e-01 -7.10889280e-01 -4.97717470e-01
7.37717569e-01 -4.74804461e-01 -1.18652570e+00 2.75353104e-01
-1.38908207e+00 -1.24541633e-01 -3.17851067e-01 3.74532521e-01
-3.97580534e-01 3.34433883e-01 -3.06220055e-01 -1.24615359e+00
3.39968532e-01 -6.99485779e-01 4.62608725e-01 -1.64567679e-01
-4.63533431e-01 -1.56237721e+00 3.07098001e-01 -2.84948885e-01
2.00553849e-01 7.56588876e-01 9.03088808e-01 -3.66745621e-01
-5.10703266e-01 -1.93360776e-01 -1.10111021e-01 -7.40758926e-02
3.05842251e-01 5.56558728e-01 -9.37905252e-01 3.79209034e-02
3.04259360e-01 4.49866027e-01 6.27571821e-01 4.03391451e-01
6.08136415e-01 -2.51152664e-01 -4.33060259e-01 -4.17790413e-02
1.52141571e+00 6.81362301e-02 5.56364536e-01 -2.50531472e-02
5.43502152e-01 1.16854954e+00 1.95194006e-01 4.67785209e-01
8.07338208e-02 6.40435040e-01 5.48385441e-01 -1.54268220e-01
8.17213282e-02 -9.33211148e-02 3.13552916e-01 9.50614631e-01
-5.58533192e-01 -2.92565107e-01 -9.65660393e-01 3.96451890e-01
-1.81119454e+00 -1.44022667e+00 -1.13097799e+00 2.37990880e+00
6.60606384e-01 2.56848216e-01 2.22139791e-01 5.94777167e-01
7.89885938e-01 1.06256247e-01 -1.14656910e-02 -1.72548443e-01
-3.71314555e-01 5.77247553e-02 4.92803603e-01 6.50130153e-01
-8.63642752e-01 1.24125794e-01 6.13336277e+00 5.25802970e-01
-1.17105258e+00 -9.48286578e-02 5.37730277e-01 2.76093096e-01
-4.21604604e-01 3.28292161e-01 -1.62499771e-01 6.90137088e-01
1.00345850e+00 -3.16435546e-01 3.54384810e-01 9.50726047e-02
8.79175723e-01 -4.64005470e-01 -9.87452269e-01 6.30934238e-01
-3.64350408e-01 -1.08705986e+00 -2.32992128e-01 3.29526573e-01
5.54934382e-01 -4.11276609e-01 -1.23180874e-01 -3.70582521e-01
3.70322406e-01 -1.00641418e+00 4.91123825e-01 7.19857097e-01
3.79159808e-01 -6.13528669e-01 6.98381186e-01 2.28593171e-01
-1.00122690e+00 -2.19288561e-02 -1.16892993e-01 -6.83351874e-01
4.35534716e-01 1.54700863e+00 -5.56532741e-01 9.65474367e-01
4.72257346e-01 8.66948962e-01 -3.50034684e-01 1.00451255e+00
-4.06779140e-01 1.03958821e+00 -2.67032444e-01 -2.31726199e-01
-7.86106810e-02 -5.32032371e-01 9.00207520e-01 1.15740693e+00
3.57743263e-01 -1.19146109e-01 -4.34559435e-01 1.14845562e+00
3.76032859e-01 -3.41608152e-02 -9.34627950e-01 -1.20269932e-01
1.64661452e-01 1.21204758e+00 -1.12908101e+00 -3.56832407e-02
-2.22368509e-01 5.68026662e-01 -1.03493921e-01 4.32463050e-01
-7.66487002e-01 -8.44337717e-02 4.54419196e-01 2.25393653e-01
-1.36079431e-01 -4.77185011e-01 -7.34600902e-01 -9.51848269e-01
2.90739182e-02 -4.52746660e-01 2.63567179e-01 -4.43003982e-01
-1.39825726e+00 1.72140628e-01 1.11384697e-01 -1.19346571e+00
-2.81809300e-01 -4.74143684e-01 -7.09382474e-01 9.07525539e-01
-1.38767898e+00 -5.39568663e-01 -1.20454162e-01 4.20347095e-01
-3.58516760e-02 3.53699088e-01 6.87452197e-01 3.92823488e-01
-6.48694873e-01 -1.67798072e-01 1.91831812e-01 -1.31521169e-02
6.46710932e-01 -1.41855693e+00 2.23627210e-01 1.00607359e+00
1.01749927e-01 7.68020809e-01 1.35727358e+00 -8.32338333e-01
-1.27154827e+00 -6.60700381e-01 1.38346827e+00 -5.36216557e-01
1.53312969e+00 -5.66331327e-01 -7.60920525e-01 1.55515015e-01
4.06491011e-01 -2.42613822e-01 3.79507095e-01 1.59640178e-01
8.34328961e-03 -4.91366349e-02 -4.30022508e-01 7.55422115e-01
9.89709198e-01 -1.36307478e-01 -5.70707500e-01 2.63841629e-01
4.17069584e-01 3.65932018e-01 -1.00477326e+00 2.49236286e-01
4.10333455e-01 -1.22701001e+00 7.60320902e-01 -2.21329048e-01
6.78138673e-01 -4.04969186e-01 3.05089682e-01 -1.30453026e+00
1.38149085e-02 -1.04272830e+00 2.15945914e-01 1.33516896e+00
2.85719097e-01 -9.76562142e-01 4.13934648e-01 2.78983563e-01
2.52500415e-01 -3.55043292e-01 -1.11435437e+00 -7.78357744e-01
1.16699390e-01 -4.95615423e-01 1.25439927e-01 1.33942008e+00
3.35938007e-01 4.37304527e-01 -1.54251605e-01 2.22350016e-01
9.84758019e-01 9.63964537e-02 6.45199478e-01 -1.77154815e+00
-2.93368429e-01 -7.38276541e-01 -3.74385327e-01 -5.70612192e-01
-7.99400657e-02 -6.40039742e-01 -2.68983275e-01 -1.14372194e+00
-1.54929444e-01 -4.75030482e-01 -1.16419792e-01 -4.78839912e-02
-1.66951299e-01 6.97458759e-02 -5.68020120e-02 3.40752274e-01
1.97242737e-01 3.33619386e-01 1.23718226e+00 1.16084099e-01
-2.04567909e-02 2.73984820e-01 -4.44226027e-01 5.73260069e-01
6.91275358e-01 -5.77267230e-01 -3.85074586e-01 3.49928021e-01
8.50993276e-01 4.36682016e-01 1.07426655e+00 -7.00375259e-01
3.81048411e-01 -2.22636256e-02 -5.15732169e-02 -2.14603707e-01
-1.74742848e-01 -9.19205666e-01 5.51290214e-01 5.80220103e-01
-1.54862747e-01 2.52301216e-01 -9.84441936e-02 7.71077216e-01
-3.78965825e-01 -1.12062879e-01 2.64035970e-01 -1.14068404e-01
-3.95370662e-01 -8.18337947e-02 -3.89706910e-01 -1.15334399e-01
9.03944969e-01 -1.90723926e-01 -3.39925170e-01 -4.15309846e-01
-7.91809857e-01 -3.84587161e-02 1.51401550e-01 1.54699743e-01
1.89613506e-01 -1.00421607e+00 -7.52002835e-01 -1.01356916e-01
-2.20793277e-01 -2.92371511e-01 2.89756060e-01 1.43168211e+00
-2.51393467e-01 4.72849488e-01 9.35547054e-02 -5.77158749e-01
-8.22927952e-01 7.80767977e-01 2.00355351e-01 -2.92067081e-01
-4.97502297e-01 2.32581317e-01 1.49656922e-01 1.75026909e-01
-1.49416834e-01 -3.80083084e-01 -3.38996947e-01 5.05932689e-01
3.88265103e-01 6.14734709e-01 -8.07442963e-02 -2.76488483e-01
-1.26834586e-01 5.21662176e-01 7.84140646e-01 -1.51009023e-01
1.35828125e+00 -2.31506929e-01 -6.76336348e-01 1.05590892e+00
8.39807808e-01 3.71803403e-01 -1.11228001e+00 1.80303916e-01
4.20446962e-01 -1.75264120e-01 -2.98284948e-01 -5.94849527e-01
-7.35122085e-01 8.76401961e-01 3.49736139e-02 1.17655313e+00
9.73659873e-01 -8.04010779e-02 -6.85092583e-02 -8.06079283e-02
3.93042833e-01 -5.68910122e-01 -1.19358085e-01 3.34634632e-01
9.50362861e-01 -9.19985414e-01 -1.04668140e-02 -7.25271225e-01
4.80556488e-02 1.32218516e+00 -2.89227635e-01 -3.28564912e-01
7.40838647e-01 5.37325382e-01 -3.16491395e-01 -3.33367735e-01
-6.40862584e-01 -1.22435316e-01 -1.54271290e-01 4.12014961e-01
5.33330202e-01 1.06981456e-01 -7.73455679e-01 1.49563223e-01
-2.90308654e-01 -4.81192507e-02 8.88229668e-01 5.15680313e-01
-1.65274560e-01 -1.01997197e+00 -5.23523867e-01 1.06092244e-01
-4.78421658e-01 -1.16317675e-01 -4.08069283e-01 1.30339170e+00
-8.57762173e-02 1.23653305e+00 3.61499712e-02 -1.51066393e-01
3.01286608e-01 -1.40167311e-01 1.91826299e-01 -2.17458773e-02
-5.80265403e-01 -6.02870770e-02 1.11701824e-02 -4.12675023e-01
-8.70643735e-01 -8.43688369e-01 -9.17677462e-01 -6.07646823e-01
-1.73751667e-01 3.82935256e-01 6.17758453e-01 1.03187478e+00
-1.23738676e-01 8.40693593e-01 7.18735635e-01 -6.68731391e-01
-2.41710082e-01 -6.58345163e-01 -8.05123687e-01 2.31896371e-01
4.74358857e-01 -7.12957740e-01 -9.06680703e-01 1.97326779e-01] | [7.627429008483887, 5.177863597869873] |
0009ab29-983b-4088-a5c8-f9ee02f271bd | prediction-assistant-frame-super-resolution | 2103.09455 | null | https://arxiv.org/abs/2103.09455v1 | https://arxiv.org/pdf/2103.09455v1.pdf | Prediction-assistant Frame Super-Resolution for Video Streaming | Video frame transmission delay is critical in real-time applications such as online video gaming, live show, etc. The receiving deadline of a new frame must catch up with the frame rendering time. Otherwise, the system will buffer a while, and the user will encounter a frozen screen, resulting in unsatisfactory user experiences. An effective approach is to transmit frames in lower-quality under poor bandwidth conditions, such as using scalable video coding. In this paper, we propose to enhance video quality using lossy frames in two situations. First, when current frames are too late to receive before rendering deadline (i.e., lost), we propose to use previously received high-resolution images to predict the future frames. Second, when the quality of the currently received frames is low~(i.e., lossy), we propose to use previously received high-resolution frames to enhance the low-quality current ones. For the first case, we propose a small yet effective video frame prediction network. For the second case, we improve the video prediction network to a video enhancement network to associate current frames as well as previous frames to restore high-quality images. Extensive experimental results demonstrate that our method performs favorably against state-of-the-art algorithms in the lossy video streaming environment. | ['Zhiyong Gao', 'Jerry W Hu', 'Charlie L Wang', 'Guangtao Zhai', 'Wenbo Bao', 'Wang Shen'] | 2021-03-17 | null | null | null | null | ['video-enhancement'] | ['computer-vision'] | [ 5.54076433e-01 -1.94622323e-01 -1.29722223e-01 -2.03765437e-01
-2.21385449e-01 4.02386747e-02 -2.79537946e-01 -4.08122689e-02
-5.27556181e-01 9.24738467e-01 -1.41290829e-01 -1.89476103e-01
1.25441596e-01 -9.22375798e-01 -5.51574469e-01 -6.43011808e-01
-2.98904598e-01 -2.26985335e-01 8.44676197e-01 -1.47682086e-01
3.11660856e-01 4.88023460e-01 -1.73276532e+00 4.67807800e-01
7.47018158e-01 1.18778765e+00 5.71038485e-01 7.57799983e-01
-2.35295758e-01 1.04406297e+00 -6.50140166e-01 -2.15860933e-01
2.07848504e-01 -4.05299127e-01 -5.83178878e-01 3.78585339e-01
1.59354001e-01 -1.30561626e+00 -5.29159367e-01 9.98811007e-01
2.51797855e-01 1.62276372e-01 -2.28269562e-01 -1.13611770e+00
1.86992630e-01 1.83638051e-01 -7.10324764e-01 5.73498845e-01
5.69078922e-01 -2.00787047e-03 3.82193714e-01 -5.77515364e-01
8.49505305e-01 1.18726504e+00 3.46603215e-01 5.72993279e-01
-9.29277241e-01 -8.31782401e-01 1.56737790e-01 5.89134216e-01
-1.17075872e+00 -9.45223272e-01 6.11339211e-01 3.91907692e-02
7.11626589e-01 3.11942965e-01 6.81217611e-01 2.62862474e-01
4.10060078e-01 5.52373469e-01 4.21065301e-01 -3.31994802e-01
3.06683421e-01 -2.36594111e-01 -3.91755432e-01 3.18672925e-01
-7.54639134e-02 8.28629211e-02 -6.53744638e-01 -2.71824952e-02
8.54261279e-01 2.16336563e-01 -6.13443434e-01 2.71908134e-01
-1.07266045e+00 3.17709625e-01 3.51469852e-02 3.65454674e-01
-6.26999319e-01 1.47284836e-01 4.50960308e-01 5.81249833e-01
7.23293662e-01 -2.42619678e-01 -8.88603926e-02 -3.40921015e-01
-1.39110291e+00 1.98526055e-01 5.12420774e-01 9.00578141e-01
7.74978340e-01 -2.89372192e-03 8.70650783e-02 8.51405919e-01
1.95657194e-01 3.61177474e-01 1.92115888e-01 -1.47242248e+00
7.17109203e-01 -7.86768124e-02 3.72422844e-01 -1.17764580e+00
3.96210141e-02 2.58956432e-01 -1.05019510e+00 5.94167829e-01
1.82556733e-01 -2.70244610e-02 -4.14161235e-01 1.38403857e+00
1.58356786e-01 5.79740465e-01 -5.03997058e-02 1.08734417e+00
4.77278650e-01 1.32352531e+00 -1.15078680e-01 -1.20159960e+00
8.62932324e-01 -6.38259590e-01 -1.03037310e+00 1.39942810e-01
1.95639804e-01 -1.17837608e+00 5.26330948e-01 6.91170394e-01
-1.84339535e+00 -6.98094189e-01 -1.04764199e+00 3.65561604e-01
5.06101429e-01 -2.94028699e-01 9.13940296e-02 3.33049595e-01
-1.25851619e+00 7.99879491e-01 -8.03876162e-01 -1.11748062e-01
1.35059834e-01 2.82903969e-01 -1.67520016e-01 -5.32029212e-01
-1.03999114e+00 2.66047329e-01 4.29231316e-01 -4.59597148e-02
-7.89036334e-01 -5.70072353e-01 -4.83152419e-01 2.24575877e-01
4.89335388e-01 -3.98896247e-01 1.02460718e+00 -1.38696182e+00
-1.29036176e+00 1.35032073e-01 -4.46410030e-01 -1.68369383e-01
6.03648365e-01 -4.47270498e-02 -5.24474978e-01 6.97805822e-01
-2.43760958e-01 5.32859027e-01 1.04550600e+00 -1.22163391e+00
-9.67915177e-01 -2.04115566e-02 3.70515376e-01 4.28110033e-01
-4.79636878e-01 2.16102839e-01 -6.35876298e-01 -3.01736176e-01
3.25719476e-01 -5.22503376e-01 -1.61465909e-02 7.01714337e-01
8.37104395e-02 2.23257542e-01 1.25080967e+00 -5.27686179e-01
1.62516665e+00 -2.45275927e+00 -1.75286680e-01 1.33814961e-01
3.13879102e-01 5.20077646e-01 2.26447964e-03 3.95741463e-02
4.09668032e-03 2.47998938e-01 1.53377101e-01 -1.36610091e-01
-8.16568196e-01 3.83341908e-01 -1.22575901e-01 2.50158757e-01
-1.38256595e-01 -1.77505553e-01 -9.49344575e-01 -7.53566206e-01
3.31447572e-01 4.97732937e-01 -6.63518190e-01 5.50848424e-01
1.30244434e-01 2.63852090e-01 -4.48310941e-01 4.56990838e-01
1.07397771e+00 7.27481395e-02 4.32275496e-02 -3.13484967e-01
-4.32532370e-01 1.34868816e-01 -1.38818800e+00 1.26241469e+00
-4.20017898e-01 7.18612850e-01 4.46804702e-01 -6.46588206e-01
5.28938711e-01 5.95022023e-01 7.36706257e-01 -7.00197875e-01
-1.53738707e-01 2.72507638e-01 -2.79611826e-01 -7.12574005e-01
7.51716912e-01 -2.78672636e-01 7.20279098e-01 2.04843879e-01
-5.54392159e-01 1.72324419e-01 3.61155003e-01 2.44339317e-01
1.23103702e+00 1.04491264e-01 3.70024219e-02 3.93770665e-01
6.49360657e-01 -5.16277730e-01 8.41861129e-01 3.78045201e-01
-4.26792175e-01 6.38743162e-01 4.00004745e-01 -4.40664142e-01
-1.28655159e+00 -9.59254563e-01 2.40023330e-01 8.16889226e-01
6.07253969e-01 -4.49223489e-01 -5.91529071e-01 3.75151727e-03
-5.79914093e-01 3.57776582e-01 2.63077736e-01 -1.75812602e-01
-8.51729810e-01 -2.21276522e-01 1.24150433e-01 -1.66945551e-02
7.64271319e-01 -1.25930190e+00 -8.46352637e-01 7.95527637e-01
-6.49153292e-01 -1.13058782e+00 -3.87971431e-01 -4.18089002e-01
-1.23199248e+00 -9.08664107e-01 -8.14261019e-01 -6.50755227e-01
3.90251249e-01 8.89503896e-01 9.92716968e-01 7.96913385e-01
1.56507380e-02 1.64096651e-03 -6.13027036e-01 2.45554894e-01
-5.45390964e-01 -4.93513554e-01 -1.90791357e-02 -4.29376923e-02
-2.27079481e-01 -5.81210077e-01 -8.56163621e-01 3.60467166e-01
-1.39137530e+00 2.71923661e-01 6.92129806e-02 5.99892199e-01
8.34630847e-01 7.65559554e-01 4.25276697e-01 -2.95602888e-01
4.09948558e-01 -3.36113065e-01 -4.33832735e-01 5.80222309e-02
-3.32424998e-01 -4.81253475e-01 9.66008246e-01 -3.30960155e-01
-1.19951069e+00 -3.96167725e-01 -4.60278034e-01 -5.90770483e-01
1.23407375e-02 9.90278423e-02 -1.65345624e-01 -5.13405912e-02
-4.46375348e-02 1.43716425e-01 3.16822380e-02 -1.89626068e-01
-3.98013592e-01 6.52487636e-01 3.31446320e-01 -7.11441487e-02
4.94824290e-01 5.01521587e-01 -1.16526879e-01 -9.27844286e-01
-7.53847063e-02 -4.92746979e-01 -1.16300195e-01 -7.24278927e-01
6.47350550e-01 -8.91344666e-01 -7.75909245e-01 5.72994232e-01
-1.34455276e+00 -1.24803625e-01 5.10572456e-02 6.50964677e-01
-5.95364988e-01 8.30358386e-01 -9.18796480e-01 -8.81305456e-01
-1.86656758e-01 -1.38692510e+00 7.45853484e-01 3.12251985e-01
2.35942155e-01 -7.41734207e-01 -3.69406432e-01 -2.53969710e-02
4.19651061e-01 -1.12103969e-01 6.03261530e-01 2.52023250e-01
-9.78775442e-01 1.55722991e-01 -4.67310190e-01 4.31934863e-01
1.40356302e-01 2.95836687e-01 -4.02684182e-01 -5.40789306e-01
3.75620052e-02 3.81609909e-02 4.56514865e-01 4.04759258e-01
1.45623469e+00 -3.01800668e-01 -2.06330106e-01 7.71619320e-01
1.63133180e+00 7.03934252e-01 1.30914617e+00 2.18070522e-01
2.98280209e-01 3.09696198e-01 1.17076278e+00 9.73127782e-01
1.06059767e-01 6.31717741e-01 6.59042120e-01 -3.37792411e-02
-1.23738989e-01 1.74462855e-01 6.09312713e-01 1.03627110e+00
-1.51344687e-01 -7.86993206e-01 -2.09636390e-01 3.24735403e-01
-1.70309222e+00 -1.38180554e+00 -1.35140523e-01 2.37683010e+00
7.73575723e-01 3.80963862e-01 -2.75146842e-01 4.41531956e-01
9.70031738e-01 3.06560099e-01 -4.30347264e-01 -4.52934176e-01
1.74450744e-02 -6.48683384e-02 2.58419007e-01 4.70902413e-01
-4.78953332e-01 5.87256551e-01 5.38923025e+00 1.03047061e+00
-1.04062676e+00 -1.77386720e-02 7.12709844e-01 -3.72472316e-01
-4.17066365e-01 7.49280527e-02 -5.80529690e-01 8.28202665e-01
9.14118171e-01 -1.47912309e-01 6.58110499e-01 3.17883074e-01
9.07104015e-01 -6.76277041e-01 -6.44168079e-01 1.17588198e+00
1.07478008e-01 -1.28239918e+00 -5.97103201e-02 -9.35759693e-02
2.53354222e-01 -4.45463479e-01 -2.84716129e-01 -9.38111395e-02
-5.99210620e-01 -3.79915148e-01 6.26032233e-01 4.71692652e-01
1.04496586e+00 -8.47716570e-01 7.47184098e-01 4.78789777e-01
-1.25168979e+00 1.39529958e-01 -7.49499619e-01 -1.27296001e-01
7.35212326e-01 9.23531950e-01 -1.44597605e-01 3.94537270e-01
9.51646924e-01 7.08293676e-01 5.14571182e-02 1.38237405e+00
3.17200333e-01 4.14902538e-01 -3.57140213e-01 5.55210531e-01
-3.39606106e-02 -7.18686283e-02 6.84980750e-01 7.93999255e-01
7.74167120e-01 6.63501382e-01 1.57011807e-01 2.99856275e-01
-9.17856321e-02 1.06738977e-01 -4.31896925e-01 5.82709789e-01
3.82530600e-01 8.90612483e-01 -5.87641239e-01 -7.47150362e-01
-5.57698369e-01 1.16392791e+00 -2.22558662e-01 3.69856209e-01
-8.28729868e-01 -4.49496657e-01 6.46326125e-01 3.95537049e-01
1.23535827e-01 -1.52882174e-01 1.65070325e-01 -1.25844181e+00
2.66727060e-01 -5.82179427e-01 7.58329853e-02 -1.09977603e+00
-7.36402452e-01 6.93479121e-01 -1.46256492e-01 -1.60254955e+00
-2.77083248e-01 1.39037237e-01 -6.21645808e-01 6.50508046e-01
-1.91653144e+00 -1.26930743e-01 -5.30679166e-01 1.02181363e+00
9.07639265e-01 7.76014999e-02 2.87394553e-01 8.75874460e-01
-2.40459487e-01 3.21791619e-01 2.04124913e-01 -4.51722175e-01
9.65309680e-01 -3.33648860e-01 -2.73985624e-01 1.00086796e+00
-5.25464773e-01 -6.50225282e-02 8.08474123e-01 -7.25788891e-01
-1.08826101e+00 -7.26877093e-01 8.64075899e-01 7.48119712e-01
1.15552813e-01 1.93578526e-01 -1.26608419e+00 2.72877831e-02
6.23024963e-02 2.13640824e-01 5.46455681e-01 -7.42156625e-01
2.52580315e-01 -4.71938163e-01 -1.48873127e+00 4.57136661e-01
7.27436423e-01 -2.13013306e-01 -6.17120601e-02 9.65076536e-02
7.37951279e-01 -3.38243693e-01 -8.08846056e-01 2.24238575e-01
5.97416222e-01 -1.57045329e+00 8.51278007e-01 1.00549951e-01
6.67143822e-01 -3.94225389e-01 -1.97888657e-01 -1.02224183e+00
-1.39251471e-01 -6.05200291e-01 1.81609988e-02 1.14814281e+00
-1.30662382e-01 -1.04413129e-01 7.88778782e-01 5.23484707e-01
-4.00222950e-02 -4.91013348e-01 -1.17730296e+00 -5.30795991e-01
-6.85766280e-01 -3.87752593e-01 4.26504254e-01 6.19165182e-01
1.38955072e-01 -3.01815033e-01 -7.25362062e-01 1.71094581e-01
6.35557413e-01 -1.01945899e-01 4.42548901e-01 -8.60292971e-01
-2.12927431e-01 3.18981484e-02 -4.08598214e-01 -1.41533768e+00
6.73846602e-02 -1.13381766e-01 2.14933902e-01 -1.29873252e+00
2.42422640e-01 -6.19602501e-01 -2.95338392e-01 2.66106566e-03
-1.93380013e-01 3.28085452e-01 6.80752516e-01 4.20522749e-01
-8.26760530e-01 2.17155993e-01 1.29502666e+00 1.77482009e-01
-3.37311745e-01 1.99870080e-01 8.21684599e-02 7.25806594e-01
7.10337818e-01 -3.81281912e-01 -5.51619828e-01 -4.62275058e-01
1.67808324e-01 1.20029807e+00 1.46258771e-01 -1.19606876e+00
1.11441992e-01 -3.86482716e-01 2.52612203e-01 -6.25916779e-01
4.83250946e-01 -1.24523616e+00 2.01926276e-01 6.61374450e-01
-1.78209677e-01 1.99222192e-01 1.94624159e-02 4.74192411e-01
-4.56737638e-01 -4.34684753e-01 8.88979971e-01 -1.09205864e-01
-9.61809874e-01 5.20946026e-01 -7.44638383e-01 -1.99551523e-01
9.65721250e-01 -6.31005406e-01 -1.19736008e-01 -8.21725309e-01
-7.17140198e-01 2.05293655e-01 6.61191761e-01 2.35895783e-01
1.37053525e+00 -1.14647949e+00 -5.48715055e-01 2.15376616e-01
-2.61817783e-01 -2.49001920e-01 6.50552332e-01 5.89592516e-01
-9.28356111e-01 -2.47866929e-01 -4.45032835e-01 -5.67533612e-01
-1.75995922e+00 6.24855340e-01 -1.76533256e-02 -1.98480025e-01
-6.60448790e-01 4.51166332e-01 1.07905045e-01 8.20098221e-01
1.99220121e-01 -8.69323984e-02 -1.20638989e-01 -1.78224176e-01
1.09390056e+00 5.10633290e-01 -8.65056291e-02 -5.74675739e-01
-8.43554810e-02 2.77111709e-01 -6.63518831e-02 -6.44780919e-02
1.20128453e+00 -9.29901898e-01 -2.93683201e-01 1.28418520e-01
1.26233685e+00 -1.39261678e-01 -1.49102926e+00 -9.95106325e-02
-6.12433791e-01 -1.21413970e+00 3.45749944e-01 -1.97535083e-01
-1.43294644e+00 7.12752998e-01 7.94695318e-01 2.40274802e-01
1.81031036e+00 -6.03013098e-01 1.35615134e+00 -5.01156822e-02
7.37147033e-01 -1.06917894e+00 9.95980874e-02 2.44495019e-01
3.70228678e-01 -1.03261530e+00 1.79646060e-01 -6.69736683e-01
-3.32698524e-01 1.53747213e+00 6.07058942e-01 1.38891563e-01
5.61333418e-01 4.87951487e-01 -2.98932374e-01 4.36889708e-01
-1.03631973e+00 1.75490841e-01 -4.24813509e-01 3.93456876e-01
4.46427643e-01 -1.89533278e-01 -4.54954177e-01 -2.14449853e-01
3.40364873e-01 5.05667925e-01 1.20755279e+00 9.24430132e-01
-9.10342991e-01 -1.07573986e+00 -6.13263488e-01 3.74202579e-01
-7.51434922e-01 -2.25074410e-01 5.82210362e-01 2.44048744e-01
1.88893825e-01 1.27247024e+00 5.29309094e-01 -1.97268024e-01
1.78527981e-01 -4.31938320e-01 2.84436196e-01 -1.53797537e-01
-2.82338172e-01 5.41515887e-01 3.22345868e-02 -9.32276428e-01
-7.75330782e-01 -2.51955569e-01 -1.32576096e+00 -1.05761039e+00
-2.00693130e-01 -1.29853442e-01 3.77046019e-01 6.33870661e-01
6.19756766e-02 6.45132959e-01 8.84101272e-01 -9.06325638e-01
6.75537586e-02 -2.54001439e-01 -7.12904871e-01 3.40955198e-01
6.78838491e-01 -2.99797386e-01 -3.38271797e-01 3.50404769e-01] | [11.10410213470459, -1.7639203071594238] |
7527d52a-b5a3-4397-abe3-34e817ebaa28 | on-information-plane-analyses-of-neural | 2003.09671 | null | https://arxiv.org/abs/2003.09671v3 | https://arxiv.org/pdf/2003.09671v3.pdf | On Information Plane Analyses of Neural Network Classifiers -- A Review | We review the current literature concerned with information plane analyses of neural network classifiers. While the underlying information bottleneck theory and the claim that information-theoretic compression is causally linked to generalization are plausible, empirical evidence was found to be both supporting and conflicting. We review this evidence together with a detailed analysis of how the respective information quantities were estimated. Our survey suggests that compression visualized in information planes is not necessarily information-theoretic, but is rather often compatible with geometric compression of the latent representations. This insight gives the information plane a renewed justification. Aside from this, we shed light on the problem of estimating mutual information in deterministic neural networks and its consequences. Specifically, we argue that even in feed-forward neural networks the data processing inequality need not hold for estimates of mutual information. Similarly, while a fitting phase, in which the mutual information between the latent representation and the target increases, is necessary (but not sufficient) for good classification performance, depending on the specifics of mutual information estimation such a fitting phase need not be visible in the information plane. | ['Bernhard C. Geiger'] | 2020-03-21 | null | null | null | null | ['mutual-information-estimation', 'information-plane'] | ['methodology', 'methodology'] | [ 7.18635738e-01 4.70748365e-01 -2.37852767e-01 -5.90146601e-01
8.75700787e-02 -5.58398306e-01 6.33479238e-01 4.76590186e-01
-6.97201431e-01 6.10232413e-01 1.12911411e-01 -6.82582200e-01
-1.02814853e+00 -5.53691983e-01 -5.67182124e-01 -9.41075623e-01
-2.69614935e-01 3.77997786e-01 -7.46433903e-03 9.34778750e-02
6.98841333e-01 4.99173939e-01 -2.00779319e+00 3.58061194e-02
6.99757636e-01 9.00277793e-01 2.11953476e-01 7.32076943e-01
-1.70808583e-01 6.35911167e-01 -5.38471758e-01 -5.82129717e-01
2.64525395e-02 -8.05048585e-01 -9.99715030e-01 -2.94415414e-01
1.37411073e-01 -1.94523528e-01 -3.78681630e-01 1.11166191e+00
1.96645334e-01 -9.17718336e-02 1.10823750e+00 -1.25733352e+00
-3.45366567e-01 9.80268240e-01 -2.39605695e-01 5.31766295e-01
-2.61047661e-01 -1.95604011e-01 1.11504555e+00 -2.80157238e-01
3.53401512e-01 1.04572415e+00 5.06144762e-01 3.89957070e-01
-1.08269632e+00 -5.30361474e-01 1.61647737e-01 5.85997812e-02
-1.06578624e+00 -6.18395329e-01 5.32721519e-01 -4.80318546e-01
7.88155913e-01 3.57646793e-01 9.10373032e-01 7.16425478e-01
4.56194490e-01 5.84784567e-01 7.91038036e-01 -7.35778928e-01
3.27305228e-01 1.76672474e-01 5.35208881e-01 4.27356154e-01
7.83753037e-01 3.97049308e-01 -8.46049011e-01 1.52953327e-01
7.77399302e-01 -3.25709701e-01 -2.08058015e-01 -3.88332129e-01
-8.92027080e-01 8.10784519e-01 2.91301876e-01 5.67693353e-01
-1.87625661e-01 1.64773062e-01 4.37291354e-01 7.20121264e-01
3.65046352e-01 4.33921993e-01 -4.01212901e-01 -1.47089064e-01
-9.86646652e-01 2.18380421e-01 8.78615975e-01 6.45415783e-01
4.11354810e-01 -7.26439208e-02 3.98013800e-01 4.85441118e-01
3.71894598e-01 9.65849236e-02 3.77953082e-01 -1.00883472e+00
3.35809439e-01 2.88650364e-01 -3.96946102e-01 -1.09478521e+00
-6.64128065e-01 -7.34717488e-01 -9.54305232e-01 1.24418899e-01
7.97400296e-01 -1.17680900e-01 -2.31445789e-01 2.15216899e+00
-2.98378587e-01 -6.26716793e-01 1.45380497e-01 6.10759199e-01
1.94525987e-01 1.40308842e-01 -2.42554136e-02 -7.02038825e-01
1.01776779e+00 -1.69342980e-01 -8.68202507e-01 -4.23073471e-01
8.11506093e-01 -4.38571185e-01 4.50562537e-01 3.33697140e-01
-1.24664855e+00 -7.02434853e-02 -1.37482190e+00 1.48683161e-01
-2.84084290e-01 -4.87197697e-01 9.36457098e-01 8.86782169e-01
-9.62152123e-01 9.02322948e-01 -7.83504367e-01 -3.68399382e-01
3.73464763e-01 3.08380276e-01 -3.78149360e-01 1.78714991e-01
-1.10674715e+00 1.13610184e+00 8.48317742e-01 1.51424155e-01
-3.18941206e-01 -3.46155256e-01 -4.59179640e-01 2.12505624e-01
5.46958931e-02 -7.61492670e-01 9.24104095e-01 -1.10309637e+00
-1.07001841e+00 6.13053143e-01 -2.86875311e-02 -5.99783659e-01
4.22829717e-01 -1.23333402e-01 7.85236806e-02 1.85349166e-01
-5.20952404e-01 6.83405697e-01 5.13308048e-01 -1.02606618e+00
-3.54930788e-01 -7.30767131e-01 -1.76226228e-01 4.47914809e-01
-2.44862616e-01 1.34007046e-02 2.40893736e-01 -2.89385021e-01
7.42641747e-01 -5.54473281e-01 6.88555762e-02 2.53506526e-02
-1.77430287e-01 -2.36310810e-01 2.67929167e-01 -2.21311495e-01
1.16636074e+00 -2.05816889e+00 1.48640936e-02 2.76615500e-01
5.35710514e-01 -3.12559843e-01 4.20344844e-02 4.56294984e-01
-5.19852936e-01 3.02468210e-01 -3.73381257e-01 2.07039207e-01
-2.16288175e-02 1.38169587e-01 -1.73527196e-01 8.61065388e-01
1.37923166e-01 5.76488078e-01 -5.00248134e-01 -2.98115104e-01
-3.91919240e-02 4.80629951e-01 -7.26921201e-01 -2.80538619e-01
1.69956952e-01 -2.39470266e-02 -1.75981283e-01 2.99384147e-02
5.86527288e-01 -1.78877413e-01 3.78905445e-01 1.78028122e-01
-4.13907230e-01 5.19168317e-01 -9.07769144e-01 1.44271159e+00
1.52791709e-01 1.24685276e+00 -1.17756560e-01 -1.42981637e+00
7.17159033e-01 2.54852474e-01 4.25529599e-01 -5.95429361e-01
4.07293558e-01 1.44620329e-01 7.62568772e-01 -2.97208995e-01
3.19282711e-01 -2.88976163e-01 2.71247029e-01 7.67275274e-01
3.36763501e-01 2.19817236e-02 1.69931993e-01 2.32887089e-01
1.01213026e+00 -1.04672320e-01 3.38811725e-01 -6.60810709e-01
1.62572414e-02 -3.43968242e-01 2.99055487e-01 9.16902661e-01
-1.99930668e-01 3.15059364e-01 9.34429944e-01 -1.43973470e-01
-1.26024961e+00 -1.01223326e+00 -6.06449783e-01 8.60779822e-01
-2.28537116e-02 -4.47430223e-01 -1.05813444e+00 -1.65646181e-01
-2.78441757e-01 8.41689587e-01 -6.80402160e-01 -4.32854950e-01
-6.19086735e-02 -9.98815298e-01 5.15896261e-01 1.22950181e-01
4.28966075e-01 -7.60653615e-01 -1.21939564e+00 -1.48526862e-01
-1.76756606e-01 -5.00271320e-01 3.07352364e-01 8.64103496e-01
-1.32478309e+00 -1.13274145e+00 -4.75172877e-01 -3.28456134e-01
6.17894709e-01 2.01767743e-01 8.76346409e-01 6.11112528e-02
-6.62631020e-02 3.11005861e-01 1.83723811e-02 -5.58036029e-01
-4.41696644e-01 1.47103816e-01 1.12476893e-01 -6.88589394e-01
5.71630061e-01 -8.49117577e-01 -4.85017955e-01 2.72967249e-01
-1.07302523e+00 2.60008931e-01 7.34800100e-01 6.95095479e-01
-1.81514502e-01 3.42548341e-01 5.82485974e-01 -5.65297425e-01
8.37423623e-01 -3.24392378e-01 -4.77007955e-01 2.06503376e-01
-9.52939272e-01 4.43834215e-01 -6.36510029e-02 -1.63572937e-01
-9.36267316e-01 -2.71200508e-01 4.02704813e-02 8.76105800e-02
-2.28457198e-01 5.89685559e-01 1.22724399e-01 1.26962379e-01
8.86365831e-01 1.19142666e-01 3.41442317e-01 -2.24913821e-01
3.10870409e-01 3.78697485e-01 2.05018982e-01 -2.60399729e-01
3.38256627e-01 7.57151991e-02 1.76741362e-01 -8.86441588e-01
-6.09773874e-01 -1.56650051e-01 -9.61344659e-01 -4.44091111e-01
6.37481987e-01 -3.45234126e-01 -9.89643097e-01 2.42301792e-01
-1.23763990e+00 3.95943038e-03 -3.32356572e-01 8.71739447e-01
-1.00555444e+00 1.02982454e-01 -5.32088280e-01 -1.24476254e+00
1.00505576e-01 -1.04667842e+00 2.82269474e-02 -2.21341439e-02
-3.26024413e-01 -1.03685522e+00 -1.26597479e-01 -6.86309487e-02
4.66076016e-01 -1.21200129e-01 1.16919041e+00 -9.98320520e-01
-2.91653723e-01 -2.03496650e-01 -3.83965909e-01 4.69415598e-02
3.94858699e-03 1.85380094e-02 -1.12441587e+00 6.93744794e-02
5.03502131e-01 -2.02392980e-01 1.03307009e+00 7.11846113e-01
8.18050206e-01 -5.73547423e-01 -2.44041473e-01 2.57459700e-01
1.33763111e+00 2.18256041e-01 5.42963445e-01 2.82961398e-01
1.09558456e-01 1.28913760e+00 -5.19178137e-02 2.59154171e-01
-5.00921765e-03 1.80133000e-01 3.51118773e-01 2.93670118e-01
1.70001462e-01 -2.47362275e-02 3.57561037e-02 1.11280870e+00
-4.96979877e-02 -2.55517423e-01 -8.89700055e-01 1.28806859e-01
-1.53433442e+00 -1.13080394e+00 -8.34130645e-02 2.39066958e+00
9.05044913e-01 4.91150975e-01 1.49722263e-01 6.84654474e-01
6.13970578e-01 2.46058013e-02 -5.33279300e-01 -4.55187798e-01
-1.90656394e-01 -5.31077325e-01 6.60188675e-01 6.31515443e-01
-5.57716370e-01 3.57979864e-01 7.86262798e+00 5.88120103e-01
-6.73026741e-01 -9.79367867e-02 7.10170150e-01 -1.95863575e-01
-4.20290321e-01 -2.89220363e-02 -4.46900636e-01 1.51209027e-01
1.33648729e+00 -3.99091959e-01 3.00738037e-01 4.60113317e-01
-7.90852755e-02 -5.36090612e-01 -1.23893642e+00 7.48647273e-01
-1.65774569e-01 -1.01832366e+00 -8.32559690e-02 3.28259379e-01
2.00995281e-01 -3.51653844e-02 2.10376039e-01 -2.02363357e-01
1.95094064e-01 -8.97329032e-01 7.44588256e-01 5.13050199e-01
5.53957701e-01 -9.08179700e-01 6.98341310e-01 8.54023814e-01
-5.90597153e-01 2.40298267e-02 -5.41545212e-01 -4.20678586e-01
-1.84572011e-01 7.40491390e-01 -4.19783831e-01 2.63288200e-01
2.77401268e-01 4.65657085e-01 -4.18587714e-01 9.97659981e-01
6.08396567e-02 5.06289601e-01 -4.64153647e-01 -8.13555941e-02
7.13975653e-02 -1.50228351e-01 4.48892832e-01 9.95677352e-01
4.98764701e-02 4.86320071e-02 -8.70646298e-01 9.50039566e-01
4.41722214e-01 3.36214863e-02 -9.76443887e-01 -3.14709604e-01
4.34306145e-01 7.06084073e-01 -1.05655169e+00 4.22290973e-02
-2.95108497e-01 6.17122531e-01 3.29967707e-01 2.54948556e-01
-3.17991495e-01 -3.79779547e-01 4.28355128e-01 -8.98708701e-02
-2.60608554e-01 -2.86146700e-01 -7.25276768e-01 -7.33559489e-01
-8.75734016e-02 -4.48004425e-01 1.66817188e-01 -3.23311210e-01
-8.33543241e-01 2.83139467e-01 4.38636631e-01 -6.61435246e-01
-4.66667861e-01 -6.48438454e-01 -1.92592934e-01 7.93856263e-01
-9.59846377e-01 -2.61564821e-01 1.84407234e-01 1.70156583e-01
2.43790954e-01 -6.01096712e-02 8.58312249e-01 -4.00054269e-02
-5.59186757e-01 5.14380276e-01 2.75886565e-01 -1.33704156e-01
1.96973935e-01 -1.00152969e+00 2.81454951e-01 5.66845059e-01
1.00310802e-01 1.09202528e+00 1.13046682e+00 -4.58661258e-01
-1.10162795e+00 -2.69110680e-01 9.51658130e-01 -2.44991317e-01
4.99307960e-01 -2.40699679e-01 -9.19806182e-01 5.02038658e-01
2.57505745e-01 -6.36082053e-01 9.11875963e-01 5.27207315e-01
-4.82943356e-01 1.77608401e-01 -7.87877679e-01 5.21928191e-01
1.16175914e+00 -4.19607311e-01 -7.05512583e-01 1.27000839e-01
4.60284621e-01 2.00882792e-01 -5.95413744e-01 3.71198744e-01
9.44102705e-01 -1.42061317e+00 6.70265973e-01 -5.76051414e-01
4.61274922e-01 3.50572199e-01 -2.23231852e-01 -9.77323532e-01
-2.97140032e-01 -2.46507183e-01 3.04586232e-01 8.69106829e-01
5.05166411e-01 -6.95776105e-01 7.27957189e-01 8.72940779e-01
7.40778744e-02 -5.64299703e-01 -1.10453606e+00 -7.03508973e-01
4.86724526e-01 -8.73714626e-01 -3.16782780e-02 8.45326960e-01
6.89015687e-01 5.30639529e-01 1.20523892e-01 -3.70036960e-01
7.64763832e-01 -5.46098232e-01 1.16214715e-01 -1.66729128e+00
-1.70166120e-02 -1.11870027e+00 -5.71730137e-01 -8.92480373e-01
5.08829299e-03 -8.97928059e-01 7.92868584e-02 -1.36450279e+00
3.40357602e-01 -3.99198979e-01 -3.92730474e-01 1.81001201e-02
3.40848774e-01 -3.41466099e-01 1.20418511e-01 4.12026078e-01
-1.51571840e-01 2.63301611e-01 8.40836823e-01 2.82612592e-01
-1.93017609e-02 1.89052895e-01 -7.84093499e-01 8.82538855e-01
9.94528949e-01 -5.91546178e-01 -8.39755237e-01 -3.58881354e-01
8.83123457e-01 1.52409762e-01 2.44128004e-01 -1.16005433e+00
5.06326020e-01 5.76190129e-02 4.34145272e-01 -6.50864303e-01
3.08251381e-01 -8.00118625e-01 -2.94073112e-02 7.05381215e-01
-1.12794209e+00 9.05440822e-02 2.20459253e-01 6.30612910e-01
1.02630503e-01 -7.11408615e-01 7.34346628e-01 -1.29833385e-01
-2.36074030e-02 -7.77997375e-02 -6.43636286e-01 -4.18704711e-02
6.10728323e-01 -2.68876672e-01 -5.11200368e-01 -4.82480228e-01
-5.35047531e-01 -2.61352867e-01 2.17999756e-01 1.31580174e-01
6.25400066e-01 -8.91300142e-01 -4.66085196e-01 3.22641104e-01
-7.70448595e-02 -4.49296862e-01 1.46972537e-01 9.33263183e-01
-2.91803837e-01 9.05401886e-01 -2.82118708e-01 -5.09945095e-01
-9.88640547e-01 6.21101081e-01 3.30823898e-01 2.40146399e-01
-4.95667905e-01 9.69342113e-01 3.94018233e-01 1.03153773e-01
5.74590623e-01 -2.24076644e-01 -1.88344583e-01 3.50222468e-01
5.88439405e-01 2.07012296e-01 -6.04194216e-02 -2.22600415e-01
-3.08028311e-01 3.11419994e-01 -2.09175795e-01 -6.21968806e-01
1.09271705e+00 -2.64204472e-01 -1.92964360e-01 1.02087843e+00
1.12154138e+00 -5.09407699e-01 -1.13037241e+00 -1.97038770e-01
2.59796619e-01 -1.80293217e-01 1.89080372e-01 -5.11176646e-01
-7.28627443e-01 1.22200787e+00 4.88143235e-01 6.34816945e-01
1.00362551e+00 -1.10927328e-01 -3.62111807e-01 5.11354268e-01
9.49869677e-02 -1.08171451e+00 -1.20010920e-01 4.56370085e-01
8.32495511e-01 -8.85097980e-01 7.21922740e-02 -2.12408379e-01
-3.05247217e-01 1.27512980e+00 1.46314532e-01 1.46537706e-01
8.91951859e-01 3.58433366e-01 -4.93712723e-01 -2.52604008e-01
-1.14881766e+00 -2.99734287e-02 5.25176115e-02 5.34508944e-01
8.08204234e-01 -2.09885329e-01 -4.71451879e-01 1.85208946e-01
-7.04180956e-01 -1.52816966e-01 1.30041450e-01 7.54633009e-01
-8.73711050e-01 -5.80535471e-01 -1.45397335e-01 5.82765937e-01
-4.69452351e-01 -2.68246233e-01 -3.51326972e-01 8.48136425e-01
-2.16933206e-01 9.74058688e-01 5.60472250e-01 -3.46518636e-01
-3.01154226e-01 2.77779281e-01 6.96630657e-01 -8.46747085e-02
-1.76903978e-01 -6.33126423e-02 -6.29017428e-02 -3.00448924e-01
-4.91311312e-01 -7.83988476e-01 -7.91437387e-01 -5.68156898e-01
-6.73047841e-01 2.14748472e-01 1.10987699e+00 1.32027400e+00
-2.43603498e-01 6.39322937e-01 1.55497119e-01 -5.03988922e-01
-4.35824722e-01 -1.02166641e+00 -5.87949872e-01 -3.38059157e-01
4.48119998e-01 -6.21663868e-01 -7.15685368e-01 -1.52049869e-01] | [8.038516998291016, 3.497786521911621] |
87c7acc2-854d-4446-a4de-80ff7ca41d08 | cfr-icl-cascade-forward-refinement-with | 2303.05620 | null | https://arxiv.org/abs/2303.05620v1 | https://arxiv.org/pdf/2303.05620v1.pdf | CFR-ICL: Cascade-Forward Refinement with Iterative Click Loss for Interactive Image Segmentation | The click-based interactive segmentation aims to extract the object of interest from an image with the guidance of user clicks. Recent work has achieved great overall performance by employing the segmentation from the previous output. However, in most state-of-the-art approaches, 1) the inference stage involves inflexible heuristic rules and a separate refinement model; and 2) the training cannot balance the number of user clicks and model performance. To address the challenges, we propose a click-based and mask-guided interactive image segmentation framework containing three novel components: Cascade-Forward Refinement (CFR), Iterative Click Loss (ICL), and SUEM image augmentation. The proposed ICL allows model training to improve segmentation and reduce user interactions simultaneously. The CFR offers a unified inference framework to generate segmentation results in a coarse-to-fine manner. The proposed SUEM augmentation is a comprehensive way to create large and diverse training sets for interactive image segmentation. Extensive experiments demonstrate the state-of-the-art performance of the proposed approach on five public datasets. Remarkably, our model achieves an average of 2.9 and 7.5 clicks of NoC@95 on the Berkeley and DAVIS sets, respectively, improving by 33.2% and 15.5% over the previous state-of-the-art results. The code and trained model are available at https://github.com/TitorX/CFR-ICL-Interactive-Segmentation. | ['Luca Capriotti', 'Tiankai Yao', 'Fei Xu', 'Min Xian', 'Shoukun Sun'] | 2023-03-09 | null | null | null | null | ['image-augmentation', 'interactive-segmentation'] | ['computer-vision', 'computer-vision'] | [ 3.48633081e-01 -1.41449690e-01 -2.30590001e-01 -4.89183903e-01
-1.14294720e+00 -4.87491310e-01 2.64439344e-01 -1.10477738e-01
-5.91431260e-01 4.81122941e-01 -3.80145520e-01 -4.41695809e-01
2.16489151e-01 -6.69776261e-01 -9.32169318e-01 -5.35193861e-01
4.14839447e-01 4.14021343e-01 8.28537822e-01 9.80817303e-02
3.64433914e-01 2.91372895e-01 -1.42505121e+00 2.48549521e-01
1.28517330e+00 1.12694657e+00 4.40207303e-01 8.17113817e-01
-1.43620133e-01 2.73962647e-01 -3.45451564e-01 -4.38150942e-01
3.34294498e-01 -1.68762431e-01 -7.62135208e-01 1.40314266e-01
4.22596961e-01 -3.94913644e-01 -1.19740702e-01 9.89931345e-01
4.37950045e-01 -2.15142146e-02 3.96079123e-01 -1.20447087e+00
-5.21723509e-01 4.79367733e-01 -1.08807683e+00 1.70293152e-01
-1.75969838e-03 3.31248015e-01 8.93135607e-01 -9.15960491e-01
4.20751482e-01 1.06628919e+00 4.01284903e-01 4.55449939e-01
-1.39658308e+00 -1.01942873e+00 2.91603595e-01 6.22767210e-02
-1.56820631e+00 -1.61219060e-01 4.10814583e-01 -4.62830186e-01
3.75408471e-01 4.47009176e-01 5.24795771e-01 6.28256917e-01
-2.14181721e-01 1.18993998e+00 1.13549495e+00 -2.25170001e-01
1.01611629e-01 1.45485178e-01 1.73574150e-01 8.68161917e-01
1.27648130e-01 -3.28183062e-02 -2.46291578e-01 8.26038420e-02
1.01958287e+00 8.86150524e-02 -1.82496365e-02 -8.46808031e-02
-8.97042215e-01 7.69350290e-01 6.76530480e-01 -4.93553169e-02
-1.78125128e-01 2.19671324e-01 1.49718508e-01 -3.16857576e-01
4.88629907e-01 3.19690228e-01 -5.35871267e-01 -3.80993239e-03
-1.35966623e+00 3.10359806e-01 6.00565076e-01 1.01153255e+00
9.04480159e-01 -2.51376659e-01 -5.44382751e-01 7.79839277e-01
2.53312588e-01 4.66701567e-01 8.78540128e-02 -1.05021775e+00
5.16685069e-01 5.56071877e-01 1.87137976e-01 -7.25156546e-01
-6.73994571e-02 -6.22087121e-01 -6.45885944e-01 1.50060520e-01
4.33365285e-01 -2.60172576e-01 -1.45029747e+00 1.54193354e+00
5.98952055e-01 4.03553098e-01 -6.61332607e-01 7.71950781e-01
7.68688560e-01 5.73084414e-01 2.17646658e-01 1.13884732e-01
1.31708610e+00 -1.39032221e+00 -6.14145994e-01 -2.55421072e-01
4.44783211e-01 -9.97815430e-01 1.37867129e+00 3.91188264e-01
-1.23690057e+00 -6.83084369e-01 -9.10346091e-01 -6.72725663e-02
-2.38620043e-01 3.15529168e-01 6.29417956e-01 5.36432207e-01
-7.97234118e-01 4.72288460e-01 -9.76122260e-01 6.54032752e-02
8.80511045e-01 4.36545819e-01 2.23677948e-01 -6.19802922e-02
-8.07668984e-01 1.51529640e-01 3.26745600e-01 -6.83063120e-02
-7.75299430e-01 -1.01188481e+00 -4.86978561e-01 8.07233155e-02
9.06381011e-01 -6.20872319e-01 1.47729075e+00 -6.36125147e-01
-1.31743467e+00 7.23906636e-01 -3.61213952e-01 -4.50906992e-01
9.21260595e-01 -6.08674467e-01 3.88613902e-02 1.91739991e-01
3.21875364e-01 9.70007539e-01 7.12740958e-01 -1.44845116e+00
-9.79572356e-01 -2.89743304e-01 3.12857032e-02 7.17069805e-02
-7.79577866e-02 -7.60208145e-02 -1.35139310e+00 -6.72578931e-01
3.30451503e-02 -9.73138452e-01 -5.74412286e-01 1.28386086e-02
-7.15390742e-01 -1.90687537e-01 8.18721950e-01 -5.92129409e-01
1.55136454e+00 -2.10008717e+00 -2.81157434e-01 9.43245962e-02
2.90754348e-01 4.84744579e-01 6.64631231e-03 -2.78389547e-02
2.16897100e-01 5.90452969e-01 -2.96249479e-01 -5.19826651e-01
-7.82382265e-02 -1.48044556e-01 -3.48833762e-02 5.11664078e-02
8.06695223e-02 9.54239309e-01 -5.70519507e-01 -8.76866162e-01
2.56886959e-01 3.30629826e-01 -8.62209797e-01 2.83176482e-01
-4.10869420e-01 5.73120654e-01 -5.54951489e-01 6.98105693e-01
8.90574157e-01 -5.12947619e-01 -3.18179071e-01 -2.43827716e-01
-2.41804853e-01 4.98551056e-02 -1.26308191e+00 1.67726946e+00
-2.24999726e-01 2.53227293e-01 1.48807511e-01 -4.96354580e-01
5.97715259e-01 -3.19176354e-02 2.28350118e-01 -7.02293932e-01
1.57192335e-01 1.20149128e-01 -2.07683802e-01 -2.55364031e-01
4.07839417e-01 4.21473235e-01 -2.84276493e-02 3.69407922e-01
-1.96209058e-01 -3.53670977e-02 6.10377371e-01 4.15407777e-01
8.75239670e-01 2.08510384e-01 -1.59898162e-01 -4.02780920e-02
4.47512925e-01 -7.29325339e-02 7.21430182e-01 1.01547563e+00
-2.48705342e-01 7.16736853e-01 3.73344600e-01 5.89870429e-03
-9.28891003e-01 -1.02735901e+00 -1.11778334e-01 1.13320708e+00
4.87613082e-01 -4.33715314e-01 -1.16645539e+00 -7.35134602e-01
-1.89738974e-01 7.17862964e-01 -4.93565053e-01 3.36443275e-01
-5.22770941e-01 -5.32861054e-01 3.59770060e-01 6.11865640e-01
9.60361719e-01 -9.18693960e-01 -2.83394068e-01 -1.09997317e-02
-2.22213879e-01 -1.31160808e+00 -8.87524605e-01 -2.78362781e-01
-8.58563125e-01 -9.59503591e-01 -5.78999221e-01 -6.39526665e-01
8.87108862e-01 3.51376534e-01 9.38873887e-01 3.55078727e-01
-5.86337984e-01 -3.28752398e-02 -2.61174113e-01 -2.44160533e-01
-3.45792286e-02 3.43868196e-01 -5.26342034e-01 9.08458680e-02
1.12463847e-01 -4.38932955e-01 -1.17811966e+00 5.78581691e-01
-9.30611789e-01 5.41644514e-01 8.17581356e-01 6.00120425e-01
1.09073675e+00 2.21892190e-03 2.71470338e-01 -1.37518191e+00
3.05554599e-01 -2.24188179e-01 -8.89034808e-01 1.17058843e-01
-7.54588366e-01 -4.20285426e-02 4.29654688e-01 -2.93402702e-01
-1.24857676e+00 3.20500731e-01 -2.84077615e-01 -3.85996759e-01
-3.46680135e-01 2.30840713e-01 -6.67760447e-02 -4.61490452e-02
3.77386123e-01 1.61708698e-01 -3.94250184e-01 -6.72981560e-01
5.47369063e-01 6.18640125e-01 6.16783321e-01 -6.11731350e-01
8.69077981e-01 4.73834932e-01 -3.42186838e-01 -4.59411204e-01
-1.13561237e+00 -6.68665230e-01 -6.80631280e-01 -4.96355034e-02
1.06715560e+00 -8.10714126e-01 -6.92898154e-01 4.33821410e-01
-9.28202212e-01 -5.60703516e-01 3.25812139e-02 1.13722891e-01
-3.96177441e-01 2.83596903e-01 -7.35386431e-01 -6.96758628e-01
-5.94441891e-01 -1.49437058e+00 9.87970769e-01 6.77226663e-01
1.07646868e-01 -4.86499846e-01 -4.43396598e-01 9.10641074e-01
3.17464560e-01 1.64933145e-01 5.21669149e-01 -5.67992628e-01
-1.12057281e+00 -2.87561893e-01 -6.28760934e-01 2.90249705e-01
-3.69705707e-02 9.36799198e-02 -8.40695620e-01 -1.13042958e-01
-3.83402914e-01 -1.66857019e-01 9.06900048e-01 4.97072339e-01
1.74373066e+00 -1.66078284e-01 -5.10319710e-01 7.48409927e-01
1.51784635e+00 2.93931633e-01 7.36159205e-01 1.53395042e-01
8.05019319e-01 -8.99238233e-03 9.94122028e-01 4.53835487e-01
2.24891633e-01 6.60477936e-01 3.46052766e-01 -5.22247791e-01
-1.90876916e-01 -3.87291670e-01 -2.64022380e-01 3.92047286e-01
-2.98212208e-02 -4.81798984e-02 -8.21351230e-01 5.86531758e-01
-1.87005532e+00 -6.28040671e-01 -2.60222524e-01 2.21843314e+00
1.14182103e+00 6.67219400e-01 1.59665927e-01 7.09464774e-03
8.94886971e-01 6.63442090e-02 -7.67059028e-01 -6.48719743e-02
5.54917514e-01 3.41429621e-01 6.97160363e-01 6.42745912e-01
-1.24330020e+00 1.27067256e+00 5.05670738e+00 1.31696427e+00
-8.87711167e-01 1.46889463e-01 9.48886931e-01 -8.51277709e-02
3.39306779e-02 7.83783719e-02 -1.27032828e+00 5.94194770e-01
5.06839216e-01 1.69106331e-02 3.90773624e-01 8.86551619e-01
4.98938084e-01 -2.76311904e-01 -7.62407720e-01 7.69812107e-01
-3.04484516e-01 -1.34099054e+00 3.47657390e-02 -6.29002899e-02
8.12489510e-01 -4.98381853e-02 1.98935896e-01 2.36044392e-01
3.16055864e-01 -8.68561327e-01 5.89789808e-01 3.16959530e-01
8.73710811e-01 -7.05874622e-01 4.43770647e-01 3.47818643e-01
-1.29726112e+00 1.59648925e-01 1.30686775e-01 2.14548260e-01
2.66116083e-01 5.69295168e-01 -8.28037083e-01 3.16808909e-01
9.25707221e-01 3.10963035e-01 -7.71111250e-01 1.34777546e+00
-3.81730169e-01 9.37940240e-01 -3.78798664e-01 9.04828236e-02
3.86294633e-01 -1.51676834e-01 3.38618994e-01 1.32842708e+00
-1.46750122e-01 8.80035236e-02 3.96122962e-01 1.16758263e+00
-2.37031117e-01 6.26037195e-02 2.26091161e-01 4.71839169e-03
7.57389784e-01 1.46836245e+00 -1.11732912e+00 -4.52412158e-01
-2.79049933e-01 9.20183539e-01 2.09831193e-01 4.40136909e-01
-1.29663265e+00 -6.33112967e-01 3.51769507e-01 4.06313896e-01
7.15578914e-01 -1.54434294e-01 -6.90022707e-01 -8.38702679e-01
-1.12246983e-02 -7.09456146e-01 2.64167070e-01 -5.71284592e-01
-8.29300761e-01 3.46656710e-01 -7.16707408e-02 -7.85512507e-01
3.06632221e-01 -2.32604221e-01 -6.74290299e-01 8.02462041e-01
-1.57156205e+00 -1.16018951e+00 -6.06467903e-01 3.95275086e-01
1.00577629e+00 3.51166844e-01 3.75159740e-01 6.02119029e-01
-8.70955229e-01 8.55425715e-01 -1.08319372e-01 2.88879603e-01
6.56051040e-01 -1.32497394e+00 5.72614968e-01 9.75775778e-01
-3.52277718e-02 5.42602360e-01 5.92208743e-01 -6.78933144e-01
-8.83429945e-01 -1.22758019e+00 3.92544597e-01 -2.23860860e-01
3.89576972e-01 -3.93140495e-01 -7.50522494e-01 6.46418869e-01
5.98704070e-02 -2.03924365e-02 5.45254827e-01 -4.93499041e-02
-1.54127538e-01 -1.06305420e-01 -1.00702190e+00 9.22219634e-01
9.40920293e-01 -7.77980387e-02 -9.10072103e-02 3.16872090e-01
1.02091634e+00 -6.73506796e-01 -6.42422497e-01 4.20181334e-01
5.36102533e-01 -9.78302181e-01 9.61362481e-01 -2.43907526e-01
5.45506418e-01 -4.42139804e-01 9.61093605e-02 -6.70719624e-01
-3.04612517e-01 -8.50050449e-01 1.63480900e-02 1.34545910e+00
7.59652793e-01 -2.82079816e-01 9.86033916e-01 8.79955769e-01
-1.24614283e-01 -1.24692690e+00 -4.18101311e-01 -4.09365833e-01
-2.15056151e-01 -5.26498914e-01 5.40575802e-01 3.43094945e-01
-7.29785025e-01 3.76555651e-01 -1.08368583e-01 2.54425585e-01
7.81669617e-01 2.64969319e-01 9.72531140e-01 -9.66107070e-01
-4.17634279e-01 -4.05686200e-01 1.73640713e-01 -1.54803550e+00
-3.45750719e-01 -6.20581150e-01 8.33901688e-02 -1.66022575e+00
4.36872214e-01 -7.23534763e-01 -2.29999676e-01 4.65189815e-01
-5.97166657e-01 4.43127245e-01 4.05437052e-01 1.09476142e-01
-9.82890308e-01 2.12693840e-01 1.56629169e+00 1.23527274e-01
-4.14470136e-01 3.65700364e-01 -7.81744719e-01 9.23200905e-01
8.59168887e-01 -6.44019485e-01 -3.25821787e-01 -3.70699108e-01
-1.44437641e-01 -1.05208650e-01 3.45967263e-01 -8.49394917e-01
3.71909171e-01 -1.80039883e-01 3.75287116e-01 -9.37416971e-01
3.23655903e-01 -6.00238740e-01 -1.05534747e-01 5.04478097e-01
-3.33645433e-01 -3.35319251e-01 2.76656508e-01 7.04125226e-01
-5.12017012e-02 -8.43471214e-02 1.02218115e+00 -2.60277808e-01
-6.16306484e-01 6.22867763e-01 1.12006836e-01 3.02687436e-01
1.02101791e+00 -2.10635453e-01 -1.59435093e-01 -2.22962335e-01
-8.06527138e-01 5.68269908e-01 2.90475607e-01 1.82321042e-01
4.37882900e-01 -9.56057131e-01 -5.85408151e-01 1.64064050e-01
-2.07043156e-01 4.29966986e-01 3.56820613e-01 9.69429910e-01
-6.08920813e-01 2.71268755e-01 2.28853449e-01 -6.52139068e-01
-1.50779903e+00 3.45979869e-01 7.93101341e-02 -5.70169747e-01
-4.21783388e-01 1.06580174e+00 5.21527469e-01 -1.54734746e-01
3.63817543e-01 -2.02544481e-01 1.51598960e-01 -3.12522799e-01
3.75486642e-01 4.73731279e-01 -2.05557346e-01 -1.18712962e-01
-8.05506036e-02 4.85642791e-01 -5.66907465e-01 -1.35625020e-01
1.11401641e+00 -1.35677069e-01 1.94995850e-01 5.62445894e-02
9.51850295e-01 -1.34662926e-01 -1.54106534e+00 -2.02076331e-01
-2.48518467e-01 -7.13809311e-01 1.23757541e-01 -1.03336084e+00
-1.18477917e+00 8.20824206e-01 5.47835886e-01 1.44997463e-01
1.16836548e+00 -2.55031623e-02 1.15935266e+00 -7.42570907e-02
1.21284589e-01 -1.24904644e+00 1.37760267e-01 1.81033269e-01
6.50137544e-01 -1.30816960e+00 5.30144423e-02 -9.61660266e-01
-6.57870710e-01 5.77547610e-01 8.74295175e-01 -1.35680288e-01
7.00652540e-01 3.65336359e-01 -1.51050612e-01 -6.46802261e-02
-4.55302119e-01 -2.97709733e-01 3.60853434e-01 5.95576875e-02
3.15140158e-01 1.20176427e-01 -5.68738580e-01 8.37481916e-01
-4.50229198e-02 2.52886772e-01 1.19673997e-01 9.41315532e-01
-5.00762761e-01 -1.01914895e+00 -1.90584600e-01 6.47648275e-01
-6.88300550e-01 -2.38775536e-01 -1.69455856e-01 7.43656218e-01
7.34811798e-02 9.41214979e-01 -1.36925891e-01 -3.22058678e-01
2.74408966e-01 -1.28841877e-01 1.75515130e-01 -6.32346451e-01
-4.87958372e-01 5.21701455e-01 -1.26902163e-01 -7.99437523e-01
-1.05995737e-01 -3.70851189e-01 -1.41966903e+00 -2.29037017e-01
-7.92855501e-01 6.44302890e-02 6.40625596e-01 7.48531699e-01
6.29948437e-01 6.49697065e-01 5.27137816e-01 -8.73153090e-01
-3.24697465e-01 -8.50825250e-01 -4.04773802e-01 4.26728159e-01
-3.99840996e-02 -5.17636240e-01 -3.09167117e-01 3.03986758e-01] | [9.444198608398438, -0.059335965663194656] |
b5440c21-f9c9-4bdb-995d-c12d868e92db | image-segmentation-via-probabilistic-graph | 2305.07954 | null | https://arxiv.org/abs/2305.07954v1 | https://arxiv.org/pdf/2305.07954v1.pdf | Image Segmentation via Probabilistic Graph Matching | This work presents an unsupervised and semi-automatic image segmentation approach where we formulate the segmentation as a inference problem based on unary and pairwise assignment probabilities computed using low-level image cues. The inference is solved via a probabilistic graph matching scheme, which allows rigorous incorporation of low level image cues and automatic tuning of parameters. The proposed scheme is experimentally shown to compare favorably with contemporary semi-supervised and unsupervised image segmentation schemes, when applied to contemporary state-of-the-art image sets. | ['Yosi Keller', 'Ayelet Heimowitz'] | 2023-05-13 | null | null | null | null | ['graph-matching'] | ['graphs'] | [ 9.32408512e-01 4.84507918e-01 -2.30489433e-01 -6.23789608e-01
-8.03739190e-01 -6.09179139e-01 6.18294597e-01 3.49290729e-01
-7.76727140e-01 4.65439975e-01 -2.82361805e-01 -1.79990843e-01
-4.25657362e-01 -5.83561182e-01 -3.07282329e-01 -7.06743717e-01
4.52105515e-02 9.80843186e-01 6.13550544e-01 1.10150225e-01
4.75344539e-01 5.17446935e-01 -1.58450282e+00 -1.80594221e-01
8.63876522e-01 7.37290800e-01 2.17529714e-01 8.42998564e-01
-8.10737605e-04 2.59507865e-01 -2.44951949e-01 -2.69439250e-01
3.28551114e-01 -3.88985187e-01 -1.20451772e+00 7.33278036e-01
7.30943680e-01 1.75526977e-01 6.89078942e-02 1.34445369e+00
2.17911601e-01 2.32402116e-01 1.05062258e+00 -9.45608377e-01
-3.92140508e-01 3.24065030e-01 -7.67301202e-01 2.35931620e-01
4.32762980e-01 -8.00045654e-02 1.21918881e+00 -4.46260571e-01
8.85405064e-01 1.01247621e+00 6.58989131e-01 -3.28417458e-02
-1.75117540e+00 -1.50716975e-01 -1.35103896e-01 -1.30075336e-01
-1.41082275e+00 3.34164761e-02 8.63689482e-01 -6.76060736e-01
7.17166722e-01 9.07393843e-02 4.70153034e-01 1.86093286e-01
3.40223163e-02 6.24049127e-01 1.45468318e+00 -8.08804750e-01
3.63033444e-01 7.45345205e-02 3.24687272e-01 1.01395941e+00
1.85825840e-01 7.86753297e-02 -2.67191768e-01 -1.13215178e-01
8.36648583e-01 -4.04705048e-01 1.38835728e-01 -6.64921403e-01
-1.13778436e+00 8.64389122e-01 5.10306537e-01 4.94647145e-01
-5.09486556e-01 -1.10216983e-01 5.18648066e-02 -2.83742040e-01
4.04060721e-01 2.72372812e-01 -2.27635652e-01 4.02145386e-01
-1.74695194e+00 5.48964068e-02 7.68638551e-01 9.30766761e-01
1.21543443e+00 -1.95284113e-01 -7.10017756e-02 8.07497084e-01
8.04822683e-01 3.57761294e-01 2.54556593e-02 -1.27318728e+00
2.49289367e-02 7.10505247e-01 -1.96570873e-01 -1.13175416e+00
-3.83252859e-01 -3.04192789e-02 -4.11495060e-01 3.52460027e-01
4.53058779e-01 2.35462263e-01 -1.47117603e+00 1.18038929e+00
3.78906667e-01 1.19809300e-01 -2.11978644e-01 7.33620703e-01
8.19709957e-01 2.10110202e-01 2.26926997e-01 -2.68072933e-01
1.34232819e+00 -7.67763674e-01 -7.10482478e-01 2.40759570e-02
2.83773318e-02 -1.10937977e+00 5.16646862e-01 3.56481165e-01
-9.27399993e-01 -6.91949487e-01 -1.01817012e+00 -7.42467195e-02
-4.70490098e-01 6.64852634e-02 7.14435220e-01 9.84617472e-01
-1.25193608e+00 5.18683732e-01 -7.70020843e-01 -5.28732002e-01
5.67206442e-01 8.28930080e-01 -4.28369135e-01 2.19702721e-03
-6.74133122e-01 6.25709593e-01 8.86426985e-01 3.54287386e-01
-5.75555205e-01 -1.85343787e-01 -1.24050152e+00 -4.91639823e-01
2.09912792e-01 -6.56880260e-01 8.40568364e-01 -9.78696883e-01
-1.63106871e+00 1.67452419e+00 -1.13873072e-01 -5.12376666e-01
4.44106877e-01 1.38513073e-01 2.03135726e-03 8.35093081e-01
1.91725388e-01 1.18481207e+00 9.76155221e-01 -1.72157109e+00
-4.55829114e-01 -3.43291551e-01 -9.37438384e-02 2.20874578e-01
2.94818878e-01 5.98139539e-02 -7.42370009e-01 -6.04075849e-01
6.77119017e-01 -9.35195327e-01 -6.81123197e-01 -1.38238892e-01
-7.98423469e-01 1.12113748e-02 7.80098140e-01 -5.79667032e-01
7.65074015e-01 -1.65519929e+00 2.77499706e-01 7.70306826e-01
2.46565014e-01 -1.15156267e-02 1.30236119e-01 1.07604630e-01
1.65884607e-02 1.83589179e-02 -1.01689482e+00 -5.23796380e-01
-1.16009481e-01 6.33141398e-01 3.70431483e-01 8.05803359e-01
9.23697650e-02 7.84255207e-01 -9.50391531e-01 -1.27317643e+00
7.55337656e-01 3.57225180e-01 -4.69932020e-01 2.09688753e-01
-1.56769425e-01 4.68098134e-01 -4.70572561e-01 7.65991747e-01
7.87309051e-01 -1.29634425e-01 1.98241159e-01 -1.34742007e-01
-7.09677637e-02 -3.09830308e-01 -1.24001372e+00 1.87347078e+00
2.11173594e-01 5.98227978e-01 1.97664738e-01 -1.22867751e+00
9.06805396e-01 1.23220772e-01 7.25266397e-01 -1.23634927e-01
4.73402172e-01 -2.54873857e-02 -1.47221491e-01 -3.10609907e-01
5.62508821e-01 -1.71721578e-01 -2.82195862e-02 2.70656019e-01
6.02717936e-01 -7.78409421e-01 4.16129678e-01 4.17035818e-01
5.86935639e-01 4.51899290e-01 3.54323238e-01 -8.36653650e-01
8.73223960e-01 2.26782396e-01 2.93459743e-01 9.57900524e-01
-4.13002253e-01 8.84054005e-01 2.32471928e-01 9.21712443e-03
-8.00281167e-01 -1.17063177e+00 -5.13164997e-01 7.65822470e-01
5.44496775e-01 -7.33412355e-02 -1.35127532e+00 -5.54162800e-01
-1.01003066e-01 2.39787400e-01 -7.32900560e-01 4.73104179e-01
-2.97749698e-01 -8.73491228e-01 3.41352701e-01 2.62168109e-01
5.10689259e-01 -1.20575714e+00 -3.71755689e-01 1.75057054e-01
1.85159534e-01 -1.49800992e+00 -2.60818005e-01 1.34072185e-01
-1.08240092e+00 -1.21271062e+00 -6.24191880e-01 -9.82379317e-01
1.12526548e+00 -7.72449300e-02 1.22057545e+00 2.86276013e-01
-8.25147986e-01 6.99404895e-01 -2.57669866e-01 6.91800267e-02
-2.69854277e-01 8.11358243e-02 -3.30381811e-01 1.24728106e-01
3.73223037e-01 -6.13860130e-01 -5.64166784e-01 3.54929507e-01
-1.06983149e+00 -9.00350958e-02 5.64430296e-01 7.41786778e-01
1.15312707e+00 1.40988261e-01 -2.26853445e-01 -1.40593684e+00
3.36309463e-01 4.54951636e-02 -8.92038882e-01 1.94918990e-01
-7.13076472e-01 1.83279455e-01 -7.18790069e-02 1.47710703e-02
-1.13049078e+00 7.16016710e-01 -9.84152257e-02 -9.75879580e-02
-7.69871294e-01 1.32004201e-01 -1.31764770e-01 -7.14405537e-01
3.81510854e-01 1.41817078e-01 2.68062074e-02 -3.02033037e-01
7.89337635e-01 5.27848363e-01 9.65388536e-01 -8.11859429e-01
8.25617671e-01 7.16259181e-01 1.43617928e-01 -9.79642928e-01
-6.54067516e-01 -1.04377902e+00 -1.61199582e+00 -3.95056486e-01
1.39384282e+00 -4.52043921e-01 -3.94152611e-01 5.43590069e-01
-8.69820416e-01 -2.17035204e-01 -8.20581019e-02 3.58076870e-01
-9.89040375e-01 7.62009740e-01 -5.84293306e-01 -7.75710166e-01
-3.65332663e-02 -1.45882380e+00 1.36362791e+00 3.65489036e-01
-1.56097218e-01 -1.41618526e+00 7.20683187e-02 9.17403042e-01
-1.67705521e-01 4.68487471e-01 5.71634233e-01 -3.90143275e-01
-6.38311386e-01 -2.35126764e-01 -4.57829982e-01 3.48122329e-01
-8.02238006e-03 4.30397809e-01 -9.10539567e-01 1.02438450e-01
-3.87009978e-01 -7.93933496e-02 9.72832799e-01 6.82278574e-01
8.84333968e-01 2.83447385e-01 -4.39810574e-01 4.82757211e-01
1.70849240e+00 -1.32962167e-01 6.73165202e-01 7.74979964e-02
8.77399206e-01 7.66480803e-01 5.51784813e-01 6.49873614e-02
4.04259264e-01 4.77225482e-01 2.42595956e-01 -3.51061165e-01
-1.84216976e-01 -6.91833049e-02 -3.54336321e-01 7.98475623e-01
-1.96328819e-01 7.95483310e-03 -1.10653710e+00 6.21795058e-01
-2.02432680e+00 -6.66608512e-01 -4.74168777e-01 1.93245852e+00
7.75909126e-01 4.20552671e-01 5.89101687e-02 3.08387965e-01
9.57002699e-01 9.60281938e-02 -1.38043463e-01 -3.70963812e-01
-4.34398837e-02 4.99164283e-01 9.23120439e-01 8.96519780e-01
-1.38602185e+00 1.20077753e+00 8.42417336e+00 8.47960055e-01
-5.52316666e-01 2.29683515e-04 6.63479090e-01 6.94297969e-01
-1.44128829e-01 2.76615590e-01 -5.25780320e-01 1.14313751e-01
4.46153879e-01 3.01554918e-01 1.33925796e-01 4.98549581e-01
-3.26238386e-02 -7.19877422e-01 -8.80702436e-01 8.30503821e-01
3.06127101e-01 -1.20481980e+00 5.82038658e-03 5.00986911e-02
9.41545248e-01 -3.82692255e-02 -1.39889389e-01 -4.33753043e-01
3.84681612e-01 -9.39256907e-01 5.69901824e-01 4.28812712e-01
3.22149724e-01 -6.14881277e-01 6.67850792e-01 -1.96664332e-04
-1.20070517e+00 4.80788201e-01 8.56664106e-02 8.38463977e-02
3.63720626e-01 5.44496894e-01 -5.16627729e-01 6.85079813e-01
5.53339303e-01 6.11336172e-01 -7.18089163e-01 1.09948981e+00
-5.18329918e-01 6.00767910e-01 -4.94622946e-01 2.64426500e-01
4.96550441e-01 -6.98804080e-01 5.15475631e-01 1.45565677e+00
-5.46574056e-01 1.51797652e-01 5.33399224e-01 8.54424238e-01
1.91498280e-01 1.67769283e-01 -4.28992003e-01 1.57079846e-01
1.18447572e-01 1.53646970e+00 -1.77161610e+00 -3.46871644e-01
-6.84933141e-02 1.01563704e+00 1.35213971e-01 3.47765952e-01
-2.70985425e-01 -2.53273517e-01 3.25371996e-02 -9.10629258e-02
2.87226766e-01 -3.37378949e-01 -6.28035784e-01 -7.06673741e-01
-4.10754472e-01 -2.79098958e-01 5.81555426e-01 -6.62321031e-01
-1.19615471e+00 4.76425260e-01 5.18366873e-01 -7.27631807e-01
-2.72514939e-01 -8.23002338e-01 -6.02594316e-01 6.14098370e-01
-1.42528498e+00 -1.29769063e+00 3.15390117e-02 4.14382666e-01
3.01697820e-01 3.99288498e-02 7.21747339e-01 1.10358149e-01
-3.75974208e-01 2.19921380e-01 -1.67159349e-01 1.80185169e-01
2.12175295e-01 -1.64812899e+00 -8.86967257e-02 1.06692147e+00
4.69569355e-01 5.69674134e-01 9.85515773e-01 -4.83763516e-01
-9.68801796e-01 -6.26943111e-01 5.50425231e-01 -4.69214976e-01
5.56039155e-01 -3.73063952e-01 -6.98093593e-01 3.86566550e-01
5.22784054e-01 2.87744671e-01 8.09768498e-01 -1.01548538e-01
-6.73495010e-02 3.35520476e-01 -1.48137498e+00 2.11572587e-01
8.04167271e-01 -6.54881597e-01 -7.66808033e-01 4.20268595e-01
3.72367412e-01 -3.00664753e-01 -1.03068042e+00 4.51521337e-01
3.68187934e-01 -8.13381135e-01 1.07363272e+00 -8.95541161e-03
6.93052486e-02 -4.49669719e-01 -1.01836070e-01 -6.19570673e-01
1.85341164e-02 -8.90767694e-01 4.70687270e-01 1.20728171e+00
5.22139668e-01 -3.87533575e-01 1.05969572e+00 6.29528522e-01
2.39315644e-01 -4.52782094e-01 -9.36050415e-01 -3.75063092e-01
-2.55469710e-01 -4.60301667e-01 -1.06031507e-01 7.52118707e-01
-1.92784742e-01 6.56823814e-02 3.44148576e-02 2.63354063e-01
1.35336375e+00 3.07247132e-01 4.25799787e-01 -1.37463832e+00
-2.21888587e-01 -6.43370986e-01 -9.87931669e-01 -7.78240919e-01
4.28812712e-01 -7.66840756e-01 5.92272937e-01 -1.54811108e+00
1.88891292e-01 -3.72864604e-01 -2.78613240e-01 2.61963934e-01
-1.40287533e-01 8.72986674e-01 -1.59427494e-01 -9.29606892e-03
-7.43087173e-01 1.23391822e-01 1.04668820e+00 -6.43600523e-02
-1.84186678e-02 -1.97210908e-01 -3.51102591e-01 9.49349761e-01
5.16335249e-01 -6.09788179e-01 -3.17304641e-01 6.38903528e-02
-1.71716869e-01 -1.73482746e-01 3.00710231e-01 -7.52185285e-01
4.11846608e-01 -4.40113917e-02 1.39579639e-01 -7.75081456e-01
1.70992091e-01 -7.60216951e-01 -6.97004348e-02 1.80916101e-01
-2.64216989e-01 -2.32593730e-01 5.08076623e-02 6.59138739e-01
-4.76362944e-01 -5.02913594e-01 8.95087600e-01 -2.19222337e-01
-1.01266050e+00 2.44529232e-01 -3.32870871e-01 -2.89553050e-02
1.14137864e+00 -7.01831639e-01 4.00499046e-01 -6.91800611e-03
-1.07473528e+00 1.69455141e-01 4.91835952e-01 -1.03596924e-02
5.09351254e-01 -8.81015718e-01 -3.63247693e-01 1.28704846e-01
1.84990495e-01 6.86891750e-02 2.55586766e-02 8.02483082e-01
-8.91583860e-01 1.32346526e-01 -1.97319806e-01 -1.07125926e+00
-1.35136831e+00 2.80080974e-01 1.57604888e-01 -1.42964318e-01
-5.19273162e-01 8.62112641e-01 -2.12070271e-01 -7.24370360e-01
7.10496902e-02 -2.08135143e-01 -3.56236875e-01 5.48598655e-02
-2.20948026e-01 1.66370437e-01 -2.12458149e-01 -1.34168780e+00
-4.23801959e-01 9.97459888e-01 1.66313529e-01 -3.75126123e-01
1.01298916e+00 -2.76729882e-01 -4.48151022e-01 2.79884309e-01
1.10189724e+00 -3.72696370e-01 -1.06759214e+00 -2.55647928e-01
2.70877540e-01 -4.52055395e-01 5.24486363e-01 -5.57066262e-01
-1.01007640e+00 7.11447120e-01 6.82426751e-01 1.17886335e-01
1.01402175e+00 1.15619808e-01 3.21148217e-01 1.76070914e-01
3.44505847e-01 -1.30806613e+00 -3.18736285e-01 6.56650066e-02
3.46814513e-01 -1.46224284e+00 5.17272592e-01 -9.81818974e-01
-4.00326610e-01 1.02081406e+00 1.95806831e-01 -5.29587984e-01
9.46442008e-01 2.76835531e-01 1.48144439e-01 -3.68294686e-01
1.48267925e-01 -7.27530837e-01 8.10193658e-01 6.91697240e-01
3.14435869e-01 1.73778921e-01 -4.46392894e-01 -1.46331683e-01
-3.56433354e-02 -1.41610697e-01 1.68603629e-01 9.08195436e-01
-3.48472565e-01 -1.29023468e+00 -3.13476473e-01 3.11129302e-01
-4.82730776e-01 -2.85442155e-02 -4.95720416e-01 8.35877299e-01
3.16647887e-01 8.68340135e-01 5.22603840e-02 1.09666973e-01
-1.08236730e-01 -1.43823698e-01 8.87281477e-01 -7.17371523e-01
-5.56760192e-01 2.73392379e-01 -1.09598376e-01 -4.15012628e-01
-1.25906873e+00 -8.17826688e-01 -1.46958160e+00 4.15513396e-01
-6.22298360e-01 6.10731728e-02 5.92273355e-01 1.20623374e+00
-1.42457366e-01 2.87002265e-01 2.45692208e-01 -1.36216915e+00
1.43436998e-01 -6.34381175e-01 -6.87561870e-01 5.10319412e-01
3.35990340e-02 -7.53125012e-01 -2.92505622e-01 6.31947041e-01] | [9.47188949584961, 0.26979753375053406] |
e5f364e4-1405-4561-9d06-5549ecec9b0f | semantic-scene-completion-from-a-single-depth | 1611.08974 | null | http://arxiv.org/abs/1611.08974v1 | http://arxiv.org/pdf/1611.08974v1.pdf | Semantic Scene Completion from a Single Depth Image | This paper focuses on semantic scene completion, a task for producing a
complete 3D voxel representation of volumetric occupancy and semantic labels
for a scene from a single-view depth map observation. Previous work has
considered scene completion and semantic labeling of depth maps separately.
However, we observe that these two problems are tightly intertwined. To
leverage the coupled nature of these two tasks, we introduce the semantic scene
completion network (SSCNet), an end-to-end 3D convolutional network that takes
a single depth image as input and simultaneously outputs occupancy and semantic
labels for all voxels in the camera view frustum. Our network uses a
dilation-based 3D context module to efficiently expand the receptive field and
enable 3D context learning. To train our network, we construct SUNCG - a
manually created large-scale dataset of synthetic 3D scenes with dense
volumetric annotations. Our experiments demonstrate that the joint model
outperforms methods addressing each task in isolation and outperforms
alternative approaches on the semantic scene completion task. | ['Angel X. Chang', 'Thomas Funkhouser', 'Manolis Savva', 'Fisher Yu', 'Shuran Song', 'Andy Zeng'] | 2016-11-28 | semantic-scene-completion-from-a-single-depth-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Song_Semantic_Scene_Completion_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Song_Semantic_Scene_Completion_CVPR_2017_paper.pdf | cvpr-2017-7 | ['3d-semantic-scene-completion'] | ['computer-vision'] | [ 4.42483068e-01 4.45267558e-01 1.31925747e-01 -7.52319932e-01
-6.98795319e-01 -6.25562310e-01 6.62330270e-01 1.46489948e-01
-2.80281097e-01 3.12816054e-01 4.15629953e-01 -1.50083289e-01
2.75074631e-01 -7.42931962e-01 -8.64620805e-01 -1.73210338e-01
1.60135739e-02 5.45968890e-01 4.66052651e-01 3.54145974e-01
2.52296478e-01 6.78984165e-01 -1.50065506e+00 3.38916689e-01
4.14736241e-01 1.00846410e+00 6.05954051e-01 7.92434990e-01
-2.62122959e-01 9.44458723e-01 -4.05454496e-03 1.78120196e-01
4.14585948e-01 -3.83136384e-02 -9.90938127e-01 5.15927613e-01
8.47740948e-01 -7.60909736e-01 -6.18791759e-01 6.80020750e-01
2.47107163e-01 6.89965039e-02 6.09268785e-01 -1.03637528e+00
-1.85413778e-01 1.21777982e-01 -5.27750850e-01 2.27134019e-01
4.46829081e-01 4.25275527e-02 8.18619251e-01 -1.03432894e+00
8.32502663e-01 1.32022738e+00 4.85885859e-01 4.07848060e-01
-1.37693024e+00 -3.21425617e-01 4.05399978e-01 -5.11828303e-01
-1.15111685e+00 -2.50763595e-01 9.05072451e-01 -6.52806997e-01
1.30607116e+00 -1.13464631e-01 8.66360188e-01 8.46309125e-01
9.54281241e-02 6.91107988e-01 1.21382582e+00 -1.25448599e-01
5.67181051e-01 -3.17773551e-01 -1.54348999e-01 8.10849786e-01
-4.58230637e-02 2.49360465e-02 -4.98523682e-01 5.30849583e-02
1.30417478e+00 3.43510538e-01 1.30614927e-02 -9.15684283e-01
-1.27889633e+00 7.59112000e-01 7.41760552e-01 -2.59824693e-01
-2.70548582e-01 7.32037902e-01 1.45966709e-01 -2.08581716e-01
7.06668854e-01 2.93801099e-01 -4.46215868e-01 2.19925359e-01
-7.44717836e-01 5.80616236e-01 5.10098219e-01 1.04129267e+00
9.38240528e-01 -2.96720117e-01 -3.99066024e-02 5.05876839e-01
3.32166553e-01 4.39817816e-01 -2.43647024e-01 -1.49609745e+00
3.42406511e-01 7.02938855e-01 2.89187338e-02 -4.11799043e-01
-6.70773149e-01 -2.73652911e-01 -2.90387571e-01 4.40981060e-01
3.29719990e-01 2.26824492e-01 -1.48609972e+00 1.70090663e+00
7.20784783e-01 3.16880733e-01 -3.63293439e-02 1.01586378e+00
1.01360202e+00 2.60498583e-01 2.25034446e-01 4.43278342e-01
1.16759527e+00 -1.07420492e+00 -1.42768949e-01 -6.26853764e-01
5.16977012e-01 -5.27135491e-01 8.84629071e-01 6.12783059e-02
-1.24920774e+00 -3.44665974e-01 -1.06088543e+00 -7.72710443e-01
-2.18564868e-01 -3.35257113e-01 9.92368996e-01 1.02598861e-01
-1.24036086e+00 3.42487574e-01 -1.16110003e+00 -3.28776926e-01
7.99772739e-01 1.45978540e-01 -5.19557953e-01 -5.61375976e-01
-5.21691859e-01 5.96319020e-01 3.29533458e-01 -2.91200906e-01
-1.66420197e+00 -8.71625185e-01 -1.26476455e+00 -1.42543375e-01
2.82923341e-01 -1.12228274e+00 1.50740743e+00 -3.90605778e-01
-9.18103456e-01 1.27782965e+00 -1.48516536e-01 -1.51861683e-01
3.71393234e-01 -1.02526397e-01 4.39444095e-01 3.56597781e-01
5.33639371e-01 1.25286460e+00 4.69406486e-01 -1.52508879e+00
-6.48714364e-01 -6.77283227e-01 3.43122929e-01 5.42463660e-01
4.72276151e-01 -5.62712312e-01 -7.64159679e-01 -4.11795795e-01
6.75368786e-01 -5.50437629e-01 -5.23812890e-01 2.33634904e-01
-5.13554692e-01 7.88789690e-02 5.06895781e-01 -4.53273535e-01
3.50204647e-01 -2.01031089e+00 3.52297753e-01 1.78944051e-01
5.68202913e-01 -6.07355833e-01 -1.50924474e-01 1.69423983e-01
-6.96559176e-02 -1.37631342e-01 -5.32930970e-01 -1.01689708e+00
-1.39078617e-01 4.04960483e-01 -2.38887638e-01 5.43302774e-01
1.65438309e-01 1.03756177e+00 -1.24089217e+00 -3.23926330e-01
6.48989916e-01 6.60337508e-01 -9.15815294e-01 4.75616813e-01
-6.19003296e-01 8.32790136e-01 -5.87129951e-01 7.06114173e-01
6.58146977e-01 -3.85247201e-01 7.41596799e-03 -3.42415236e-02
-4.02300209e-02 3.78204226e-01 -8.04574609e-01 2.82883000e+00
-7.17240691e-01 4.56707239e-01 1.35008723e-01 -6.55675650e-01
6.73494399e-01 5.63257560e-03 7.35714316e-01 -6.40252233e-01
8.81803781e-02 1.75517991e-01 -7.40150094e-01 -3.82618308e-01
4.84108567e-01 -1.81838587e-01 -1.80356249e-01 5.21942258e-01
2.50349194e-01 -1.00668979e+00 -3.27948600e-01 4.09342259e-01
1.36354291e+00 4.92992014e-01 -2.38364954e-02 -5.10196686e-02
1.08210936e-01 9.48498324e-02 2.48822019e-01 5.21656573e-01
-3.82540748e-02 9.43061352e-01 5.40998399e-01 -5.42593062e-01
-1.50522625e+00 -1.60281134e+00 -1.39520183e-01 7.35110939e-01
4.10643011e-01 -2.76073009e-01 -6.91167831e-01 -6.93555295e-01
2.80025989e-01 4.62322831e-01 -7.93390334e-01 2.94189811e-01
-4.10182416e-01 -1.23000868e-01 1.84654906e-01 6.90894842e-01
3.52192461e-01 -8.77502918e-01 -9.73847866e-01 2.72279121e-02
-2.70909518e-02 -1.68181086e+00 -2.38280520e-01 4.89334136e-01
-1.05233395e+00 -1.22920787e+00 -5.78968346e-01 -7.99942493e-01
8.96473706e-01 4.68975514e-01 1.43684876e+00 -7.72439502e-03
-5.52960992e-01 6.14513993e-01 -1.09666973e-01 -1.33849010e-01
-1.92770287e-01 1.14137471e-01 -3.93670887e-01 -4.52837288e-01
8.78297985e-02 -1.05256534e+00 -8.75426471e-01 -7.86834732e-02
-1.03950047e+00 6.82217717e-01 2.36451402e-01 5.00262678e-01
1.03381503e+00 -3.47419888e-01 1.99485242e-01 -8.68087411e-01
3.25083882e-02 -4.36225027e-01 -7.30354309e-01 -2.35694155e-01
-1.29156068e-01 3.37199233e-02 9.24818069e-02 3.01931530e-01
-9.30596590e-01 6.09915495e-01 -2.79656023e-01 -7.11795330e-01
-4.81279612e-01 2.06109276e-03 -2.55466551e-01 5.94050214e-02
4.18998718e-01 1.30973697e-01 -1.69004485e-01 -4.57674384e-01
7.12195754e-01 3.15665126e-01 9.03347194e-01 -5.54055572e-01
5.41825056e-01 1.06549954e+00 1.97054639e-01 -4.70294714e-01
-1.24165440e+00 -5.97348452e-01 -1.07340717e+00 -1.15109883e-01
1.36010039e+00 -1.49352419e+00 -5.28682888e-01 2.70978242e-01
-1.31031787e+00 -8.61074746e-01 -4.86247897e-01 4.40625846e-01
-1.03756344e+00 2.05488026e-01 -4.64642495e-01 -4.58481759e-01
-3.33850533e-02 -1.23629165e+00 1.85632634e+00 -5.31974733e-02
-4.74229231e-02 -9.99035656e-01 9.60640162e-02 4.09395456e-01
5.26656024e-02 7.72644937e-01 8.79043102e-01 -1.23395748e-01
-1.06949139e+00 1.29447654e-01 -5.31377852e-01 1.99791715e-01
-5.82540259e-02 -6.13137424e-01 -1.47660792e+00 -1.31189615e-01
-6.98025897e-02 -6.03237152e-01 1.00381744e+00 5.47072470e-01
1.44254887e+00 3.15706670e-01 -3.36991131e-01 1.19884348e+00
1.60877371e+00 -2.59264052e-01 5.01669168e-01 1.18821792e-01
9.34471190e-01 7.33230889e-01 3.15774083e-01 4.40263391e-01
7.90427983e-01 3.48271877e-01 9.42556739e-01 -3.88682872e-01
-4.89152461e-01 -5.47850072e-01 -1.96025565e-01 3.53470504e-01
2.78469533e-01 -6.60388768e-02 -9.80485976e-01 5.34755051e-01
-1.53566682e+00 -5.53430200e-01 1.37116536e-01 1.99524534e+00
4.90059972e-01 1.72277793e-01 -1.87603831e-01 -2.03527451e-01
2.42036209e-01 3.59734118e-01 -9.27904189e-01 -1.77762151e-01
1.23305947e-01 3.49584937e-01 6.56542361e-01 9.41969633e-01
-1.14947629e+00 1.11377645e+00 6.48882580e+00 3.54466707e-01
-7.92456448e-01 3.34139317e-01 8.47905219e-01 -4.87950414e-01
-7.81527400e-01 5.52580729e-02 -4.46501851e-01 -9.07879323e-02
3.70274425e-01 4.21363831e-01 6.22873127e-01 8.83756042e-01
1.55566961e-01 -5.65563321e-01 -1.52376831e+00 1.21897280e+00
9.79119688e-02 -1.35004926e+00 -5.91032347e-03 2.21836135e-01
1.05733287e+00 6.50501072e-01 -2.12229878e-01 -9.32448506e-02
6.01335704e-01 -1.27897549e+00 8.74683857e-01 3.80681902e-01
1.12518251e+00 -5.50590575e-01 2.07477763e-01 1.99198261e-01
-1.35671186e+00 7.79782161e-02 -2.22606733e-01 -1.07706957e-01
4.82567251e-01 7.08632648e-01 -8.37092459e-01 3.65549475e-01
5.84446430e-01 8.04758012e-01 -5.47310650e-01 9.32721615e-01
-4.15734023e-01 -1.06395386e-01 -3.68017703e-01 6.06098652e-01
3.67610335e-01 2.04136357e-01 3.44472528e-01 9.40311790e-01
1.06674366e-01 9.07679349e-02 3.83172303e-01 1.38823545e+00
-1.67046800e-01 -4.34577763e-01 -8.84758532e-01 2.50415236e-01
4.33276534e-01 1.28614080e+00 -1.12690258e+00 -4.23906446e-01
-3.64397287e-01 1.24350631e+00 5.93625903e-01 4.02350724e-01
-6.51969016e-01 6.52401447e-02 7.41251230e-01 2.83248156e-01
3.01768392e-01 -5.78543007e-01 -7.57160127e-01 -1.00792074e+00
-4.53988016e-02 8.66417401e-03 8.23410451e-02 -1.17792022e+00
-9.91205156e-01 4.28134084e-01 4.70568128e-02 -8.55364859e-01
-2.15302244e-01 -5.55213630e-01 -3.17803651e-01 9.90314722e-01
-1.80739582e+00 -1.22406566e+00 -9.49063241e-01 6.98367357e-01
6.81249797e-01 4.06860054e-01 6.83679104e-01 1.48058653e-01
3.80987786e-02 -1.08509675e-01 -3.68079066e-01 -4.00201455e-02
1.95024192e-01 -1.38269353e+00 8.22487593e-01 5.50774157e-01
-1.72789425e-01 1.23819135e-01 2.92524785e-01 -7.81720757e-01
-1.34900188e+00 -1.38671017e+00 6.14196956e-01 -9.81419206e-01
9.37172100e-02 -8.15267622e-01 -4.54998195e-01 8.55963290e-01
-1.73790157e-01 3.46661299e-01 2.85611570e-01 -1.81410372e-01
-4.62671012e-01 3.33256721e-01 -1.08742678e+00 3.64307553e-01
1.67129731e+00 -8.45670998e-01 -5.23000896e-01 4.47288513e-01
1.18660736e+00 -8.83156300e-01 -8.83081198e-01 3.76581073e-01
3.20790023e-01 -1.14111984e+00 1.31042933e+00 -2.18383834e-01
8.83908093e-01 -3.98497492e-01 -6.23001993e-01 -1.04196405e+00
6.94248453e-02 -1.38186570e-03 3.90180573e-02 5.56460798e-01
9.64986235e-02 -1.45433977e-01 1.06177354e+00 6.06099069e-01
-5.81840873e-01 -8.00681889e-01 -9.74014282e-01 -3.86718333e-01
1.22989550e-01 -8.43896329e-01 6.00672245e-01 7.29815304e-01
-2.83804923e-01 4.59749579e-01 1.74320489e-01 1.75183684e-01
9.21617150e-01 2.51223862e-01 9.72342849e-01 -1.07242298e+00
-1.51463106e-01 -1.95817724e-01 -3.72641802e-01 -1.30944884e+00
6.10410199e-02 -1.30110645e+00 2.07407996e-01 -1.92118168e+00
4.06067580e-01 -5.96004784e-01 1.62073672e-01 4.03152287e-01
1.48539871e-01 4.06478643e-01 -7.54718706e-02 1.16688684e-01
-7.21351981e-01 6.30070090e-01 1.59471560e+00 3.32708769e-02
-2.33999386e-01 -5.54925203e-01 -5.46920180e-01 7.78225839e-01
4.11675185e-01 -3.34480524e-01 -7.99397707e-01 -8.58682811e-01
1.04648970e-01 2.44282037e-01 8.52869868e-01 -9.98497486e-01
2.81701679e-03 -6.52573407e-02 6.80725992e-01 -9.54468668e-01
7.35018790e-01 -7.49919116e-01 -3.03384922e-02 1.60160527e-01
-3.50579560e-01 -1.19720556e-01 7.24083781e-02 7.23391354e-01
8.02350417e-02 2.83234686e-01 5.50662518e-01 -6.20964229e-01
-8.56496871e-01 7.34616399e-01 -2.17387732e-02 1.95882112e-01
1.01705229e+00 -2.99718022e-01 -3.38094272e-02 -6.69954941e-02
-9.17129755e-01 3.60387892e-01 9.48085606e-01 3.35737079e-01
9.71087217e-01 -1.17941880e+00 -4.63631123e-01 3.92186671e-01
2.84023285e-01 1.03564584e+00 4.47743863e-01 4.84851241e-01
-8.78149092e-01 4.82470393e-01 -8.82720277e-02 -1.03696036e+00
-6.64253414e-01 5.36053061e-01 5.34656525e-01 -1.52004331e-01
-1.03842616e+00 1.03419864e+00 8.75107050e-01 -7.65594006e-01
4.87003028e-01 -7.26731479e-01 1.74553454e-01 -5.54642320e-01
2.92107701e-01 -1.29815236e-01 3.76845128e-03 -4.35238153e-01
-2.90295511e-01 5.62223673e-01 1.99901775e-01 -4.07332540e-01
1.47685969e+00 -3.74553412e-01 -2.93312091e-02 3.46783817e-01
1.39837587e+00 -3.29816043e-01 -1.92132473e+00 -2.54862696e-01
-4.04451817e-01 -6.92090631e-01 2.04346180e-01 -6.72336280e-01
-1.00890362e+00 9.21963334e-01 4.55924600e-01 -3.10093880e-01
8.14004838e-01 4.65587854e-01 7.68181741e-01 5.51037826e-02
5.15368819e-01 -8.22672129e-01 4.30799782e-01 4.94477391e-01
7.87260354e-01 -1.23362195e+00 1.12335540e-01 -7.70914853e-01
-2.67244041e-01 8.56082439e-01 8.80502939e-01 -3.23399186e-01
6.56459868e-01 3.09605896e-01 -1.95216522e-01 -6.95834279e-01
-6.56155825e-01 -2.61518508e-01 6.32142797e-02 7.36937761e-01
2.13933900e-01 1.91390105e-02 4.50363696e-01 1.54826194e-01
-2.76379824e-01 1.89971253e-02 4.17048275e-01 1.16700017e+00
-4.64873999e-01 -7.74216056e-01 -2.49224171e-01 2.30906606e-01
-1.13540284e-01 -8.62296075e-02 -2.61090338e-01 5.15524626e-01
2.06844866e-01 6.03483915e-01 5.58207095e-01 -3.64369899e-01
2.72734821e-01 -1.87209785e-01 8.22264731e-01 -1.04912174e+00
-3.40935998e-02 2.08274126e-02 -4.19369377e-02 -1.05911398e+00
-4.09735113e-01 -4.34108973e-01 -1.61340535e+00 -1.08977385e-01
3.85588765e-01 -4.59446937e-01 9.89022911e-01 9.18904901e-01
1.94034159e-01 8.07149947e-01 6.08630657e-01 -1.42348623e+00
7.79519230e-02 -6.54688597e-01 -7.23603427e-01 4.73224044e-01
4.24513310e-01 -7.58892357e-01 -1.45463869e-01 3.52749452e-02] | [8.504201889038086, -2.942798137664795] |
c4a44ec6-92eb-4c09-a7ae-ef2336e0bb08 | mots-r-cnn-cosine-margin-triplet-loss-for | 2102.03512 | null | https://arxiv.org/abs/2102.03512v1 | https://arxiv.org/pdf/2102.03512v1.pdf | MOTS R-CNN: Cosine-margin-triplet loss for multi-object tracking | One of the central tasks of multi-object tracking involves learning a distance metric that is consistent with the semantic similarities of objects. The design of an appropriate loss function that encourages discriminative feature learning is among the most crucial challenges in deep neural network-based metric learning. Despite significant progress, slow convergence and a poor local optimum of the existing contrastive and triplet loss based deep metric learning methods necessitates a better solution. In this paper, we propose cosine-margin-contrastive (CMC) and cosine-margin-triplet (CMT) loss by reformulating both contrastive and triplet loss functions from the perspective of cosine distance. The proposed reformulation as a cosine loss is achieved by feature normalization which distributes the learned features on a hypersphere. We then propose the MOTS R-CNN framework for joint multi-object tracking and segmentation, particularly targeted at improving the tracking performance. Specifically, the tracking problem is addressed through deep metric learning based on the proposed loss functions. We propose a scale-invariant tracking by using a multi-layer feature aggregation scheme to make the model robust against object scale variations and occlusions. The MOTS R-CNN achieves the state-of-the-art tracking performance on the KITTI MOTS dataset. We show that the MOTS R-CNN reduces the identity switching by $62\%$ and $61\%$ on cars and pedestrians, respectively in comparison to Track R-CNN. | ['Renu M. Rameshan', 'Amit Satish Unde'] | 2021-02-06 | null | null | null | null | ['multi-object-tracking-and-segmentation'] | ['computer-vision'] | [-1.78840920e-01 -6.38248265e-01 5.86597212e-02 -5.92133701e-01
-8.15115988e-01 -4.17827874e-01 3.93155456e-01 2.55799890e-02
-7.55561113e-01 5.92186749e-01 -2.91645378e-01 7.16400966e-02
-5.62331080e-01 -6.56414330e-01 -7.14295983e-01 -8.03061724e-01
1.29695982e-01 4.85777080e-01 4.17121559e-01 -1.92882076e-01
3.71409021e-02 8.61459970e-01 -1.61630785e+00 -2.92836308e-01
8.15007746e-01 1.36138821e+00 6.09392375e-02 4.72763538e-01
1.65256754e-01 4.81133014e-01 -4.85937178e-01 -4.85507607e-01
6.06902003e-01 -2.13019401e-01 -3.19105327e-01 -2.45460615e-01
1.03486788e+00 -8.11501071e-02 -3.94747823e-01 1.19551969e+00
6.13602996e-01 4.24561739e-01 7.42992997e-01 -1.54787648e+00
-6.56183898e-01 7.12956041e-02 -7.08563209e-01 3.47597301e-01
-2.40591958e-01 1.07988164e-01 1.16031551e+00 -6.56372845e-01
3.48272085e-01 1.26992404e+00 1.05821395e+00 4.93011355e-01
-9.11979318e-01 -7.72378266e-01 9.19564962e-02 4.19303596e-01
-1.69136107e+00 -1.55420667e-02 7.32051551e-01 -4.63504344e-01
5.09262204e-01 2.87112504e-01 5.48959434e-01 5.27806342e-01
2.74830759e-01 5.59994638e-01 7.40138292e-01 -8.27743784e-02
-1.10357799e-01 -1.17941601e-02 2.85955612e-02 9.13154125e-01
5.42647123e-01 3.58658373e-01 -2.02534527e-01 6.03002310e-02
5.29707313e-01 2.17070609e-01 -1.14847636e-02 -6.07737184e-01
-1.00608695e+00 9.25884545e-01 8.18077385e-01 2.97689080e-01
-5.61293699e-02 3.30983490e-01 3.83928031e-01 8.58383477e-02
3.23143452e-01 2.24393994e-01 -3.58820856e-01 4.04648706e-02
-1.12128162e+00 4.57427204e-01 3.28695655e-01 1.03992999e+00
7.73498297e-01 3.67310159e-02 -4.73652095e-01 6.44866347e-01
6.53753400e-01 8.64005506e-01 1.05565228e-01 -8.99740636e-01
3.04285675e-01 6.64676726e-01 2.32672617e-01 -1.21301270e+00
-5.97012937e-01 -7.47528374e-01 -6.98377132e-01 3.93841624e-01
6.36111975e-01 -1.39657140e-01 -5.73722303e-01 2.04063678e+00
6.90193832e-01 2.25815400e-01 -2.50945270e-01 1.01469755e+00
6.94620430e-01 2.64572918e-01 4.22546156e-02 9.38504562e-02
1.26523614e+00 -8.25303197e-01 -5.36881566e-01 9.44167748e-02
6.24318123e-01 -8.55301142e-01 6.33024037e-01 -1.19676203e-01
-8.92929435e-01 -7.75869966e-01 -1.32074714e+00 -2.53662448e-02
-5.08260548e-01 2.22787336e-01 3.12212169e-01 7.25231767e-01
-8.74587536e-01 7.65607595e-01 -8.51631165e-01 -3.98557305e-01
5.68004131e-01 5.33132136e-01 3.36393528e-02 7.55892470e-02
-9.29592669e-01 1.10019779e+00 2.54232615e-01 2.31333867e-01
-6.42359734e-01 -9.60986793e-01 -8.12396288e-01 -7.20640793e-02
7.51100257e-02 -6.85658574e-01 8.31576526e-01 -4.31048900e-01
-1.29169965e+00 7.87212372e-01 1.17541127e-01 -4.62283939e-01
8.45041990e-01 -2.68237025e-01 -5.03875136e-01 -7.80209675e-02
1.18520483e-01 6.85804904e-01 6.51498199e-01 -8.36332262e-01
-9.80692744e-01 -5.22013903e-01 1.94285382e-02 1.03057325e-01
-4.01756883e-01 2.06505373e-01 -2.57392585e-01 -6.15013540e-01
-9.35318470e-02 -1.08103478e+00 -3.62041108e-02 8.00126433e-01
-1.18231662e-01 -4.13814157e-01 1.26815212e+00 -3.71796966e-01
9.67667699e-01 -2.06521964e+00 1.70707390e-01 -4.72766235e-02
2.80451834e-01 5.59430897e-01 -2.29835585e-01 -1.41885877e-01
2.79114097e-01 -4.38943326e-01 -5.98396286e-02 -4.64398563e-01
3.55459541e-01 -4.22564708e-02 2.34247133e-01 1.07488096e+00
2.32508734e-01 9.53587174e-01 -7.72345245e-01 -7.52654314e-01
5.24719656e-01 8.11537921e-01 -4.31830168e-01 1.59638152e-01
-1.32224306e-01 5.28572440e-01 -3.89165491e-01 6.63198233e-01
9.45754766e-01 -1.18739635e-01 -4.41952527e-01 -6.13157749e-01
-5.41744292e-01 -2.56769776e-01 -1.33510590e+00 1.68726337e+00
-1.84440121e-01 7.37543285e-01 5.29578812e-02 -9.92109895e-01
1.13744068e+00 -6.63864017e-02 9.18260694e-01 -7.60407567e-01
3.72454911e-01 3.50237936e-01 5.31306267e-02 -7.97652602e-02
5.67691386e-01 -1.38222605e-01 4.17331830e-02 1.30206004e-01
-3.42658535e-02 1.96500555e-01 3.37389827e-01 -7.06911758e-02
9.04024601e-01 2.07302153e-01 -5.95540889e-02 -4.92236674e-01
8.63913417e-01 -1.42923832e-01 8.47976208e-01 4.98721570e-01
-6.37499750e-01 4.65694577e-01 -1.93524599e-01 -5.64301610e-01
-9.41269696e-01 -9.67488706e-01 -3.90801162e-01 1.01994252e+00
4.31893796e-01 2.86046378e-02 -4.43552136e-01 -6.61580324e-01
4.00487095e-01 1.96523950e-01 -6.13533258e-01 -3.06153238e-01
-9.30792034e-01 -7.31163621e-01 5.94098508e-01 7.57445037e-01
8.13805044e-01 -7.82205284e-01 -6.64656222e-01 1.59102783e-01
8.15909579e-02 -1.16579437e+00 -8.66962969e-01 5.31680658e-02
-5.28400302e-01 -1.18985724e+00 -8.17145646e-01 -7.18197703e-01
5.02838433e-01 3.80396992e-01 6.94846988e-01 1.92207156e-03
-5.02047420e-01 3.48377109e-01 -1.55484110e-01 -2.63474762e-01
4.81280275e-02 5.97953014e-02 2.32271299e-01 1.81554675e-01
5.08600056e-01 -3.72289956e-01 -8.34357202e-01 6.47935987e-01
-5.27218819e-01 -4.16418105e-01 5.28462291e-01 6.22293711e-01
6.65823519e-01 -2.26251826e-01 3.41904461e-01 -2.04784647e-01
-7.25403428e-02 -2.06655011e-01 -1.11085963e+00 3.31974208e-01
-4.63019550e-01 4.48118336e-02 5.31098962e-01 -5.23591518e-01
-6.51503682e-01 6.64253086e-02 -2.49524042e-02 -5.50328910e-01
1.36247739e-01 -1.19462542e-01 -2.23073527e-01 -8.29827189e-01
4.35511768e-01 -6.53927773e-02 -2.22284004e-01 -3.32228094e-01
4.27667558e-01 4.01740313e-01 3.91141534e-01 -5.50350249e-01
1.26482296e+00 7.46198237e-01 4.63792980e-01 -6.18233144e-01
-1.07711756e+00 -6.40787482e-01 -7.26538718e-01 -5.26074409e-01
1.12672591e+00 -8.08101892e-01 -1.12383401e+00 4.76413697e-01
-1.09552109e+00 -2.61068027e-02 -3.69348288e-01 7.63579667e-01
-5.04727483e-01 2.30485007e-01 -3.29817414e-01 -6.40803397e-01
-5.44219971e-01 -1.16223121e+00 1.12560177e+00 5.82265437e-01
1.41456991e-01 -1.04937387e+00 2.25132033e-01 3.29958111e-01
7.18045354e-01 6.61678791e-01 4.80334729e-01 -5.66677213e-01
-7.96575189e-01 -1.34723052e-01 -7.50745535e-01 3.29056263e-01
1.36976331e-01 -2.75925174e-02 -6.68051541e-01 -6.75479472e-01
-1.80771813e-01 -1.54361427e-01 7.58917928e-01 6.20084465e-01
8.30131531e-01 8.12474415e-02 -3.47151577e-01 1.13327849e+00
1.61259544e+00 1.86640024e-01 1.90876320e-01 5.36779761e-01
1.03971791e+00 2.29039803e-01 8.73078346e-01 4.53980953e-01
4.42775071e-01 9.70521927e-01 6.34786487e-01 -7.15030506e-02
-3.70834082e-01 1.76611930e-01 8.36202055e-02 7.19854295e-01
1.15783773e-01 4.97970060e-02 -5.88629007e-01 4.56104189e-01
-1.95346475e+00 -8.49062026e-01 -1.39142811e-01 2.16449857e+00
5.23838639e-01 1.19221583e-01 3.20026517e-01 -2.03902781e-01
1.00537217e+00 2.50793785e-01 -7.45745182e-01 -9.67961699e-02
-1.84544697e-01 1.19205445e-01 6.93828583e-01 3.34658325e-01
-1.47596657e+00 7.52363563e-01 4.91983652e+00 8.39105606e-01
-1.13835669e+00 2.53272206e-01 -4.46735024e-02 -2.51765132e-01
1.89082965e-01 -2.75779843e-01 -1.45610631e+00 5.14754891e-01
7.79067099e-01 -1.17253177e-01 8.68460760e-02 7.89023876e-01
6.37483448e-02 3.58492076e-01 -1.01666439e+00 1.14808404e+00
9.05512422e-02 -1.24613523e+00 -1.88234553e-01 2.04805844e-02
6.68416798e-01 2.26722240e-01 3.27040225e-01 3.51434469e-01
9.79251415e-02 -6.12684131e-01 1.03227103e+00 4.42735761e-01
6.23532116e-01 -7.95558989e-01 5.46414077e-01 -9.10368189e-02
-1.90145361e+00 -8.98435563e-02 -4.07749653e-01 4.02544111e-01
2.11922243e-01 3.30639929e-01 -2.18592852e-01 6.66826904e-01
8.13488543e-01 9.24636483e-01 -6.05877876e-01 1.53845465e+00
3.20337772e-01 -3.28605212e-02 -3.87565315e-01 -1.38769463e-01
2.89566666e-01 -3.84763569e-01 7.81058729e-01 1.32251191e+00
3.58348042e-01 -2.89664537e-01 3.34898382e-01 9.65096056e-01
-1.67271271e-01 5.99736534e-02 -2.89945662e-01 3.22507143e-01
4.33936566e-01 1.45279753e+00 -6.22454822e-01 -7.71790892e-02
-5.25971174e-01 4.54846889e-01 3.66352141e-01 7.76550220e-03
-1.30890632e+00 -5.16762853e-01 9.98108864e-01 -2.27987161e-03
7.92227268e-01 -2.32785627e-01 3.00528016e-02 -8.27205300e-01
2.06525251e-01 -4.10635859e-01 4.55893666e-01 -2.41960451e-01
-1.32195652e+00 4.83650655e-01 -9.52936262e-02 -1.50242746e+00
1.63445950e-01 -7.86835849e-01 -5.38644075e-01 6.91086531e-01
-1.96518207e+00 -1.44375956e+00 -6.05553567e-01 7.34229326e-01
1.77926034e-01 -1.47605404e-01 2.92255104e-01 9.65387166e-01
-5.81374466e-01 9.00766909e-01 2.85093129e-01 2.64208764e-01
7.30223298e-01 -1.24411881e+00 1.32158786e-01 8.03957582e-01
-2.92925555e-02 3.91806513e-01 6.43830597e-01 -4.68741179e-01
-1.28447747e+00 -1.62763751e+00 5.82302809e-01 -2.85384923e-01
7.03850091e-01 -1.51300490e-01 -7.06646264e-01 5.27667999e-01
-1.23358399e-01 6.00746810e-01 5.30599892e-01 -2.77733088e-01
-3.92900616e-01 -5.45326173e-01 -1.31480479e+00 2.84180671e-01
1.12891984e+00 -3.08262318e-01 -2.70352691e-01 3.94867688e-01
6.38132691e-01 -3.45882207e-01 -1.15294695e+00 6.04946375e-01
6.67774975e-01 -7.15805709e-01 1.07817471e+00 -4.68000472e-01
-3.74533534e-01 -7.06524909e-01 -3.92861277e-01 -8.82437289e-01
-4.20046240e-01 -6.70750141e-01 -1.35268018e-01 1.16621399e+00
8.63577519e-03 -4.70810354e-01 8.49415183e-01 2.66113609e-01
-3.57855946e-01 -6.37842774e-01 -1.24547517e+00 -1.23929513e+00
2.30214074e-01 -8.34403783e-02 4.64586645e-01 7.17782319e-01
-7.75197744e-01 -2.41503008e-02 -2.73772955e-01 2.34042510e-01
1.28573465e+00 1.83508754e-01 7.72237480e-01 -1.58110392e+00
5.94587810e-02 -5.73549926e-01 -7.94312119e-01 -1.01964271e+00
7.00260997e-02 -8.67738783e-01 4.22034413e-02 -1.22547245e+00
1.37467772e-01 -6.42117381e-01 -5.58060944e-01 7.50112087e-02
-1.81342866e-02 3.13409090e-01 4.25487339e-01 8.70057009e-03
-1.14681196e+00 8.57895255e-01 1.29580903e+00 -3.16343367e-01
5.56575544e-02 1.41319245e-01 -4.39782739e-01 4.82599646e-01
5.73425055e-01 -7.31788516e-01 9.60301682e-02 -4.54945982e-01
-5.23199793e-03 -3.70954365e-01 4.96965915e-01 -1.41247118e+00
5.82894683e-01 5.91129297e-03 2.65384525e-01 -9.64028060e-01
5.41356564e-01 -9.45436656e-01 1.06438972e-01 5.75808525e-01
-1.70759305e-01 2.30706289e-01 2.37364858e-01 5.29221237e-01
-1.92116257e-02 -4.43012081e-02 1.39829338e+00 2.20617682e-01
-5.13023555e-01 7.54330575e-01 3.70698124e-01 3.35019916e-01
1.10371745e+00 -3.07212383e-01 -3.31502974e-01 2.24516168e-01
-3.88249934e-01 3.55689675e-01 2.62564629e-01 5.16019464e-01
4.78743672e-01 -1.78987384e+00 -7.76082575e-01 -4.49022688e-02
8.67509022e-02 -1.46395326e-01 1.17123924e-01 1.11888063e+00
-4.14009094e-01 6.00836635e-01 -4.89805341e-01 -8.12564492e-01
-1.39956677e+00 2.97873944e-01 6.15755081e-01 -3.02378356e-01
-4.72234875e-01 9.10112321e-01 -4.15536352e-02 -3.91334385e-01
5.31722844e-01 -3.42997015e-01 -1.71877891e-01 8.25712085e-03
5.03551602e-01 7.83257902e-01 1.40458625e-02 -1.02030253e+00
-6.92127526e-01 1.14485466e+00 -1.63432300e-01 1.88755944e-01
1.36291540e+00 -7.83308223e-02 1.11475393e-01 6.54441789e-02
1.55478930e+00 -4.50217336e-01 -1.60472667e+00 -5.37007570e-01
8.88387039e-02 -4.99330074e-01 5.66143282e-02 -4.18440998e-01
-1.38720393e+00 7.82708943e-01 1.14143789e+00 -1.29550949e-01
7.10137308e-01 -1.54847920e-01 1.14514387e+00 2.29702696e-01
1.93610430e-01 -1.13617945e+00 2.41238758e-01 5.01487195e-01
6.01918221e-01 -1.46320879e+00 1.49198592e-01 7.44756088e-02
-1.03049554e-01 9.57109511e-01 7.59801686e-01 -2.49683708e-01
8.43948185e-01 3.50495726e-02 -2.57728938e-02 -3.26743364e-01
-1.16718672e-01 -5.92700362e-01 7.36612797e-01 5.88724136e-01
2.04286501e-01 -1.03136577e-01 -2.75857329e-01 1.04881644e-01
4.55142930e-02 -3.07627976e-01 -1.96065336e-01 9.07327235e-01
-5.71413994e-01 -8.53344619e-01 -3.95281851e-01 2.15613022e-01
-3.07304829e-01 3.19334328e-01 -8.33984744e-03 8.79892051e-01
3.99320632e-01 7.84845114e-01 6.46115169e-02 -3.26508343e-01
6.11384571e-01 -3.99768621e-01 6.84581876e-01 -2.10584864e-01
-7.11009026e-01 -1.65016279e-01 -5.35189629e-01 -6.19327605e-01
-6.48361444e-01 -9.43867147e-01 -1.19687235e+00 -3.83967817e-01
-6.81369245e-01 -5.11374660e-02 8.23855639e-01 9.66511190e-01
2.07283780e-01 5.09356558e-01 6.75079882e-01 -8.31705213e-01
-7.88875103e-01 -8.03426445e-01 -4.38521564e-01 5.05294323e-01
4.82279032e-01 -1.11580861e+00 -3.81736457e-01 -4.45927888e-01] | [6.312716960906982, -2.1321845054626465] |
7debb7fa-eba1-48ca-b843-64c500ca1cd6 | scir-qa-at-semeval-2017-task-3-cnn-model | null | null | https://aclanthology.org/S17-2049 | https://aclanthology.org/S17-2049.pdf | SCIR-QA at SemEval-2017 Task 3: CNN Model Based on Similar and Dissimilar Information between Keywords for Question Similarity | We describe a method of calculating the similarity of questions in community QA. Question in cQA are usually very long and there are a lot of useless information about calculating the similarity of questions. Therefore,we implement a CNN model based on similar and dissimilar information between question{'}s keywords. We extract the keywords of questions, and then model the similar and dissimilar information between the keywords, and use the CNN model to calculate the similarity. | ['Ting Liu', 'Le Qi', 'Yu Zhang'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['graph-ranking', 'question-similarity'] | ['graphs', 'natural-language-processing'] | [-6.96313202e-01 -4.49859411e-01 3.68570745e-01 -6.40245140e-01
-6.42073154e-01 -6.81326687e-01 3.24446052e-01 6.46701574e-01
-7.45515883e-01 2.53855765e-01 7.39245534e-01 -1.85270328e-02
-3.83477211e-01 -1.08765042e+00 -1.92883953e-01 -2.09694684e-01
3.16344649e-01 4.96781677e-01 5.91623008e-01 -8.32519293e-01
6.39815569e-01 1.54984757e-01 -1.27751791e+00 6.16698802e-01
1.23414052e+00 1.09831679e+00 3.21720302e-01 7.39120066e-01
-9.06621337e-01 9.99159455e-01 -1.15931714e+00 -4.39287335e-01
-2.26342455e-01 -8.42487872e-01 -1.55772007e+00 -3.54465157e-01
3.81323040e-01 -3.76629740e-01 -7.08717763e-01 1.11560261e+00
7.92902172e-01 3.14782739e-01 5.72514653e-01 -1.12964761e+00
-1.07901812e+00 4.52243239e-01 2.17244700e-01 7.52759755e-01
8.58539283e-01 -1.98587716e-01 1.19295776e+00 -7.08274126e-01
5.49122870e-01 1.43828595e+00 7.96021521e-01 2.48316437e-01
-2.25369915e-01 -3.93012643e-01 -3.84393841e-01 6.90105140e-01
-1.51181221e+00 8.78071412e-02 5.03304958e-01 -2.94246357e-02
8.70979071e-01 3.22620362e-01 5.00462413e-01 4.30313408e-01
9.58795547e-02 8.11490357e-01 6.15464866e-01 -3.61640677e-02
-1.35392249e-01 2.67689656e-02 4.19437021e-01 5.58223188e-01
-3.00608963e-01 -7.64676392e-01 1.09631807e-01 -3.16367209e-01
2.74581075e-01 3.08825254e-01 -3.61998856e-01 4.89217520e-01
-9.98300016e-01 1.31929779e+00 8.32235456e-01 8.21523428e-01
-2.64507264e-01 -7.47230798e-02 5.24893880e-01 6.42537832e-01
9.57397521e-02 6.39661372e-01 -3.50173354e-01 7.76963606e-02
-8.88138175e-01 7.78650701e-01 1.04541254e+00 1.16098559e+00
8.96957040e-01 -8.55114698e-01 -6.70643389e-01 8.82602990e-01
4.24451679e-01 4.01723146e-01 6.20860338e-01 -1.33197773e+00
4.85995203e-01 9.01317656e-01 -4.77696583e-02 -1.65430439e+00
-4.13099140e-01 -2.35503033e-01 -9.22344744e-01 -8.75068247e-01
2.25794539e-01 -1.48323789e-01 -4.14998472e-01 1.17125547e+00
2.19124660e-01 -3.50120813e-01 5.76753244e-02 7.12476254e-01
2.08255768e+00 9.17554259e-01 -2.44541764e-01 -5.92640527e-02
1.45666337e+00 -1.06116021e+00 -1.20745218e+00 4.06984895e-01
7.54453838e-01 -1.22756743e+00 9.90796447e-01 -2.43810460e-01
-1.21311486e+00 -7.14149237e-01 -4.35798436e-01 -7.79100239e-01
-7.16774285e-01 -1.40374690e-01 1.79935306e-01 1.18929148e-01
-1.35486925e+00 8.90786767e-01 6.21963143e-02 -5.87786019e-01
3.79281372e-01 2.17360854e-01 4.47827987e-02 -3.26489717e-01
-1.89618635e+00 9.42833364e-01 3.85268778e-01 -6.25070632e-02
-5.91500461e-01 -3.26026201e-01 -8.88876557e-01 4.61275160e-01
-2.78482065e-02 -8.14753652e-01 1.48799694e+00 -3.26796383e-01
-7.92014897e-01 4.39240664e-01 -2.35935599e-01 -1.08955450e-01
5.71530908e-02 1.13037303e-01 -6.45114839e-01 4.46966201e-01
3.33295435e-01 8.64159405e-01 2.14674234e-01 -7.99207032e-01
-7.19835699e-01 -1.26307994e-01 5.86587727e-01 3.63064170e-01
-5.53304374e-01 4.78761107e-01 -5.68265676e-01 -2.63691783e-01
1.69576690e-01 -2.77529806e-01 -3.29035163e-01 -1.28069103e-01
-2.45714724e-01 -1.11533356e+00 7.50111759e-01 -9.66737926e-01
1.46997190e+00 -1.77215409e+00 -4.37570125e-01 -3.66161466e-02
6.80331826e-01 3.06343675e-01 -3.87896508e-01 8.96705985e-01
6.56938255e-01 1.30265847e-01 -2.23589048e-01 2.46420160e-01
-1.34085342e-01 3.04605901e-01 -9.39925015e-02 -1.42325312e-02
-5.89190833e-02 1.06016254e+00 -1.29343390e+00 -1.05511379e+00
-4.57861394e-01 3.76966596e-01 -5.07602751e-01 6.73261821e-01
-2.76478469e-01 2.08419070e-01 -8.67754579e-01 2.94897676e-01
6.79784596e-01 -5.18644869e-01 -4.03065205e-01 -3.73003870e-01
2.89349109e-01 6.60775840e-01 -7.48338163e-01 1.62850010e+00
-2.87326515e-01 8.54429841e-01 -2.04467386e-01 -7.65281320e-01
9.45714056e-01 4.55406278e-01 2.62843877e-01 -7.31361330e-01
5.12503386e-01 1.91714719e-01 6.52046129e-02 -1.34211886e+00
7.46228933e-01 1.55567065e-01 1.59054741e-01 7.39737272e-01
1.69501722e-01 -7.32531369e-01 3.33129317e-01 7.01329172e-01
1.07410538e+00 -5.71874917e-01 -5.51885739e-02 -3.67543519e-01
8.90434504e-01 7.01545775e-02 1.48293361e-01 7.17008889e-01
-3.54145437e-01 7.53214300e-01 3.55614901e-01 -3.44363064e-01
-7.87571609e-01 -7.54226327e-01 1.30715728e-01 9.83483493e-01
2.09658161e-01 -5.48690498e-01 -8.52799654e-01 -8.01583648e-01
-2.55442667e-03 2.02978969e-01 -7.18251407e-01 -2.64095187e-01
-5.85876286e-01 -5.20248935e-02 8.22391272e-01 4.31941420e-01
1.05393648e+00 -1.24920154e+00 7.23415688e-02 2.67881006e-01
-9.03429925e-01 -7.19494700e-01 -1.00916457e+00 -4.49505210e-01
-8.55190635e-01 -1.22127867e+00 -7.74217784e-01 -1.47620285e+00
3.05239290e-01 7.77336538e-01 1.60808265e+00 8.09198081e-01
1.51416669e-02 3.10012281e-01 -5.12730479e-01 -1.56825498e-01
-1.15821473e-01 1.81761831e-01 -4.35113579e-01 -3.90130758e-01
7.65969932e-01 -4.36624646e-01 -1.00714219e+00 1.76513955e-01
-1.03977549e+00 -9.09734070e-01 3.70033607e-02 6.12113416e-01
1.08853064e-01 2.09944174e-01 7.07715869e-01 -3.49702030e-01
1.72979379e+00 -6.66598141e-01 7.50638023e-02 4.61646676e-01
-4.61672157e-01 -1.57033145e-01 7.15381980e-01 -3.21668804e-01
-8.97952139e-01 -7.70121455e-01 -4.71152872e-01 3.82853150e-02
-3.43326442e-02 7.10810483e-01 -1.39874667e-01 -1.31234184e-01
4.94005054e-01 3.09576154e-01 -2.36757576e-01 -6.98759556e-01
6.73312783e-01 1.30069864e+00 3.81959200e-01 -1.75362766e-01
4.34097022e-01 4.49751407e-01 -5.95557749e-01 -4.34878796e-01
-1.09638119e+00 -9.55317140e-01 -4.22551662e-01 -1.09366789e-01
9.67154324e-01 -5.17942786e-01 -1.23468494e+00 2.95514435e-01
-1.85461509e+00 6.09779894e-01 -1.84477925e-01 5.62382042e-01
-1.20790675e-01 8.59490395e-01 -1.12605882e+00 -1.07304908e-01
-7.47360587e-01 -8.86769652e-01 5.92452705e-01 8.58128071e-01
-1.67650551e-01 -1.26744783e+00 4.68694627e-01 6.58777416e-01
7.48221397e-01 -5.72767437e-01 9.53075290e-01 -1.24388230e+00
-4.73926105e-02 -2.72381306e-01 -4.85448867e-01 4.33075637e-01
3.17497820e-01 -5.45098446e-02 -6.77574635e-01 -9.27273110e-02
5.13117731e-01 -4.17050660e-01 7.61531472e-01 8.02985430e-02
1.35039997e+00 -4.81421828e-01 6.39750287e-02 -1.22542595e-02
1.06457233e+00 4.32054810e-02 7.57213652e-01 2.71478631e-02
5.69310844e-01 8.73938084e-01 5.86855829e-01 2.77993977e-01
1.07225347e+00 1.81397989e-01 2.82591820e-01 1.16258375e-01
1.01065420e-01 -1.52776599e-01 -1.91786170e-01 1.69550622e+00
1.88827366e-01 -4.67984200e-01 -8.13118219e-01 9.07313943e-01
-1.46749139e+00 -1.08596265e+00 -8.35304976e-01 1.54742944e+00
1.08086467e+00 -4.01924402e-01 -1.82701111e-01 -1.75226882e-01
9.47952688e-01 1.11503258e-01 -1.19044527e-01 -4.95548159e-01
-1.68721288e-01 2.45749399e-01 -1.40834421e-01 4.22681302e-01
-6.18657947e-01 6.61626995e-01 7.36162281e+00 9.30000842e-01
-5.18455148e-01 1.54536128e-01 3.02133441e-01 1.90631047e-01
-7.98611224e-01 1.62220262e-02 -6.26267195e-01 6.09812677e-01
8.00607085e-01 -4.11941707e-01 1.67406023e-01 4.83911246e-01
2.12845542e-02 8.40639994e-02 -8.24441135e-01 8.87568295e-01
5.36267757e-01 -1.42083466e+00 3.95842731e-01 -4.11374539e-01
6.43240929e-01 -1.62683974e-03 -3.19595724e-01 4.89334106e-01
5.06134272e-01 -1.09107494e+00 -1.27729893e-01 8.05379689e-01
-1.54353261e-01 -6.92659557e-01 1.34564638e+00 5.09419262e-01
-1.15912187e+00 1.05936311e-01 -1.03048086e+00 -1.11798562e-01
-5.45517579e-02 5.10011494e-01 -4.82201666e-01 6.60702050e-01
1.16154826e+00 7.32201159e-01 -8.55832577e-01 1.41906357e+00
-1.20206706e-01 4.27195042e-01 -4.23187874e-02 -5.14492691e-01
6.81546211e-01 -3.37338060e-01 -1.06271543e-01 1.13168871e+00
3.71625841e-01 2.69007146e-01 -1.67958468e-01 7.73790002e-01
-6.04031324e-01 6.43150449e-01 -3.38263273e-01 -5.61817698e-02
7.93398201e-01 1.41050577e+00 -3.70122045e-01 -7.02081561e-01
-5.50516725e-01 8.92405987e-01 2.22753465e-01 5.42374738e-02
-5.04741907e-01 -1.12431312e+00 1.88869759e-01 -3.29421729e-01
1.69863477e-01 1.55603066e-01 1.45036504e-01 -8.26959789e-01
1.28567100e-01 -9.11796033e-01 9.59817410e-01 -1.19741404e+00
-1.83179307e+00 6.36028230e-01 -2.57383466e-01 -1.33448744e+00
-2.71202214e-02 -1.09084293e-01 -1.15669882e+00 1.07062364e+00
-1.39537525e+00 -7.18999386e-01 -6.71076655e-01 1.08633459e+00
2.71030784e-01 1.13895781e-01 5.51067293e-01 5.27104855e-01
9.84119102e-02 3.05883348e-01 4.41615164e-01 6.12778604e-01
6.40143335e-01 -1.07662845e+00 5.01815796e-01 -3.24754491e-02
-5.09707592e-02 9.82111752e-01 4.64910001e-01 -5.42001426e-01
-8.24850261e-01 -7.95501709e-01 1.82362926e+00 -4.76736337e-01
5.49163222e-01 1.73432946e-01 -1.35834789e+00 3.54111820e-01
1.02704763e+00 -4.84635979e-02 9.86936510e-01 -2.88450480e-01
-3.51981014e-01 -3.23801562e-02 -1.31968772e+00 5.15394211e-01
7.26439238e-01 -9.57786322e-01 -1.33720160e+00 7.72365928e-01
1.39817595e+00 1.44672528e-01 -1.22968495e+00 1.26999930e-01
1.39290437e-01 -7.36298800e-01 9.88516629e-01 -1.08594525e+00
4.83356088e-01 -2.61970639e-01 -1.14672147e-01 -1.28525794e+00
-3.99447680e-01 -5.99696487e-02 2.33197540e-01 1.32690358e+00
2.89887667e-01 -4.09076214e-01 5.52829146e-01 1.67055920e-01
-1.44525379e-01 -7.07224965e-01 -7.81288564e-01 -5.37579000e-01
5.41844308e-01 2.16499805e-01 1.15618742e+00 1.36331761e+00
3.70160453e-02 6.54989541e-01 2.41198912e-01 -2.42226392e-01
-7.14335516e-02 3.83760899e-01 1.91966802e-01 -1.29214919e+00
2.18646571e-01 -4.30812120e-01 -1.47775024e-01 -1.39069068e+00
-1.20276570e-01 -8.33449721e-01 1.63676292e-01 -2.10488200e+00
5.11395097e-01 -1.42423719e-01 -2.34581977e-01 -2.52914459e-01
-6.21326387e-01 -1.43757295e-02 6.30584732e-02 5.24631143e-01
-1.14767218e+00 9.83805299e-01 1.76991582e+00 -3.95955294e-01
2.15002254e-01 -1.09781809e-01 -6.03953183e-01 5.56844532e-01
7.28019655e-01 -6.42181635e-01 -2.93791205e-01 -7.93027103e-01
6.02046251e-01 7.94340968e-02 5.39721213e-02 -7.50702262e-01
7.67259240e-01 -6.44823536e-02 2.50295073e-01 -1.09550977e+00
-1.45628816e-02 -6.95805132e-01 -7.13476777e-01 5.95535517e-01
-9.10324275e-01 6.03184819e-01 -5.31033278e-01 2.25339144e-01
-6.66288853e-01 -8.20712745e-01 3.47035825e-01 -6.52676523e-01
-5.11671007e-02 4.97167170e-01 -5.11624217e-01 6.81450009e-01
1.66529104e-01 1.73278823e-01 -4.07789558e-01 -1.05364525e+00
-5.18216252e-01 8.25276732e-01 3.88949849e-02 4.80215490e-01
8.98035824e-01 -1.66373229e+00 -9.25429940e-01 -5.57562947e-01
8.02761391e-02 4.24111187e-02 4.08000559e-01 5.28129697e-01
-6.64114177e-01 5.85041285e-01 -1.66741863e-01 -2.40835682e-01
-1.25654292e+00 6.45086706e-01 5.89612484e-01 -3.02183598e-01
9.04283822e-02 9.57264602e-01 -2.28064671e-01 -1.12621784e+00
-5.58663206e-03 -1.98479816e-01 -1.34000182e+00 3.65689367e-01
8.10278296e-01 3.76966923e-01 9.48404819e-02 -6.23445928e-01
-1.93067193e-01 5.75701058e-01 -3.47162068e-01 4.85907635e-03
1.09173465e+00 -3.89233440e-01 -9.11029458e-01 1.46973541e-03
1.78988969e+00 -2.81359851e-01 -2.18722358e-01 -6.66611791e-01
1.04614804e-02 -4.00797755e-01 -2.36376733e-01 -3.10902685e-01
-1.18296599e+00 1.43753433e+00 6.37682021e-01 6.03907108e-01
9.30545688e-01 1.78940445e-01 1.53363180e+00 1.00650609e+00
-1.94325626e-01 -1.26584601e+00 5.09988606e-01 1.09087157e+00
8.86740446e-01 -1.00702691e+00 -6.04857877e-02 -2.68356502e-01
-2.90402263e-01 1.07317054e+00 6.14430368e-01 1.30884303e-03
7.85326242e-01 -5.61576903e-01 4.12595868e-01 -8.41974139e-01
-4.97410327e-01 -4.87302035e-01 3.84503394e-01 6.73918724e-01
5.25369644e-01 -5.66826820e-01 -8.17805171e-01 5.84443808e-01
-3.66809428e-01 -1.68507412e-01 4.98875856e-01 9.77188826e-01
-1.01464403e+00 -7.20329881e-01 -1.67856157e-01 6.28779233e-01
-5.97477078e-01 -5.83330452e-01 -6.18617654e-01 3.02173913e-01
1.55929357e-01 1.60692668e+00 3.59986037e-01 -6.18286133e-01
1.67073905e-01 -1.74602598e-01 7.07112029e-02 -4.93217885e-01
-1.28266883e+00 -6.71350181e-01 -1.37450397e-02 -3.60213757e-01
-3.54889870e-01 -2.39734977e-01 -1.34023583e+00 -6.30243361e-01
-4.71423119e-01 8.33323598e-01 4.07900006e-01 1.33503711e+00
3.49860966e-01 1.25998825e-01 1.08143997e+00 4.03098673e-01
-5.61770499e-01 -1.36531484e+00 -3.03128958e-01 8.40632379e-01
-7.72534590e-03 -3.78300808e-02 -4.16240394e-01 -3.78377289e-01] | [11.243217468261719, 8.010148048400879] |
0824c9a6-0a37-4f18-b201-ea5e6855cb5d | cut-inner-layers-a-structured-pruning | 2206.14658 | null | https://arxiv.org/abs/2206.14658v1 | https://arxiv.org/pdf/2206.14658v1.pdf | Cut Inner Layers: A Structured Pruning Strategy for Efficient U-Net GANs | Pruning effectively compresses overparameterized models. Despite the success of pruning methods for discriminative models, applying them for generative models has been relatively rarely approached. This study conducts structured pruning on U-Net generators of conditional GANs. A per-layer sensitivity analysis confirms that many unnecessary filters exist in the innermost layers near the bottleneck and can be substantially pruned. Based on this observation, we prune these filters from multiple inner layers or suggest alternative architectures by completely eliminating the layers. We evaluate our approach with Pix2Pix for image-to-image translation and Wav2Lip for speech-driven talking face generation. Our method outperforms global pruning baselines, demonstrating the importance of properly considering where to prune for U-Net generators. | ['Hancheol Park', 'Shinkook Choi', 'Bo-Kyeong Kim'] | 2022-06-29 | null | null | null | null | ['talking-face-generation'] | ['computer-vision'] | [ 5.83363533e-01 6.96471930e-01 -1.75169762e-02 -3.96826476e-01
-9.60497499e-01 -4.37656432e-01 6.02150798e-01 -5.87606847e-01
-1.70647383e-01 9.30396855e-01 4.93584424e-01 -5.34967005e-01
1.96372807e-01 -8.01076829e-01 -8.54811609e-01 -5.23297012e-01
3.86730552e-01 4.73682135e-01 -9.97398570e-02 2.11030334e-01
-2.03035817e-01 4.53140676e-01 -1.28741217e+00 5.08930266e-01
5.01968741e-01 7.10257649e-01 1.63707197e-01 8.19509029e-01
-9.54563543e-02 7.96465516e-01 -9.42894340e-01 -8.85167301e-01
3.67799282e-01 -8.57017577e-01 -5.79892635e-01 7.35477656e-02
6.68126822e-01 -7.23515570e-01 -1.37442052e-01 7.47615695e-01
7.68434644e-01 1.35406675e-02 6.16402984e-01 -9.97643054e-01
-4.13765043e-01 1.09426892e+00 -6.65826380e-01 2.50187010e-01
-1.07198030e-01 2.46022135e-01 9.99844193e-01 -5.85080504e-01
6.51063323e-01 1.35867739e+00 6.60677075e-01 8.21138620e-01
-1.63094532e+00 -8.55141163e-01 1.49803042e-01 -3.28990817e-01
-1.33416057e+00 -9.18543935e-01 5.53459466e-01 -1.97634116e-01
1.11425614e+00 3.70081127e-01 5.97151518e-01 1.44935524e+00
-4.69155908e-02 6.18741274e-01 8.61075103e-01 -5.10163069e-01
8.53854343e-02 3.93362679e-02 -2.91255265e-01 7.33063877e-01
5.04816413e-01 4.23121080e-02 -6.44345641e-01 -1.76998869e-01
9.99742508e-01 -6.52121007e-01 2.68644970e-02 2.07514048e-01
-5.84480941e-01 8.08018565e-01 -6.52519893e-03 -2.28295177e-02
-2.61372298e-01 6.73229277e-01 2.17047945e-01 5.68848476e-02
6.23624504e-01 5.18302441e-01 -2.85966277e-01 -3.66487414e-01
-1.52229774e+00 3.00319791e-01 5.99680603e-01 1.30932534e+00
6.25022411e-01 5.64794779e-01 -5.58400333e-01 9.56009388e-01
-9.42329094e-02 2.62523174e-01 1.16916239e-01 -1.16273224e+00
3.82751763e-01 2.16833457e-01 -2.22810209e-01 -4.12735999e-01
2.02083196e-02 -6.12741530e-01 -6.20709479e-01 1.86893240e-01
9.22798663e-02 -7.19810307e-01 -1.30150962e+00 1.80217230e+00
5.39229577e-03 1.96001008e-01 -1.68125853e-01 4.99353677e-01
6.41012847e-01 5.90915084e-01 2.64193505e-01 -5.31528406e-02
1.11388063e+00 -8.94191325e-01 -4.04650867e-01 -4.46840197e-01
2.80738205e-01 -8.81965160e-01 1.07694209e+00 3.31781209e-01
-1.40728128e+00 -4.56750631e-01 -9.27796543e-01 -2.11900651e-01
2.36099675e-01 5.65687656e-01 7.81602085e-01 9.20044363e-01
-1.34384692e+00 5.51475286e-01 -1.00374210e+00 -2.10176826e-01
8.22910309e-01 3.64662200e-01 -2.50100829e-02 9.64130461e-03
-7.48707473e-01 7.82429874e-01 1.66834101e-01 -2.22087372e-02
-1.16924357e+00 -6.91997349e-01 -5.95548451e-01 3.42923850e-01
6.14382401e-02 -1.11962938e+00 1.31379604e+00 -8.97241354e-01
-1.62059820e+00 4.74105269e-01 -4.68117505e-01 -7.90928006e-01
5.51023901e-01 -2.46368110e-01 -7.03447834e-02 1.76012412e-01
-1.85840219e-01 1.33375180e+00 1.09803534e+00 -1.14567375e+00
-5.42932928e-01 1.90608039e-01 7.76749700e-02 1.36003673e-01
-3.44130337e-01 -2.56458428e-02 -5.89921534e-01 -1.04335427e+00
-1.65750653e-01 -7.31592357e-01 -2.04936936e-01 -4.56761658e-01
-6.98890924e-01 1.39798969e-01 9.03970063e-01 -7.10827231e-01
1.19840610e+00 -1.84056222e+00 -1.57404363e-01 1.13337949e-01
9.49900448e-02 3.83563265e-02 -3.93468350e-01 6.56869486e-02
2.26947125e-02 5.07918477e-01 -1.41542971e-01 -9.05456543e-01
-1.89848572e-01 3.02166164e-01 -4.46866333e-01 -1.03006132e-01
4.44178790e-01 9.60744083e-01 -4.50181961e-01 -5.60048878e-01
-4.45395708e-04 7.92211592e-01 -9.53993738e-01 2.39557233e-02
-4.42160994e-01 1.82980299e-01 -1.36580333e-01 6.29807115e-01
5.73942721e-01 -7.86046404e-03 1.58760935e-01 -6.79258853e-02
1.96216136e-01 6.67925358e-01 -7.17340052e-01 1.36575949e+00
-5.26419222e-01 8.47688198e-01 1.50978670e-01 -5.46349823e-01
8.44606340e-01 2.22734973e-01 1.64896354e-01 -3.62852991e-01
1.09715834e-01 -1.11905232e-01 8.42080414e-02 -1.66711994e-02
5.39646685e-01 -3.47074538e-01 2.82363713e-01 4.15272117e-01
3.99430543e-01 -3.16018730e-01 2.86419094e-01 1.80026069e-01
1.17114735e+00 3.30207407e-01 -1.78390115e-01 -2.18160182e-01
-7.58819953e-02 -2.67978698e-01 6.38368130e-01 8.51116478e-01
2.56502867e-01 1.07901394e+00 8.32488418e-01 -3.60432677e-02
-1.21816981e+00 -1.12637579e+00 -6.40421063e-02 9.86077487e-01
-6.19759977e-01 -8.26335490e-01 -1.21159661e+00 -5.97094417e-01
-2.84606516e-01 1.09963155e+00 -5.74334085e-01 -6.72309101e-02
-6.55069947e-01 -7.66652882e-01 1.07148588e+00 6.46340549e-01
5.30902803e-01 -8.99227440e-01 -6.78860366e-01 1.16339222e-01
-1.92324489e-01 -1.10716808e+00 -3.17462653e-01 2.93711036e-01
-1.04937971e+00 -7.32325554e-01 -7.00458229e-01 -4.26122606e-01
7.38672376e-01 -1.74828276e-01 1.33944023e+00 -1.93169475e-01
-1.90717325e-01 2.07430914e-01 -2.33238917e-02 -5.40817499e-01
-7.00093985e-01 5.62596619e-01 -2.44692922e-01 -3.60424966e-01
2.73772895e-01 -7.61219978e-01 -5.07415056e-01 7.20510706e-02
-5.65443814e-01 4.16541964e-01 7.97071874e-01 6.98799133e-01
6.86830640e-01 -7.04721408e-03 5.04523158e-01 -1.11908793e+00
7.63073027e-01 -1.25343442e-01 -6.14621878e-01 2.05194354e-01
-4.10380542e-01 1.87915072e-01 4.63299632e-01 -3.28671515e-01
-1.33579123e+00 -1.00996412e-01 -2.31164336e-01 -6.16566062e-01
-1.35315225e-01 1.26389250e-01 -3.57389003e-01 2.05840781e-01
5.72612345e-01 -1.94639340e-01 -1.34027347e-01 -4.50684011e-01
4.95841026e-01 2.99851686e-01 4.99435067e-01 -8.62431228e-01
6.35928154e-01 3.15026075e-01 -5.48515245e-02 -1.00212550e+00
-4.58768070e-01 3.99003774e-01 -1.58421889e-01 1.70542542e-02
7.25381613e-01 -1.03188324e+00 -1.84383035e-01 -5.69269173e-02
-1.14247024e+00 -6.09167516e-01 -6.99303746e-01 2.29527146e-01
-5.06905556e-01 -7.96885490e-02 -7.13264287e-01 -7.55370319e-01
-4.59391624e-01 -1.07596147e+00 9.72163796e-01 1.33217320e-01
-7.31376052e-01 -6.15384161e-01 -1.25429094e-01 3.86293530e-01
4.30168360e-01 1.25634119e-01 9.91767406e-01 -3.49227279e-01
-6.40395641e-01 5.58630340e-02 -2.47859024e-02 6.88050210e-01
9.30008814e-02 3.40601653e-01 -1.19385517e+00 -1.27907559e-01
-6.15322031e-02 -2.54897237e-01 1.19656539e+00 7.62435079e-01
1.27789271e+00 -6.13028705e-01 -3.76110703e-01 9.34795678e-01
1.03622198e+00 7.02264830e-02 8.34467649e-01 -1.16250791e-01
6.46014750e-01 3.93947840e-01 -1.50433302e-01 3.17273200e-01
-1.45526631e-02 3.24802935e-01 1.56699702e-01 -2.23961994e-01
-5.02019942e-01 -5.30400097e-01 4.41997111e-01 4.16619390e-01
-8.00760984e-02 -5.01742423e-01 -6.93022549e-01 5.63664556e-01
-1.20312560e+00 -8.61931205e-01 3.72787327e-01 1.88149762e+00
8.77536774e-01 3.81084889e-01 -1.14682555e-01 -2.42167950e-01
6.59421027e-01 2.60948956e-01 -3.57065320e-01 -6.61597192e-01
-2.76002824e-01 7.22926259e-01 5.36555827e-01 6.85641348e-01
-8.40446055e-01 1.18398881e+00 7.53733397e+00 1.01331449e+00
-9.01502669e-01 1.90457910e-01 1.27050984e+00 -9.25680578e-01
-6.40547931e-01 2.18229681e-01 -1.20233583e+00 3.01562518e-01
9.50840235e-01 1.73811659e-01 5.33725381e-01 8.37508142e-01
1.99762061e-01 -2.75196463e-01 -1.13717353e+00 7.04600930e-01
3.94665673e-02 -1.51979840e+00 3.90151381e-01 1.53244302e-01
1.00579202e+00 -1.99479098e-03 2.78336227e-01 2.49919817e-01
7.29819417e-01 -1.36823010e+00 7.81372488e-01 2.64862627e-01
1.08475876e+00 -1.00334525e+00 2.87825704e-01 1.03528788e-02
-7.40259111e-01 1.41521707e-01 -5.20669043e-01 1.23678043e-01
3.45497489e-01 7.75514185e-01 -1.22229707e+00 2.21965745e-01
5.32154739e-01 1.19171545e-01 -5.98685443e-01 6.09375358e-01
-5.26300490e-01 1.28882277e+00 -5.93136907e-01 4.24727082e-01
1.49686530e-01 -1.53769240e-01 5.25784671e-01 1.40360057e+00
5.31450748e-01 -8.19252618e-03 -6.05154157e-01 1.26491368e+00
-3.57768327e-01 -3.02877992e-01 -5.88829577e-01 -2.67407805e-01
4.79907870e-01 1.00578141e+00 -7.37396717e-01 -2.73500949e-01
-2.20082775e-01 8.03476870e-01 1.98202401e-01 6.28428578e-01
-1.04091942e+00 -6.46658391e-02 8.16861808e-01 4.07153338e-01
6.11182094e-01 -1.29926771e-01 -7.39905238e-01 -8.26030076e-01
-1.77716970e-01 -9.22215343e-01 2.26175070e-01 -7.26378918e-01
-7.45112717e-01 6.90269649e-01 2.20510617e-01 -4.80670661e-01
-4.42685962e-01 -1.73041508e-01 -8.00837815e-01 9.93332744e-01
-1.03571534e+00 -1.24313045e+00 -7.76063534e-04 3.28531832e-01
4.01669115e-01 -1.73482910e-01 7.32262552e-01 1.18262723e-01
-6.71069741e-01 1.10351992e+00 -2.46479109e-01 1.49709117e-02
4.18787628e-01 -7.97206044e-01 7.80215025e-01 1.22787321e+00
2.72724777e-01 8.56252789e-01 5.82560003e-01 -8.53797257e-01
-7.34663486e-01 -1.12062836e+00 6.91991329e-01 -3.55310202e-01
6.51220679e-02 -6.39929056e-01 -5.42308033e-01 9.71138656e-01
4.77872849e-01 -5.34007668e-01 6.81860626e-01 1.92832783e-01
-1.43397704e-01 -1.24744996e-01 -1.06138217e+00 8.10166419e-01
1.12358546e+00 -3.49256843e-01 -1.01983771e-01 2.07200367e-02
7.35155523e-01 -2.80527294e-01 -3.87457281e-01 4.38582271e-01
4.05585796e-01 -9.71095622e-01 9.02173162e-01 -3.70818675e-01
7.49838233e-01 1.83796007e-02 -1.07182205e-01 -1.15872276e+00
-1.98568210e-01 -9.04201686e-01 -2.33169019e-01 1.64542818e+00
6.36929333e-01 -2.71383464e-01 1.32412434e+00 4.79313523e-01
-1.84691191e-01 -8.64141583e-01 -1.00168598e+00 -4.71629411e-01
1.11315213e-01 -4.42335665e-01 7.66100466e-01 4.18089449e-01
-6.18345261e-01 4.81679738e-01 -4.75461543e-01 -1.65399268e-01
4.19614971e-01 -3.34647894e-01 8.42457950e-01 -6.38353407e-01
-6.03732347e-01 -6.92076325e-01 -9.70480312e-03 -1.18388104e+00
9.87558439e-02 -6.45932555e-01 -5.51944003e-02 -1.65635788e+00
5.62533438e-02 -3.76562148e-01 2.06268296e-01 7.11768389e-01
-1.36189952e-01 4.95970458e-01 1.41804069e-01 -1.53332487e-01
1.85606301e-01 4.83014643e-01 1.08197832e+00 -1.70206681e-01
-6.91939369e-02 -6.50172979e-02 -9.75698292e-01 7.67695963e-01
9.34276044e-01 -3.35805684e-01 -8.63615632e-01 -7.49368608e-01
1.50726676e-01 -2.99291134e-01 3.60627145e-01 -1.06519043e+00
-1.75342515e-01 -4.36171219e-02 5.87900400e-01 -6.75538003e-01
4.48306859e-01 -5.43927550e-01 3.96293491e-01 1.60315230e-01
-2.72795707e-01 8.66279304e-02 4.95205164e-01 1.83088422e-01
-1.01935178e-01 -2.62118548e-01 8.77018511e-01 -1.32877886e-01
-1.79709837e-01 6.87388256e-02 -4.31965202e-01 1.22451097e-01
6.26817644e-01 -5.06107450e-01 -2.78353423e-01 -5.88166118e-01
-6.95125639e-01 -9.10257921e-02 4.24172044e-01 8.82612541e-02
5.29378653e-01 -1.03196883e+00 -7.57683933e-01 4.01122123e-01
-6.04842484e-01 3.22229564e-01 1.60101801e-01 4.85474944e-01
-6.69594884e-01 4.18443948e-01 -6.15011416e-02 -3.17571968e-01
-1.27818537e+00 5.51427640e-02 3.39140415e-01 -3.33834082e-01
-6.44001067e-01 1.40630054e+00 3.96808445e-01 1.03485234e-01
1.87050298e-01 -4.11495626e-01 3.41940463e-01 6.51951581e-02
1.41810641e-01 5.29540002e-01 7.08397850e-02 -2.81540424e-01
-1.00522175e-01 5.04320264e-02 -1.50500879e-01 -4.69902754e-01
1.19302773e+00 1.10270113e-01 -9.35577229e-02 2.34863386e-02
9.57139909e-01 2.61410177e-01 -1.54468036e+00 3.18463355e-01
-5.34003437e-01 -2.43198141e-01 -3.93972434e-02 -8.35834026e-01
-1.29868782e+00 9.48211074e-01 1.50778130e-01 -3.76841992e-01
1.23538125e+00 -3.38105187e-02 6.94480479e-01 2.36092627e-01
1.11546494e-01 -1.05104232e+00 -1.89917862e-01 4.41780806e-01
8.21629882e-01 -4.45837528e-01 3.21532339e-01 -4.06180799e-01
-5.74844122e-01 7.62877285e-01 8.25077057e-01 1.84913054e-02
2.65693426e-01 8.83899212e-01 -1.28737718e-01 1.03171617e-01
-1.01458549e+00 1.50945708e-01 -6.53889626e-02 6.14947081e-01
5.48169494e-01 8.96500126e-02 -2.51922965e-01 4.92864341e-01
-8.71488750e-01 -1.89897478e-01 4.43107843e-01 6.63033247e-01
-1.70568272e-01 -1.22910428e+00 -2.39157438e-01 8.09474230e-01
-5.46689093e-01 -5.30275047e-01 -4.37689126e-01 7.49939740e-01
3.51547658e-01 7.91686356e-01 3.23752105e-01 -3.84611875e-01
9.84204374e-03 4.32518244e-01 7.15472877e-01 -6.22399032e-01
-7.39561677e-01 3.30815077e-01 3.46034884e-01 -3.34214002e-01
-7.21504539e-02 -6.97687209e-01 -7.36317813e-01 -5.06554484e-01
-3.65033537e-01 -5.53126931e-02 3.79712164e-01 6.05851948e-01
6.32339001e-01 8.09062541e-01 -4.93827611e-02 -6.09390616e-01
-3.24368000e-01 -1.11038971e+00 -1.93669230e-01 -1.60269260e-01
-3.15342518e-03 -4.03137326e-01 -1.81512013e-01 2.29901895e-01] | [11.372920989990234, -0.1926930695772171] |
8ccbd1f9-5637-4bb6-a13e-caf8ca8a2471 | sentiarabic-a-sentiment-analyzer-for-standard | null | null | https://aclanthology.org/L18-1195 | https://aclanthology.org/L18-1195.pdf | SentiArabic: A Sentiment Analyzer for Standard Arabic | null | ['Ramy er', 'Esk'] | 2018-05-01 | sentiarabic-a-sentiment-analyzer-for-standard-1 | https://aclanthology.org/L18-1195 | https://aclanthology.org/L18-1195.pdf | lrec-2018-5 | ['arabic-sentiment-analysis'] | ['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] | [-7.2081990242004395, 3.594397783279419] |
72713e3c-082d-471e-8d4d-5813b6f53f3e | high-quality-task-division-for-large-scale | 2208.10366 | null | https://arxiv.org/abs/2208.10366v1 | https://arxiv.org/pdf/2208.10366v1.pdf | High-quality Task Division for Large-scale Entity Alignment | Entity Alignment (EA) aims to match equivalent entities that refer to the same real-world objects and is a key step for Knowledge Graph (KG) fusion. Most neural EA models cannot be applied to large-scale real-life KGs due to their excessive consumption of GPU memory and time. One promising solution is to divide a large EA task into several subtasks such that each subtask only needs to match two small subgraphs of the original KGs. However, it is challenging to divide the EA task without losing effectiveness. Existing methods display low coverage of potential mappings, insufficient evidence in context graphs, and largely differing subtask sizes. In this work, we design the DivEA framework for large-scale EA with high-quality task division. To include in the EA subtasks a high proportion of the potential mappings originally present in the large EA task, we devise a counterpart discovery method that exploits the locality principle of the EA task and the power of trained EA models. Unique to our counterpart discovery method is the explicit modelling of the chance of a potential mapping. We also introduce an evidence passing mechanism to quantify the informativeness of context entities and find the most informative context graphs with flexible control of the subtask size. Extensive experiments show that DivEA achieves higher EA performance than alternative state-of-the-art solutions. | ['Xia Zhang', 'Genghong Zhao', 'Guido Zuccon', 'Wen Hua', 'Bing Liu'] | 2022-08-22 | null | null | null | null | ['entity-alignment', 'entity-alignment'] | ['knowledge-base', 'natural-language-processing'] | [ 2.04380214e-01 3.46796602e-01 -9.41709504e-02 -8.55088979e-02
-7.63031363e-01 -4.30645019e-01 5.17334819e-01 4.99935180e-01
-3.65919709e-01 7.74205804e-01 2.75501937e-01 -3.04987133e-01
-3.86218339e-01 -1.08574986e+00 -9.75817502e-01 -4.22563374e-01
-1.12906866e-01 6.86946750e-01 5.84902763e-01 -2.05347717e-01
1.91983040e-02 2.06167012e-01 -1.57735419e+00 3.25050265e-01
1.00820005e+00 8.79280984e-01 5.08646548e-01 3.65785658e-01
-4.22735304e-01 6.97980940e-01 -4.71210569e-01 -7.95188665e-01
1.29525840e-01 -1.74622700e-01 -1.07672000e+00 -3.70483100e-01
8.36902380e-01 9.59168747e-02 -1.87954083e-01 1.20285654e+00
5.68205535e-01 1.04916625e-01 3.84227693e-01 -1.42064130e+00
-3.86564702e-01 1.01118314e+00 -5.18649101e-01 3.39468420e-01
2.09001273e-01 -3.16835076e-01 1.34551191e+00 -8.64814878e-01
9.54539359e-01 7.89641201e-01 8.37501764e-01 2.14171678e-01
-1.11922455e+00 -5.47730803e-01 3.83601964e-01 5.23284912e-01
-1.52158546e+00 -2.94876993e-01 6.45113707e-01 -1.52326688e-01
1.45847714e+00 3.82828444e-01 7.44743466e-01 8.10227752e-01
2.39371099e-02 7.34591126e-01 7.43137658e-01 -5.03635228e-01
2.49269411e-01 -5.74722253e-02 2.92144150e-01 9.04756129e-01
5.69811821e-01 -4.11046982e-01 -7.12964118e-01 -3.42206359e-01
4.11312938e-01 -3.23582351e-01 -3.58094990e-01 -5.06083071e-01
-1.27149713e+00 4.75829899e-01 6.42250121e-01 3.66378844e-01
-4.99752730e-01 2.63079822e-01 2.92081386e-01 1.28239721e-01
1.91043153e-01 6.62153721e-01 -5.64261615e-01 8.61538574e-02
-8.19910944e-01 4.29812461e-01 9.14952338e-01 1.20656359e+00
8.42358053e-01 -4.54047441e-01 -3.23315293e-01 4.97666478e-01
8.10279250e-02 -3.31956297e-02 9.76584628e-02 -4.74653155e-01
7.45824933e-01 1.02135825e+00 -2.28574932e-01 -1.12431490e+00
-3.81242812e-01 -7.46701121e-01 -7.01713622e-01 -1.40805006e-01
3.14536452e-01 1.05879709e-01 -7.68935621e-01 1.88993049e+00
6.08462691e-01 3.64343017e-01 -9.05233398e-02 5.66776693e-01
8.35719824e-01 2.54674345e-01 1.85234070e-01 2.07689237e-02
1.57300723e+00 -1.03742313e+00 -5.33343017e-01 -6.09432101e-01
8.35746586e-01 -4.81073737e-01 8.16490471e-01 1.01215988e-01
-1.04941154e+00 -3.03719908e-01 -1.11877048e+00 -2.61344820e-01
-6.83660150e-01 1.54906511e-02 7.74079144e-01 3.64100993e-01
-1.12257981e+00 6.07287943e-01 -4.36061889e-01 -3.22402120e-01
3.34800363e-01 3.90446991e-01 -4.24316972e-01 -1.35066718e-01
-1.39630604e+00 1.03760707e+00 8.88469338e-01 1.43509477e-01
-2.95400977e-01 -8.00261736e-01 -7.35449374e-01 3.55130792e-01
9.62179244e-01 -1.15576172e+00 9.12758529e-01 -5.58754206e-01
-6.51580095e-01 6.78386331e-01 -1.01488307e-01 -3.57033342e-01
3.49591702e-01 -2.47147039e-01 -3.69137824e-01 -2.16581762e-01
1.61590427e-01 5.69600105e-01 5.81224203e-01 -9.29106057e-01
-9.78986323e-01 -4.07175869e-01 2.11923167e-01 4.06025797e-01
-3.91637623e-01 -2.50262827e-01 -1.02096391e+00 -5.65606594e-01
1.06867701e-02 -8.38178873e-01 -1.46434948e-01 -5.37270069e-01
-3.67504627e-01 -3.25251997e-01 5.18401563e-01 -7.14345813e-01
1.62527537e+00 -1.93585825e+00 4.69571114e-01 3.31723452e-01
5.44935226e-01 1.66022345e-01 -4.84643430e-02 4.51592356e-01
1.33913875e-01 -4.60192338e-02 -1.30945206e-01 -2.01091632e-01
1.72455072e-01 3.07365268e-01 -2.95314133e-01 -1.09722897e-01
1.65026546e-01 1.31628752e+00 -1.12683523e+00 -5.73875070e-01
-2.63376921e-01 1.46981210e-01 -4.21244353e-01 -1.44785091e-01
-3.92656505e-01 -7.35027716e-02 -4.71285343e-01 6.28106952e-01
4.41058725e-01 -6.60794199e-01 4.86510992e-01 -5.29240668e-01
2.53794909e-01 5.16660333e-01 -1.44008493e+00 1.77965117e+00
-1.70764789e-01 4.24061775e-01 -6.10511787e-02 -8.19527984e-01
6.09389782e-01 -1.44142946e-02 2.60187715e-01 -6.40057802e-01
-2.01391354e-01 4.15970355e-01 -1.19736634e-01 -1.93845794e-01
9.77352440e-01 2.05139935e-01 -9.79821011e-02 8.36342350e-02
2.21374795e-01 2.37598091e-01 4.20172185e-01 5.10412335e-01
1.43780947e+00 7.56858215e-02 6.18757069e-01 -2.36633882e-01
2.02497497e-01 8.41935202e-02 5.30092299e-01 8.88506532e-01
1.24259166e-01 2.05438614e-01 4.17320728e-01 -3.40218931e-01
-7.60489345e-01 -7.01825857e-01 1.67500556e-01 1.17850339e+00
3.21201771e-01 -7.95151711e-01 -6.20769083e-01 -1.08572507e+00
8.24299455e-02 6.26260519e-01 -7.06362247e-01 -1.74574494e-01
-6.20337307e-01 -9.86508965e-01 5.57619572e-01 7.52596796e-01
4.02659446e-01 -9.30289388e-01 -5.61231256e-01 4.22513932e-01
-3.88614148e-01 -1.23326015e+00 -2.33761534e-01 3.11746478e-01
-6.16024971e-01 -1.23336744e+00 -2.52361447e-01 -7.35807478e-01
6.23969376e-01 1.68963417e-01 1.60388339e+00 4.51998487e-02
-1.98840246e-01 1.76559970e-01 -2.45383099e-01 -3.86983931e-01
-2.80981332e-01 3.95902902e-01 -1.96813315e-01 -3.62335891e-01
5.89883506e-01 -5.91841042e-01 -3.97531748e-01 2.30116650e-01
-6.67173207e-01 4.22946483e-01 8.99829924e-01 7.86082685e-01
8.74360800e-01 2.39887744e-01 5.99078000e-01 -8.86363506e-01
7.21033633e-01 -5.43192089e-01 -5.11555672e-01 8.60511780e-01
-9.11027074e-01 2.59564131e-01 3.21889699e-01 -3.31601322e-01
-7.91024685e-01 -5.49200103e-02 1.23454615e-01 -3.31042796e-01
1.79660723e-01 9.63097870e-01 -2.34177008e-01 -8.28285515e-02
6.55017495e-01 8.17924589e-02 -4.32433486e-01 -2.78787673e-01
5.51797688e-01 1.65953770e-01 7.43838906e-01 -7.42744803e-01
4.49490696e-01 1.20125733e-01 2.42793679e-01 -2.57548749e-01
-6.79947436e-01 -5.91804862e-01 -5.22878945e-01 -1.43247619e-01
6.15709066e-01 -9.94745255e-01 -4.81052011e-01 1.46570012e-01
-1.03966212e+00 -1.21569343e-01 -3.41740757e-01 2.54209548e-01
-2.15223730e-01 4.51751232e-01 -2.94556707e-01 -3.15796286e-01
-5.87206006e-01 -9.65282321e-01 1.18420517e+00 1.40226513e-01
-2.10448563e-01 -8.17748547e-01 9.63834673e-02 2.24739596e-01
3.60334247e-01 1.28790453e-01 1.12037385e+00 -8.03031862e-01
-9.28924322e-01 -2.20384657e-01 -4.16086018e-01 -2.78845340e-01
4.27975692e-02 -2.31306627e-01 -8.38835537e-01 -1.71096206e-01
-4.61574912e-01 -1.58353925e-01 1.09629357e+00 2.04984710e-01
9.25984621e-01 -2.47305959e-01 -7.90741503e-01 3.98486644e-01
1.33693492e+00 -1.26805410e-01 5.05456805e-01 4.25638974e-01
1.03867710e+00 6.44143879e-01 5.29264510e-01 -1.92048475e-02
7.73076177e-01 9.63648319e-01 4.71340477e-01 3.52338701e-02
-2.18312338e-01 -3.23704004e-01 1.08428352e-01 9.70892966e-01
-1.79366961e-01 -4.50993657e-01 -1.11053717e+00 8.52425218e-01
-2.27406240e+00 -9.08275008e-01 -3.19003552e-01 2.23444653e+00
8.10501933e-01 1.65419206e-01 -7.40853176e-02 -5.71838282e-02
6.40982449e-01 -3.10309771e-02 -3.22870642e-01 -4.70918864e-02
-1.84102461e-01 -7.16388077e-02 4.56251740e-01 2.37182394e-01
-1.11539352e+00 8.10756266e-01 5.59501076e+00 1.11512530e+00
-5.18823564e-01 4.83278893e-02 2.69606918e-01 -5.22747263e-02
-4.11448568e-01 8.98979381e-02 -1.13100946e+00 2.60267675e-01
7.59395897e-01 -2.81778634e-01 3.28769684e-01 7.98040032e-01
-7.20265388e-01 -1.31282955e-01 -1.32343972e+00 7.39189267e-01
-6.02862313e-02 -1.56940448e+00 1.11351714e-01 1.70338690e-01
6.46607280e-01 1.98676839e-01 -4.04422015e-01 5.31847179e-01
4.28190231e-01 -7.18625784e-01 6.34152293e-01 4.76968408e-01
4.71663654e-01 -5.32499373e-01 8.22432160e-01 2.71765798e-01
-1.64014721e+00 -6.01683930e-02 -2.89216161e-01 2.06330746e-01
1.60542905e-01 6.54242873e-01 -1.01653528e+00 1.28229904e+00
7.80732691e-01 4.09033328e-01 -8.59415829e-01 9.60275352e-01
-1.74673826e-01 2.26825103e-01 -5.58911979e-01 6.12603091e-02
2.19369948e-01 3.10854963e-03 6.89897537e-01 1.34460568e+00
3.29320997e-01 -2.00249627e-01 1.05132461e-01 8.41036201e-01
-4.28152055e-01 1.60855204e-01 -6.44949377e-01 -4.89954464e-02
7.40483522e-01 1.30715501e+00 -6.93059266e-01 -4.77464378e-01
-4.75400776e-01 7.73235202e-01 9.27849948e-01 8.06554481e-02
-7.45594800e-01 -2.04619259e-01 3.42864990e-01 -1.46644875e-01
5.19908965e-01 1.65865988e-01 -2.18827412e-01 -1.06949544e+00
4.08338755e-01 -6.72302783e-01 8.37360024e-01 -6.97071910e-01
-1.34984756e+00 6.20633364e-01 2.10182220e-01 -7.23215461e-01
-3.15281451e-01 -3.43575180e-01 -4.56760794e-01 8.04977119e-01
-1.41682923e+00 -1.41264296e+00 -5.33178508e-01 4.80773419e-01
1.87652379e-01 5.27213700e-02 7.59766400e-01 4.61105853e-01
-5.00342667e-01 5.68423748e-01 -8.55505019e-02 -5.02970554e-02
6.43146396e-01 -1.54761732e+00 8.03106070e-01 1.10826492e+00
4.04924333e-01 7.35714555e-01 5.95160306e-01 -1.11714923e+00
-1.33634746e+00 -1.22521794e+00 1.19462061e+00 -5.19683123e-01
7.43796229e-01 -3.94496232e-01 -1.14409077e+00 7.61738300e-01
-2.50446983e-02 1.50472924e-01 4.40452218e-01 5.56815386e-01
-4.02488351e-01 -5.94970165e-03 -7.03119338e-01 7.24851251e-01
1.52001166e+00 -5.41691840e-01 -7.16248691e-01 3.98023091e-02
7.26118028e-01 -5.56752086e-01 -9.11668301e-01 6.67239308e-01
5.17414689e-01 -6.26065671e-01 1.05809653e+00 -6.42031968e-01
2.46909156e-01 -4.61836219e-01 -1.07927889e-01 -1.14036286e+00
-4.91940379e-01 -3.23808581e-01 -6.89777017e-01 1.37473547e+00
4.65199888e-01 -4.40729946e-01 7.04571664e-01 6.65748417e-01
-1.15503274e-01 -9.13926840e-01 -8.99242401e-01 -7.53582120e-01
-3.80865723e-01 -3.35448891e-01 8.18222523e-01 1.26499093e+00
5.15538268e-02 4.33097631e-01 -9.88075584e-02 5.31192422e-01
6.36179745e-01 4.29094285e-01 7.18451023e-01 -1.39691579e+00
-4.80259120e-01 -4.83041257e-01 -4.75743473e-01 -5.78548014e-01
6.97850296e-03 -1.18753588e+00 -1.32556260e-01 -1.79379368e+00
4.50411916e-01 -5.62925875e-01 -5.89132607e-01 8.14659655e-01
-6.80447221e-01 -1.06392264e-01 1.93423103e-03 9.09404159e-02
-7.75309622e-01 3.03144306e-01 6.48871362e-01 -1.52237222e-01
-2.09827945e-01 -4.47585076e-01 -6.94719076e-01 4.94100630e-01
4.89434749e-01 -4.98991430e-01 -6.18746638e-01 -4.31380093e-01
9.70517039e-01 -1.94369227e-01 5.94109535e-01 -8.78908932e-01
7.02496827e-01 5.42450510e-03 1.09777533e-01 -5.43818355e-01
1.91197228e-02 -7.39184439e-01 7.23728836e-01 -1.22812269e-02
-1.73382446e-01 4.39325906e-02 3.74780715e-01 9.42885995e-01
-1.41290963e-01 -9.16290656e-02 5.17074540e-02 -3.46065052e-02
-1.08414412e+00 1.46802664e-01 2.30370820e-01 8.04096460e-02
8.78530681e-01 -2.93803453e-01 -6.94783390e-01 2.03120232e-01
-6.29314840e-01 3.79611284e-01 4.24266011e-01 4.73629385e-01
3.70457172e-01 -1.21857989e+00 -5.46025574e-01 -1.31045476e-01
4.14216280e-01 3.79609793e-01 2.49761283e-01 9.46886361e-01
-3.71876992e-02 2.95225292e-01 -2.26693437e-01 -2.53997684e-01
-1.44322062e+00 6.47449076e-01 2.57313192e-01 -9.58657861e-01
-7.01767743e-01 9.04032707e-01 2.30603248e-01 -2.35466540e-01
1.69203088e-01 -2.51072943e-01 -2.28788257e-01 1.44678161e-01
4.74596351e-01 3.29958975e-01 5.63695192e-01 -3.16390991e-01
-4.84555185e-01 2.45672569e-01 -3.51409853e-01 3.47642779e-01
1.30618680e+00 -7.69303963e-02 -2.66431898e-01 1.44011453e-02
6.08009338e-01 1.65648982e-01 -8.15283537e-01 -4.15795088e-01
3.37304920e-01 -3.26678693e-01 1.56612322e-01 -1.02489400e+00
-9.46245492e-01 3.95001411e-01 1.70540482e-01 1.80195823e-01
1.28006005e+00 1.26544446e-01 6.21467471e-01 6.23648942e-01
6.44176245e-01 -1.06683099e+00 -3.38312626e-01 2.41224542e-01
7.38003314e-01 -1.00411952e+00 2.50200897e-01 -7.08763123e-01
-5.25577545e-01 8.64738226e-01 6.87165797e-01 3.42592567e-01
2.00067550e-01 3.01043391e-01 -5.26005745e-01 -6.83803618e-01
-9.32322204e-01 -3.69503647e-01 7.85869360e-01 3.54230374e-01
-3.61329541e-02 -1.29670417e-02 -3.05092394e-01 7.31248796e-01
-1.21923657e-02 1.04494870e-01 9.03496295e-02 9.53677952e-01
-2.21402779e-01 -1.06066215e+00 7.65011087e-02 5.54774225e-01
-4.00302798e-01 -3.31754237e-01 -5.78327239e-01 7.73687124e-01
1.58887416e-01 4.50692117e-01 -2.33773366e-01 -5.04803479e-01
3.56994987e-01 4.11856502e-01 5.02494812e-01 -2.85409063e-01
-7.95539021e-01 -1.70671090e-01 5.45509100e-01 -7.22577155e-01
-4.31256533e-01 -3.91052186e-01 -1.13163519e+00 -2.46069580e-01
-6.52266800e-01 -2.54426878e-02 4.80368614e-01 1.02290142e+00
8.75533402e-01 5.87664187e-01 -1.37349024e-01 -3.20053905e-01
-1.54083222e-01 -7.08328068e-01 -3.71356785e-01 5.22097945e-01
-1.50460675e-01 -7.58345544e-01 1.25972673e-01 -2.41352946e-01] | [8.985014915466309, 8.221659660339355] |
77ec96f6-57ab-4653-b5fd-9426871ed7d4 | face-age-progression-with-attribute | 2106.07696 | null | https://arxiv.org/abs/2106.07696v1 | https://arxiv.org/pdf/2106.07696v1.pdf | Face Age Progression With Attribute Manipulation | Face is one of the predominant means of person recognition. In the process of ageing, human face is prone to many factors such as time, attributes, weather and other subject specific variations. The impact of these factors were not well studied in the literature of face aging. In this paper, we propose a novel holistic model in this regard viz., ``Face Age progression With Attribute Manipulation (FAWAM)", i.e. generating face images at different ages while simultaneously varying attributes and other subject specific characteristics. We address the task in a bottom-up manner, as two submodules i.e. face age progression and face attribute manipulation. For face aging, we use an attribute-conscious face aging model with a pyramidal generative adversarial network that can model age-specific facial changes while maintaining intrinsic subject specific characteristics. For facial attribute manipulation, the age processed facial image is manipulated with desired attributes while preserving other details unchanged, leveraging an attribute generative adversarial network architecture. We conduct extensive analysis in standard large scale datasets and our model achieves significant performance both quantitatively and qualitatively. | ['Anurag Mittal', 'Athira Nambiar', 'Sinzith Tatikonda'] | 2021-06-14 | null | null | null | null | ['person-recognition'] | ['computer-vision'] | [ 3.37261021e-01 1.51156023e-01 1.92020416e-01 -6.75534308e-01
-3.67735699e-02 -3.61121982e-01 6.53439164e-01 -3.88111591e-01
-1.22452609e-01 6.33930147e-01 1.67243287e-01 2.62850702e-01
2.37383008e-01 -8.91997755e-01 -6.94691718e-01 -8.82545173e-01
3.98785695e-02 1.50760248e-01 -3.97663921e-01 -2.29489133e-01
-5.08200489e-02 5.52265346e-01 -1.84705949e+00 -1.97023198e-01
9.20126140e-01 1.04660213e+00 -5.97650647e-01 5.17488539e-01
1.21600837e-01 4.62113351e-01 -7.01369286e-01 -7.76522160e-01
2.61510938e-01 -3.44619274e-01 -4.56631780e-01 2.51502812e-01
8.73162985e-01 -3.90680969e-01 -4.26750749e-01 1.03966343e+00
8.59700203e-01 6.24164008e-03 9.07718480e-01 -1.77986300e+00
-1.25108743e+00 3.16736400e-01 -7.37844408e-01 -1.65679485e-01
2.49024943e-01 3.45718235e-01 2.54831344e-01 -9.83942211e-01
3.34883988e-01 1.87433231e+00 7.40512252e-01 1.25025535e+00
-1.28023982e+00 -1.10005724e+00 4.40327495e-01 3.19651067e-02
-1.24792123e+00 -6.75151765e-01 9.99542236e-01 -4.54838812e-01
1.54266044e-01 5.44039980e-02 6.09363854e-01 1.45715392e+00
5.12580693e-01 3.34200621e-01 1.32968664e+00 -1.80229574e-01
1.46496996e-01 -1.07304350e-01 -1.91437185e-01 6.51687086e-01
1.96157351e-01 2.94618309e-01 -6.65911019e-01 -3.20555083e-02
6.19735003e-01 6.65077642e-02 -1.99212972e-02 -1.79226413e-01
-8.17629755e-01 4.89548445e-01 3.55336308e-01 -1.82793289e-01
-4.95605707e-01 3.97585601e-01 3.05103928e-01 5.47334909e-01
6.90822542e-01 -5.75583056e-02 -3.74802887e-01 5.30129611e-01
-6.86305046e-01 4.39646691e-01 3.13702673e-01 9.27775383e-01
8.31144691e-01 4.52173322e-01 -3.45534861e-01 7.72487223e-01
5.00528276e-01 8.77420545e-01 2.48424187e-01 -9.77275252e-01
-7.98772126e-02 5.80481648e-01 -2.39900813e-01 -8.49508345e-01
-2.30818897e-01 -2.41009489e-01 -1.18560684e+00 6.75995827e-01
3.12212497e-01 -1.76277608e-01 -1.37103713e+00 2.54017091e+00
6.58049703e-01 2.49757111e-01 -5.53896613e-02 3.45885247e-01
8.59641612e-01 3.88037950e-01 8.95074069e-01 -3.84666145e-01
1.74420452e+00 -4.96940911e-01 -9.77870464e-01 -2.95200557e-01
-1.72320545e-01 -6.35248244e-01 9.20238018e-01 1.31473422e-01
-1.16110706e+00 -7.12317288e-01 -1.03628910e+00 -1.13980815e-01
-5.27563691e-01 -2.93415952e-02 6.01509631e-01 9.37307894e-01
-1.34517157e+00 4.51177120e-01 -4.58812118e-01 -4.29721266e-01
8.32630932e-01 6.75741017e-01 -5.54526389e-01 1.65278912e-02
-1.31808376e+00 7.26170957e-01 -6.14713766e-02 1.64646596e-01
-1.24003160e+00 -1.22759318e+00 -8.99874032e-01 -2.29635283e-01
1.85584411e-01 -1.22206783e+00 1.00409877e+00 -1.18001020e+00
-1.34213483e+00 1.03661573e+00 -1.95959717e-01 -5.66545241e-02
5.63488364e-01 -3.47451746e-01 -6.45847380e-01 -1.94808677e-01
-1.52276903e-01 8.24597120e-01 1.64740741e+00 -1.45729673e+00
-1.16878510e-01 -1.01739013e+00 -7.39080757e-02 1.31507784e-01
-5.98485708e-01 3.18992525e-01 5.24435379e-03 -9.05922771e-01
-2.90858537e-01 -9.32545602e-01 -1.09617924e-03 6.46549225e-01
3.14246789e-02 -6.34017363e-02 1.11086357e+00 -1.03830719e+00
1.28040993e+00 -2.15318942e+00 3.71956646e-01 -6.77442625e-02
5.88575065e-01 7.10692257e-02 -1.51509315e-01 5.68810813e-02
-3.17903012e-01 2.23079562e-01 -3.08552295e-01 -6.26681089e-01
-3.52577940e-02 1.04788102e-01 5.39125949e-02 4.35649574e-01
5.32362998e-01 9.49045837e-01 -7.88266957e-01 -6.20242715e-01
-2.85343021e-01 9.39351976e-01 -4.87266123e-01 2.36547679e-01
-1.66643113e-01 4.81227875e-01 -3.70770961e-01 1.08921826e+00
9.90793765e-01 5.91253459e-01 -2.65018612e-01 -5.41689813e-01
2.05407754e-01 -7.12695718e-01 -8.79540622e-01 1.36093545e+00
-4.35860872e-01 1.04608901e-01 4.34263825e-01 -3.08011979e-01
1.07552814e+00 4.72406566e-01 4.43056017e-01 -3.76660585e-01
3.18702132e-01 -2.61069424e-02 -2.43431516e-02 -1.83285281e-01
2.92899847e-01 -3.75268310e-01 -3.34365331e-02 4.77893889e-01
1.78938434e-01 7.37191439e-02 -2.93913722e-01 1.72069706e-02
8.28531921e-01 1.74317405e-01 2.14883178e-01 -3.49213302e-01
8.26348901e-01 -8.43306482e-01 6.95565403e-01 1.50271848e-01
-6.98601484e-01 4.17231828e-01 2.94030994e-01 -4.95859116e-01
-1.26228559e+00 -1.24370885e+00 6.27706125e-02 1.15622616e+00
-2.21050784e-01 -8.89949650e-02 -1.13220775e+00 -6.19767129e-01
1.67920545e-01 2.30948836e-01 -1.29346144e+00 -8.40814769e-01
-5.49005628e-01 -8.11688900e-01 7.82310963e-01 6.29015505e-01
7.17890263e-01 -1.20704138e+00 -1.52081594e-01 -3.94512899e-02
2.22723678e-01 -7.45205522e-01 -8.17843974e-01 -4.82873529e-01
-7.19668150e-01 -8.28557849e-01 -7.71657288e-01 -6.66998148e-01
1.00045121e+00 -2.76650667e-01 1.21352541e+00 1.40529886e-01
-4.46428537e-01 5.69087207e-01 8.12891405e-03 -9.16208386e-01
-5.84129453e-01 -7.66693801e-02 6.18494809e-01 5.37258983e-01
2.06367001e-01 -9.42464352e-01 -9.19147789e-01 1.47667438e-01
-9.53095436e-01 -4.08073187e-01 6.17982328e-01 7.60046959e-01
3.14330995e-01 2.48236023e-02 1.07722306e+00 -8.97479117e-01
6.33988202e-01 -3.56367469e-01 1.70577206e-02 1.97096974e-01
-8.51644576e-01 -5.97466826e-02 4.48378474e-01 -7.06516623e-01
-1.30535996e+00 5.02499007e-02 -3.22738575e-04 -4.24151689e-01
-1.63801536e-01 -1.83521301e-01 -9.62162197e-01 -2.95335531e-01
6.67434692e-01 2.21546456e-01 6.24060988e-01 -1.81440979e-01
3.94471258e-01 4.80279654e-01 7.50025809e-01 -9.22306538e-01
1.40897787e+00 4.84238386e-01 3.81221205e-01 -5.70260465e-01
-6.70455098e-01 3.39616925e-01 -7.23450243e-01 -6.76715493e-01
8.53888929e-01 -9.79349017e-01 -6.57776237e-01 1.37200522e+00
-9.03016686e-01 -5.02735935e-02 -1.29508972e-01 -1.96763605e-01
-6.11268282e-01 2.00479388e-01 -5.34424722e-01 -8.12629580e-01
-8.30680490e-01 -8.43599617e-01 1.02066767e+00 4.52646703e-01
-1.61228254e-01 -7.81001687e-01 -1.78929433e-01 1.80390388e-01
5.66090524e-01 7.48824239e-01 1.02072155e+00 -1.08199008e-01
-1.77746445e-01 -9.27853510e-02 1.07679218e-02 6.53653204e-01
5.93058765e-01 2.06753984e-01 -1.22293198e+00 -4.54848200e-01
-6.96806684e-02 -1.73532844e-01 6.07972205e-01 1.47353947e-01
1.27308249e+00 -4.67655241e-01 -8.35041553e-02 6.67491972e-01
1.30972207e+00 2.09154844e-01 9.57318246e-01 -5.13870418e-02
8.50983202e-01 7.56601214e-01 4.77409750e-01 4.38728809e-01
1.81928605e-01 2.56702930e-01 5.05467772e-01 -1.86561495e-01
-2.64246047e-01 -1.10652700e-01 3.83142442e-01 2.34643936e-01
-4.13105696e-01 -1.34367552e-02 -5.34476221e-01 3.01538736e-01
-1.34425330e+00 -9.87074018e-01 3.90706569e-01 2.06342840e+00
8.65137041e-01 -1.99734047e-01 2.18197420e-01 2.18474910e-01
9.49816763e-01 1.32403091e-01 -8.69611084e-01 -3.83171856e-01
-1.16648369e-01 3.91791224e-01 1.73750564e-01 2.10936502e-01
-1.01053822e+00 7.92519987e-01 6.29608202e+00 2.10794240e-01
-8.93969178e-01 -6.96809366e-02 8.50075722e-01 -5.66000156e-02
-3.14279169e-01 -3.41940135e-01 -6.83692515e-01 5.86397946e-01
7.48610139e-01 -4.46813971e-01 5.31803429e-01 6.21574461e-01
1.11540206e-01 4.68223304e-01 -1.10934103e+00 8.40659380e-01
3.24287891e-01 -5.54741323e-01 4.01307821e-01 9.65982080e-02
5.65815806e-01 -9.70676899e-01 6.31561816e-01 3.88125956e-01
3.15781944e-02 -1.26339054e+00 7.29723692e-01 7.76501000e-01
1.14986730e+00 -9.02307391e-01 3.54088455e-01 -3.08197200e-01
-1.31171346e+00 -3.11430037e-01 1.31592959e-01 2.91246593e-01
-4.34213281e-02 2.65533596e-01 -8.37428942e-02 3.94938737e-01
7.20774233e-01 5.18490553e-01 -7.63689101e-01 4.75239098e-01
5.64743318e-02 3.28733921e-01 4.96626878e-03 5.86908400e-01
-4.05552506e-01 -1.69404566e-01 3.30871314e-01 5.83692908e-01
5.21743357e-01 1.78578496e-01 -3.08767766e-01 6.94233835e-01
-4.30357158e-01 -1.67921502e-02 -5.10139763e-01 6.42458722e-02
7.65789926e-01 1.17338991e+00 -1.90193683e-01 -1.74746618e-01
-1.82735398e-01 9.91151154e-01 9.65333730e-02 3.01722586e-01
-8.33876252e-01 7.76898442e-03 1.25705814e+00 3.84518325e-01
-2.15461493e-01 1.10864393e-01 -1.90036744e-01 -7.67709374e-01
2.79138256e-02 -1.13401318e+00 2.70378083e-01 -6.40369475e-01
-1.46903968e+00 5.63413143e-01 -1.52909428e-01 -8.69450092e-01
2.12507144e-01 -5.85555315e-01 -7.98413515e-01 1.11245215e+00
-1.34733081e+00 -1.88688457e+00 -6.96165979e-01 8.52360606e-01
4.07538265e-01 -3.35490763e-01 7.24978447e-01 5.64565718e-01
-8.05858791e-01 1.09354436e+00 -5.11669755e-01 -1.20860323e-01
1.17741632e+00 -1.19936693e+00 4.90485787e-01 8.24649632e-01
-7.83089459e-01 7.95059562e-01 8.44703972e-01 -8.00852299e-01
-1.28676045e+00 -1.43483329e+00 7.45912254e-01 -5.75538933e-01
2.96090692e-01 -4.52511936e-01 -1.01938844e+00 6.92024231e-01
1.86084032e-01 2.91457176e-02 6.02907717e-01 -1.02607481e-01
-6.44038856e-01 -5.28087258e-01 -1.63530612e+00 7.64167905e-01
1.39266300e+00 -4.48035985e-01 -1.71764761e-01 -1.77501529e-01
7.80378520e-01 -6.01110123e-02 -1.17848933e+00 6.23014271e-01
8.08788121e-01 -4.91485834e-01 1.29635251e+00 -5.74015319e-01
4.23069000e-01 -3.06440860e-01 6.38424903e-02 -1.44083560e+00
-4.67660993e-01 -8.83292019e-01 -5.15208125e-01 1.90921414e+00
5.00020338e-03 -7.28342831e-01 7.31347084e-01 6.60930693e-01
-4.01135795e-02 -7.35363901e-01 -6.86989784e-01 -6.52733803e-01
3.58674735e-01 2.36523479e-01 1.36125493e+00 7.46494651e-01
-8.88742745e-01 2.06183478e-01 -5.93393505e-01 2.83062249e-01
1.05889070e+00 -6.21023715e-01 7.14975536e-01 -1.52003467e+00
3.41811568e-01 -4.23180401e-01 -5.24777651e-01 4.49039228e-02
4.23688501e-01 -4.18404877e-01 -5.16321175e-02 -1.03038704e+00
2.38740817e-01 -3.14229161e-01 -3.25734645e-01 3.78552049e-01
-5.75466275e-01 2.95229137e-01 -6.50899205e-03 -4.40727845e-02
2.68331468e-01 8.42552185e-01 1.53551972e+00 -2.47896865e-01
2.33181238e-01 -1.07087538e-01 -9.64483798e-01 4.51977253e-01
8.67722452e-01 -1.72128782e-01 -6.04613125e-01 -1.12843424e-01
-4.32906672e-02 -4.04340386e-01 3.43373120e-01 -1.07547116e+00
-9.62726921e-02 -2.48097196e-01 8.71734798e-01 -8.61063898e-02
5.14338553e-01 -7.85592675e-01 3.39181453e-01 4.93265331e-01
-5.75797632e-02 4.30604368e-01 2.04965100e-01 6.15024567e-01
1.01665948e-02 3.32352996e-01 1.18532383e+00 -6.75411429e-03
-6.75268292e-01 1.10327661e+00 -1.10683873e-01 -1.75688982e-01
1.31046498e+00 -2.87418216e-01 -3.55077714e-01 -2.97150850e-01
-8.41732085e-01 1.61490113e-01 6.60361528e-01 6.64023042e-01
4.65288967e-01 -1.65194142e+00 -9.48991418e-01 3.80421162e-01
1.95576623e-01 -1.49911925e-01 5.68162143e-01 3.96166682e-01
-6.93049431e-02 -3.46607894e-01 -7.48634040e-01 -1.33556679e-01
-1.58348906e+00 9.53149676e-01 5.29225588e-01 1.13522217e-01
-2.82176994e-02 8.82131696e-01 4.57825989e-01 -4.68126774e-01
4.69705537e-02 2.39934683e-01 -3.30116898e-01 2.75943667e-01
5.45299888e-01 4.07422096e-01 -2.80189663e-01 -8.13627362e-01
-1.94847673e-01 8.45435381e-01 -2.44486183e-01 2.07451284e-01
9.92486000e-01 -2.66943783e-01 -3.02433223e-01 5.67378476e-02
8.31943214e-01 -2.60357946e-01 -1.36991358e+00 -1.84350625e-01
-5.66359580e-01 -2.91270375e-01 -2.21875370e-01 -7.40038216e-01
-1.47908664e+00 6.34244025e-01 9.68221426e-01 -2.73577958e-01
1.65032434e+00 -1.18751623e-01 6.34592414e-01 -2.74573475e-01
1.07598826e-01 -1.05198085e+00 3.49548042e-01 2.72372514e-02
1.11466956e+00 -1.11135101e+00 5.15951449e-03 -6.17634594e-01
-3.77996862e-01 7.35122681e-01 1.08618319e+00 4.49026711e-02
8.29028964e-01 1.84845150e-01 2.51121789e-01 -1.79434344e-01
-6.64867222e-01 1.89640000e-01 2.28525341e-01 9.39971626e-01
4.21464682e-01 1.14991451e-02 -1.96299896e-01 4.92078394e-01
-2.86972940e-01 -7.21303746e-03 1.92204371e-01 9.08796251e-01
-1.29017517e-01 -1.35708451e+00 -5.29098392e-01 4.65506524e-01
-5.64967632e-01 5.49170412e-02 -3.68366808e-01 6.02976203e-01
4.60965902e-01 5.32304943e-01 2.98001748e-02 -4.96118098e-01
3.89363438e-01 4.34134573e-01 6.86550260e-01 -3.23660761e-01
-6.44354641e-01 -6.03436947e-01 -2.95705020e-01 -3.52106184e-01
-4.32988495e-01 -1.02658224e+00 -7.28976011e-01 -6.55267775e-01
1.85419649e-01 -5.06959021e-01 5.46086729e-01 5.78985393e-01
3.56043875e-01 6.51593506e-01 9.18346345e-01 -6.78757548e-01
-6.61163211e-01 -1.02477551e+00 -5.60499012e-01 6.66161358e-01
4.25239742e-01 -8.97883058e-01 -4.05838728e-01 4.71591532e-01] | [13.224575996398926, 0.5192817449569702] |
4068d51d-d146-4643-8dbb-a43ba5ed579c | semantically-aligned-universal-tree | 2010.06823 | null | https://arxiv.org/abs/2010.06823v1 | https://arxiv.org/pdf/2010.06823v1.pdf | Semantically-Aligned Universal Tree-Structured Solver for Math Word Problems | A practical automatic textual math word problems (MWPs) solver should be able to solve various textual MWPs while most existing works only focused on one-unknown linear MWPs. Herein, we propose a simple but efficient method called Universal Expression Tree (UET) to make the first attempt to represent the equations of various MWPs uniformly. Then a semantically-aligned universal tree-structured solver (SAU-Solver) based on an encoder-decoder framework is proposed to resolve multiple types of MWPs in a unified model, benefiting from our UET representation. Our SAU-Solver generates a universal expression tree explicitly by deciding which symbol to generate according to the generated symbols' semantic meanings like human solving MWPs. Besides, our SAU-Solver also includes a novel subtree-level semanticallyaligned regularization to further enforce the semantic constraints and rationality of the generated expression tree by aligning with the contextual information. Finally, to validate the universality of our solver and extend the research boundary of MWPs, we introduce a new challenging Hybrid Math Word Problems dataset (HMWP), consisting of three types of MWPs. Experimental results on several MWPs datasets show that our model can solve universal types of MWPs and outperforms several state-of-the-art models. | ['Liang Lin', 'Rumin Zhang', 'Xiaodan Liang', 'Lihui Lin', 'Jinghui Qin'] | 2020-10-14 | null | https://aclanthology.org/2020.emnlp-main.309 | https://aclanthology.org/2020.emnlp-main.309.pdf | emnlp-2020-11 | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [ 2.01700971e-01 1.30354524e-01 -1.33538038e-01 -3.15520525e-01
-8.82591486e-01 -4.53392684e-01 -6.22033775e-02 -3.12179089e-01
5.98053746e-02 8.22337985e-01 4.73590381e-02 -4.29846406e-01
-2.57341444e-01 -1.22796977e+00 -8.59064877e-01 -2.95213312e-01
5.67446709e-01 5.81586361e-01 -1.00160055e-02 -4.82631832e-01
2.20878661e-01 -7.11993352e-02 -9.58439469e-01 6.33435488e-01
1.59928560e+00 9.74909723e-01 7.16244698e-01 2.47813806e-01
-9.18013930e-01 1.06178451e+00 -6.03636384e-01 -7.86157489e-01
2.81191319e-01 -4.96934980e-01 -1.07778525e+00 8.09149295e-02
2.65014112e-01 -1.58470884e-01 -1.69862136e-02 1.00309205e+00
1.74793571e-01 1.27650350e-01 5.31112313e-01 -1.23518050e+00
-1.09496140e+00 1.15263367e+00 -4.24770415e-01 -1.49745494e-01
7.29795218e-01 1.55532002e-01 1.45667720e+00 -6.64866447e-01
7.66583443e-01 1.37881041e+00 5.49400508e-01 3.63427997e-01
-9.69066143e-01 -8.46720099e-01 6.61006272e-01 5.39445281e-01
-1.66880631e+00 1.81770593e-01 7.74086118e-01 -1.21419564e-01
1.11957121e+00 4.69434679e-01 4.10974771e-01 1.08752823e+00
8.35306793e-02 1.00566673e+00 9.46146607e-01 -1.64582193e-01
-5.49045801e-02 4.64959033e-02 3.09474200e-01 1.01407301e+00
-9.88760889e-02 -5.33916175e-01 -3.57053936e-01 9.70952064e-02
7.14645863e-01 -6.99117184e-01 -2.16589004e-01 1.69231370e-02
-1.10097086e+00 7.84264505e-01 2.96116471e-01 3.57657313e-01
1.62018538e-01 4.44268212e-02 2.31614605e-01 2.21963376e-01
2.47513384e-01 7.44352698e-01 -8.00326049e-01 5.96387833e-02
-8.35834146e-01 5.90486169e-01 8.44513893e-01 1.56588578e+00
6.81559443e-01 1.42089188e-01 -4.74780023e-01 7.54828811e-01
5.37151247e-02 3.44724655e-01 5.40056229e-01 -6.84136689e-01
1.01828349e+00 1.03397882e+00 -1.04052037e-01 -1.34335089e+00
-3.87075663e-01 -7.41532564e-01 -5.36876380e-01 -7.52500117e-01
1.31134465e-01 -1.37007952e-01 -5.62084854e-01 1.87596118e+00
2.43164897e-01 4.83235449e-01 1.63580611e-01 9.51148689e-01
1.24875998e+00 1.24799180e+00 3.02623659e-02 -3.55503112e-01
1.35068774e+00 -1.26219714e+00 -7.49394774e-01 -3.35568428e-01
9.68176067e-01 -5.25791883e-01 1.27559340e+00 5.45722067e-01
-1.14045310e+00 -3.94134820e-01 -9.42336440e-01 -4.42295074e-01
-2.78168738e-01 4.22243267e-01 7.19244003e-01 3.98995340e-01
-4.47372258e-01 3.18848848e-01 -2.16414258e-01 2.00574979e-01
1.77711233e-01 1.74753904e-01 -8.80478024e-02 -1.82610840e-01
-1.65329528e+00 1.22338295e+00 7.41860211e-01 2.58200735e-01
-6.16693616e-01 -1.10899401e+00 -1.18821359e+00 2.74772286e-01
1.03389239e+00 -1.00299203e+00 1.18208969e+00 -9.25403893e-01
-1.62905860e+00 6.12198234e-01 -3.88082325e-01 -2.37842768e-01
4.08073246e-01 -5.86932823e-02 -3.71223390e-01 -1.78198576e-01
2.92231977e-01 2.72111475e-01 4.80937630e-01 -1.16664529e+00
-4.40260708e-01 9.39695910e-02 3.75006497e-01 1.27614081e-01
-2.02557385e-01 1.33235127e-01 -6.06637418e-01 -8.38737965e-01
1.01038992e-01 -4.36133772e-01 -2.59835422e-01 -4.11338836e-01
-4.89214301e-01 -5.48388422e-01 3.79359245e-01 -8.70782077e-01
1.75652802e+00 -1.72752798e+00 1.03081501e+00 1.34894580e-01
1.75287873e-01 8.56266171e-02 -4.25564229e-01 3.73326540e-01
-1.33385062e-01 6.53387159e-02 -5.23217678e-01 -3.33588719e-01
5.25788069e-01 7.25606441e-01 -7.05712795e-01 -1.96130008e-01
3.51281762e-01 1.29346502e+00 -8.24262977e-01 -8.07373464e-01
1.12586766e-01 -2.06441030e-01 -7.14328289e-01 3.39438796e-01
-8.17651987e-01 2.34101579e-01 -7.45491982e-01 5.34925759e-01
9.14851546e-01 -1.70187086e-01 3.02596211e-01 -2.78556377e-01
-2.72362605e-02 2.45367900e-01 -1.48677754e+00 1.98471975e+00
-8.31000566e-01 7.62549192e-02 -1.33936599e-01 -1.26514268e+00
9.12087500e-01 -1.03666894e-01 1.41870886e-01 -5.14741361e-01
3.72323811e-01 2.91571647e-01 -1.88619420e-01 -8.14318895e-01
4.64806259e-01 -1.43511087e-01 -3.22527826e-01 9.65444371e-02
3.42569828e-01 -4.40758169e-01 5.34547687e-01 3.91724765e-01
7.85254478e-01 5.88162661e-01 1.51411891e-01 -3.01925778e-01
1.02106833e+00 -8.59829970e-03 1.05752945e+00 5.66643000e-01
4.16384488e-01 4.33854550e-01 7.18896151e-01 -4.68779147e-01
-5.16025782e-01 -7.08609998e-01 -1.48661852e-01 9.35873091e-01
4.48127270e-01 -1.10465503e+00 -9.35855865e-01 -6.20692492e-01
-3.72635484e-01 1.17083073e+00 -3.82688165e-01 3.25024650e-02
-8.51207852e-01 -8.19855154e-01 6.11898184e-01 4.58553314e-01
7.39852548e-01 -8.29323947e-01 -8.19587260e-02 3.59462738e-01
-8.20031047e-01 -1.65758681e+00 -3.40791523e-01 -1.23489968e-01
-3.13471079e-01 -9.13718700e-01 -2.92105138e-01 -9.11056697e-01
7.52227783e-01 -1.19732529e-01 1.10780931e+00 1.42030835e-01
-2.89223611e-01 -7.32426941e-02 -5.06856441e-01 -1.32877514e-01
-2.06206828e-01 3.22252274e-01 -2.23547101e-01 5.45174303e-03
4.45454791e-02 -4.98882174e-01 1.07098609e-01 1.47975624e-01
-7.67573237e-01 6.26950085e-01 3.61848384e-01 5.78675568e-01
4.01270866e-01 2.31722340e-01 4.06551868e-01 -8.95747542e-01
8.71425986e-01 -6.61795139e-01 -7.49541700e-01 7.39516318e-01
-2.55222976e-01 3.00150633e-01 1.05460393e+00 -5.22946715e-01
-1.23561835e+00 -1.66566640e-01 -2.01718330e-01 -3.59148294e-01
2.93371618e-01 9.35020745e-01 -5.93680739e-01 -1.25746392e-02
3.66517693e-01 5.39994359e-01 -7.45630980e-01 -2.29629278e-01
6.27090573e-01 2.23520368e-01 5.23908257e-01 -1.58705926e+00
9.27943408e-01 -4.35802281e-01 3.94094735e-02 -4.48350430e-01
-1.31965828e+00 1.21164680e-01 -3.67816865e-01 7.23415688e-02
8.79350901e-01 -7.30170131e-01 -6.01369500e-01 3.28232408e-01
-1.70412326e+00 -2.44802281e-01 2.93514132e-03 -5.24094403e-02
-4.63484049e-01 7.17275023e-01 -6.73686564e-01 -6.44581318e-01
-2.79065847e-01 -1.37959409e+00 1.04554367e+00 9.45105776e-02
-4.10238296e-01 -9.00943995e-01 -2.88493782e-01 7.35724807e-01
2.63292432e-01 1.91810369e-01 1.47285676e+00 -2.67684937e-01
-7.60605156e-01 2.73361772e-01 -4.59726095e-01 1.27031565e-01
-6.90571889e-02 -1.24431089e-01 -4.77681130e-01 4.25054789e-01
3.67980897e-02 -3.10258508e-01 5.64596176e-01 1.59485504e-01
1.77086520e+00 -6.72429204e-01 -7.71025121e-02 9.13954675e-01
1.44488502e+00 -1.39926136e-01 6.16774321e-01 2.45460793e-01
6.41748846e-01 4.70804572e-01 5.56455135e-01 3.16921949e-01
7.72171497e-01 7.06806302e-01 3.43545467e-01 2.50776231e-01
2.62270391e-01 -5.12604237e-01 3.17303151e-01 1.07576704e+00
-7.40971789e-02 -3.31723779e-01 -7.11715102e-01 3.35228473e-01
-2.00794077e+00 -5.74649453e-01 -4.37215447e-01 1.40141702e+00
1.27297187e+00 -5.04678451e-02 -4.59305912e-01 -2.74647474e-01
2.84060717e-01 8.88800323e-02 -3.00912201e-01 -5.79326212e-01
-3.26783180e-01 6.03592575e-01 1.79251909e-01 7.33452737e-01
-8.48103762e-01 1.50252545e+00 5.23443460e+00 1.62753594e+00
-7.83978224e-01 1.03589240e-03 2.47673765e-01 3.57099585e-02
-5.47225058e-01 -4.59732190e-02 -8.29697490e-01 4.51855510e-01
4.56550002e-01 -3.85500014e-01 6.93253398e-01 8.41462672e-01
-9.45152342e-02 1.57667935e-01 -9.77697372e-01 9.65953410e-01
6.06642812e-02 -1.51041579e+00 4.15872991e-01 -5.06422102e-01
7.71427453e-01 -7.15402186e-01 -3.05994619e-02 7.39147604e-01
3.19173098e-01 -1.14120615e+00 7.91306138e-01 3.07674110e-01
4.58294570e-01 -7.25023627e-01 6.11286163e-01 7.47177005e-01
-1.53419292e+00 -7.90291578e-02 -5.79036117e-01 -3.45220745e-01
1.82024255e-01 3.72138888e-01 -2.45065466e-01 1.37433076e+00
3.18872064e-01 8.03698778e-01 -3.43813270e-01 4.33325291e-01
-7.49783158e-01 2.39919931e-01 -2.01624617e-01 -2.49630623e-02
5.81266165e-01 -5.54603696e-01 4.77035224e-01 1.33030033e+00
4.91507679e-01 7.33099103e-01 2.67217875e-01 1.58809257e+00
7.37064183e-02 5.02811432e-01 -1.36182234e-01 -4.23266590e-02
3.88116777e-01 1.40738189e+00 -2.83932358e-01 -3.49750698e-01
-2.52564609e-01 1.02793908e+00 7.09718227e-01 2.34063923e-01
-1.40454066e+00 -2.40631223e-01 5.53066015e-01 -2.25233555e-01
2.14145124e-01 -3.98522198e-01 -4.67433989e-01 -1.69358444e+00
2.67786801e-01 -9.87977505e-01 3.64447504e-01 -1.11693430e+00
-1.33132529e+00 4.99946952e-01 1.71571314e-01 -9.61292326e-01
1.79830551e-01 -8.78525436e-01 -6.25279784e-01 1.02588952e+00
-1.48610640e+00 -1.35470247e+00 -2.66551435e-01 4.13093746e-01
8.01978350e-01 9.40590352e-03 7.76064277e-01 1.93736821e-01
-1.06593561e+00 6.57387972e-01 -6.89909160e-01 5.72372079e-02
2.89232790e-01 -1.03843784e+00 5.54839522e-02 9.13236856e-01
-6.14657104e-02 1.02309978e+00 6.82997227e-01 -6.36026025e-01
-1.76102996e+00 -1.12800241e+00 9.88127232e-01 -4.21099037e-01
8.84441674e-01 -3.90372574e-01 -1.00314128e+00 9.32964206e-01
3.16324860e-01 -4.21732038e-01 5.73904634e-01 8.54603294e-03
-4.20918971e-01 2.04007566e-01 -9.62181270e-01 7.18273699e-01
1.37716317e+00 -7.27335736e-02 -1.00677443e+00 7.30751574e-01
1.21493196e+00 -1.12346601e+00 -7.72935748e-01 5.01084089e-01
1.62326843e-01 -2.99378932e-01 1.05504525e+00 -9.14431930e-01
1.31540036e+00 -4.42183733e-01 -2.52425730e-01 -1.27358234e+00
-3.99968326e-01 -4.12137181e-01 -1.79303646e-01 1.34222126e+00
5.64182639e-01 -5.65690815e-01 4.05098975e-01 7.26757765e-01
-3.90994132e-01 -1.15938425e+00 -8.88329506e-01 -7.42807388e-01
3.13939452e-01 -8.57080042e-01 9.85925078e-01 1.23783016e+00
1.33016855e-01 3.74320269e-01 -6.40467227e-01 3.33546638e-01
3.10663730e-01 5.81668973e-01 5.93983293e-01 -6.70880973e-01
-6.11977339e-01 -5.46393275e-01 7.22215772e-02 -1.37185931e+00
7.84274817e-01 -1.41664386e+00 -1.70256913e-01 -1.65110612e+00
9.37697142e-02 -3.29191476e-01 3.08254063e-02 6.23393655e-01
-2.72443682e-01 -2.84970194e-01 2.61488140e-01 -4.73291934e-01
-6.75564289e-01 7.66684651e-01 1.64064121e+00 -3.66338640e-01
1.09437287e-01 -5.76115131e-01 -9.65259433e-01 7.51828909e-01
4.69447196e-01 -1.44013762e-01 -5.33372521e-01 -9.74037349e-01
7.59250641e-01 2.05176443e-01 2.86958441e-02 -6.34640694e-01
3.50106239e-01 -8.50867569e-01 8.70785397e-03 -3.29608798e-01
1.19212657e-01 -7.56023228e-01 1.50017232e-01 1.09956749e-01
-2.89733529e-01 -1.97110146e-01 3.62366259e-01 -1.40948100e-02
-1.62098855e-01 -6.52207196e-01 3.19288969e-01 -4.43826973e-01
-8.48244488e-01 2.46861875e-01 -2.80562341e-01 4.10233051e-01
1.15784466e+00 5.48239350e-02 -3.81533444e-01 -3.89039144e-02
-4.97180998e-01 8.56378555e-01 -2.27704495e-01 3.18349183e-01
6.84426308e-01 -1.33143902e+00 -7.67145693e-01 3.70236002e-02
4.61798795e-02 4.12447453e-01 4.10495162e-01 5.50273180e-01
-6.78116381e-01 4.24641490e-01 1.04057882e-02 -1.35983571e-01
-9.53721464e-01 6.61578894e-01 5.15355408e-01 -7.53038883e-01
-6.19447589e-01 1.11282253e+00 3.97308767e-01 -8.63379955e-01
2.34824307e-02 -8.98242712e-01 -4.44240123e-02 -1.75177455e-01
3.72990310e-01 8.99044871e-02 -1.21753007e-01 -4.61328685e-01
-1.85252890e-01 8.06399584e-01 1.98826790e-02 3.98760617e-01
1.23970556e+00 1.91448361e-01 -7.07201719e-01 4.41950783e-02
7.51276374e-01 -1.95833500e-02 -4.96295810e-01 -3.44839722e-01
-1.41744092e-01 -3.37109625e-01 -2.67228633e-01 -7.66924679e-01
-1.10103762e+00 7.83261895e-01 -4.65130508e-01 -2.44450301e-01
1.24556017e+00 -1.87245727e-01 1.18988991e+00 4.63536322e-01
6.23057842e-01 -1.02950299e+00 -1.48511142e-01 7.88735509e-01
1.09710491e+00 -6.68828607e-01 -4.93481569e-02 -1.16151786e+00
-8.29051852e-01 1.33312619e+00 1.04654193e+00 -4.25345870e-03
-8.33734944e-02 4.06014085e-01 -3.59544456e-01 6.36883155e-02
-7.73609638e-01 -4.81196307e-02 3.75348628e-01 1.81194663e-01
1.84950858e-01 3.17008048e-02 -7.55309582e-01 1.73219705e+00
-6.85326219e-01 1.81141272e-01 2.89628088e-01 5.60754657e-01
-3.48561943e-01 -1.35382962e+00 -7.31407404e-02 5.29725961e-02
-9.92792286e-03 -3.92109841e-01 -4.09746468e-01 5.74459434e-01
5.89481175e-01 7.51465082e-01 -4.06926304e-01 -3.75744879e-01
4.02342260e-01 1.40562624e-01 8.65562022e-01 -6.98337138e-01
-6.09862387e-01 -3.56915027e-01 3.22768450e-01 -6.06662571e-01
-1.18591733e-01 -6.02143332e-02 -1.48540211e+00 -2.84163356e-01
-2.13884413e-01 6.58579767e-02 1.12763517e-01 1.43940544e+00
-4.59342077e-02 9.53707039e-01 2.13564888e-01 -3.36582482e-01
-4.83850479e-01 -6.01900995e-01 -2.77714491e-01 2.35662594e-01
-3.44929248e-01 -5.99348664e-01 -9.50227156e-02 -7.66457170e-02] | [9.728458404541016, 7.480113983154297] |
c2739af3-85ef-45a4-bc57-814cd7824cfb | cross-lingual-pretraining-methods-for-spoken | null | null | https://openreview.net/forum?id=c1oDhu_hagR | https://openreview.net/pdf?id=c1oDhu_hagR | Cross-Lingual Pretraining Methods for Spoken Dialog | There has been an increasing interest among NLP researchers towards learning generic representations. However, in the field of multilingual spoken dialogue systems, this problem remains overlooked. Indeed most of the pre-training methods focus on learning representations for written and non-conversational data or are restricted to the monolingual setting. In this work we (1) generalise existing losses to the multilingual setting, (2) develop a new set of losses to leverage parallel conversations when available. These losses improve the learning of representations by fostering the deep encoder to better learn contextual dependencies. The pre-training relies on \texttt{OpenSubtitles}, a huge multilingual corpus that is composed of $24.3$G tokens; a by-product of the pre-processing includes multilingual aligned conversations. We also introduce two new multilingual tasks and a new benchmark on multilingual dialogue act labels called \texttt{MIAM}. We validate our pre-training on the three aforementioned tasks and show that our model using our newly designed losses achieves better performances than existing models. Our implementation will be available on \url{github.com} and preprocessed data will be available in Datasets | ['Anonymous'] | 2021-03-17 | null | null | null | null | ['spoken-dialogue-systems'] | ['speech'] | [-2.82955337e-02 3.47024739e-01 -6.52666092e-02 -7.22322941e-01
-1.18906474e+00 -7.16799915e-01 8.77663732e-01 1.15975693e-01
-6.56173885e-01 1.19137073e+00 6.09079719e-01 -5.46881735e-01
1.75804973e-01 -5.40861428e-01 -6.44697309e-01 -4.45157111e-01
-1.27914965e-01 7.90686190e-01 -1.76329628e-01 -6.61443651e-01
-1.93293437e-01 -4.36138064e-02 -9.53448534e-01 4.00634319e-01
1.01493025e+00 6.87410176e-01 1.58229887e-01 7.52939999e-01
-1.28347009e-01 9.54777837e-01 -6.90022051e-01 -7.07905233e-01
1.53249860e-01 -4.53590631e-01 -1.22055233e+00 -1.35329813e-01
2.54864842e-01 -2.33325496e-01 -3.46081674e-01 7.38643706e-01
8.13336253e-01 1.58736572e-01 4.68614042e-01 -9.49344575e-01
-5.25641918e-01 1.21744788e+00 -1.70971945e-01 1.84704959e-01
3.90999526e-01 2.80377045e-02 1.45919561e+00 -8.08057845e-01
6.36151135e-01 1.47646797e+00 6.59678102e-01 5.29730618e-01
-1.07698214e+00 -5.55514872e-01 2.67228544e-01 1.71904519e-01
-9.86365497e-01 -6.08092487e-01 5.62192976e-01 -1.54168934e-01
1.24646354e+00 1.52379096e-01 1.56978786e-01 1.56624079e+00
-1.21235520e-01 1.17560792e+00 1.17586660e+00 -7.02278912e-01
-2.40114957e-01 1.50295973e-01 2.31351480e-01 5.16849458e-01
-5.03960073e-01 -1.14015453e-01 -3.22083145e-01 4.98345010e-02
3.41459006e-01 -5.10008156e-01 -4.16455120e-01 -4.71857339e-02
-1.24587345e+00 1.16261089e+00 2.04236880e-01 4.93903279e-01
1.42259002e-02 -1.52966604e-01 7.87491143e-01 6.80460632e-01
7.20874250e-01 3.79893363e-01 -7.91836858e-01 -5.37362635e-01
-3.92615438e-01 1.84246838e-01 1.26742148e+00 8.81970048e-01
7.65738547e-01 -1.71897233e-01 -3.02679259e-02 1.57710636e+00
9.07096416e-02 4.43119794e-01 5.03735065e-01 -8.29386115e-01
9.60772455e-01 3.47199082e-01 -1.85803220e-01 -3.57003748e-01
-5.60830057e-01 -3.12583804e-01 -7.53794730e-01 -3.67941201e-01
5.95436454e-01 -6.87245250e-01 -3.01866442e-01 1.90703917e+00
1.33036941e-01 -2.14731172e-01 4.36763018e-01 7.08773077e-01
9.03217733e-01 8.45543265e-01 -8.99688751e-02 5.56046469e-03
1.21057045e+00 -1.39709651e+00 -8.64123821e-01 -1.71831384e-01
9.58973825e-01 -9.08453882e-01 1.28458655e+00 3.34813029e-01
-1.02231979e+00 -3.43603045e-01 -9.89272475e-01 -4.09382135e-01
-5.26163936e-01 2.53698826e-01 5.60904145e-01 5.25480151e-01
-1.12959719e+00 3.01598072e-01 -4.30579990e-01 -4.53836709e-01
-9.81472880e-02 3.10832858e-01 -3.13020468e-01 -2.56864607e-01
-1.83632421e+00 1.18472266e+00 9.94533822e-02 2.07736537e-01
-7.76577830e-01 -2.32075050e-01 -1.13619971e+00 -1.59953713e-01
4.18817878e-01 -2.75965512e-01 1.56835592e+00 -1.01883173e+00
-2.00844455e+00 9.75942373e-01 3.90784256e-02 -5.68494618e-01
6.46816969e-01 -2.96816647e-01 -2.91088104e-01 -1.28019586e-01
7.93788508e-02 5.45127332e-01 1.80599749e-01 -1.13642895e+00
-6.70713663e-01 -8.48085433e-02 4.93049115e-01 4.60460037e-01
-2.70820230e-01 6.78015500e-02 -3.97924453e-01 -5.82392216e-01
-5.68732679e-01 -9.87814367e-01 -6.67134747e-02 -8.32979441e-01
-5.32798290e-01 -5.93434095e-01 4.65558767e-01 -7.56737113e-01
1.05691051e+00 -1.77907956e+00 2.51502335e-01 -2.59278089e-01
-1.83512136e-01 3.62339258e-01 -2.91890472e-01 9.23522770e-01
5.81907928e-02 7.64006376e-02 -3.86615962e-01 -8.35330784e-01
3.01544070e-01 4.45784599e-01 -2.15683341e-01 3.73760939e-01
1.75625667e-01 7.10601449e-01 -9.73447561e-01 -3.29967916e-01
2.33386993e-01 4.48538601e-01 -5.04891753e-01 4.82405543e-01
-3.67163748e-01 8.58722627e-01 -1.85539871e-01 2.17611000e-01
4.59870756e-01 -3.85143086e-02 3.90549511e-01 1.82618886e-01
-1.72436953e-01 1.04847598e+00 -6.76383674e-01 2.07414436e+00
-1.03564084e+00 6.46664917e-01 2.72022486e-01 -1.04111385e+00
8.41159701e-01 6.45664036e-01 4.55919892e-01 -6.61943793e-01
2.31083483e-01 2.93504864e-01 8.80209953e-02 -5.27420163e-01
6.64378703e-01 -1.87976472e-02 -4.46951121e-01 8.94542396e-01
3.80515367e-01 -1.94428325e-01 4.60480809e-01 1.23437099e-01
8.84681523e-01 1.10970624e-01 1.83474779e-01 -2.75357991e-01
7.80398786e-01 -2.93359280e-01 4.93315965e-01 5.77454865e-01
-9.66506377e-02 3.52007419e-01 7.35093355e-01 -3.62991333e-01
-9.14760172e-01 -8.24577928e-01 -4.29208964e-01 1.64237678e+00
-3.12906295e-01 -4.12511766e-01 -7.21890867e-01 -9.41722155e-01
-2.07970232e-01 7.21567214e-01 -5.04187405e-01 2.24552885e-01
-8.83632720e-01 -9.67095196e-01 7.93915749e-01 2.45975062e-01
5.12550831e-01 -1.22732842e+00 1.27465859e-01 2.95621246e-01
-6.13282561e-01 -1.24729276e+00 -5.21206319e-01 3.37263525e-01
-3.17263395e-01 -1.00014162e+00 -6.39340878e-01 -8.80945563e-01
2.56994098e-01 -3.68095636e-01 1.34204769e+00 -1.37467951e-01
6.42989799e-02 3.41845393e-01 -6.61506891e-01 -3.05597305e-01
-7.63851464e-01 7.25899339e-01 -1.25591502e-01 -6.76494911e-02
3.89894515e-01 -4.00096476e-01 -2.17558920e-01 4.06909697e-02
-6.82201862e-01 2.35648118e-02 3.35196137e-01 1.14724207e+00
1.23570189e-01 -4.65121806e-01 1.02494740e+00 -1.19240499e+00
9.59600568e-01 -6.27375782e-01 -3.43504518e-01 2.29682684e-01
-2.72329867e-01 2.13503152e-01 7.30538249e-01 -1.55761406e-01
-1.19333196e+00 -2.47449100e-01 -7.52186656e-01 2.16340363e-01
-1.41077995e-01 7.43807971e-01 -3.25467139e-01 5.62631965e-01
3.78246099e-01 -7.40350178e-03 -1.13325775e-01 -6.59232616e-01
5.86183369e-01 9.90845680e-01 2.15513974e-01 -1.02903509e+00
1.08594678e-01 8.23357608e-03 -5.75839698e-01 -8.93709838e-01
-8.43995273e-01 -2.45838404e-01 -7.05505550e-01 4.96139470e-03
7.05877662e-01 -9.30742383e-01 -7.43827701e-01 4.68821019e-01
-1.19633830e+00 -8.38375986e-01 -9.50037688e-02 5.59202254e-01
-6.18762016e-01 4.52428222e-01 -1.14178121e+00 -7.21786082e-01
-4.29591447e-01 -1.40581477e+00 9.32150066e-01 -1.96090788e-01
-1.66959509e-01 -1.41500831e+00 4.02870715e-01 5.51325858e-01
3.59658837e-01 -1.44181460e-01 1.03767085e+00 -1.09975016e+00
-9.91334021e-02 2.78464183e-02 6.64368458e-03 7.51311183e-01
1.38385713e-01 -2.51094043e-01 -1.19706440e+00 -4.23153102e-01
-2.59039193e-01 -1.05324543e+00 7.46789634e-01 8.85841623e-02
7.50247121e-01 -3.28079462e-01 5.77665344e-02 3.27983081e-01
8.92840505e-01 5.43099223e-03 2.64520586e-01 3.16857636e-01
5.68349481e-01 1.04194140e+00 3.35353911e-01 4.69946504e-01
1.05990446e+00 7.52678692e-01 2.86750704e-01 -1.62670150e-01
1.72874853e-01 -1.07434206e-01 4.36534226e-01 1.59209967e+00
4.57757749e-02 -4.74768817e-01 -9.27874923e-01 6.71830773e-01
-1.96134734e+00 -6.27916574e-01 5.35946973e-02 2.00872993e+00
1.50823581e+00 -1.96076185e-01 1.12374008e-01 -2.28714615e-01
5.53682029e-01 2.78286070e-01 -2.73388326e-01 -6.74158335e-01
-2.54631162e-01 3.95686924e-01 1.88817471e-01 1.04918408e+00
-1.24259210e+00 1.23319077e+00 5.38369274e+00 5.80208957e-01
-1.04569817e+00 4.02450383e-01 7.47358382e-01 4.96427491e-02
-2.78716773e-01 -7.89420754e-02 -1.04549682e+00 3.03654969e-01
1.24050057e+00 -1.55462623e-01 3.46392512e-01 6.32224262e-01
1.49502948e-01 1.39420435e-01 -1.34647310e+00 6.08268976e-01
4.67209369e-02 -9.30684268e-01 -9.65319425e-02 -5.32825179e-02
6.50574625e-01 4.04051989e-01 -9.05777514e-02 9.17922676e-01
9.96259451e-01 -1.17546535e+00 3.78084809e-01 2.46886298e-01
6.47114813e-01 -9.42645431e-01 1.08600175e+00 4.05263394e-01
-9.52804565e-01 1.27019048e-01 -2.35892072e-01 -2.60388032e-02
2.74826348e-01 4.94951159e-01 -1.10483980e+00 7.88150847e-01
5.78971982e-01 8.87069225e-01 -1.88459188e-01 4.71229076e-01
-3.19873273e-01 6.31844938e-01 -2.00729430e-01 4.24770713e-02
5.09282351e-01 -2.30690196e-01 2.56075919e-01 1.61787307e+00
-8.72642770e-02 -4.02813137e-01 3.63184035e-01 4.76159394e-01
-4.79075372e-01 4.42859560e-01 -6.04222119e-01 2.79536787e-02
4.26875502e-01 1.05435777e+00 -1.00534819e-02 -2.16088116e-01
-6.99500740e-01 8.29486847e-01 7.54785955e-01 3.34339142e-01
-6.31271303e-01 -4.94339794e-01 6.95939362e-01 -3.92007023e-01
5.15901484e-02 -2.89834410e-01 1.60654873e-01 -1.31714749e+00
-5.16582318e-02 -1.28687298e+00 2.78095245e-01 -2.12434679e-01
-1.60673499e+00 1.08421171e+00 3.03507340e-03 -7.20395446e-01
-9.19701934e-01 -6.40423656e-01 -2.18705252e-01 1.09563434e+00
-1.73375440e+00 -1.29289591e+00 1.32811323e-01 7.30986059e-01
8.13301921e-01 -2.50146270e-01 1.13934326e+00 5.94353974e-01
-6.27318203e-01 8.45235407e-01 2.45520607e-01 6.02665484e-01
1.18925953e+00 -1.47664654e+00 3.23803246e-01 3.95298392e-01
1.05149977e-01 5.59950531e-01 4.00873303e-01 -1.46323487e-01
-9.63866949e-01 -9.69111621e-01 1.31768811e+00 -5.55883884e-01
9.26873982e-01 -8.09843123e-01 -8.62367868e-01 1.19875515e+00
9.92598891e-01 -3.91260743e-01 7.98043072e-01 4.42998141e-01
-2.03066558e-01 -3.11837625e-02 -7.82090366e-01 5.03925383e-01
7.08012342e-01 -6.25099421e-01 -5.97315192e-01 5.62568903e-01
8.22532415e-01 -5.35718143e-01 -1.13024175e+00 3.66262734e-01
2.75070727e-01 -8.73437047e-01 6.83644772e-01 -3.68280619e-01
3.11410338e-01 3.77524793e-01 -1.43010199e-01 -1.65548611e+00
3.36299121e-01 -7.75666535e-01 4.02930081e-01 1.42824411e+00
7.73318410e-01 -8.60763609e-01 2.62917191e-01 1.55089706e-01
-4.56069380e-01 -7.65912831e-01 -1.01362443e+00 -4.32348520e-01
6.78851247e-01 -3.28703612e-01 5.02991498e-01 1.26200032e+00
4.29099798e-01 8.48614275e-01 -6.13344431e-01 -2.35909313e-01
2.81639814e-01 3.20429057e-02 8.61311495e-01 -9.73258972e-01
-4.33070838e-01 -3.66254538e-01 2.72110641e-01 -1.50660527e+00
5.95960677e-01 -1.10112679e+00 2.24241406e-01 -1.34184921e+00
-1.90412387e-01 -7.30178475e-01 -2.03402862e-02 6.43736303e-01
-8.44705924e-02 -1.58565879e-01 2.87039354e-02 5.44588082e-04
-6.79547191e-01 9.03681159e-01 1.15943575e+00 -1.45844847e-01
-6.57800436e-02 2.28089336e-02 -4.50103819e-01 3.71899217e-01
7.86516666e-01 -2.30140477e-01 -4.08611566e-01 -5.97301185e-01
1.51395842e-01 1.81562483e-01 -2.29540676e-01 -6.14699781e-01
-5.39273508e-02 2.49759853e-01 -2.72902429e-01 -3.56567264e-01
5.42397380e-01 -4.52323496e-01 -6.04601860e-01 -2.59642415e-02
-7.33566523e-01 6.67771921e-02 2.28487760e-01 2.67075777e-01
-5.94473600e-01 -3.56708527e-01 5.86336493e-01 -2.35167906e-01
-2.84152627e-01 1.32838592e-01 -5.34075916e-01 4.81028646e-01
6.04871631e-01 5.73150635e-01 -4.78487670e-01 -7.29374647e-01
-8.16985190e-01 5.83214879e-01 9.12281051e-02 5.77904761e-01
3.03231850e-02 -1.00821078e+00 -1.02263081e+00 2.35714894e-02
2.12652534e-02 7.90624321e-02 3.57893072e-02 8.40654254e-01
-4.62050468e-01 8.19692254e-01 -6.86150193e-02 -4.15743560e-01
-1.05552065e+00 3.63189951e-02 4.91198719e-01 -8.18147779e-01
-4.19894665e-01 8.66468370e-01 2.55789906e-01 -1.39874887e+00
4.88863140e-01 -4.12556440e-01 -3.56329113e-01 1.29875302e-01
1.79784283e-01 1.23169936e-01 1.79321751e-01 -7.14612484e-01
-1.38124272e-01 2.98317592e-03 -4.30590749e-01 -3.47331583e-01
1.44148719e+00 -3.27752024e-01 -2.38357186e-01 7.98893511e-01
1.42633104e+00 1.10823475e-01 -1.29339588e+00 -6.74589276e-01
1.41299009e-01 6.63074777e-02 -3.23338419e-01 -8.92221570e-01
-9.26338315e-01 1.08964467e+00 2.36346573e-01 1.77301869e-01
7.61291087e-01 1.08083919e-01 8.91545773e-01 6.70620680e-01
4.08745378e-01 -1.13233638e+00 1.19436793e-01 1.31605053e+00
1.06854069e+00 -1.44661295e+00 -4.58232611e-01 -8.85146707e-02
-8.81595671e-01 1.01827598e+00 5.55342734e-01 3.76473204e-03
5.80534995e-01 3.49450260e-01 4.66400534e-01 1.09618694e-01
-8.69540989e-01 -1.54205993e-01 -2.03998297e-01 4.50232804e-01
1.08886147e+00 1.11423962e-01 -4.47836727e-01 8.32548559e-01
-5.17658412e-01 -3.09235930e-01 5.26979268e-01 7.87247598e-01
-1.34904524e-02 -1.69800854e+00 -1.93737876e-02 1.26737639e-01
-6.78121507e-01 -5.48502862e-01 -4.25138980e-01 8.84917140e-01
-7.77445808e-02 1.14893806e+00 -8.89517963e-02 -1.46748215e-01
2.97279775e-01 3.32809448e-01 2.80790925e-01 -7.90414631e-01
-9.16231990e-01 -3.28092114e-03 7.45530307e-01 -2.59670496e-01
-5.13170004e-01 -6.66609287e-01 -1.08885920e+00 -3.86509329e-01
-1.29399478e-01 5.28117359e-01 5.73529959e-01 1.02326322e+00
1.10199630e-01 5.21907091e-01 6.03144526e-01 -7.97731578e-01
-5.71162164e-01 -1.61359632e+00 -4.75344718e-01 3.55272502e-01
3.81957412e-01 -5.16615629e-01 -4.56406921e-01 -2.90444288e-02] | [12.423635482788086, 8.362923622131348] |
1e3ac908-1324-4d66-bde2-306e8f683e20 | improving-few-and-zero-shot-reaction-template | null | null | https://pubs.acs.org/doi/abs/10.1021/acs.jcim.1c01065 | https://pubs.acs.org/doi/pdf/10.1021/acs.jcim.1c01065 | Improving Few- and Zero-Shot Reaction Template Prediction Using Modern Hopfield Networks | Finding synthesis routes for molecules of interest is essential in the discovery of new drugs and materials. To find such routes, computer-assisted synthesis planning (CASP) methods are employed, which rely on a single-step model of chemical reactivity. In this study, we introduce a template-based single-step retrosynthesis model based on Modern Hopfield Networks, which learn an encoding of both molecules and reaction templates in order to predict the relevance of templates for a given molecule. The template representation allows generalization across different reactions and significantly improves the performance of template relevance prediction, especially for templates with few or zero training examples. With inference speed up to orders of magnitude faster than baseline methods, we improve or match the state-of-the-art performance for top-k exact match accuracy for k ≥ 3 in the retrosynthesis benchmark USPTO-50k. Code to reproduce the results is available at github.com/ml-jku/mhn-react. | ['and Günter Klambauer', 'Sepp Hochreiter', 'Marwin Segler', 'Jörg K. Wegner', 'Jonas Verhoeven', 'Paulo Neves', 'Natalia Dyubankova', 'Philipp Renz', 'Philipp Seidl'] | 2022-01-15 | null | null | null | journal-of-chemical-information-and-modeling-1 | ['retrosynthesis'] | ['medical'] | [ 4.53079492e-01 -3.39450827e-03 -4.58764017e-01 -1.23125732e-01
-5.94304502e-01 -1.09514463e+00 8.87959540e-01 5.62464833e-01
-3.04733634e-01 9.86269593e-01 2.15398967e-01 -4.96496350e-01
-1.96113423e-01 -8.12436640e-01 -8.49551499e-01 -8.20791721e-01
5.51566929e-02 4.20189619e-01 8.09448361e-02 -3.53414446e-01
6.60325289e-01 8.90577435e-01 -1.02387786e+00 5.30329227e-01
5.27737081e-01 9.73270178e-01 3.06520075e-01 4.92170364e-01
1.05652973e-01 4.56571400e-01 -1.15803041e-01 -5.23712099e-01
2.26962551e-01 -2.43412942e-01 -1.09279740e+00 -8.57316971e-01
1.52592212e-01 -1.71000399e-02 -5.31646132e-01 7.59720147e-01
6.98157370e-01 3.82010370e-01 9.42693233e-01 -6.14510894e-01
-4.07009959e-01 7.79861629e-01 4.52581704e-01 1.87607944e-01
3.74006301e-01 3.29428464e-01 1.00422502e+00 -1.28857791e+00
8.01175892e-01 8.53556573e-01 3.85494322e-01 5.89404464e-01
-1.18017817e+00 -9.50733602e-01 -6.19342700e-02 3.15215409e-01
-1.56881762e+00 -6.97858632e-01 3.69720131e-01 -3.63029957e-01
1.73057628e+00 -1.92520430e-03 6.86494112e-01 1.19166481e+00
7.20683038e-01 2.89182186e-01 5.16403735e-01 -1.72019213e-01
5.60510159e-01 -3.10192406e-01 -2.79301673e-01 6.77450776e-01
7.41292834e-02 4.96031374e-01 -7.98586011e-01 5.30264005e-02
5.49941540e-01 -8.78179893e-02 -2.23356947e-01 4.13618721e-02
-1.31056166e+00 9.47667897e-01 9.05542612e-01 2.57478565e-01
-5.60811102e-01 1.77559316e-01 2.79816300e-01 -1.35548472e-01
1.85667068e-01 1.40261638e+00 -6.35862231e-01 1.61928192e-01
-8.24784815e-01 4.50968504e-01 8.28881443e-01 1.06949985e+00
5.95624328e-01 1.42042441e-02 -1.88063473e-01 1.43201664e-01
-7.84195811e-02 -9.88161340e-02 9.90134329e-02 -8.91298115e-01
4.01447386e-01 3.78687650e-01 1.28482863e-01 -4.50374305e-01
-8.04386914e-01 -4.00569439e-01 -7.86409259e-01 -1.53680280e-01
3.81186604e-01 1.26281843e-01 -1.01928794e+00 1.43339360e+00
3.77962798e-01 -2.43332520e-01 2.98213810e-01 5.42749047e-01
1.03147960e+00 1.10076499e+00 2.79725283e-01 -4.85165238e-01
1.02704692e+00 -1.15855694e+00 -5.64204156e-01 2.50703245e-01
6.28255486e-01 -8.92758429e-01 3.03910971e-01 7.35559225e-01
-1.17062843e+00 -3.23889166e-01 -1.27774847e+00 -1.41453624e-01
-1.12358153e+00 -1.42985091e-01 8.08753014e-01 4.22779649e-01
-8.18318784e-01 1.60347879e+00 -4.06924099e-01 -5.56154214e-02
5.49951494e-01 9.65858102e-01 -5.06974101e-01 -6.37979880e-02
-1.44579732e+00 1.13200271e+00 1.01257515e+00 1.67809337e-01
-1.37273729e+00 -1.10526299e+00 -5.84393501e-01 1.33785009e-01
7.36724317e-01 -5.68524897e-01 1.37548983e+00 -5.38829625e-01
-1.86248839e+00 1.83202729e-01 -5.82103394e-02 -3.50572407e-01
1.15862310e-01 1.31151855e-01 -3.79710346e-01 2.09906429e-01
-1.77556768e-01 1.15482354e+00 5.26793718e-01 -6.07617021e-01
-5.03080226e-02 3.49249728e-02 2.74301823e-02 9.36795622e-02
3.33620280e-01 5.60609391e-03 5.32553419e-02 -4.61061984e-01
-2.03343585e-01 -8.25760126e-01 -5.74443042e-01 -2.67294854e-01
-5.20402193e-01 -3.44573230e-01 -2.26072297e-01 -5.94356596e-01
9.26234782e-01 -1.46226180e+00 4.08674896e-01 3.47380966e-01
5.69002219e-02 3.49738419e-01 -2.33480558e-01 1.05071640e+00
-6.11183345e-01 1.97437599e-01 -1.06389999e-01 4.08509165e-01
-1.69450536e-01 -3.83318156e-01 -2.23024726e-01 2.56474197e-01
4.13927048e-01 9.96581256e-01 -9.84285176e-01 1.76421419e-01
1.09995581e-01 5.39053082e-01 -5.43895125e-01 5.84148355e-02
-7.49130189e-01 4.91316110e-01 -3.52199256e-01 6.27740979e-01
1.54147699e-01 -4.00762945e-01 3.42085451e-01 -4.77824062e-01
-5.26145995e-01 7.65313983e-01 -8.62020969e-01 1.89797759e+00
-3.24764073e-01 -6.40790761e-02 -8.20526063e-01 -5.49428225e-01
7.37444699e-01 7.11679637e-01 1.81383178e-01 -7.67075062e-01
2.70719945e-01 4.27988648e-01 2.99206823e-01 1.31262839e-01
4.76221442e-01 -2.35654697e-01 2.55703956e-01 1.99848503e-01
3.29214007e-01 -2.20224842e-01 4.79775995e-01 1.76343933e-01
8.71902883e-01 3.27178150e-01 7.57326126e-01 -3.92230690e-01
5.86493611e-01 -5.73489070e-02 2.99498409e-01 5.63674092e-01
2.11343810e-01 2.16547728e-01 2.96155244e-01 -9.23442364e-01
-1.30099547e+00 -5.50365448e-01 1.11566009e-02 8.17542613e-01
-1.99176535e-01 -7.61305451e-01 -6.82589173e-01 -3.71576875e-01
-3.83946121e-01 8.70758891e-01 -6.26707137e-01 -2.70969987e-01
-3.41016591e-01 -5.05048752e-01 5.00450790e-01 5.27765810e-01
5.08922935e-02 -1.35874331e+00 1.08754234e-02 7.61903107e-01
2.30046451e-01 -8.39792252e-01 -5.92897058e-01 7.11301506e-01
-6.30044103e-01 -1.21468890e+00 -6.17777586e-01 -4.95329142e-01
4.80380237e-01 -2.92759063e-03 7.09766150e-01 -1.73884824e-01
-4.24583018e-01 -4.24937159e-01 -2.43122410e-02 -5.24752975e-01
-6.87980950e-01 3.85600150e-01 1.12273104e-01 -5.60099900e-01
-5.78383431e-02 -4.77646887e-01 -8.64981771e-01 1.33224279e-01
-7.80743897e-01 1.44233972e-01 4.70296979e-01 7.79142857e-01
7.00502932e-01 -6.85789958e-02 7.16000915e-01 -4.87223923e-01
2.64343768e-01 -2.88977832e-01 -1.12890661e+00 3.85375738e-01
-8.36449325e-01 5.71852565e-01 1.01194119e+00 -3.39267135e-01
-7.55780995e-01 4.74681020e-01 -2.99108297e-01 -7.45697916e-02
-1.92729943e-02 5.38699746e-01 -1.84325740e-01 -3.09912145e-01
6.81620419e-01 2.84025878e-01 -4.69790488e-01 -1.97429284e-01
5.60101509e-01 1.53166354e-01 7.23048300e-02 -6.52534485e-01
2.68192261e-01 1.17396691e-03 5.65460742e-01 -4.17585403e-01
-5.44802845e-01 -3.15784812e-01 -6.92080379e-01 6.94886595e-02
7.98512638e-01 -8.33244920e-01 -1.25366926e+00 1.81371674e-01
-1.36400151e+00 -5.61252058e-01 1.82412758e-01 5.28646171e-01
-7.55829334e-01 1.49746150e-01 -7.91629970e-01 -2.71361232e-01
-8.31743300e-01 -1.65105164e+00 7.84877539e-01 1.56335443e-01
-1.94366902e-01 -7.25883305e-01 9.42717940e-02 8.19928274e-02
1.76051795e-01 -2.43169758e-02 1.34516084e+00 -9.37641621e-01
-9.30970907e-01 -1.43223882e-01 2.62649413e-02 -1.76918104e-01
1.19883113e-01 -6.31306916e-02 -1.02459526e+00 -3.46649252e-02
-7.57086039e-01 -2.85268098e-01 9.28664625e-01 3.60714316e-01
1.27099681e+00 -5.58146834e-01 -5.13778985e-01 3.65966886e-01
1.34511936e+00 7.10047483e-01 7.88112879e-01 1.98236376e-01
6.56351030e-01 6.84277117e-01 4.82641995e-01 3.45427871e-01
-1.49368182e-01 5.88975370e-01 5.44933915e-01 2.39529461e-01
1.71909437e-01 -3.77091229e-01 1.33521169e-01 3.75131935e-01
-4.00272697e-01 -4.83315527e-01 -7.81875849e-01 1.29872784e-01
-1.48953259e+00 -1.02921641e+00 1.75563931e-01 2.38114667e+00
1.12839246e+00 1.56994328e-01 2.24241555e-01 1.04266711e-01
5.20028830e-01 -1.46066025e-01 -6.54940963e-01 -5.95541835e-01
2.57301211e-01 9.05670464e-01 7.84948289e-01 7.01861739e-01
-8.61549199e-01 1.52088404e+00 6.75388479e+00 1.13261092e+00
-1.30489874e+00 -3.67397219e-01 7.02784300e-01 -8.57604668e-02
1.23726099e-03 1.51119471e-01 -9.96102452e-01 1.89583525e-01
1.41213691e+00 -4.89302129e-02 9.71870840e-01 4.84407663e-01
2.79622853e-01 2.32250273e-01 -1.53745401e+00 5.95057309e-01
-4.59512413e-01 -2.45911765e+00 3.21736097e-01 2.60912180e-02
7.59815991e-01 -1.49667382e-01 -1.28603846e-01 -3.20241563e-02
3.89937907e-02 -1.36647797e+00 7.06979454e-01 3.26893508e-01
7.33389199e-01 -1.15581429e+00 2.33150244e-01 -3.93002704e-02
-1.10395348e+00 3.39459851e-02 -3.39863807e-01 1.52397037e-01
-1.73017547e-01 2.32972309e-01 -1.31107569e+00 8.28727424e-01
1.03106178e-01 7.79097676e-01 -1.43867329e-01 9.13568735e-01
-3.77345890e-01 1.88998654e-01 -2.16739401e-01 -3.47059667e-01
5.06112576e-01 -1.22772284e-01 7.86957070e-02 1.06058908e+00
3.37609738e-01 3.34147066e-01 1.12906024e-01 9.17278111e-01
-2.99672604e-01 2.27956876e-01 -5.88196218e-01 -4.98815179e-01
6.08034551e-01 1.14511323e+00 -9.48142827e-01 -2.98797697e-01
5.68803959e-03 7.23679602e-01 1.34154141e-01 4.07849908e-01
-7.53320158e-01 -4.50700879e-01 4.01795208e-01 -1.13319114e-01
6.47042394e-01 -1.59411937e-01 1.81783721e-01 -4.47047025e-01
-3.97834897e-01 -1.04962468e+00 2.16673315e-01 -7.42991626e-01
-6.57872915e-01 5.44874191e-01 -1.46483794e-01 -7.85699904e-01
7.50958100e-02 -1.02745664e+00 -4.55989659e-01 7.65495956e-01
-1.37160599e+00 -8.94546926e-01 4.07003671e-01 2.77071208e-01
6.24509692e-01 -1.57471061e-01 1.22929072e+00 5.09409718e-02
-5.26053429e-01 4.61050421e-01 2.30109662e-01 -4.27906930e-01
6.97751462e-01 -9.50928569e-01 6.83108926e-01 4.66868192e-01
-3.23391296e-02 8.42240453e-01 7.01587081e-01 -6.83780313e-01
-1.48143125e+00 -1.14140558e+00 9.66390312e-01 -2.94498295e-01
5.72472155e-01 -4.53744262e-01 -4.60763305e-01 2.84216732e-01
2.11260200e-01 -2.78022438e-01 7.70490885e-01 -1.83485612e-01
-6.53916359e-01 6.69346377e-02 -9.84715700e-01 8.45431209e-01
1.07422435e+00 -4.13892120e-01 2.70608682e-02 8.74619365e-01
1.03844082e+00 -5.04317164e-01 -1.36379409e+00 2.15874732e-01
7.09138870e-01 -4.71143216e-01 1.17292988e+00 -8.15064132e-01
6.31842136e-01 -3.45067799e-01 -1.18861876e-01 -1.22986674e+00
-5.86305797e-01 -9.97906625e-01 1.30851632e-02 3.42161596e-01
9.60194111e-01 -5.19539595e-01 5.09945989e-01 3.98005843e-01
-4.37886715e-01 -9.17500019e-01 -8.16056490e-01 -7.59932458e-01
1.38649106e-01 -3.19407731e-02 8.59782934e-01 8.17719936e-01
3.64929527e-01 6.81683898e-01 -1.34864256e-01 2.28827566e-01
9.06311944e-02 -2.82111224e-02 3.34236063e-02 -8.79254699e-01
-3.36606145e-01 -6.54185653e-01 -5.07698534e-03 -5.74249446e-01
1.07185870e-01 -1.24419475e+00 -4.46875719e-03 -1.33460379e+00
1.02953807e-01 -9.57495645e-02 -3.06177080e-01 7.35410869e-01
1.23564415e-01 2.28262898e-02 1.78578030e-02 -5.62613755e-02
-5.41556954e-01 6.02662921e-01 1.23351336e+00 -3.62357438e-01
-3.10028553e-01 -1.74699441e-01 -6.77973509e-01 1.18730165e-01
9.02107418e-01 -6.68892086e-01 -3.55504483e-01 1.94135591e-01
7.40319729e-01 3.00557077e-01 -1.81748345e-02 -9.34968293e-01
3.23668212e-01 -3.44758868e-01 6.08333468e-01 -6.96369529e-01
5.56065798e-01 -4.45344985e-01 4.53015059e-01 8.01739037e-01
-7.37479985e-01 -6.47947267e-02 3.18198323e-01 5.95294595e-01
3.03167582e-01 -2.90716439e-01 4.61459398e-01 -4.59716320e-01
-3.11439246e-01 7.95132995e-01 -5.80924034e-01 -6.39686882e-01
6.68749392e-01 -9.04423147e-02 -5.08558333e-01 -9.12277773e-02
-6.78896248e-01 -6.31387532e-02 1.64454058e-01 2.16252685e-01
5.02920210e-01 -1.07910955e+00 -2.80419350e-01 -2.38316104e-01
3.26025248e-01 1.22766308e-01 9.94054228e-02 6.57008231e-01
-7.24191844e-01 1.21433425e+00 -2.00894415e-01 5.71079552e-02
-9.40326154e-01 1.12624359e+00 5.22557795e-01 -3.18568826e-01
-1.54800899e-03 7.95230031e-01 2.89146930e-01 -2.39166513e-01
2.29134232e-01 -5.22131562e-01 -1.24904983e-01 -1.04965486e-01
7.08537281e-01 2.31676772e-01 6.00624025e-01 -2.38890231e-01
-4.16168094e-01 3.76207173e-01 -5.85350096e-01 3.25938344e-01
1.34080470e+00 7.76151240e-01 1.72298014e-01 -1.37511000e-01
9.28625464e-01 -5.19829452e-01 -1.12388074e+00 -4.17390466e-02
-6.38927817e-02 1.37212619e-01 1.57328621e-01 -1.22597206e+00
-5.67944348e-01 8.27969909e-01 3.34528357e-01 -3.84982735e-01
6.40434206e-01 -2.58517802e-01 5.10111034e-01 1.07497156e+00
2.80193537e-01 -9.38064158e-01 1.34734929e-01 5.87774336e-01
1.09359515e+00 -9.58872974e-01 3.17220241e-01 -4.41655874e-01
-2.83669502e-01 1.48752737e+00 5.94221950e-02 1.31287575e-01
3.74336839e-01 -3.00495297e-01 -5.89229107e-01 -3.29429835e-01
-8.21536541e-01 4.32655290e-02 2.46196330e-01 3.51786047e-01
5.83122551e-01 5.53564131e-02 -2.80601621e-01 3.43655914e-01
5.22349812e-02 -9.33063626e-02 3.85241479e-01 8.29304934e-01
-3.21728706e-01 -1.46783412e+00 -2.63555604e-03 -5.45741841e-02
-4.44813192e-01 -8.21622074e-01 -7.59710252e-01 5.21195292e-01
-1.11357115e-01 8.15094650e-01 -5.61486125e-01 -2.96981215e-01
1.40501633e-01 3.13939393e-01 1.08734369e+00 -4.49647725e-01
-9.31252182e-01 3.16072889e-02 3.51355344e-01 -7.15027630e-01
-3.12743634e-01 -1.70646459e-01 -1.37168443e+00 -4.46174741e-01
-5.29104710e-01 2.64355361e-01 7.96205640e-01 9.91017818e-01
6.39273643e-01 3.67968410e-01 5.16211629e-01 -1.24597597e+00
-4.21756536e-01 -6.01290047e-01 -1.45411313e-01 -4.51905966e-01
9.36703011e-02 -5.43056011e-01 2.38692448e-01 -7.78427497e-02] | [4.510021686553955, 6.088146686553955] |
96323041-cb70-4c20-94e2-641e4c8b9677 | deep-implicit-statistical-shape-models-for-3d | 2104.02847 | null | https://arxiv.org/abs/2104.02847v2 | https://arxiv.org/pdf/2104.02847v2.pdf | Deep Implicit Statistical Shape Models for 3D Medical Image Delineation | 3D delineation of anatomical structures is a cardinal goal in medical imaging analysis. Prior to deep learning, statistical shape models that imposed anatomical constraints and produced high quality surfaces were a core technology. Prior to deep learning, statistical shape models that imposed anatomical constraints and produced high quality surfaces were a core technology. Today fully-convolutional networks (FCNs), while dominant, do not offer these capabilities. We present deep implicit statistical shape models (DISSMs), a new approach to delineation that marries the representation power of convolutional neural networks (CNNs) with the robustness of SSMs. DISSMs use a deep implicit surface representation to produce a compact and descriptive shape latent space that permits statistical models of anatomical variance. To reliably fit anatomically plausible shapes to an image, we introduce a novel rigid and non-rigid pose estimation pipeline that is modelled as a Markov decision process(MDP). We outline a training regime that includes inverted episodic training and a deep realization of marginal space learning (MSL). Intra-dataset experiments on the task of pathological liver segmentation demonstrate that DISSMs can perform more robustly than three leading FCN models, including nnU-Net: reducing the mean Hausdorff distance (HD) by 7.7-14.3mm and improving the worst case Dice-Sorensen coefficient (DSC) by 1.2-2.3%. More critically, cross-dataset experiments on a dataset directly reflecting clinical deployment scenarios demonstrate that DISSMs improve the mean DSC and HD by 3.5-5.9% and 12.3-24.5mm, respectively, and the worst-case DSC by 5.4-7.3%. These improvements are over and above any benefits from representing delineations with high-quality surface. | ['Adam P. Harrison', 'Junzhou Huang', 'Dakai Jin', 'Le Lu', 'Shun Miao', 'Ashwin Raju'] | 2021-04-07 | null | null | null | null | ['liver-segmentation'] | ['medical'] | [ 1.91038936e-01 5.85929871e-01 2.30421975e-01 -3.37292314e-01
-1.03118336e+00 -6.55918062e-01 7.02776849e-01 2.14197859e-01
-3.46321076e-01 4.08852547e-01 1.90725029e-01 -5.41156530e-01
-3.25769395e-01 -3.81456643e-01 -6.76418841e-01 -8.53914320e-01
-5.00975311e-01 5.75215876e-01 2.01653570e-01 1.42618820e-01
5.90309314e-02 8.90363932e-01 -5.98886013e-01 2.34368771e-01
8.24155331e-01 9.37044859e-01 -1.21006712e-01 8.17904949e-01
-1.13150086e-02 2.23854527e-01 -2.41903782e-01 -1.15853235e-01
3.78552884e-01 -1.95062876e-01 -9.54689920e-01 3.00438087e-02
4.30948764e-01 -4.18296784e-01 -1.70059264e-01 6.40017092e-01
5.19870400e-01 -5.30904233e-02 9.58053946e-01 -6.85287118e-01
-3.77153903e-01 3.40280741e-01 -3.65466267e-01 1.25009984e-01
-1.26589179e-01 3.81774038e-01 3.32254201e-01 -6.41655564e-01
5.60121775e-01 1.00791311e+00 1.31969750e+00 5.32320678e-01
-1.42566776e+00 -1.11324310e-01 -3.60606402e-01 -6.64537907e-01
-1.24996209e+00 -3.95939231e-01 2.50678778e-01 -6.99170530e-01
8.85453582e-01 6.17852390e-01 5.92066228e-01 6.37085319e-01
8.11034560e-01 4.82173145e-01 1.15168536e+00 -1.48949802e-01
1.12708397e-01 -2.15061858e-01 -2.60457933e-01 8.67140234e-01
3.44617128e-01 1.68522790e-01 2.13315710e-01 -2.96765000e-01
1.37648332e+00 -1.33022279e-01 -2.63073206e-01 -5.58793843e-01
-1.23285055e+00 7.50693321e-01 6.54074490e-01 2.98707396e-01
-4.99456197e-01 4.04605150e-01 3.21834654e-01 -2.36247092e-01
3.31890047e-01 5.82946002e-01 -3.18598807e-01 -5.60242310e-02
-1.32601583e+00 -8.23077336e-02 7.65622854e-01 6.61842465e-01
9.41532850e-02 -1.51724312e-02 -2.10545227e-01 4.36901897e-01
4.67892349e-01 3.70539725e-01 2.86350369e-01 -1.07610714e+00
-2.12503135e-01 4.69973505e-01 -1.87151115e-02 -9.05320525e-01
-8.38122010e-01 -8.32387269e-01 -1.12768674e+00 2.88140476e-01
6.74666703e-01 1.31772896e-02 -1.38310111e+00 1.40558577e+00
3.66342098e-01 -3.81424697e-03 -2.69583493e-01 9.21769261e-01
8.21114480e-01 1.37992084e-01 2.82171875e-01 -8.96505266e-02
1.26823056e+00 -5.07124066e-01 -4.04700160e-01 1.02731891e-01
8.62630844e-01 -8.52368295e-01 8.59860659e-01 3.07424992e-01
-1.41640270e+00 -2.15981767e-01 -8.60405803e-01 7.32695907e-02
8.27939957e-02 -1.59517959e-01 5.03792107e-01 6.68114007e-01
-1.44249296e+00 1.02903187e+00 -1.45823765e+00 -1.65005803e-01
8.31278443e-01 6.38992131e-01 -3.56678963e-01 1.26919463e-01
-6.70389235e-01 1.10963714e+00 7.33947828e-02 1.60007447e-01
-9.70077395e-01 -1.09549642e+00 -7.92337656e-01 -1.66889355e-01
-1.15006333e-02 -9.47660327e-01 1.11025882e+00 -7.41414368e-01
-1.41986310e+00 1.00644851e+00 2.01057419e-01 -7.20706701e-01
1.00616527e+00 -1.42210215e-01 -2.60144961e-03 4.51803982e-01
-1.07776292e-01 7.26705372e-01 5.87393701e-01 -1.31421411e+00
1.35045469e-01 -2.21742004e-01 -3.11693788e-01 9.64705721e-02
3.89363289e-01 -9.61102545e-02 -2.46003091e-01 -7.27609336e-01
4.78199303e-01 -1.12270892e+00 -6.15789831e-01 3.25538576e-01
-4.31544781e-01 2.58537591e-01 4.43579346e-01 -1.03463888e+00
6.85120225e-01 -1.82944596e+00 -1.84183568e-02 3.44070643e-01
5.21715581e-01 2.65839994e-01 1.94919676e-01 -1.40442386e-01
-1.80974212e-02 2.77894139e-01 -8.60150158e-01 -4.13460046e-01
-1.35606810e-01 3.55934888e-01 2.36937374e-01 9.69440699e-01
1.52969524e-01 1.15170860e+00 -7.94458747e-01 -5.08062482e-01
3.10315073e-01 8.10578346e-01 -5.41848242e-01 3.75276897e-03
1.60781652e-01 7.04095542e-01 -5.75119480e-02 5.42421639e-01
7.98008382e-01 -3.81069243e-01 2.60569304e-01 -3.71177912e-01
-1.14022091e-01 8.14752057e-02 -9.03109014e-01 1.98706222e+00
-5.27738750e-01 3.90181690e-01 3.95567983e-01 -6.63245738e-01
6.47219181e-01 3.61838579e-01 9.98667240e-01 -3.86988521e-01
2.23765463e-01 5.50026596e-01 2.59296626e-01 -3.00085962e-01
-1.86155625e-02 -5.39295673e-01 2.42010981e-01 2.15889886e-01
1.19606845e-01 -3.65096420e-01 -3.89950424e-01 -9.83033851e-02
1.06244385e+00 3.67278069e-01 1.52407050e-01 -1.03250396e+00
3.32135081e-01 5.90733886e-02 5.01074553e-01 5.32825291e-01
-4.24158216e-01 1.02151203e+00 4.20414567e-01 -7.81142235e-01
-1.14352512e+00 -1.49473190e+00 -6.24584258e-01 4.80977893e-01
-1.78044960e-01 2.05160622e-02 -8.51432621e-01 -7.84015775e-01
1.14473686e-01 4.50052291e-01 -6.69350743e-01 -6.39216825e-02
-7.30412304e-01 -8.89363825e-01 6.91659749e-01 5.43950319e-01
7.51001760e-02 -8.54887307e-01 -9.45498168e-01 2.28094250e-01
1.87263973e-02 -8.81880999e-01 -3.96042168e-01 2.21579999e-01
-1.21961892e+00 -1.13013005e+00 -1.01745224e+00 -4.89932507e-01
8.99083138e-01 -2.74627596e-01 1.25915921e+00 2.00811341e-01
-7.43622184e-01 1.52471349e-01 1.83997806e-02 -2.45861322e-01
-7.31634855e-01 2.41664425e-02 1.49349391e-01 -3.81585836e-01
-2.86023527e-01 -5.11247575e-01 -1.07890189e+00 1.73195899e-01
-9.80689287e-01 2.19566360e-01 8.96245599e-01 6.74985588e-01
7.35134363e-01 -4.70573694e-01 2.35025033e-01 -8.05817008e-01
1.83901906e-01 -3.58630508e-01 -3.47682953e-01 1.73136331e-02
-6.31972909e-01 1.61445931e-01 3.80771637e-01 -2.45067656e-01
-7.94591367e-01 9.27650556e-02 -1.75745308e-01 -5.58945417e-01
-3.80562961e-01 3.99757564e-01 4.19443965e-01 -2.87407666e-01
9.04832959e-01 5.96300550e-02 5.89028835e-01 -3.95490557e-01
3.20078135e-01 2.38001108e-01 5.72576940e-01 -5.11686504e-01
6.62841558e-01 9.02378857e-01 5.08295178e-01 -6.21618748e-01
-6.39741540e-01 -3.61533999e-01 -1.04089427e+00 1.69254746e-03
1.01083553e+00 -5.98742962e-01 -5.09814620e-01 1.88221887e-01
-7.59770572e-01 -7.61682093e-01 -4.50987756e-01 5.59152424e-01
-8.25206101e-01 5.26131034e-01 -9.08834040e-01 -4.93082404e-01
-7.62856662e-01 -1.42854941e+00 1.01280856e+00 -7.29075493e-03
-5.29978216e-01 -1.31394541e+00 -1.93970963e-01 1.29064351e-01
8.98760021e-01 9.48364913e-01 7.77586877e-01 -6.55385375e-01
-3.54515284e-01 9.33538526e-02 -2.64591008e-01 2.50260264e-01
2.31345579e-01 -8.26538429e-02 -8.18415642e-01 -5.69210649e-01
5.16765676e-02 2.08743542e-01 6.05551064e-01 1.05617428e+00
1.14541924e+00 -2.92772114e-01 -3.00268203e-01 1.05450308e+00
1.37456930e+00 6.39500767e-02 6.26126945e-01 2.54750073e-01
5.83938479e-01 4.48318005e-01 5.09966165e-02 1.89121008e-01
1.06872402e-01 4.46213752e-01 5.92664003e-01 -6.29472733e-01
-5.06954551e-01 1.08548164e-01 4.84033450e-02 9.15812969e-01
-1.07768215e-01 5.26398242e-01 -1.44584119e+00 6.20465040e-01
-1.35369873e+00 -4.20694083e-01 -5.19415081e-01 2.14404702e+00
8.99845362e-01 2.30195925e-01 8.27306733e-02 -2.88623512e-01
3.04411560e-01 -2.23824590e-01 -4.59629267e-01 -4.99751776e-01
1.79619044e-01 4.04420912e-01 7.42941201e-01 6.18313849e-01
-1.23399913e+00 5.20602643e-01 5.75466394e+00 5.75880110e-01
-1.21235323e+00 1.44585088e-01 9.18968260e-01 3.77638824e-02
-1.57022238e-01 -1.30266041e-01 -1.91318870e-01 2.65086830e-01
1.07528305e+00 2.66427815e-01 2.19251752e-01 5.19911289e-01
2.86567360e-01 -1.48172349e-01 -1.02618504e+00 8.65408838e-01
-2.35164881e-01 -1.52566981e+00 -1.36849999e-01 3.80399406e-01
7.85699964e-01 2.65291303e-01 1.59896046e-01 2.30002422e-02
1.65636331e-01 -1.59889042e+00 5.47404110e-01 8.19294274e-01
1.38886917e+00 -5.17233372e-01 9.75796759e-01 2.42300928e-01
-9.07025456e-01 3.45266670e-01 -1.66773245e-01 5.03559411e-01
1.57097012e-01 6.00213051e-01 -1.12391901e+00 6.75614357e-01
6.30744278e-01 4.14440960e-01 -4.05951291e-01 1.23833084e+00
2.71326602e-01 4.45185155e-01 -5.35962939e-01 5.85194528e-01
5.09521008e-01 -1.54842794e-01 5.96352816e-01 1.52622235e+00
2.12975755e-01 1.63957193e-01 -2.08365042e-02 1.06424296e+00
1.01549409e-01 1.81057695e-02 -2.86504537e-01 2.92688042e-01
-2.02865768e-02 1.44588935e+00 -1.04206765e+00 -1.61529973e-01
2.76245922e-02 8.20271075e-01 -1.26658022e-01 -6.27688691e-02
-8.37597847e-01 2.77814306e-02 4.47734654e-01 4.85288322e-01
-5.97450137e-02 -2.34308645e-01 -6.70239329e-01 -7.66874611e-01
-2.70518303e-01 -6.34876549e-01 3.50745738e-01 -4.15474892e-01
-1.24693608e+00 5.37618041e-01 -1.12223335e-01 -1.00830483e+00
-4.21829289e-03 -5.62575817e-01 -5.74223340e-01 1.16845930e+00
-1.33711529e+00 -1.24938154e+00 -2.66627103e-01 7.47670978e-02
3.12563360e-01 2.11256206e-01 1.06200981e+00 9.89575684e-03
-1.27810761e-01 4.19220448e-01 2.35727936e-01 3.45530868e-01
4.09664392e-01 -1.51118112e+00 3.77734721e-01 5.64776897e-01
-2.16804221e-01 8.19405794e-01 5.23601294e-01 -6.07400477e-01
-1.29006445e+00 -1.15790093e+00 4.27778572e-01 -6.06748939e-01
3.44896287e-01 -5.79497684e-03 -1.02794778e+00 6.19374931e-01
2.37074941e-02 5.27650476e-01 7.26118147e-01 -2.04149976e-01
-2.21915673e-02 3.72770846e-01 -1.51651502e+00 4.68649685e-01
8.60692620e-01 -2.95401812e-01 -6.25228107e-01 2.49029070e-01
3.63344878e-01 -8.46500337e-01 -1.47296715e+00 8.09991717e-01
6.22790337e-01 -8.42160225e-01 1.07433331e+00 -5.64222515e-01
3.34878087e-01 -1.72583893e-01 -3.48593034e-02 -1.27169907e+00
-4.59701627e-01 -8.39423478e-01 -2.27107868e-01 5.10263145e-01
2.82027185e-01 -4.52062160e-01 8.10474098e-01 8.10260594e-01
-6.96638286e-01 -1.26655567e+00 -1.18335760e+00 -6.62833929e-01
7.50630915e-01 -3.63694400e-01 2.16883704e-01 1.01387775e+00
-2.63310581e-01 -3.65792841e-01 3.74102205e-01 1.49703085e-01
8.34040165e-01 -9.59322602e-02 2.46118993e-01 -1.21392751e+00
-7.26155713e-02 -8.17524195e-01 -3.26261133e-01 -6.65695667e-01
-2.91587353e-01 -1.28300524e+00 -5.25457226e-02 -1.89800179e+00
2.20628113e-01 -7.44096935e-01 -3.65630835e-01 3.38320434e-01
1.50398925e-01 2.25771099e-01 9.43038613e-02 3.51999044e-01
-2.08969101e-01 1.24832101e-01 1.47462773e+00 2.51151770e-01
-2.41834596e-01 -4.41008247e-02 -4.00775254e-01 8.87724280e-01
1.00695622e+00 -3.12882036e-01 -1.73247784e-01 -2.79861510e-01
-2.10734382e-01 2.16102004e-01 5.67219794e-01 -1.00075161e+00
3.60263400e-02 1.44302681e-01 7.84664750e-01 -3.79391819e-01
9.93442088e-02 -7.84649968e-01 4.01846290e-01 9.43278432e-01
-1.78268805e-01 -1.70768678e-01 5.84589303e-01 3.01973075e-01
3.25902104e-01 1.69057101e-01 1.15180254e+00 -3.76969695e-01
3.37044001e-02 4.51128095e-01 -2.20390901e-01 1.12196080e-01
9.25198674e-01 -3.58939648e-01 4.22102585e-02 -4.82370593e-02
-1.21727705e+00 -6.89579770e-02 4.52203363e-01 -2.83888951e-02
6.73136532e-01 -1.13694930e+00 -8.57290685e-01 2.63889253e-01
-3.80141824e-01 5.31292021e-01 3.63986641e-01 1.66622305e+00
-1.10035563e+00 5.49553216e-01 -1.01326630e-01 -1.06833851e+00
-9.60642457e-01 1.79618984e-01 1.02796650e+00 -3.85530591e-01
-1.00817037e+00 9.74284053e-01 3.20219874e-01 -5.11996031e-01
-1.43140391e-01 -7.54184365e-01 2.97482193e-01 -3.68336231e-01
3.29167135e-02 2.62033105e-01 2.96858281e-01 -7.24329174e-01
-4.38947499e-01 4.02302533e-01 7.30063766e-02 1.14031330e-01
1.48795068e+00 3.88650671e-02 -9.23183933e-02 9.95494872e-02
1.28671777e+00 -1.95132479e-01 -1.63109469e+00 -5.65015413e-02
1.32816106e-01 -1.47171453e-01 3.14133555e-01 -1.18652380e+00
-1.00077760e+00 8.78311872e-01 7.34281600e-01 1.20902762e-01
7.81328022e-01 -5.47694266e-02 8.36061299e-01 -4.37692553e-01
1.24809943e-01 -7.35830843e-01 -1.80548832e-01 2.68702000e-01
1.03971589e+00 -1.03074491e+00 1.22865573e-01 -4.30783063e-01
-5.52747250e-01 1.30484617e+00 2.46686295e-01 -1.78142980e-01
8.31069589e-01 4.90780592e-01 1.44045219e-01 -4.89669591e-01
-2.05432579e-01 1.82397991e-01 7.59975433e-01 4.48558152e-01
6.85765505e-01 4.45183516e-01 -5.57627119e-02 4.98160660e-01
-9.88878310e-02 -1.30232587e-01 2.39046872e-01 8.08428228e-01
-2.06727475e-01 -6.49143159e-01 -3.41862172e-01 5.23457110e-01
-8.18192065e-01 -1.13552593e-01 2.19388112e-01 9.06649768e-01
-1.90233514e-02 3.70303363e-01 1.13830268e-01 8.63582641e-02
2.14955464e-01 1.35110309e-02 5.74940085e-01 -5.30578613e-01
-7.72386074e-01 3.64222139e-01 -1.34066164e-01 -6.74047709e-01
-1.58103392e-01 -7.86544561e-01 -1.58350527e+00 -3.54049325e-01
-5.44527955e-02 -2.14377016e-01 7.96907961e-01 8.08809519e-01
3.68719250e-01 6.45834982e-01 2.03669831e-01 -9.30998921e-01
-7.79926121e-01 -7.86900342e-01 -3.66146684e-01 3.86133224e-01
3.94150943e-01 -4.86732244e-01 -3.00073564e-01 1.12670645e-01] | [14.22671890258789, -2.4550859928131104] |
6416121d-3bd7-46b0-9ff9-ab30cf2eafb5 | neural-chronos-ode-unveiling-temporal | 2307.01023 | null | https://arxiv.org/abs/2307.01023v1 | https://arxiv.org/pdf/2307.01023v1.pdf | Neural Chronos ODE: Unveiling Temporal Patterns and Forecasting Future and Past Trends in Time Series Data | This work introduces Neural Chronos Ordinary Differential Equations (Neural CODE), a deep neural network architecture that fits a continuous-time ODE dynamics for predicting the chronology of a system both forward and backward in time. To train the model, we solve the ODE as an initial value problem and a final value problem, similar to Neural ODEs. We also explore two approaches to combining Neural CODE with Recurrent Neural Networks by replacing Neural ODE with Neural CODE (CODE-RNN), and incorporating a bidirectional RNN for full information flow in both time directions (CODE-BiRNN), and variants with other update cells namely GRU and LSTM: CODE-GRU, CODE-BiGRU, CODE-LSTM, CODE-BiLSTM. Experimental results demonstrate that Neural CODE outperforms Neural ODE in learning the dynamics of a spiral forward and backward in time, even with sparser data. We also compare the performance of CODE-RNN/-GRU/-LSTM and CODE-BiRNN/-BiGRU/-BiLSTM against ODE-RNN/-GRU/-LSTM on three real-life time series data tasks: imputation of missing data for lower and higher dimensional data, and forward and backward extrapolation with shorter and longer time horizons. Our findings show that the proposed architectures converge faster, with CODE-BiRNN/-BiGRU/-BiLSTM consistently outperforming the other architectures on all tasks. | ['L. L. Ferrás', 'M. Fernanda P. Costa', 'C. Coelho'] | 2023-07-03 | null | null | null | null | ['imputation', 'imputation', 'imputation'] | ['computer-vision', 'miscellaneous', 'time-series'] | [-6.07881427e-01 -2.18640938e-01 6.71578273e-02 1.74724385e-01
-1.21342167e-01 -5.11842191e-01 3.73107582e-01 -3.70290816e-01
-1.19799383e-01 8.41359913e-01 3.12931597e-01 -1.09340513e+00
-2.13938847e-01 -3.73633862e-01 -8.45111489e-01 -6.07563853e-01
-6.47372127e-01 4.69770133e-01 -2.26518571e-01 -5.17364800e-01
6.19787723e-02 1.69897690e-01 -8.42956245e-01 1.74077883e-01
5.95207870e-01 1.08737159e+00 -1.14192553e-01 1.05049658e+00
5.39189354e-02 1.85511732e+00 -2.11889029e-01 -2.82897621e-01
3.16815645e-01 -5.41349471e-01 -6.50413513e-01 -9.27680850e-01
-4.75698829e-01 -4.67179030e-01 -9.35782373e-01 3.38935077e-01
3.75459045e-01 2.50360310e-01 8.26317668e-01 -1.20354545e+00
-1.12599134e+00 6.09483004e-01 -3.35636020e-01 5.76205015e-01
2.04332232e-01 4.84127700e-01 5.67108870e-01 -7.85504520e-01
4.41154510e-01 1.08800030e+00 1.70626605e+00 5.22503257e-01
-1.23547387e+00 -5.80899835e-01 1.05085649e-01 -2.02878732e-02
-1.05078208e+00 3.41326296e-02 5.67405343e-01 -7.55049884e-01
1.60827720e+00 7.16530830e-02 8.38114500e-01 1.44513595e+00
1.03161085e+00 5.04706025e-01 8.59185755e-01 2.94894606e-01
1.88137665e-01 -6.53745949e-01 1.92393333e-01 5.36948144e-01
-2.53188133e-01 6.56649888e-01 -6.80406839e-02 -1.72138959e-01
1.15516722e+00 4.38980609e-01 3.37608345e-02 3.10861677e-01
-1.22864616e+00 6.43363893e-01 7.36602843e-01 4.75775182e-01
-5.84363759e-01 7.51047075e-01 9.09290552e-01 9.09136891e-01
6.92361057e-01 5.55007577e-01 -9.10583913e-01 -5.48759699e-01
-8.46828163e-01 1.67923674e-01 1.13311243e+00 7.28218853e-01
3.33215564e-01 8.10116053e-01 9.82110351e-02 6.38676703e-01
-1.64945930e-01 3.70788962e-01 1.01780856e+00 -1.12021291e+00
2.15097278e-01 3.86184126e-01 2.85281278e-02 -6.72912478e-01
-7.72528231e-01 -4.43647861e-01 -1.57656360e+00 3.78187634e-02
2.72122324e-01 -8.21924925e-01 -1.29230547e+00 1.82700884e+00
-7.68034682e-02 6.17985129e-01 2.14737445e-01 6.13409579e-01
5.55753887e-01 1.35733235e+00 -3.71523350e-01 -6.97234124e-02
7.09970951e-01 -1.11399376e+00 -6.93806648e-01 1.19016603e-01
1.10004652e+00 -2.71860719e-01 6.31310582e-01 1.49758279e-01
-1.14385164e+00 -6.29601300e-01 -6.33929193e-01 -2.68847615e-01
-5.56798279e-01 6.55452814e-03 4.91076142e-01 -1.96661308e-01
-1.55639505e+00 9.89099622e-01 -1.17481959e+00 -2.10373864e-01
-3.20266604e-01 5.29995739e-01 -8.09136033e-02 4.17639881e-01
-1.53913999e+00 1.14706075e+00 1.15414418e-01 6.75778687e-01
-1.20152056e+00 -1.11442864e+00 -1.02931237e+00 -1.99697316e-02
-3.43585044e-01 -8.96105707e-01 1.63637924e+00 -7.91651428e-01
-1.63635671e+00 1.02902733e-01 -1.58846810e-01 -1.22792697e+00
6.31396413e-01 -7.62358680e-02 -5.16939700e-01 -5.93620539e-01
-1.48502856e-01 3.68491411e-01 4.38350648e-01 -7.63079166e-01
-3.65028977e-02 7.51643330e-02 -2.40432575e-01 -1.50910079e-01
3.41966562e-02 -1.94573820e-01 1.85123488e-01 -9.37748909e-01
-8.08015615e-02 -1.10975015e+00 -4.90468174e-01 -4.05160546e-01
-2.61688679e-01 -1.88464656e-01 9.79642689e-01 -1.29494083e+00
1.74502790e+00 -1.59648204e+00 1.95516512e-01 -2.82711953e-01
7.63460174e-02 4.16494310e-01 -7.51987174e-02 9.62590218e-01
-5.07156014e-01 -8.64925757e-02 -5.20467579e-01 -3.40776026e-01
-3.35049003e-01 5.62696636e-01 -8.21544409e-01 1.98776662e-01
1.19597279e-01 1.33817613e+00 -7.81372786e-01 2.63783097e-01
5.30438051e-02 5.81336141e-01 -3.97116840e-01 1.74840286e-01
-4.05596584e-01 6.13583267e-01 -1.35697439e-01 3.04603428e-01
2.59898216e-01 -3.62232774e-01 -2.81767726e-01 4.09473836e-01
-5.62568665e-01 3.07607591e-01 -7.94209301e-01 1.33652711e+00
-8.58676553e-01 9.69713986e-01 -3.24428439e-01 -1.00542068e+00
1.00011086e+00 5.54821312e-01 6.53558314e-01 -1.06990361e+00
1.49767488e-01 4.56550241e-01 9.49721187e-02 -8.32740188e-01
3.67360383e-01 -1.23617813e-01 -2.00741142e-01 6.90039515e-01
-5.27851954e-02 1.32570416e-01 2.56186575e-01 1.55246798e-02
1.38205183e+00 4.19704556e-01 -1.10550925e-01 -2.33707987e-02
3.52482527e-01 5.72312027e-02 6.80424690e-01 5.64818382e-01
9.67831463e-02 3.58685642e-01 8.76316190e-01 -1.09210563e+00
-1.53091931e+00 -8.51137698e-01 8.31473172e-02 9.00570333e-01
-4.09182906e-01 -1.16717622e-01 -4.20787305e-01 1.52631953e-01
1.89828992e-01 6.53664589e-01 -1.09844363e+00 -2.58658022e-01
-1.05989540e+00 -7.89282024e-01 9.83968556e-01 8.66154671e-01
4.32714224e-01 -1.29175019e+00 -5.24294436e-01 7.45281994e-01
3.67190875e-02 -5.78475952e-01 -5.89445353e-01 3.25708300e-01
-1.09481227e+00 -7.89097786e-01 -1.10563540e+00 -6.36124969e-01
4.35335219e-01 -4.26480293e-01 9.32940841e-01 -1.53848529e-01
-1.42477274e-01 -2.95098983e-02 -3.34473103e-02 -1.77404881e-01
-6.29289687e-01 1.05052494e-01 1.96653351e-01 -2.53211677e-01
-3.22115928e-01 -8.52875173e-01 -6.41385376e-01 2.59168565e-01
-6.20631218e-01 2.77544796e-01 1.49825618e-01 1.02071357e+00
8.94299895e-03 -4.75054353e-01 7.21065760e-01 -4.01919931e-01
1.08181536e+00 -9.06423628e-01 -4.01436239e-01 -9.31351855e-02
-8.07188690e-01 3.90293330e-01 1.19481719e+00 -7.04713047e-01
-8.25891733e-01 -2.13737577e-01 -4.35250968e-01 -8.05785596e-01
7.34604120e-01 1.07672060e+00 9.80597377e-01 2.79444724e-01
5.68100452e-01 4.70227033e-01 1.10291965e-01 -4.62569445e-01
2.14675307e-01 4.76051688e-01 9.36265171e-01 -3.41217190e-01
2.12341174e-01 2.87528366e-01 -7.72451013e-02 -3.23690236e-01
-5.47129452e-01 -2.22519711e-01 -5.96084356e-01 -2.62275338e-01
6.63569391e-01 -1.15923834e+00 -1.00953507e+00 1.09122121e+00
-1.29640543e+00 -1.15251923e+00 -3.02726626e-01 3.91290098e-01
-6.95085168e-01 -2.72150964e-01 -1.51710653e+00 -8.92249942e-01
-5.19790828e-01 -6.92613721e-01 5.34463823e-01 3.02725900e-02
-3.69597226e-01 -1.67816460e+00 6.97989821e-01 -6.34515584e-01
7.07150102e-01 9.11762476e-01 7.17744529e-01 -4.45061713e-01
-1.10659048e-01 -2.57058531e-01 -6.13440759e-02 3.51300895e-01
-1.19496129e-01 2.98689753e-01 -5.53130209e-01 -2.15036809e-01
1.91154227e-01 -1.12613134e-01 9.98810351e-01 7.97757149e-01
8.77357900e-01 -8.57410312e-01 -1.88787535e-01 8.98140788e-01
1.41619217e+00 6.13934457e-01 5.57945251e-01 1.62747011e-01
8.64984453e-01 1.43107980e-01 -1.90862808e-02 5.81104875e-01
9.00295079e-01 5.64290360e-02 4.58742350e-01 -2.12211743e-01
1.34771064e-01 -3.32310081e-01 9.24743354e-01 1.58883023e+00
-3.41050774e-01 6.32320791e-02 -1.34427881e+00 7.85034060e-01
-2.20627284e+00 -1.02827477e+00 -6.84653699e-01 1.91073787e+00
8.53098810e-01 8.40077773e-02 3.95010620e-01 1.05182000e-01
5.72441816e-01 -2.51110154e-03 -1.15447962e+00 -1.15464664e+00
6.74606785e-02 -6.98382631e-02 5.57963550e-01 4.75816995e-01
-7.51897395e-01 6.51387751e-01 6.44704866e+00 1.13629617e-01
-1.53040326e+00 2.41330788e-01 7.87857294e-01 -2.00175345e-02
5.26420511e-02 1.11417793e-01 -3.34805816e-01 9.39399898e-01
1.92569256e+00 -3.65673900e-01 7.15889990e-01 6.06858492e-01
5.08563936e-01 3.17558765e-01 -1.19120145e+00 8.13055158e-01
-4.06883121e-01 -1.48637724e+00 -3.25152129e-01 -4.28885408e-02
1.15438402e+00 7.07866073e-01 9.14056152e-02 7.93808639e-01
1.13867569e+00 -1.01258695e+00 8.62329900e-01 9.72836375e-01
7.80986130e-01 -4.38806593e-01 7.38415778e-01 4.54982340e-01
-1.37510073e+00 -7.29991734e-01 -2.14903757e-01 -7.35193610e-01
3.49178880e-01 5.75680435e-01 -3.08607191e-01 5.13317227e-01
8.55656505e-01 1.32953942e+00 -1.46170914e-01 9.84035373e-01
1.53760344e-01 6.66879952e-01 -4.68854219e-01 2.94554681e-02
6.62527978e-01 -1.97295010e-01 3.25461775e-01 1.28868377e+00
8.12638581e-01 -1.84613526e-01 -2.76020408e-01 9.79883730e-01
8.27376395e-02 -7.76505232e-01 -7.55385458e-01 -1.87032968e-01
2.10180074e-01 6.92240834e-01 -1.98097095e-01 -2.70349950e-01
-7.17680529e-02 8.94888580e-01 1.67561963e-01 6.96267664e-01
-1.10576999e+00 -3.66601348e-01 6.06365263e-01 -2.49350816e-02
1.97390646e-01 -6.13335133e-01 -2.31078818e-01 -1.10541046e+00
-2.16690674e-01 -4.62971836e-01 6.58720076e-01 -1.27794266e+00
-1.11092818e+00 8.38299334e-01 -2.36778572e-01 -1.29306185e+00
-1.10237122e+00 -7.25026429e-01 -9.13487434e-01 9.30988789e-01
-1.19560492e+00 -9.23107505e-01 1.03200480e-01 6.02759123e-01
4.95442420e-01 9.53156278e-02 5.89534640e-01 3.15353572e-01
-8.67022038e-01 4.77563553e-02 9.93833125e-01 3.59976679e-01
2.50019073e-01 -1.21956992e+00 1.22248840e+00 6.47881031e-01
-5.72934449e-01 7.10772932e-01 7.32392073e-01 -7.02719271e-01
-1.47739947e+00 -1.27206481e+00 8.08023214e-01 -5.35284996e-01
1.09228456e+00 -4.12976265e-01 -1.09780836e+00 1.16397977e+00
2.67711461e-01 2.18795519e-02 -9.74038988e-02 -2.70417243e-01
-2.57502317e-01 7.77306631e-02 -7.21275806e-01 3.96579385e-01
9.05491471e-01 -6.91740692e-01 -5.32350123e-01 5.68839349e-03
9.23271298e-01 -8.60328674e-01 -1.16941595e+00 4.44244117e-01
6.92251623e-01 -9.67060745e-01 7.70973384e-01 -5.92958450e-01
8.97148669e-01 -1.26352198e-02 3.90498161e-01 -1.50985777e+00
-3.48223865e-01 -1.09183300e+00 -8.80840540e-01 8.49196970e-01
3.05959523e-01 -1.06317437e+00 2.10025668e-01 4.65029508e-01
-4.80481416e-01 -1.08706868e+00 -1.22925794e+00 -1.03658092e+00
6.30850792e-01 -3.35292965e-01 6.03177071e-01 1.02712226e+00
-1.09603301e-01 -9.19549540e-02 -7.23647833e-01 -2.47829139e-01
-7.50919878e-02 -1.09071441e-01 4.66440260e-01 -1.12455988e+00
-9.50630978e-02 -4.74511802e-01 -1.24516664e-02 -8.95202875e-01
1.49056643e-01 -7.85481095e-01 -1.94279879e-01 -1.86159968e+00
-2.82136738e-01 -1.62562996e-01 -4.89828348e-01 6.12253606e-01
3.27354550e-01 -1.93676263e-01 7.38940202e-03 5.98854601e-01
1.09412922e-02 6.16993964e-01 9.76546049e-01 -6.66914582e-02
-4.59373653e-01 -2.81607676e-02 -2.81118751e-01 4.80885953e-01
5.69948077e-01 -5.52026093e-01 -1.94947362e-01 -5.63131213e-01
6.91215754e-01 6.86722159e-01 6.63498580e-01 -1.15453303e+00
5.90443671e-01 2.13768870e-01 3.49028111e-01 -6.61512315e-01
1.79080233e-01 -4.65286374e-01 5.95896900e-01 1.15911448e+00
-3.48404199e-01 1.07383430e+00 4.52674896e-01 3.53612095e-01
-1.12796113e-01 2.74178892e-01 3.68988574e-01 -2.29357347e-01
-4.50611264e-01 4.41713899e-01 -6.89503312e-01 3.61475423e-02
6.84047163e-01 -4.02057737e-01 -3.94699812e-01 -6.92299843e-01
-9.21661913e-01 7.05287814e-01 -8.00636858e-02 5.59246778e-01
4.42739755e-01 -1.47434449e+00 -4.73138183e-01 1.57553077e-01
-5.41451573e-01 -5.54436147e-02 3.41451257e-01 1.26138830e+00
-7.02334642e-01 7.25277960e-01 -2.08735168e-01 -3.87594312e-01
-2.71532565e-01 7.01279223e-01 9.09850776e-01 -5.91981590e-01
-6.31896377e-01 6.07187927e-01 -1.07533233e-02 -8.68255734e-01
6.04197308e-02 -9.80804563e-01 2.57442534e-01 2.89754607e-02
1.46236882e-01 6.55324876e-01 -1.12917572e-01 -3.78494948e-01
4.85522151e-02 6.17323279e-01 1.73242688e-01 -8.96950364e-02
1.63789809e+00 -7.06753209e-02 -3.50443602e-01 1.13036323e+00
1.30305266e+00 -1.09031844e+00 -1.63827467e+00 -5.40690590e-03
5.15224710e-02 5.83409369e-01 -1.80859715e-01 -8.08017015e-01
-9.55885649e-01 9.52623308e-01 3.62118185e-01 3.99321467e-01
9.99986887e-01 -5.67606807e-01 1.23602068e+00 3.43263358e-01
1.41338587e-01 -9.56037283e-01 -3.87330651e-02 1.31496978e+00
1.08693349e+00 -6.59891188e-01 -5.81525147e-01 7.76685417e-01
-5.43239176e-01 1.33587015e+00 5.64101934e-01 -4.72591847e-01
9.57862377e-01 6.39140189e-01 7.35843778e-02 -8.21603090e-02
-1.47612095e+00 7.12618306e-02 2.47695073e-02 2.05488759e-04
4.76532936e-01 -3.14946592e-01 3.97884473e-02 5.01274109e-01
-3.32686663e-01 5.48795164e-01 6.98693633e-01 7.82433748e-01
2.52608776e-01 -5.48430264e-01 -2.08494574e-01 6.57809436e-01
-3.09753954e-01 -1.43671617e-01 -2.39840522e-01 8.61693323e-01
-2.21937507e-01 4.55216825e-01 4.31448013e-01 -6.08015180e-01
1.49471387e-01 2.32220769e-01 -6.08390346e-02 1.83467902e-02
-1.12729061e+00 -2.26933569e-01 -2.50725925e-01 -4.06104982e-01
5.13024367e-02 -8.19841862e-01 -1.72091997e+00 -7.46842980e-01
-4.36964594e-02 4.10883613e-02 6.32771134e-01 8.67062986e-01
6.53675199e-01 9.25668895e-01 6.76416278e-01 -9.10541892e-01
-5.08471012e-01 -1.03756106e+00 -2.29436457e-01 -4.54990938e-02
1.02176595e+00 -2.48745292e-01 -4.50302124e-01 1.46017462e-01] | [7.303421974182129, 3.379690408706665] |
a2d08ddf-368f-4370-902f-f6365e5a100b | adversarial-and-safely-scaled-question | 2210.09467 | null | https://arxiv.org/abs/2210.09467v1 | https://arxiv.org/pdf/2210.09467v1.pdf | Adversarial and Safely Scaled Question Generation | Question generation has recently gained a lot of research interest, especially with the advent of large language models. In and of itself, question generation can be considered 'AI-hard', as there is a lack of unanimously agreed sense of what makes a question 'good' or 'bad'. In this paper, we tackle two fundamental problems in parallel: on one hand, we try to solve the scaling problem, where question-generation and answering applications have to be applied to a massive amount of text without ground truth labeling. The usual approach to solve this problem is to either downsample or summarize. However, there are critical risks of misinformation with these approaches. On the other hand, and related to the misinformation problem, we try to solve the 'safety' problem, as many public institutions rely on a much higher level of accuracy for the content they provide. We introduce an adversarial approach to tackle the question generation safety problem with scale. Specifically, we designed a question-answering system that specifically prunes out unanswerable questions that may be generated, and further increases the quality of the answers that are generated. We build a production-ready, easily-plugged pipeline that can be used on any given body of text, that is scalable and immune from generating any hate speech, profanity, or misinformation. Based on the results, we are able to generate more than six times the number of quality questions generated by the abstractive approach, with a perceived quality being 44% higher, according to a survey of 168 participants. | ['Zhihang Dong', 'Sreehari Sankar'] | 2022-10-17 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 2.9179075e-01 6.0719067e-01 5.4796749e-01 -4.7686324e-02
-1.1921796e+00 -9.3489861e-01 7.4762106e-01 5.5717921e-01
-4.4752771e-01 9.1502142e-01 2.6233265e-01 -5.8108324e-01
4.2222708e-02 -1.1454009e+00 -6.9601673e-01 -6.5200455e-02
5.2448398e-01 5.5315077e-01 2.9209462e-01 -7.0746905e-01
3.6448875e-01 2.6676374e-02 -1.5465040e+00 3.4249777e-01
1.3121516e+00 8.1647170e-01 -1.6621061e-01 7.8086162e-01
-4.0089434e-01 1.0753462e+00 -1.1798260e+00 -9.7963077e-01
-5.8218423e-02 -6.0379577e-01 -1.3756655e+00 -2.9600838e-01
4.0134373e-01 -4.0977737e-01 1.8498436e-01 1.1800297e+00
4.5879176e-01 -1.7816432e-01 4.3043306e-01 -1.3127730e+00
-7.1143365e-01 6.5745449e-01 6.3487902e-02 -2.7781927e-03
8.8280970e-01 3.2022110e-01 9.5791686e-01 -5.4387385e-01
6.9486445e-01 1.2062154e+00 3.6392379e-01 6.9824213e-01
-1.0296671e+00 -5.3927892e-01 -2.7207559e-01 2.2704189e-01
-1.0548720e+00 -1.5867181e-01 5.0689089e-01 -5.6446940e-01
4.5602041e-01 7.5231594e-01 3.1603393e-01 1.4140128e+00
8.1740946e-02 3.7723523e-01 1.1737775e+00 -5.3264582e-01
5.1436806e-01 5.5133206e-01 7.3007382e-02 3.9125314e-01
4.2182940e-01 -2.3428182e-01 -1.5753005e-01 -3.1478500e-01
2.8704830e-02 -4.3769202e-01 -2.7531922e-01 2.8298518e-01
-9.6722990e-01 1.1413454e+00 2.1125597e-01 6.2972695e-01
-1.9217353e-01 -1.8711978e-01 2.3708887e-01 3.5240439e-01
5.4686105e-01 1.0758899e+00 -2.7656007e-01 -1.8840200e-01
-8.7147534e-01 7.2721624e-01 1.2940620e+00 5.9017307e-01
7.3550904e-01 -3.7258086e-01 -3.5548526e-01 3.9273086e-01
9.6701391e-02 5.1066136e-01 4.9406356e-01 -8.9266497e-01
6.8100798e-01 7.8486878e-01 3.5208640e-01 -1.4940819e+00
-9.3454860e-02 -1.2528941e-01 -8.0546504e-01 1.6907997e-01
6.5691370e-01 -5.6693971e-01 -5.3285843e-01 1.7169794e+00
4.0055263e-01 -2.0850828e-01 2.3567833e-02 8.0472600e-01
7.7680969e-01 7.0353514e-01 -6.0841002e-02 9.7652853e-02
1.7801825e+00 -5.9630287e-01 -7.5776303e-01 -3.3435047e-01
4.7096732e-01 -7.6488113e-01 1.2623473e+00 3.3250290e-01
-8.5301000e-01 -2.9782531e-01 -9.0259367e-01 -9.9232391e-02
-6.0495132e-01 -3.1988358e-01 2.6015103e-01 1.0286549e+00
-9.7610956e-01 5.2830178e-01 5.4334216e-03 -3.0821604e-01
2.7584943e-01 -4.0421911e-02 -3.8835937e-01 -1.7915231e-01
-1.6238465e+00 1.0794827e+00 2.8379557e-01 -9.0628549e-02
-5.8109057e-01 -7.1840066e-01 -7.0302498e-01 4.9544118e-02
7.8687185e-01 -7.1872103e-01 1.2657803e+00 -7.5982279e-01
-1.3057697e+00 6.6065037e-01 -9.9377269e-03 -3.7156042e-01
6.5311426e-01 -3.4286034e-01 -2.6376334e-01 2.0957069e-01
3.0943292e-01 3.5917279e-01 8.2973474e-01 -1.1721056e+00
-2.7900663e-01 -2.6519248e-01 5.1472425e-01 -6.8862163e-02
-2.9405728e-01 2.0260672e-01 1.7956087e-01 -5.6175315e-01
-4.0332639e-01 -7.7255106e-01 -3.8456079e-01 -2.2558562e-01
-5.1959783e-01 -2.5514519e-01 6.0908574e-01 -1.1088674e+00
1.2341225e+00 -1.8178915e+00 -1.7360799e-01 4.7990516e-02
3.1798151e-01 7.2113115e-01 -8.6101212e-02 6.6777152e-01
2.2195764e-01 8.0144787e-01 -4.4322363e-01 -4.0184781e-02
8.8711344e-02 7.7844190e-04 -6.9152379e-01 -1.6104937e-01
5.2653283e-01 9.2880660e-01 -1.1206194e+00 -3.1278235e-01
-1.9450593e-01 2.6224536e-01 -7.5583643e-01 4.3777648e-01
-5.7687420e-01 3.3225739e-01 -4.6188903e-01 6.8844825e-02
5.4271638e-01 -9.6765213e-02 -8.8028900e-02 1.6838743e-01
2.0131069e-01 5.4314828e-01 -1.1690032e+00 1.3631094e+00
-4.3979755e-01 4.9285668e-01 2.3052707e-02 -5.3945482e-01
9.3031877e-01 4.2469558e-01 -1.6815868e-01 -4.7548670e-01
1.8640365e-01 2.7803132e-01 -2.4679267e-01 -7.2850740e-01
7.2847503e-01 -4.1541666e-01 -3.2489696e-01 7.8554374e-01
1.5115081e-02 -5.4805225e-01 1.7513543e-01 4.3500525e-01
1.4110465e+00 -1.7302375e-01 4.4317227e-02 1.9709485e-02
8.3598405e-01 1.8636170e-01 7.8738555e-02 8.1824458e-01
-3.8628362e-03 7.2157961e-01 8.7329024e-01 -1.8408228e-01
-8.9188647e-01 -8.0510229e-01 3.9079931e-01 7.4571067e-01
-2.2717577e-01 -4.4870850e-01 -1.3175517e+00 -8.8427550e-01
-3.0168894e-01 1.1354902e+00 -5.2492166e-01 -1.5696967e-01
-3.9857560e-01 -3.8943130e-01 9.1272211e-01 -2.2213702e-01
4.4372872e-01 -1.2104102e+00 -6.9675368e-01 2.8988716e-01
-7.3253196e-01 -1.1546736e+00 -2.1447130e-01 -4.2022368e-01
-2.6872742e-01 -1.1169190e+00 -5.8942449e-01 -1.5875924e-01
5.7811034e-01 -8.7429032e-02 1.1595210e+00 3.1143275e-01
-1.8736970e-01 2.8219038e-01 -6.9770873e-01 -5.4123348e-01
-1.0541084e+00 3.0827281e-01 -3.0253434e-01 1.7704348e-01
1.0538918e-01 -2.7210423e-01 -4.5269930e-01 6.9373295e-02
-1.6022898e+00 -2.2298816e-01 2.6529443e-01 3.9904091e-01
-1.8060915e-01 4.8426196e-02 9.2503119e-01 -1.1636844e+00
1.2819237e+00 -6.5253508e-01 -3.4886193e-01 2.9437384e-01
-5.0770259e-01 2.1692787e-01 9.6710205e-01 -2.0858148e-01
-8.5552496e-01 -4.8876035e-01 -5.8611113e-01 1.5138137e-01
-4.3611050e-01 4.6096975e-01 -3.0415946e-01 8.0722898e-02
1.0412201e+00 7.3747806e-02 1.4560400e-01 -3.6132184e-01
7.7628267e-01 8.0313224e-01 3.7676933e-01 -4.6679389e-01
1.2163508e+00 2.1088524e-01 -2.6704067e-01 -7.5765502e-01
-9.8794997e-01 -1.6310732e-01 -7.3391825e-02 -2.2290935e-01
9.1962320e-01 -3.8538319e-01 -7.0244801e-01 4.6699077e-01
-1.4640236e+00 -6.2253244e-02 -4.9854800e-01 -1.3028583e-01
-2.4483286e-01 6.8212265e-01 -4.1191229e-01 -8.9181960e-01
-7.0669287e-01 -8.9794666e-01 7.2356331e-01 1.7418963e-01
-5.9865826e-01 -7.7636433e-01 1.2701911e-01 8.8013101e-01
7.7282542e-01 4.4697735e-01 1.0139292e+00 -9.9033588e-01
-4.2141089e-01 -3.7474480e-01 -1.5686028e-01 6.3209969e-01
2.8626073e-02 -2.0973198e-01 -9.3706548e-01 -9.2418000e-02
3.5039356e-01 -4.6689072e-01 2.3415372e-01 -5.6094044e-01
6.9464338e-01 -1.0301963e+00 3.5532501e-01 -2.6280019e-01
1.2211148e+00 -2.4742924e-01 8.9531869e-01 2.2606006e-01
4.2196301e-01 1.1090120e+00 4.3815354e-01 2.8389388e-01
6.0349369e-01 5.2239597e-01 4.5187166e-01 2.6226133e-01
9.6029788e-02 -6.6674805e-01 2.9317611e-01 5.6458592e-01
3.0097929e-01 -2.9374561e-01 -8.8665658e-01 5.6046188e-01
-1.4110473e+00 -1.1135209e+00 -3.4919584e-01 2.1305778e+00
1.0219245e+00 2.7091602e-02 6.8019517e-02 2.7154979e-01
6.0538965e-01 4.2785298e-02 -6.5932207e-02 -6.9846404e-01
1.4174727e-01 3.9158294e-01 2.3151701e-02 6.5638292e-01
-6.5264362e-01 8.5345286e-01 5.8537345e+00 8.6189526e-01
-8.6280316e-01 1.3791324e-01 6.4724988e-01 2.9310977e-01
-8.7289196e-01 2.3172052e-01 -5.8734393e-01 7.5141120e-01
1.2418925e+00 -4.4419262e-01 4.0239093e-01 6.5691274e-01
-3.5754929e-03 -2.1331550e-01 -8.1662369e-01 6.2667972e-01
3.4944859e-01 -1.2389045e+00 3.1210223e-01 -8.7719135e-02
3.9039913e-01 -4.6218991e-01 -1.3591503e-01 3.7950239e-01
4.2284340e-01 -1.1989001e+00 8.0087572e-01 4.7525206e-01
5.5931598e-01 -7.0772278e-01 9.1580874e-01 9.6305281e-01
-4.9534616e-01 -5.0071310e-02 -2.5677750e-01 -2.4226697e-01
2.6426357e-01 9.1039366e-01 -1.0124347e+00 6.5363967e-01
4.3507144e-01 -3.9061061e-01 -8.8817340e-01 8.0461895e-01
-6.3676167e-01 6.4902622e-01 -2.8382450e-01 -3.5067645e-01
1.0992878e-01 -5.5419192e-02 3.8620853e-01 1.0175217e+00
4.2313197e-01 1.6294667e-01 -1.8327250e-01 1.1518508e+00
-2.2440487e-01 2.4082102e-01 -6.9872075e-01 -4.7563113e-02
3.5055450e-01 1.4017369e+00 -3.4786388e-01 -2.5968724e-01
1.7598979e-02 9.7290617e-01 2.2960560e-01 9.3720816e-02
-6.9593292e-01 -5.9544122e-01 2.4359602e-01 4.1716775e-01
-6.2426489e-02 9.4604023e-02 -1.6090333e-01 -1.2276595e+00
2.6484334e-01 -1.5232852e+00 2.2538263e-01 -7.9544103e-01
-1.2598377e+00 7.3641223e-01 -2.5543511e-01 -6.6960526e-01
-6.5897429e-01 -3.4041595e-01 -5.8013207e-01 1.0258512e+00
-1.4081646e+00 -8.4401459e-01 -2.1743803e-01 2.8788158e-01
1.7324321e-01 2.0703809e-01 9.3578452e-01 2.6908219e-01
-1.7033729e-01 5.2347404e-01 -4.6183446e-01 1.9751766e-01
5.2183658e-01 -1.1559221e+00 5.6152123e-01 9.6368790e-01
1.8065788e-02 5.7489187e-01 8.5275757e-01 -5.6366718e-01
-1.1236284e+00 -1.0998932e+00 1.6073794e+00 -8.7069219e-01
8.7569612e-01 -5.4658842e-01 -1.1328791e+00 1.7810878e-01
3.2369289e-01 -5.6765103e-01 7.5007653e-01 -2.4221404e-01
-5.2480328e-01 1.2412173e-01 -1.5087067e+00 6.3150555e-01
4.3096387e-01 -7.7734059e-01 -9.3099940e-01 4.3061465e-01
1.0076358e+00 -1.7847919e-01 -6.2292337e-01 8.2197152e-02
4.6205960e-02 -1.1141624e+00 5.7672673e-01 -5.8754176e-01
7.6345634e-01 -2.7275762e-01 9.6854158e-02 -1.2280360e+00
9.8613255e-02 -9.2475778e-01 2.8959563e-01 1.7255698e+00
5.6755644e-01 -7.1789020e-01 5.9618586e-01 9.9488670e-01
3.1817424e-01 -5.5130243e-01 -1.0703641e+00 -5.5899668e-01
4.3175241e-01 -4.0353891e-01 7.7533573e-01 8.7997109e-01
3.8850989e-02 7.3711950e-01 -3.0397215e-01 1.6561275e-02
2.7573347e-01 -1.5914908e-01 9.0293616e-01 -1.1112001e+00
-3.1229532e-01 -1.7016016e-01 -1.6939254e-01 -6.4675361e-01
-4.8064385e-02 -8.2290417e-01 1.4713542e-01 -1.7158122e+00
-9.1090217e-02 -1.8358645e-01 3.3053124e-01 2.7490368e-01
-5.5225039e-01 7.6605208e-02 5.2750307e-01 -1.4886506e-01
-4.3803847e-01 3.4579739e-01 1.0713819e+00 -4.7316283e-02
1.7574888e-01 -5.1898502e-02 -1.2633315e+00 5.2667570e-01
8.1212747e-01 -5.8614808e-01 -2.8313872e-01 -3.2098430e-01
8.0360985e-01 4.9956437e-02 6.1175066e-01 -1.0759575e+00
7.4084021e-02 3.0620152e-02 -3.0508775e-01 -2.6042143e-01
1.3445675e-02 -6.2738639e-01 3.6423471e-02 4.9201304e-01
-4.1226494e-01 -2.4781961e-02 3.2155540e-02 3.4812978e-01
-1.5211448e-01 -8.0841869e-01 5.6203884e-01 -3.0294564e-01
-8.6849958e-02 -2.0092724e-02 -6.1552578e-01 5.6976575e-01
9.2929977e-01 2.6492372e-01 -6.1175895e-01 -8.3912075e-01
-2.2553253e-01 1.7289233e-01 2.5660092e-01 4.8595116e-01
5.1122421e-01 -9.3951261e-01 -1.0577286e+00 -8.0150805e-02
7.0039190e-02 -3.1598084e-02 2.9074398e-01 2.2868112e-01
-5.9002358e-01 3.2535201e-01 1.3874421e-01 2.5307403e-03
-7.7678621e-01 6.6446143e-01 1.1153611e-01 -6.1982393e-01
-2.2380833e-01 6.7002690e-01 -2.8547338e-01 -6.9920927e-01
-1.6012377e-01 -1.6270742e-01 -4.3124604e-01 3.0044687e-01
8.5764474e-01 4.0465245e-01 1.7758901e-01 -5.1438224e-01
-8.7640725e-02 2.8936725e-02 3.2504354e-02 -4.2263457e-01
1.0400525e+00 4.9001530e-02 -3.8645819e-01 1.1962310e-01
9.6612567e-01 2.6287362e-01 -5.2597255e-01 6.6590540e-02
1.8429759e-01 -4.0441951e-01 -4.5900154e-01 -1.0068294e+00
-5.3638148e-01 8.5813874e-01 1.7842972e-01 8.7838477e-01
8.3931077e-01 -6.1041683e-02 1.1613945e+00 2.6586935e-01
4.2160782e-01 -8.4677625e-01 1.1713205e-01 5.6523031e-01
1.1503795e+00 -1.0805451e+00 -3.6334708e-01 -4.2482972e-01
-6.0688889e-01 8.3408821e-01 3.6882961e-01 1.4543717e-01
1.7194577e-01 -2.7058689e-02 2.2723833e-01 1.1831555e-02
-5.9812248e-01 -3.1164700e-02 1.4030123e-01 6.6333377e-01
1.3140027e-01 2.0002509e-03 -5.4320836e-01 8.6219162e-01
-7.2222185e-01 9.9974900e-02 9.4693798e-01 6.3382560e-01
-4.9682635e-01 -1.3291160e+00 -6.4900225e-01 3.5542765e-01
-8.1245565e-01 -2.4652867e-01 -7.3894930e-01 2.8630584e-01
1.7717524e-01 1.6380358e+00 -4.6750876e-01 -4.9655813e-01
4.4434255e-01 2.1144992e-01 -5.1076300e-02 -8.1716287e-01
-1.0091599e+00 -7.9227024e-01 4.2723581e-01 -3.7591237e-01
3.7553135e-02 -3.3828741e-01 -8.6764652e-01 -5.3420883e-01
-2.5396219e-01 5.8058590e-01 7.0121425e-01 1.1656923e+00
5.4969674e-01 1.5570872e-01 6.2952584e-01 -1.8658192e-01
-1.0570232e+00 -8.2405323e-01 -1.6407287e-01 7.8685713e-01
1.6145763e-01 8.8717952e-02 -5.2437526e-01 -2.0913240e-01] | [11.5302095413208, 8.312457084655762] |
c02da36d-881d-4922-a55f-26c72fd31aa0 | dynamical-hyperspectral-unmixing-with | 2303.10566 | null | https://arxiv.org/abs/2303.10566v1 | https://arxiv.org/pdf/2303.10566v1.pdf | Dynamical Hyperspectral Unmixing with Variational Recurrent Neural Networks | Multitemporal hyperspectral unmixing (MTHU) is a fundamental tool in the analysis of hyperspectral image sequences. It reveals the dynamical evolution of the materials (endmembers) and of their proportions (abundances) in a given scene. However, adequately accounting for the spatial and temporal variability of the endmembers in MTHU is challenging, and has not been fully addressed so far in unsupervised frameworks. In this work, we propose an unsupervised MTHU algorithm based on variational recurrent neural networks. First, a stochastic model is proposed to represent both the dynamical evolution of the endmembers and their abundances, as well as the mixing process. Moreover, a new model based on a low-dimensional parametrization is used to represent spatial and temporal endmember variability, significantly reducing the amount of variables to be estimated. We propose to formulate MTHU as a Bayesian inference problem. However, the solution to this problem does not have an analytical solution due to the nonlinearity and non-Gaussianity of the model. Thus, we propose a solution based on deep variational inference, in which the posterior distribution of the estimated abundances and endmembers is represented by using a combination of recurrent neural networks and a physically motivated model. The parameters of the model are learned using stochastic backpropagation. Experimental results show that the proposed method outperforms state of the art MTHU algorithms. | ['Pau Closas', 'Tales Imbiriba', 'Ricardo Augusto Borsoi'] | 2023-03-19 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 4.65400577e-01 -7.10622489e-01 3.56707215e-01 3.00271176e-02
-1.38166696e-01 -3.87301922e-01 6.20428860e-01 -7.57014379e-02
-1.13414332e-01 8.21558297e-01 -1.82773676e-02 1.53445723e-02
-3.55732590e-01 -7.76636362e-01 -5.96517086e-01 -1.41235399e+00
3.35936427e-01 2.35063523e-01 -1.60182908e-01 -8.99785459e-02
-1.89693883e-01 3.81072670e-01 -1.74455047e+00 -1.88504592e-01
1.25021374e+00 9.79381204e-01 5.38461447e-01 3.61152887e-01
-2.42818147e-01 7.33515024e-01 -2.70384610e-01 2.31904387e-01
4.36847061e-02 -8.68294418e-01 -2.47125864e-01 3.83146167e-01
1.35232657e-01 -1.42552435e-01 -4.30635840e-01 1.49346840e+00
1.04253404e-01 7.57181942e-01 1.05802691e+00 -7.36255109e-01
-4.15892273e-01 6.22544408e-01 -5.82102716e-01 -1.60494536e-01
-4.95471954e-01 -2.41948906e-02 7.91105747e-01 -5.91546893e-01
3.51099581e-01 1.01340473e+00 4.98725593e-01 1.13909096e-01
-1.53150868e+00 -4.17948067e-01 2.12706149e-01 1.86742783e-01
-1.65379012e+00 -3.13212365e-01 1.03253067e+00 -8.41965199e-01
3.75165462e-01 2.65427470e-01 7.71447837e-01 8.31707776e-01
-2.00441126e-02 6.36023581e-01 1.13173258e+00 -4.19650823e-01
4.36723083e-01 -5.01142852e-02 3.81611496e-01 3.95010412e-01
1.64313883e-01 1.04317814e-01 -2.90168613e-01 -9.06604305e-02
6.17613375e-01 3.30334604e-01 -5.12869418e-01 -4.09435093e-01
-8.12682629e-01 8.09693635e-01 5.24865270e-01 4.07859296e-01
-8.59998405e-01 9.52076316e-02 -1.03672929e-01 -2.64851004e-01
8.70445371e-01 1.00933172e-01 -8.19511339e-02 3.77538890e-01
-1.31340170e+00 8.19553435e-02 4.65030372e-01 3.40034574e-01
1.11941719e+00 3.03489715e-01 -8.78039524e-02 9.98619914e-01
6.13599777e-01 9.60459948e-01 4.70378846e-01 -8.51432443e-01
-1.35576949e-01 4.34707880e-01 4.40313280e-01 -1.02456486e+00
-2.61406809e-01 -7.19043970e-01 -1.27376723e+00 2.22342655e-01
3.18335623e-01 -2.34560534e-01 -9.54341114e-01 1.83090067e+00
2.08687857e-01 4.86343622e-01 8.90501514e-02 8.60423446e-01
5.27063787e-01 1.21592283e+00 -4.90026698e-02 -7.54118860e-01
9.10035491e-01 -6.99350476e-01 -1.03110111e+00 -2.26453930e-01
-2.44184695e-02 -4.48675036e-01 3.85768622e-01 1.94878638e-01
-7.74357140e-01 -4.31867301e-01 -1.02779162e+00 2.91118264e-01
-2.15607852e-01 2.84495264e-01 2.31941894e-01 3.71413857e-01
-6.98290169e-01 8.69968295e-01 -1.05120754e+00 -1.88233316e-01
3.74360159e-02 -5.50269783e-02 1.94545284e-01 1.09531529e-01
-1.12918544e+00 6.34524465e-01 6.50787354e-01 7.94338465e-01
-8.35579097e-01 -4.43550795e-01 -7.51216233e-01 1.49040908e-01
3.89688671e-01 -6.76631272e-01 9.68312919e-01 -9.31841791e-01
-1.75196719e+00 1.74070761e-01 -6.34327114e-01 -2.47676864e-01
4.29834127e-01 3.10664400e-02 -2.91569889e-01 1.53966904e-01
-2.65643388e-01 1.62767917e-01 1.18228567e+00 -1.53736198e+00
-3.57362837e-01 -3.65928501e-01 -3.51632029e-01 1.70597658e-01
-4.54591811e-01 -4.94207919e-01 -1.06307454e-01 -6.31855249e-01
5.31581223e-01 -9.10102487e-01 -2.41459385e-01 -1.58736452e-01
-2.77230352e-01 2.57452518e-01 7.30177283e-01 -8.71004343e-01
1.13112366e+00 -2.23447418e+00 6.03376567e-01 2.87374139e-01
2.15727359e-01 2.61385620e-01 1.13929793e-01 2.82457083e-01
7.97137711e-03 -2.19322816e-01 -9.93635595e-01 -4.82008874e-01
4.74419957e-03 2.88860947e-01 -4.70912606e-01 5.72551191e-01
1.49029484e-02 5.83091795e-01 -9.95229065e-01 3.37135531e-02
5.43385684e-01 8.84771109e-01 -1.08215675e-01 2.18728334e-01
-5.95529139e-01 7.01144695e-01 -2.49805331e-01 2.91529268e-01
9.65104461e-01 -3.50061536e-01 4.01986361e-01 -2.64495522e-01
-4.76556122e-01 -1.83988616e-01 -1.30978072e+00 1.38717103e+00
-4.13077950e-01 4.48051691e-01 2.95682669e-01 -1.10196698e+00
8.83194923e-01 2.72955090e-01 5.54396033e-01 -9.49146673e-02
2.04458997e-01 5.10965109e-01 -2.56177559e-02 -3.63734424e-01
3.81128103e-01 -6.19858861e-01 6.17051363e-01 3.71640116e-01
-1.36279717e-01 -9.46472660e-02 3.77891213e-01 -3.06378007e-01
2.86447316e-01 2.92010576e-01 5.71396947e-02 -3.32893938e-01
6.51997209e-01 -1.46537215e-01 5.33794940e-01 7.49973893e-01
1.19113475e-01 4.11738575e-01 1.00848086e-01 -6.51055649e-02
-8.15133750e-01 -8.43283713e-01 -3.94979477e-01 4.80577797e-01
9.67826322e-02 1.60627261e-01 -7.28320777e-01 1.67559698e-01
-6.69158250e-02 8.24893594e-01 -5.77412784e-01 -4.90327515e-02
-9.81727168e-02 -1.52960515e+00 1.46138100e-02 3.76023091e-02
5.35089076e-01 -8.47237229e-01 -5.88729024e-01 4.92719829e-01
-5.04098296e-01 -9.15910125e-01 -8.92275125e-02 1.84677511e-01
-1.03956223e+00 -8.00560892e-01 -9.27695274e-01 -2.38106534e-01
4.87011254e-01 2.86727965e-01 5.51187158e-01 -6.37962580e-01
-2.34880093e-02 1.07030556e-01 -1.44536436e-01 -3.14374268e-01
-5.38355350e-01 -2.25168854e-01 2.03287631e-01 8.07295561e-01
1.11542739e-01 -9.23551798e-01 -4.61786389e-01 1.30491093e-01
-1.25024879e+00 2.08155528e-01 4.83965307e-01 7.58665562e-01
7.68339157e-01 5.43709099e-01 7.16540739e-02 -6.65677488e-01
2.59051174e-01 -5.86344957e-01 -9.68005955e-01 3.03752810e-01
-4.48653430e-01 1.40918553e-01 4.46502000e-01 -4.70509112e-01
-1.45673418e+00 4.91150729e-02 2.00071454e-01 -7.10911572e-01
-1.43760443e-01 1.03033078e+00 -4.38701510e-02 1.12352155e-01
3.65823090e-01 7.02829659e-01 1.83881968e-01 -7.11914122e-01
3.00555885e-01 6.62353277e-01 5.72027266e-01 -4.72356886e-01
7.72509694e-01 7.95026243e-01 1.55318856e-01 -1.48796666e+00
-8.41565192e-01 -6.02942824e-01 -5.94686508e-01 -3.74324918e-01
7.42197514e-01 -8.95346284e-01 -5.84838390e-01 1.00417447e+00
-1.17077959e+00 -4.04321790e-01 -2.76891530e-01 6.86672449e-01
-3.61405045e-01 6.17847621e-01 -4.62800145e-01 -1.50945795e+00
-2.27088824e-01 -1.08919775e+00 5.60308874e-01 2.27198720e-01
1.89049885e-01 -1.04116559e+00 1.69867352e-01 1.24609239e-01
3.93967867e-01 5.30876100e-01 9.82176185e-01 1.63239554e-01
-5.47630429e-01 -1.95086282e-02 -2.14657053e-01 5.22724390e-01
3.80280346e-01 3.01381826e-01 -1.08570170e+00 -1.45854309e-01
6.40395284e-01 2.10601225e-01 1.20868075e+00 1.10522890e+00
8.84153366e-01 -1.22608811e-01 -1.38880834e-01 7.67433465e-01
1.50915742e+00 2.92166799e-01 3.16826552e-01 -8.60287026e-02
8.31878364e-01 9.10327792e-01 -3.32133956e-02 5.79236925e-01
-7.50645474e-02 2.99501866e-01 5.93157113e-01 1.95904881e-01
1.77461505e-01 -4.33521122e-02 3.54339868e-01 1.01776457e+00
-2.49068096e-01 -3.39924872e-01 -9.11344469e-01 3.87085587e-01
-2.28410721e+00 -1.10404503e+00 -4.86418098e-01 2.22981048e+00
4.53404754e-01 -3.88123721e-01 -1.58230513e-01 8.09065551e-02
9.95113373e-01 4.89372164e-01 -8.78928483e-01 3.10925841e-01
-5.69747269e-01 -8.35288167e-02 2.99500644e-01 6.54670060e-01
-1.07869792e+00 5.34937620e-01 5.68197060e+00 5.91012895e-01
-1.31749260e+00 1.52314112e-01 3.40847909e-01 4.46863472e-02
-3.77908438e-01 1.44118620e-02 -4.67301339e-01 4.92635757e-01
7.32799292e-01 -7.23093078e-02 7.89719462e-01 1.97093338e-01
5.81124663e-01 -1.35129705e-01 -4.94458884e-01 1.04376149e+00
8.21148828e-02 -1.17392218e+00 1.20171905e-01 1.31626993e-01
1.03855360e+00 1.68217525e-01 9.08364132e-02 -1.04581207e-01
8.79133940e-02 -6.40217006e-01 8.87091041e-01 1.18480015e+00
3.83560807e-01 -5.75959027e-01 5.63095033e-01 5.88319242e-01
-9.86261964e-01 -2.06250593e-01 -5.49411416e-01 -2.38757968e-01
2.22437814e-01 1.09600651e+00 -6.57083914e-02 7.37849414e-01
3.67912084e-01 1.07648647e+00 -7.68509880e-02 9.17374969e-01
-3.59287858e-01 7.20375478e-01 -5.57605088e-01 1.59696452e-02
2.24909991e-01 -1.13349307e+00 8.15013528e-01 8.71364832e-01
6.78655982e-01 8.14779922e-02 8.04162621e-02 1.44237554e+00
-4.09244150e-02 1.48507863e-01 -3.11610460e-01 -4.17710483e-01
1.25035271e-01 1.18388259e+00 -6.23799920e-01 -2.39129901e-01
-4.29341104e-03 7.81789839e-01 -4.74154390e-02 9.19509172e-01
-4.71937180e-01 7.73575157e-02 6.62432432e-01 -9.93149132e-02
3.43268961e-01 -2.97162384e-01 -1.38886170e-02 -1.47088397e+00
5.05222715e-02 -5.67455351e-01 1.50895432e-01 -8.73739719e-01
-1.23155224e+00 5.08895755e-01 5.74713945e-02 -1.16181695e+00
-2.52499372e-01 -6.68668866e-01 -6.57386959e-01 1.25623071e+00
-1.73648345e+00 -9.42208886e-01 -3.83116663e-01 3.68763894e-01
1.20156549e-01 7.97622129e-02 8.16764534e-01 -4.36034985e-02
-1.16976178e+00 -3.77077878e-01 1.08725727e+00 -3.46128881e-01
2.68454522e-01 -1.13101208e+00 -2.65703142e-01 1.22008145e+00
6.96797948e-03 6.48247719e-01 9.68468726e-01 -5.85776210e-01
-1.18816483e+00 -1.20707464e+00 4.99490142e-01 6.12535179e-02
9.96896923e-01 -1.04004323e-01 -1.15623784e+00 2.73509651e-01
-2.50672977e-02 3.84026356e-02 7.79551923e-01 -1.19542703e-01
3.02474685e-02 -2.09306970e-01 -6.43548369e-01 4.03854102e-01
4.95505482e-01 -7.08910942e-01 -2.30727851e-01 3.93365592e-01
3.31870288e-01 -1.36321671e-02 -6.67622924e-01 4.30583358e-01
5.88631988e-01 -8.35767508e-01 7.34455347e-01 -2.55044222e-01
5.25188506e-01 -6.36978984e-01 -3.13520193e-01 -1.38286448e+00
-4.74038929e-01 -4.60646689e-01 -5.96334577e-01 1.04943585e+00
1.72950104e-01 -6.75445378e-01 4.03147906e-01 3.42270076e-01
-1.33168483e-02 -2.64137238e-01 -7.06070781e-01 -7.26711631e-01
-2.44349483e-02 -4.28143203e-01 5.21963775e-01 8.96942198e-01
-4.51336116e-01 1.90488353e-01 -6.03772879e-01 6.39307618e-01
1.00947344e+00 4.70325351e-01 1.83350474e-01 -1.50512540e+00
-4.49213833e-01 -7.61215508e-01 1.34037599e-01 -8.90786588e-01
4.26305205e-01 -7.90864944e-01 3.21799338e-01 -1.46042490e+00
2.25247815e-01 -5.23263477e-02 -5.69294572e-01 5.03419340e-02
-2.99093574e-01 1.61245301e-01 -3.07585541e-02 6.67934239e-01
1.67107522e-01 1.10897005e+00 8.57084513e-01 -5.90714455e-01
-4.27552402e-01 5.53013794e-02 -1.81433544e-01 7.88261056e-01
6.50959194e-01 -3.15212995e-01 -3.96465868e-01 -4.15008307e-01
1.33463502e-01 1.94125384e-01 4.67599213e-01 -1.01123047e+00
1.20603792e-01 -2.76078463e-01 2.15199530e-01 -7.44687259e-01
5.63553035e-01 -7.12161779e-01 5.61137915e-01 4.29592878e-01
9.29234102e-02 -6.39274716e-01 2.37818703e-01 7.64910340e-01
-3.70363563e-01 -5.38154125e-01 8.40179026e-01 -3.16904671e-02
-4.93169218e-01 5.35767019e-01 -6.52081609e-01 -5.47562599e-01
4.98788625e-01 4.44993973e-02 2.97311321e-02 -3.06095779e-01
-8.47873449e-01 6.33393154e-02 3.02148402e-01 -1.71503350e-01
4.56506133e-01 -1.08452022e+00 -8.00555646e-01 9.99085531e-02
2.18114965e-02 -2.29645148e-02 7.79990435e-01 7.92377830e-01
-4.82174933e-01 1.79081619e-01 5.20973913e-02 -6.61141336e-01
-6.76964700e-01 4.65172112e-01 7.61896610e-01 -1.70158327e-01
-2.98493654e-01 5.71307838e-01 3.16078693e-01 -4.00562495e-01
-1.86313644e-01 -1.50155127e-01 -3.69859397e-01 5.40341616e-01
4.65180010e-01 4.95598078e-01 -2.06702054e-01 -9.60012913e-01
2.91154310e-02 4.67718452e-01 4.16877657e-01 -1.29335716e-01
1.52506375e+00 -3.69144917e-01 -7.95715809e-01 1.10728788e+00
1.00975847e+00 -4.14004534e-01 -1.47535217e+00 -5.29545248e-01
-2.95001775e-01 -1.00301117e-01 5.97044230e-01 -3.86691302e-01
-9.50688541e-01 1.11724842e+00 5.84875226e-01 2.64167130e-01
1.12726164e+00 -6.39736235e-01 5.48624635e-01 4.80505645e-01
-1.06021529e-02 -9.51372087e-01 -3.69352788e-01 6.27620041e-01
7.12307453e-01 -1.11007619e+00 -1.03600338e-01 -3.10729921e-01
2.26800740e-02 1.21017396e+00 1.04193568e-01 1.78230867e-01
8.91553879e-01 -3.17722499e-01 1.89789295e-01 -4.35447395e-02
-8.42965394e-02 -3.52736473e-01 3.94651860e-01 1.30262077e-01
2.55312294e-01 3.50036353e-01 -1.13203652e-01 2.11670145e-01
1.03058547e-01 -1.46637812e-01 3.91679913e-01 4.75617677e-01
-4.14398283e-01 -7.89401054e-01 -6.51227057e-01 3.02503139e-01
-4.15364504e-02 -1.29040584e-01 3.94867249e-02 7.57046491e-02
1.99169412e-01 1.08991683e+00 4.65334486e-03 -6.47572950e-02
1.33721558e-02 4.21360165e-01 2.50869423e-01 -4.27844495e-01
1.84958547e-01 5.41722536e-01 -5.46008587e-01 7.68700615e-02
-7.93327272e-01 -8.17815661e-01 -7.60101795e-01 -9.77391228e-02
-5.20742059e-01 2.33829513e-01 8.19610953e-01 1.09387934e+00
-1.62227705e-01 7.77119398e-01 7.22663641e-01 -1.18339789e+00
-6.34987056e-01 -1.19328749e+00 -1.14744174e+00 2.05798998e-01
4.85787034e-01 -6.81435049e-01 -5.82542300e-01 -2.18683500e-02] | [10.062758445739746, -2.06974720954895] |
c69718db-ff37-4d67-881a-d502fac41c0d | adaptive-generation-of-privileged | 2307.03240 | null | https://arxiv.org/abs/2307.03240v1 | https://arxiv.org/pdf/2307.03240v1.pdf | Adaptive Generation of Privileged Intermediate Information for Visible-Infrared Person Re-Identification | Visible-infrared person re-identification seeks to retrieve images of the same individual captured over a distributed network of RGB and IR sensors. Several V-I ReID approaches directly integrate both V and I modalities to discriminate persons within a shared representation space. However, given the significant gap in data distributions between V and I modalities, cross-modal V-I ReID remains challenging. Some recent approaches improve generalization by leveraging intermediate spaces that can bridge V and I modalities, yet effective methods are required to select or generate data for such informative domains. In this paper, the Adaptive Generation of Privileged Intermediate Information training approach is introduced to adapt and generate a virtual domain that bridges discriminant information between the V and I modalities. The key motivation behind AGPI^2 is to enhance the training of a deep V-I ReID backbone by generating privileged images that provide additional information. These privileged images capture shared discriminative features that are not easily accessible within the original V or I modalities alone. Towards this goal, a non-linear generative module is trained with an adversarial objective, translating V images into intermediate spaces with a smaller domain shift w.r.t. the I domain. Meanwhile, the embedding module within AGPI^2 aims to produce similar features for both V and generated images, encouraging the extraction of features that are common to all modalities. In addition to these contributions, AGPI^2 employs adversarial objectives for adapting the intermediate images, which play a crucial role in creating a non-modality-specific space to address the large domain shifts between V and I domains. Experimental results conducted on challenging V-I ReID datasets indicate that AGPI^2 increases matching accuracy without extra computational resources during inference. | ['Eric Granger', 'Rafael M. O. Cruz', 'Pourya Shamsolmoali', 'Arthur Josi', 'Mahdi Alehdaghi'] | 2023-07-06 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [ 4.47878838e-01 -3.30795272e-04 -6.86693937e-02 -5.17221689e-01
-7.17512131e-01 -8.42740476e-01 6.15854323e-01 -3.72631878e-01
-4.05137360e-01 7.14822054e-01 1.77176312e-01 2.64697045e-01
-2.51212001e-01 -9.00044084e-01 -5.97426355e-01 -8.46009374e-01
3.19419593e-01 2.28972629e-01 -4.44196552e-01 -2.39834487e-01
-2.28083536e-01 5.77491045e-01 -1.84747958e+00 1.62816569e-01
9.42836106e-01 1.02156270e+00 -2.35099912e-01 3.44695807e-01
3.69812474e-02 1.79685876e-01 -6.61302924e-01 -7.16587722e-01
8.07295859e-01 -5.71437418e-01 -5.01876414e-01 -2.29758054e-01
8.68745863e-01 -2.70465732e-01 -7.50941932e-01 8.75391603e-01
8.14181447e-01 2.95093268e-01 7.59709954e-01 -1.60134327e+00
-1.05963004e+00 1.61104381e-01 -3.02412361e-01 -1.44186631e-01
8.74657750e-01 3.60637814e-01 4.77251202e-01 -5.35056770e-01
7.22009301e-01 1.26981878e+00 8.21728945e-01 1.08752823e+00
-1.37156892e+00 -9.88357186e-01 6.21164031e-02 8.89638364e-02
-1.67612910e+00 -3.90979707e-01 1.11658835e+00 -3.26288164e-01
3.79130363e-01 6.01141095e-01 6.20895207e-01 1.68839002e+00
-1.61149591e-01 7.23052323e-01 1.20695186e+00 -1.75546721e-01
4.74304631e-02 5.52155912e-01 -1.28564879e-01 2.04139888e-01
2.11899698e-01 4.46714461e-01 -8.41995418e-01 -1.35444522e-01
6.13662183e-01 3.00150782e-01 -3.26764852e-01 -4.85033453e-01
-9.19114530e-01 5.53646147e-01 7.66815901e-01 1.38597190e-01
-2.14251086e-01 -2.89736748e-01 4.92754653e-02 3.62728924e-01
5.73265851e-02 5.73978007e-01 -4.79913205e-02 3.11923802e-01
-7.10682333e-01 2.73264647e-01 4.26376522e-01 9.99621749e-01
9.48826611e-01 -2.58609485e-02 -5.01886725e-01 6.52207613e-01
3.96331787e-01 1.08365393e+00 2.12784380e-01 -6.87793493e-01
5.59646666e-01 8.94467175e-01 6.25866950e-02 -8.68684769e-01
-1.36766121e-01 -4.94846791e-01 -1.00170624e+00 3.64012986e-01
5.89468658e-01 -1.48305111e-02 -1.00920010e+00 2.15252042e+00
5.24298072e-01 -1.47666961e-01 3.69631052e-01 1.01897550e+00
1.12852466e+00 2.42992103e-01 3.47034305e-01 4.46965843e-01
1.25780296e+00 -1.80901691e-01 -2.56104916e-01 -3.91694278e-01
1.13271929e-01 -4.98828441e-01 8.84303451e-01 -2.42372125e-01
-8.28773141e-01 -7.95347869e-01 -1.03193426e+00 -9.81194451e-02
-6.77422285e-01 6.89125210e-02 4.97388601e-01 8.66441607e-01
-1.00346470e+00 2.29352430e-01 -1.19496636e-01 -3.78555983e-01
6.10515356e-01 6.21670246e-01 -8.61723781e-01 -3.26186806e-01
-1.54037499e+00 7.23812819e-01 2.19842941e-01 1.77797645e-01
-6.63683653e-01 -8.63193095e-01 -1.17948771e+00 -2.59604156e-01
-5.08764274e-02 -1.05243015e+00 2.40871340e-01 -1.18798733e+00
-1.30911887e+00 1.19158483e+00 1.80122592e-02 -5.99335395e-02
7.41763771e-01 8.56042728e-02 -5.39944172e-01 2.37342224e-01
2.45077878e-01 7.96823144e-01 1.16865301e+00 -1.51258361e+00
-2.54052818e-01 -8.90231669e-01 8.21498707e-02 3.77101839e-01
-4.51532066e-01 -3.04729789e-01 -3.40290248e-01 -5.47670841e-01
-6.54352754e-02 -1.01890934e+00 2.52415657e-01 9.00673792e-02
-5.04873753e-01 -1.08966045e-01 6.52470350e-01 -7.22015023e-01
4.28040355e-01 -2.12866354e+00 2.98513442e-01 6.54249370e-01
2.61034787e-01 2.65925586e-01 -3.79762143e-01 1.42354667e-01
-7.96233043e-02 -1.72677100e-01 4.31641415e-02 -4.01851803e-01
2.00094387e-01 8.78833309e-02 -2.62148470e-01 7.27439404e-01
3.03758010e-02 1.18025112e+00 -6.83931887e-01 -2.97050834e-01
3.50470901e-01 9.64678526e-01 -2.69008130e-01 3.73345822e-01
2.72358567e-01 9.61056352e-01 -4.13893253e-01 8.14818323e-01
1.05883491e+00 1.56701431e-01 -2.04596035e-02 -2.98010379e-01
2.35271052e-01 -1.91279605e-01 -1.01711929e+00 1.66269338e+00
-2.49230772e-01 2.96851903e-01 8.32456127e-02 -8.77287567e-01
1.21162832e+00 -2.38913931e-02 4.92015779e-01 -9.83766019e-01
2.47728959e-01 -6.87276624e-05 -5.31007826e-01 -2.37145677e-01
4.46039766e-01 -8.44185427e-02 -3.68151844e-01 9.21598226e-02
-2.05312315e-02 3.00745249e-01 -4.04437780e-01 4.08592299e-02
6.75467491e-01 3.37983668e-01 -2.52299279e-01 1.34695128e-01
6.97091401e-01 -1.81038484e-01 5.80439687e-01 1.08008993e+00
-3.43029886e-01 7.90546536e-01 -1.53610587e-01 -3.55918467e-01
-9.50193882e-01 -1.52609611e+00 -1.72634169e-01 1.01867366e+00
7.86627173e-01 2.26846993e-01 -5.19381523e-01 -6.27594292e-01
4.31603462e-01 3.80657017e-01 -9.79893565e-01 -5.63514054e-01
-4.45897013e-01 -3.61129165e-01 8.82784486e-01 4.91433769e-01
7.89234996e-01 -6.02842152e-01 -2.66888976e-01 -3.22206438e-01
-2.58523494e-01 -9.72892642e-01 -5.43110728e-01 -2.98530042e-01
-3.48551303e-01 -1.08195364e+00 -1.02683783e+00 -6.37453079e-01
8.25681090e-01 3.21521431e-01 8.39971721e-01 -1.98441580e-01
-4.75934148e-01 9.67039764e-01 -2.53324538e-01 -2.54769981e-01
-1.81201130e-01 3.43090072e-02 2.02557951e-01 3.62850994e-01
6.95661724e-01 -4.18528199e-01 -8.18971276e-01 4.63772684e-01
-8.79462004e-01 -1.84858397e-01 5.36281645e-01 9.44044232e-01
4.10116494e-01 -4.60559696e-01 4.27022994e-01 -3.96140903e-01
2.07668036e-01 -4.98835653e-01 -4.22948599e-01 4.08078998e-01
-4.17553276e-01 1.26271307e-01 6.24490738e-01 -7.10100889e-01
-1.30052280e+00 1.22166939e-01 1.46600813e-01 -5.30418515e-01
-2.67841905e-01 -8.17922354e-02 -7.20090985e-01 -2.68016011e-01
8.15846443e-01 3.84369493e-01 2.35302240e-01 -2.52698958e-01
4.37335521e-01 6.88147485e-01 9.41616535e-01 -6.43588722e-01
1.41956770e+00 6.30564153e-01 3.29168588e-02 -5.87022126e-01
-4.50057805e-01 -4.01098460e-01 -6.00448489e-01 -1.59515962e-01
8.42912018e-01 -1.15151465e+00 -8.55268776e-01 6.39967084e-01
-7.48738587e-01 3.39351706e-02 -5.05242229e-01 4.63855684e-01
-1.92524806e-01 3.38906020e-01 -6.42849505e-02 -6.43901289e-01
-4.75646347e-01 -9.18423176e-01 1.23594046e+00 6.89631224e-01
-1.81550696e-01 -8.23803186e-01 7.30135990e-03 7.69143522e-01
3.94701332e-01 4.82966065e-01 2.90601522e-01 -5.51463068e-01
-5.90953648e-01 -5.39085269e-01 -2.85529494e-01 1.34019881e-01
1.77876487e-01 -6.30636513e-01 -1.31232202e+00 -6.44883573e-01
-4.00562555e-01 -3.29940259e-01 8.12852204e-01 -3.18551250e-02
9.92334068e-01 -2.26065561e-01 -4.50701624e-01 1.26376545e+00
1.26284266e+00 4.84375507e-02 8.79749954e-01 5.07309258e-01
8.59957159e-01 6.24250352e-01 4.15603638e-01 2.99922854e-01
5.18848300e-01 8.86590600e-01 2.84777641e-01 -2.59992063e-01
-2.82934934e-01 -5.41285872e-01 3.90396535e-01 -2.91725397e-01
-1.94787055e-01 -1.21630924e-02 -6.09642565e-01 5.15184045e-01
-1.41775453e+00 -1.13884330e+00 4.05519068e-01 2.54443789e+00
7.89476097e-01 -4.84069824e-01 1.26665816e-01 -1.82473391e-01
9.01850343e-01 -8.82546902e-02 -9.69378829e-01 1.40307501e-01
-5.52835107e-01 2.90266097e-01 5.50073683e-01 3.20684254e-01
-1.15253198e+00 6.27024055e-01 5.37266254e+00 4.75352347e-01
-1.08605182e+00 -7.96028897e-02 2.79677927e-01 -2.70887136e-01
-6.76612556e-01 -2.75813937e-01 -9.55909431e-01 5.99367440e-01
5.11762202e-01 -4.38095257e-02 4.94158715e-01 8.47865283e-01
-3.86799723e-01 2.72337824e-01 -1.27171612e+00 1.51038349e+00
2.89493024e-01 -1.01833618e+00 1.87344868e-02 2.21606046e-01
6.89061940e-01 -3.70990038e-01 6.08559310e-01 3.16883594e-01
5.54373078e-02 -1.24262512e+00 6.66735828e-01 7.49308705e-01
1.32136011e+00 -7.27432430e-01 8.01472485e-01 -9.63203087e-02
-1.21333921e+00 -5.50453700e-02 -3.53580326e-01 3.62178057e-01
-2.84945548e-01 9.73506942e-02 -5.22348046e-01 7.82980978e-01
7.97284305e-01 4.32984829e-01 -6.39755726e-01 5.51009297e-01
-2.29035974e-01 -1.99994296e-01 -3.43565196e-01 3.43221813e-01
-2.75424987e-01 -2.34392777e-01 6.81692839e-01 7.89551854e-01
2.37635046e-01 -1.02834679e-01 1.18962325e-01 1.16846371e+00
-2.15688184e-01 -3.88420731e-01 -9.58774865e-01 3.30572695e-01
6.75234079e-01 9.53997612e-01 1.03683665e-01 -1.37994364e-01
-2.89270729e-01 1.41353297e+00 7.04906657e-02 7.02079773e-01
-9.13896799e-01 -5.00649571e-01 1.19302464e+00 6.37090653e-02
1.28976470e-02 1.32832319e-01 -2.56147861e-01 -1.26260281e+00
1.81390896e-01 -7.48615563e-01 5.70870280e-01 -5.29349208e-01
-1.67194355e+00 6.39630675e-01 9.03434381e-02 -1.34232509e+00
-4.10641432e-01 -5.16607463e-01 -2.06182033e-01 1.50998628e+00
-1.58436549e+00 -1.89796627e+00 -8.53553116e-01 1.22435606e+00
-2.37506688e-01 -5.09788275e-01 8.67361605e-01 1.55082092e-01
-4.44747567e-01 1.48926067e+00 2.56193727e-01 3.99119288e-01
9.86473203e-01 -9.83124077e-01 1.20806046e-01 9.51810062e-01
-2.89549381e-01 9.81198847e-01 3.95625204e-01 -5.95166266e-01
-1.79460669e+00 -1.19045520e+00 3.59537691e-01 -6.93729341e-01
-6.22910680e-03 -3.96277934e-01 -7.20145762e-01 6.57169998e-01
-2.53935486e-01 2.67810196e-01 1.01130581e+00 -4.55368049e-02
-8.71905446e-01 -4.70965177e-01 -1.76932442e+00 3.58820558e-01
1.35468519e+00 -1.06232095e+00 -5.15153706e-01 -4.61150631e-02
1.48365051e-01 -4.00025219e-01 -1.13401508e+00 2.40321666e-01
7.61277676e-01 -7.95148969e-01 1.63832152e+00 -3.90224010e-01
9.68345404e-02 -3.14587474e-01 -1.76502571e-01 -1.12090540e+00
-7.96831548e-02 -5.16235769e-01 1.21947601e-02 1.48936892e+00
-1.68518856e-01 -1.16937315e+00 9.42711055e-01 1.23178077e+00
3.12905252e-01 -1.10637531e-01 -1.09374583e+00 -8.86696696e-01
8.48612636e-02 -5.57052605e-02 9.58991349e-01 9.01013374e-01
-3.23368400e-01 -1.61903292e-01 -4.04393852e-01 3.18335712e-01
1.19995141e+00 3.53732914e-01 1.20252466e+00 -1.31540740e+00
-1.36394933e-01 -1.03462614e-01 -7.17005730e-01 -8.49518478e-01
4.01094198e-01 -1.09995794e+00 -2.10525289e-01 -1.11664832e+00
2.73421705e-01 -8.04814875e-01 -3.55271280e-01 6.66041195e-01
-3.41545701e-01 6.44320309e-01 3.77509981e-01 3.69728804e-01
-2.37605646e-01 7.82536030e-01 1.12813556e+00 -6.18954360e-01
-3.61748576e-01 -1.58354566e-01 -1.14148223e+00 2.41614580e-01
7.04085290e-01 3.49202231e-02 -5.08744121e-01 -3.79116684e-01
-1.93788692e-01 -3.48574221e-01 8.62192690e-01 -9.30858493e-01
1.26728386e-01 -1.23038180e-01 1.13059640e+00 -2.61149108e-01
6.95408404e-01 -8.23686957e-01 4.78927970e-01 1.05819859e-01
-1.83342010e-01 -5.27218401e-01 1.67395294e-01 4.70466465e-01
-4.14823629e-02 2.94808686e-01 8.91726434e-01 -3.01880892e-02
-9.26952422e-01 5.46251118e-01 3.37910593e-01 -9.53836814e-02
1.08625400e+00 -6.66427433e-01 -5.04891932e-01 -2.83107966e-01
-6.04442298e-01 1.08304843e-01 9.82131600e-01 6.74428940e-01
7.33227968e-01 -1.53990030e+00 -6.47538662e-01 6.84516490e-01
5.18675923e-01 -1.74019694e-01 6.06682539e-01 3.31144392e-01
-2.89515685e-02 1.19443625e-01 -5.12258232e-01 -6.70849800e-01
-1.45056462e+00 5.60588896e-01 4.82981861e-01 6.33796602e-02
-4.87979472e-01 8.19869101e-01 3.10359120e-01 -7.44735360e-01
7.24051222e-02 4.81130451e-01 -1.29620552e-01 5.94003685e-02
7.50780165e-01 2.71973789e-01 -3.50841194e-01 -1.15763474e+00
-6.91119313e-01 7.74517834e-01 7.52603635e-02 6.17108978e-02
1.16850305e+00 -3.43126625e-01 4.76819426e-02 -2.64038235e-01
1.32885063e+00 2.43603569e-02 -1.36931705e+00 -3.27199131e-01
-6.53970778e-01 -9.31328833e-01 -3.43649924e-01 -8.54123473e-01
-9.61436391e-01 4.86422092e-01 1.10677791e+00 -2.61984974e-01
1.26831937e+00 1.32561475e-01 7.01801896e-01 -7.32476451e-03
5.48982859e-01 -9.34363604e-01 -1.41152777e-02 4.65425886e-02
8.11193287e-01 -1.40953028e+00 -4.94195558e-02 -3.42469752e-01
-5.77073514e-01 8.00809503e-01 7.60658562e-01 2.49704495e-01
-1.83377173e-02 -2.11329177e-01 1.68517739e-01 3.71389533e-03
3.31872702e-01 -3.04962695e-01 5.51961482e-01 1.33567345e+00
-9.40340236e-02 1.18899107e-01 1.78227991e-01 5.27913392e-01
-2.71130651e-01 -1.22478865e-01 3.73488665e-02 7.69663393e-01
2.83200175e-01 -1.32918632e+00 -9.68684912e-01 1.48294836e-01
-6.20022006e-02 2.81995296e-01 -4.95453566e-01 6.78374410e-01
3.80548358e-01 9.31483924e-01 -6.74396306e-02 -6.34129226e-01
2.49919623e-01 -4.79372516e-02 6.85600996e-01 -8.70627686e-02
-5.13298810e-01 -6.45744026e-01 -3.01086754e-01 -5.72716117e-01
-4.74064201e-01 -5.85225523e-01 -7.87859201e-01 -5.56092858e-01
1.24543294e-01 -3.87937785e-03 6.13695145e-01 6.44013166e-01
5.42808831e-01 9.95960087e-02 8.04309905e-01 -1.03315568e+00
-5.47660112e-01 -5.59070528e-01 -5.56020498e-01 1.18579805e+00
3.91076028e-01 -7.80413389e-01 -3.19981635e-01 -2.22855404e-01] | [14.706624984741211, 0.9618216753005981] |
50733e0a-5ebd-4dab-a2b9-224e845de4f3 | explaining-predictive-uncertainty-with | 2306.05724 | null | https://arxiv.org/abs/2306.05724v1 | https://arxiv.org/pdf/2306.05724v1.pdf | Explaining Predictive Uncertainty with Information Theoretic Shapley Values | Researchers in explainable artificial intelligence have developed numerous methods for helping users understand the predictions of complex supervised learning models. By contrast, explaining the $\textit{uncertainty}$ of model outputs has received relatively little attention. We adapt the popular Shapley value framework to explain various types of predictive uncertainty, quantifying each feature's contribution to the conditional entropy of individual model outputs. We consider games with modified characteristic functions and find deep connections between the resulting Shapley values and fundamental quantities from information theory and conditional independence testing. We outline inference procedures for finite sample error rate control with provable guarantees, and implement an efficient algorithm that performs well in a range of experiments on real and simulated data. Our method has applications to covariate shift detection, active learning, feature selection, and active feature-value acquisition. | ['Ido Guy', 'Richard Mudd', 'Niek Tax', "Joshua O'Hara", 'David S. Watson'] | 2023-06-09 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 4.33114856e-01 8.69837999e-01 -7.47802913e-01 -7.45112062e-01
-6.73770785e-01 -5.20655096e-01 1.86220080e-01 4.42042947e-01
-4.03509200e-01 1.29466176e+00 -1.58738017e-01 -3.66733134e-01
-8.04605246e-01 -7.37995207e-01 -5.94189465e-01 -7.07407951e-01
-2.85427868e-01 7.22690463e-01 -2.94938684e-01 1.81185052e-01
4.37259138e-01 3.40751559e-02 -1.31676257e+00 -1.79936424e-01
1.07994199e+00 1.04999530e+00 -2.11865723e-01 4.88898128e-01
1.68376490e-01 8.59472811e-01 -3.96724164e-01 -8.46780896e-01
2.03701302e-01 -6.51353836e-01 -7.15136230e-01 -1.61877230e-01
-2.21068531e-01 -1.65601298e-01 -7.17373937e-02 1.16955340e+00
2.60607600e-01 2.04786882e-01 9.77423787e-01 -1.74868810e+00
-7.31793106e-01 1.35797179e+00 -4.37390476e-01 1.26487583e-01
-4.58204523e-02 1.75118908e-01 1.58820069e+00 -2.79834241e-01
2.97629654e-01 1.07513297e+00 4.36179549e-01 6.67657316e-01
-1.70824325e+00 -8.54315162e-01 8.00276995e-02 2.08999515e-01
-1.00143230e+00 -2.47294620e-01 5.31892955e-01 -4.76970226e-01
7.13366866e-01 4.29673612e-01 7.12142110e-01 9.11128938e-01
4.23794299e-01 1.04551184e+00 8.15194666e-01 -3.51807982e-01
7.72792280e-01 4.11282092e-01 2.71464676e-01 6.34356737e-01
8.31201375e-01 5.81405640e-01 -7.30824411e-01 -5.83349943e-01
6.51996017e-01 2.82060131e-02 -2.10715964e-01 -7.98337162e-01
-8.89551580e-01 1.25876749e+00 2.53089249e-01 -3.45374912e-01
-1.74550712e-01 4.75187123e-01 -1.09420352e-01 3.97133976e-01
5.30851305e-01 7.94989169e-01 -7.02663422e-01 4.24673147e-02
-4.76097971e-01 3.34142327e-01 1.02462101e+00 8.91309440e-01
6.40444994e-01 -1.87993839e-01 -1.40501127e-01 1.57671884e-01
5.41903555e-01 3.72436851e-01 -1.66768190e-02 -1.20820105e+00
4.18752193e-01 4.89478409e-01 4.10152078e-01 -5.98741591e-01
-4.65286881e-01 -4.88571733e-01 -7.47311056e-01 3.21175605e-01
2.38826796e-01 -5.06307960e-01 -4.76447314e-01 1.90525770e+00
-1.03332996e-01 -1.22654483e-01 3.33727002e-02 5.65157473e-01
2.09606767e-01 3.37834239e-01 2.66005337e-01 -6.40169203e-01
6.68570876e-01 -3.30104768e-01 -5.49638569e-01 -3.87529731e-01
5.11103034e-01 8.01322982e-02 6.77410126e-01 2.52262056e-01
-1.09382939e+00 2.43734270e-01 -1.03508854e+00 2.28314325e-01
4.76106210e-03 -6.71910703e-01 1.19525945e+00 9.82762516e-01
-7.61762202e-01 8.28558147e-01 -7.94785142e-01 3.59033257e-01
1.14056873e+00 7.50597835e-01 -1.34294823e-01 3.24835867e-01
-1.17553675e+00 6.26796365e-01 5.24606347e-01 -3.97393107e-01
-7.83690095e-01 -9.10154641e-01 -8.49844813e-01 6.93228006e-01
6.50649428e-01 -9.46041048e-01 1.25570905e+00 -1.06153882e+00
-1.34716499e+00 5.64517379e-01 3.24426979e-01 -8.41598392e-01
4.66352940e-01 1.70268074e-01 1.45914689e-01 -2.95728862e-01
-1.80115983e-01 5.89383423e-01 5.10561585e-01 -7.89655983e-01
-4.78267401e-01 -4.27681386e-01 2.92433035e-02 2.75268853e-01
-2.68259883e-01 -3.62333357e-01 4.43516880e-01 -3.90964270e-01
-6.54465184e-02 -7.44190574e-01 -7.15060115e-01 2.86335498e-01
-4.30737883e-01 -7.35414624e-02 -9.50866640e-02 1.02914989e-01
8.93207610e-01 -1.90443504e+00 2.01744646e-01 5.31080186e-01
4.60241318e-01 -4.70036954e-01 5.03069498e-02 7.31318071e-02
1.23223923e-01 4.26636487e-01 -4.98635769e-01 -1.62904337e-01
4.78528142e-01 1.35187404e-02 -2.85155267e-01 4.06664461e-01
5.15387803e-02 1.10901034e+00 -6.73353493e-01 -2.06151411e-01
2.99780983e-02 7.74675980e-02 -8.83858800e-01 -2.61796005e-02
-3.96660149e-01 1.68954939e-01 -5.35942972e-01 2.62000978e-01
3.89045179e-01 -6.64179623e-01 2.31207728e-01 7.71899879e-01
4.71211791e-01 7.72169977e-02 -9.97252643e-01 1.05478346e+00
7.16510415e-02 5.69128036e-01 -2.19894439e-01 -8.43510985e-01
5.22153676e-01 9.34833288e-02 3.81437451e-01 -2.48876557e-01
4.07127231e-01 -1.56166106e-01 1.46531507e-01 -6.01761490e-02
2.37543896e-01 -3.74967009e-01 -3.62161160e-01 7.34059215e-01
1.06550157e-01 -2.36147866e-01 -2.25301966e-01 3.85170698e-01
1.06785834e+00 -3.30108315e-01 9.01486218e-01 -5.49452484e-01
-1.27958432e-01 -1.51913702e-01 6.41614556e-01 1.32229674e+00
-2.98720688e-01 3.40166658e-01 1.01699388e+00 -4.86866832e-02
-8.26458573e-01 -1.28393602e+00 -2.77815253e-01 1.00161815e+00
2.15905473e-01 -2.71302313e-01 -5.02981305e-01 -6.46940410e-01
4.98557627e-01 1.17628312e+00 -1.10414422e+00 -4.34093505e-01
5.57380319e-01 -8.72265160e-01 2.19112799e-01 5.35113990e-01
1.55555829e-01 -9.67436254e-01 -6.64297223e-01 -6.00734316e-02
2.28632763e-01 -2.64214724e-01 -2.33069181e-01 7.60426998e-01
-9.47026432e-01 -1.04882336e+00 -2.54502892e-01 -2.00745746e-01
5.43656051e-01 -3.42849314e-01 1.12195635e+00 -1.05165303e-01
-3.19440842e-01 4.93213147e-01 5.75002246e-02 -9.67703938e-01
-2.54373163e-01 -6.28173202e-02 3.28402631e-02 -3.70571703e-01
4.76763517e-01 -3.60750556e-01 -4.97766614e-01 1.98643491e-01
-6.55712783e-01 1.12435736e-01 4.65796053e-01 9.42243874e-01
4.98409003e-01 -9.53819305e-02 5.20366311e-01 -1.27212787e+00
5.87607324e-01 -7.52933025e-01 -1.17023885e+00 3.92549217e-01
-1.08180630e+00 5.19483030e-01 1.80792183e-01 -2.91369557e-01
-9.55614865e-01 9.68208686e-02 4.30536747e-01 9.89180058e-02
9.21767130e-02 5.87355196e-01 -3.79429847e-01 2.00746432e-01
7.67534137e-01 -1.30275279e-01 1.34911940e-01 -8.89673680e-02
4.15059298e-01 3.56736094e-01 9.81371403e-02 -3.52259606e-01
6.38875961e-01 1.30059570e-01 3.55020702e-01 -2.49078497e-01
-1.02251971e+00 2.09344760e-01 -2.80382842e-01 1.14530146e-01
5.30198872e-01 -6.73152387e-01 -1.39537644e+00 -5.05307456e-03
-5.31624973e-01 -4.18603182e-01 -7.64963627e-01 7.16576517e-01
-1.10504818e+00 -1.56407028e-01 -1.97543412e-01 -1.21876037e+00
-1.66249558e-01 -9.33266282e-01 4.25669223e-01 4.15498018e-01
-1.81920871e-01 -1.12883031e+00 2.11179957e-01 2.79973656e-01
4.23421204e-01 -9.27572772e-02 1.23259044e+00 -8.59142780e-01
-8.39087903e-01 -2.21132934e-01 1.21975064e-01 -1.17589794e-01
-1.52202651e-01 -2.00261012e-01 -8.72575462e-01 -2.10217685e-02
-3.49134132e-02 -5.20992756e-01 1.11970258e+00 9.43194628e-01
1.43255961e+00 -6.62524283e-01 -3.81792545e-01 3.76329064e-01
1.13535500e+00 2.34297663e-01 5.51814318e-01 -5.90300709e-02
-3.16700935e-02 5.28226256e-01 4.92872447e-01 9.22381401e-01
1.08041558e-02 2.00493172e-01 5.42743683e-01 3.48556429e-01
8.36725295e-01 -2.48757020e-01 1.87002793e-01 -1.18382379e-01
1.37480676e-01 -4.67674881e-01 -7.57765472e-01 5.89678884e-02
-2.04804683e+00 -1.29808950e+00 4.61481452e-01 2.44581151e+00
9.87606466e-01 4.70491767e-01 1.14700682e-01 -4.34049629e-02
7.03278422e-01 -3.49161267e-01 -1.16812813e+00 -2.67698973e-01
-2.03592271e-01 1.70169398e-01 8.29390645e-01 7.13841498e-01
-8.69669676e-01 4.56611961e-01 7.51345348e+00 6.96586728e-01
-3.91412199e-01 -1.11983068e-01 1.27021623e+00 -5.26340842e-01
-8.94079268e-01 1.47729322e-01 -5.38472235e-01 2.12928280e-01
8.31776500e-01 -9.22473431e-01 5.00721931e-01 1.20317626e+00
-1.37541413e-01 -1.93475589e-01 -1.48752224e+00 8.10796201e-01
-2.83938050e-01 -1.42626488e+00 -2.74810076e-01 1.70448780e-01
9.11565542e-01 -2.05017462e-01 4.52385277e-01 1.31987914e-01
9.54086483e-01 -1.29560006e+00 6.52400970e-01 6.01258278e-01
6.36919737e-01 -1.00916064e+00 4.83136863e-01 5.21666825e-01
-4.71071988e-01 -5.22290051e-01 -6.06862247e-01 -4.12836641e-01
-2.53605008e-01 5.67742884e-01 -7.92185426e-01 -3.62610035e-02
3.19647312e-01 5.49789131e-01 -4.58154619e-01 1.14259970e+00
-1.70130074e-01 7.39041746e-01 -3.08999836e-01 -4.20903921e-01
-2.94923782e-01 -1.22025162e-01 3.50920677e-01 3.62921625e-01
6.36919662e-02 2.56820679e-01 -4.02086616e-01 1.34549248e+00
-2.69428521e-01 2.38883682e-02 -5.95015645e-01 -2.68645674e-01
5.84465742e-01 7.47530341e-01 -7.51785278e-01 -4.17197049e-02
-1.89020470e-01 4.97937411e-01 9.76554900e-02 1.82781368e-01
-8.00462604e-01 -2.49396801e-01 8.15560102e-01 -2.58584291e-01
6.25122041e-02 3.36501420e-01 -8.55223596e-01 -1.17563128e+00
-3.67626220e-01 -4.23149347e-01 7.76596427e-01 -6.14546716e-01
-1.53507602e+00 -1.46155611e-01 2.35751987e-01 -7.22899377e-01
-4.51029718e-01 -4.76798624e-01 -5.09078801e-01 7.63004363e-01
-7.91858554e-01 -3.29577327e-01 1.83233738e-01 3.95453304e-01
9.33454260e-02 -3.60051602e-01 9.53614891e-01 -4.92178082e-01
-5.46651125e-01 8.95223856e-01 5.73118329e-01 -1.32466167e-01
1.96547940e-01 -1.44540441e+00 4.95833248e-01 4.99908030e-01
2.57415801e-01 4.61844981e-01 9.20311570e-01 -6.05899632e-01
-1.25792396e+00 -5.91590345e-01 5.08687913e-01 -6.93714559e-01
5.71753621e-01 -2.64264017e-01 -3.87035400e-01 9.49160933e-01
5.64451329e-03 -2.36832798e-01 1.07190287e+00 4.65481311e-01
-4.46541250e-01 2.78088376e-02 -1.53631425e+00 5.21429300e-01
1.10743594e+00 -1.80709869e-01 -3.79398525e-01 9.38326865e-03
7.75825799e-01 1.01615824e-02 -6.22378409e-01 2.41257668e-01
6.37634277e-01 -9.59155083e-01 6.81266427e-01 -1.11193621e+00
5.40454209e-01 3.44231844e-01 -2.05829933e-01 -1.27302682e+00
-6.41857505e-01 -8.36521864e-01 -9.00515541e-02 7.12248862e-01
1.00921190e+00 -8.54995370e-01 1.05369902e+00 1.47949803e+00
3.72841358e-01 -6.31595850e-01 -9.75903451e-01 -4.25370216e-01
4.29521978e-01 -4.14432585e-01 6.92210555e-01 7.95413673e-01
5.67631483e-01 2.56172419e-01 -2.83780485e-01 -2.17354983e-01
1.24342906e+00 1.02556832e-01 3.19473773e-01 -1.78586769e+00
-5.60834050e-01 -5.55903435e-01 -6.21988833e-01 -5.65277815e-01
3.01755726e-01 -1.01008582e+00 -5.10213859e-02 -9.85832095e-01
1.00716710e+00 -4.48757678e-01 -5.73433757e-01 7.69097269e-01
-2.75922924e-01 -6.16313815e-02 1.71171591e-01 -1.52381688e-01
-7.41145432e-01 6.12670898e-01 8.09684217e-01 -2.61160791e-01
-1.90666839e-01 4.17836666e-01 -1.38128698e+00 6.53332233e-01
9.13300335e-01 -7.67177999e-01 -7.31098175e-01 9.63640511e-02
8.69527221e-01 3.20672184e-01 3.86597961e-01 -6.08007669e-01
1.57560050e-01 -8.74277651e-01 6.43090487e-01 -2.77464926e-01
9.24481750e-02 -8.66947532e-01 2.93273777e-01 6.73303843e-01
-1.37237132e+00 -4.08286959e-01 -6.30135387e-02 8.58904362e-01
4.38358486e-01 -3.05909872e-01 9.52117085e-01 -3.50418314e-02
-1.80012703e-01 3.72328967e-01 -4.76472884e-01 3.31452638e-01
1.17311513e+00 7.48050734e-02 -5.28497458e-01 -6.79681361e-01
-7.77196765e-01 3.79257262e-01 4.11262453e-01 1.21961437e-01
7.56262541e-01 -1.09458268e+00 -6.98751986e-01 2.50852406e-01
1.69292152e-01 -3.67909908e-01 1.90829828e-01 4.91389960e-01
3.15997452e-02 4.18042392e-01 -1.71555519e-01 -2.35746816e-01
-9.02927935e-01 4.47033554e-01 3.45432103e-01 -4.48609553e-02
-1.52133912e-01 9.63880002e-01 3.79807174e-01 -1.38905898e-01
3.35807145e-01 4.10439074e-03 3.03780101e-02 -4.57366370e-02
5.81297755e-01 4.13946688e-01 -5.48479974e-01 2.81207338e-02
-2.56567955e-01 -1.51633024e-01 -1.28721938e-01 -4.31717962e-01
1.45448482e+00 -2.20631078e-01 1.55999601e-01 5.91094613e-01
8.22424471e-01 -4.02942449e-01 -1.45239365e+00 -1.72808468e-01
1.33874863e-01 -5.23823500e-01 4.86340933e-02 -9.97926712e-01
-9.90791857e-01 6.74862146e-01 4.98937577e-01 2.11772993e-01
8.61742675e-01 2.48601124e-01 -1.74478158e-01 8.25131059e-01
6.20266676e-01 -7.76185870e-01 -9.83109176e-02 1.36689484e-01
4.90069002e-01 -1.40288198e+00 3.13759357e-01 -1.19446404e-01
-8.40326130e-01 7.60290146e-01 3.41384500e-01 1.48922607e-01
9.61657584e-01 3.88942122e-01 -3.57772499e-01 6.62836134e-02
-1.28599143e+00 3.34168561e-02 1.55344546e-01 5.51279306e-01
1.35247171e-01 3.93802702e-01 -6.01011664e-02 1.25395966e+00
-1.50151417e-01 6.32750094e-02 6.73864782e-01 5.81526935e-01
-7.31767714e-01 -8.46748710e-01 6.60758317e-02 1.03621113e+00
-5.47767043e-01 -2.33011231e-01 -5.94845295e-01 4.99598920e-01
-4.24881637e-01 8.89728129e-01 1.30082414e-01 -3.16129118e-01
-3.90411615e-01 1.37194693e-01 5.68571985e-01 -5.76695204e-01
-3.40747654e-01 -3.70932639e-01 -3.36693048e-01 -4.59968060e-01
-1.22246437e-01 -7.87495971e-01 -1.13754296e+00 -5.01272559e-01
-5.76917470e-01 4.46844399e-01 3.45599264e-01 8.64119530e-01
2.91586965e-01 1.50029706e-02 7.29063451e-01 -3.42809200e-01
-1.04385495e+00 -5.95019460e-01 -1.02387607e+00 1.74469158e-01
2.05524132e-01 -6.12016261e-01 -7.84079790e-01 -4.00184244e-01] | [8.571634292602539, 5.390944957733154] |
129d19ec-3d04-4d02-a542-71df5e6dbe83 | cms-rcnn-contextual-multi-scale-region-based | 1606.05413 | null | http://arxiv.org/abs/1606.05413v1 | http://arxiv.org/pdf/1606.05413v1.pdf | CMS-RCNN: Contextual Multi-Scale Region-based CNN for Unconstrained Face Detection | Robust face detection in the wild is one of the ultimate components to
support various facial related problems, i.e. unconstrained face recognition,
facial periocular recognition, facial landmarking and pose estimation, facial
expression recognition, 3D facial model construction, etc. Although the face
detection problem has been intensely studied for decades with various
commercial applications, it still meets problems in some real-world scenarios
due to numerous challenges, e.g. heavy facial occlusions, extremely low
resolutions, strong illumination, exceptionally pose variations, image or video
compression artifacts, etc. In this paper, we present a face detection approach
named Contextual Multi-Scale Region-based Convolution Neural Network (CMS-RCNN)
to robustly solve the problems mentioned above. Similar to the region-based
CNNs, our proposed network consists of the region proposal component and the
region-of-interest (RoI) detection component. However, far apart of that
network, there are two main contributions in our proposed network that play a
significant role to achieve the state-of-the-art performance in face detection.
Firstly, the multi-scale information is grouped both in region proposal and RoI
detection to deal with tiny face regions. Secondly, our proposed network allows
explicit body contextual reasoning in the network inspired from the intuition
of human vision system. The proposed approach is benchmarked on two recent
challenging face detection databases, i.e. the WIDER FACE Dataset which
contains high degree of variability, as well as the Face Detection Dataset and
Benchmark (FDDB). The experimental results show that our proposed approach
trained on WIDER FACE Dataset outperforms strong baselines on WIDER FACE
Dataset by a large margin, and consistently achieves competitive results on
FDDB against the recent state-of-the-art face detection methods. | ['Khoa Luu', 'Yutong Zheng', 'Marios Savvides', 'Chenchen Zhu'] | 2016-06-17 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 6.39670715e-02 -2.77845144e-01 -8.69646892e-02 -4.79311496e-01
-3.84298563e-01 -2.49163702e-01 5.14510334e-01 -7.15849876e-01
-3.37103873e-01 4.20744210e-01 -1.61859378e-01 4.05524932e-02
1.31439492e-01 -5.30916572e-01 -4.07917678e-01 -9.54661489e-01
-1.68784652e-02 1.15796529e-01 4.41446677e-02 -3.49466830e-01
1.04873993e-01 1.24767482e+00 -1.85536277e+00 1.53546289e-01
1.59471497e-01 1.46904409e+00 -1.66417375e-01 1.30349591e-01
1.33513778e-01 4.13723797e-01 -3.29650879e-01 -3.36133897e-01
3.94845903e-01 -1.24509305e-01 -3.16286683e-01 3.31046164e-01
7.06264496e-01 -5.09005070e-01 -2.57056981e-01 1.17747295e+00
8.72976363e-01 5.47661111e-02 5.66692114e-01 -1.34608054e+00
-3.16604227e-01 -3.17862853e-02 -1.30427802e+00 1.51680946e-01
3.27389449e-01 3.66776139e-02 2.89099306e-01 -1.60539305e+00
6.79726183e-01 1.74264908e+00 6.38924181e-01 9.10848856e-01
-7.66165614e-01 -1.21006298e+00 7.28466958e-02 6.31977469e-02
-1.79304290e+00 -1.06557310e+00 8.40659261e-01 -1.74078062e-01
6.68073833e-01 1.64205447e-01 3.82846683e-01 1.06249619e+00
-8.09831917e-02 6.04801834e-01 9.81536150e-01 -4.53331679e-01
-1.62323058e-01 1.88584015e-01 -3.09469402e-01 1.04582548e+00
1.02136095e-04 3.31567049e-01 -2.56654561e-01 -1.14819758e-01
1.08501041e+00 -2.91794483e-02 -2.41008893e-01 -2.06485689e-01
-5.79342961e-01 6.96977973e-01 3.93032104e-01 2.46368200e-01
-2.39571139e-01 3.78800333e-02 4.98181969e-01 1.86449056e-03
5.28337240e-01 -4.71515119e-01 -3.45331877e-01 5.61460316e-01
-1.02607393e+00 2.24589244e-01 4.38522130e-01 8.19352567e-01
5.32055557e-01 2.68378258e-01 -3.23065728e-01 1.15820444e+00
6.90778553e-01 4.55571800e-01 2.38703847e-01 -5.06734788e-01
3.86126190e-01 4.86515582e-01 -4.56961133e-02 -1.09640026e+00
-5.34069657e-01 -4.04245406e-01 -9.88130927e-01 5.20791590e-01
4.92109567e-01 -1.74308047e-01 -7.39934623e-01 1.90987802e+00
8.09578836e-01 2.96803951e-01 -2.36418471e-01 9.55538988e-01
1.34178269e+00 3.53255898e-01 1.61922760e-02 -4.39002126e-01
1.74667668e+00 -7.42201030e-01 -6.47334397e-01 -1.96812108e-01
1.77418947e-01 -9.85365868e-01 3.45403671e-01 2.25980073e-01
-8.91341329e-01 -8.19749534e-01 -7.46193409e-01 1.03227995e-01
-2.98554838e-01 7.40383863e-01 6.50470018e-01 9.56039369e-01
-1.17929363e+00 1.95137989e-02 -2.63683945e-01 -5.86077213e-01
9.32996750e-01 5.83031595e-01 -8.30285728e-01 -2.19122246e-01
-8.67147744e-01 6.61791444e-01 1.96338162e-01 6.66816533e-01
-1.14205825e+00 -3.26899648e-01 -8.37375045e-01 -7.52478912e-02
6.68695629e-01 -2.95263857e-01 8.36384058e-01 -1.01117337e+00
-1.22322261e+00 1.31465757e+00 -2.11415052e-01 -2.56443080e-02
5.59319079e-01 2.33419817e-02 -5.91440797e-01 9.17063579e-02
-2.32043415e-01 8.42688978e-01 1.24612498e+00 -9.73497272e-01
-4.46786821e-01 -8.63787711e-01 -2.81667203e-01 -7.77026489e-02
-3.86284292e-01 9.64869618e-01 -8.27321291e-01 -6.21954799e-01
-9.08279046e-03 -8.49682212e-01 8.88978913e-02 6.18376017e-01
-2.44395390e-01 -6.71669066e-01 1.28454590e+00 -5.54468870e-01
8.91551137e-01 -2.18204737e+00 -1.88460231e-01 7.23155141e-02
1.12393655e-01 5.61736524e-01 -2.87208170e-01 -9.87363085e-02
-4.13929641e-01 -1.91202626e-01 3.72247063e-02 -4.10364151e-01
-1.06846392e-01 -1.90459192e-01 -9.13961604e-02 9.24694240e-01
5.10600209e-01 7.94419110e-01 -4.52069640e-01 -8.38753283e-01
2.39303544e-01 9.08881724e-01 -4.15376306e-01 6.81043193e-02
1.56634286e-01 3.26013654e-01 -4.16189581e-01 1.35214746e+00
1.20118368e+00 2.89564878e-01 -4.56408039e-02 -4.78332996e-01
-2.43515857e-02 -6.39561713e-01 -1.49447131e+00 1.39101040e+00
-1.40010148e-01 5.29781461e-01 6.82243109e-01 -9.26531255e-01
9.76833642e-01 5.15115857e-01 3.70287836e-01 -4.41453576e-01
4.23411667e-01 1.96413204e-01 4.22676280e-02 -4.51475501e-01
1.42077982e-01 -1.42499179e-01 4.24901813e-01 1.81489006e-01
3.02083284e-01 4.19719905e-01 2.46194117e-02 -1.67006895e-01
5.36143005e-01 3.28958988e-01 5.41497469e-01 -2.35125005e-01
1.13882542e+00 -7.84911394e-01 8.00813079e-01 2.05579296e-01
-7.82118618e-01 4.85970199e-01 4.35382634e-01 -6.23641849e-01
-5.26915193e-01 -5.22119045e-01 -5.25074482e-01 1.25670886e+00
-6.82904124e-02 -1.63493399e-02 -9.50349629e-01 -8.24395537e-01
7.31437355e-02 -2.00148851e-01 -8.53881121e-01 1.64877117e-01
-6.20261669e-01 -9.05946791e-01 8.10028851e-01 4.97603178e-01
7.52258658e-01 -1.35630572e+00 -2.61529118e-01 -6.63599223e-02
2.72350341e-01 -1.34686017e+00 -4.99548405e-01 -2.61722267e-01
-4.67159420e-01 -1.14802444e+00 -7.77981341e-01 -1.00577319e+00
7.79852331e-01 4.53575164e-01 8.93689871e-01 3.21628213e-01
-9.33822393e-01 1.42415211e-01 -1.40444785e-02 -5.25937378e-01
-2.15689484e-02 -5.61393976e-01 3.00737083e-01 3.63733768e-01
5.38878143e-01 -1.94386750e-01 -7.46879339e-01 7.30028391e-01
-7.34485328e-01 -4.92741555e-01 6.90221906e-01 8.20023537e-01
4.77668464e-01 1.04497887e-01 7.73176491e-01 -5.44337988e-01
3.02379817e-01 -2.20031649e-01 -7.37701595e-01 3.03807110e-01
-1.44925043e-01 -5.54654121e-01 3.92261147e-01 -3.65375876e-01
-1.31881177e+00 2.66662985e-01 -3.62208962e-01 -5.22975862e-01
-5.25676548e-01 -2.54569829e-01 -5.84618032e-01 -7.20438719e-01
5.96095383e-01 2.14070514e-01 2.33926073e-01 -3.57744932e-01
2.01625451e-01 8.06335449e-01 4.82250571e-01 -4.86519426e-01
8.49485993e-01 6.54211283e-01 3.12061191e-01 -8.61028194e-01
-5.64263701e-01 -5.01764774e-01 -7.92354345e-01 -4.50907826e-01
5.89445055e-01 -1.12841189e+00 -9.45848882e-01 6.43441141e-01
-1.15561426e+00 2.86671430e-01 3.58630657e-01 9.79625881e-02
-2.40866408e-01 3.74929488e-01 -5.30631423e-01 -1.20780432e+00
-5.67712307e-01 -1.34739447e+00 1.37886930e+00 3.82419258e-01
3.49821717e-01 -7.60344863e-01 -5.36907017e-01 3.21633726e-01
4.67324048e-01 4.87510771e-01 4.11453247e-01 -3.89030427e-01
-4.10205632e-01 -3.57720315e-01 -6.10950172e-01 3.76182169e-01
1.07128836e-01 2.41955280e-01 -1.48417187e+00 -5.31740904e-01
3.80750708e-02 -5.19633472e-01 8.54818821e-01 3.29608470e-01
1.34298539e+00 -7.54026696e-02 -4.45072412e-01 7.34532595e-01
1.30228269e+00 1.23893298e-01 6.40352845e-01 -1.74002126e-01
5.56054413e-01 9.18720126e-01 7.28590429e-01 5.32945275e-01
-1.80648357e-01 1.05587685e+00 6.64858341e-01 -3.39449525e-01
-3.46632689e-01 1.88742727e-01 4.32968318e-01 -4.11388418e-03
-2.48874322e-01 1.65130377e-01 -5.06775498e-01 1.40341714e-01
-1.54895675e+00 -1.10738409e+00 1.62281945e-01 2.09224010e+00
5.74615538e-01 -2.72482693e-01 3.08129609e-01 5.22767343e-02
9.04683411e-01 1.37854576e-01 -4.38210011e-01 -1.89427778e-01
-2.67901212e-01 3.10551614e-01 1.51647434e-01 -1.17754964e-02
-1.50118411e+00 1.08302867e+00 5.18770504e+00 1.19541180e+00
-1.26138020e+00 1.76139265e-01 9.00200784e-01 -5.54217473e-02
6.89630866e-01 -5.47212601e-01 -1.36276209e+00 2.06854716e-01
3.01487416e-01 3.85287106e-01 3.27781349e-01 1.01056206e+00
2.01575592e-01 1.50388271e-01 -9.93051410e-01 1.45679569e+00
4.86224592e-01 -1.01581275e+00 -9.53123569e-02 1.93789408e-01
5.51293194e-01 -2.71337777e-01 3.08206111e-01 2.65608549e-01
-4.37175423e-01 -1.43375671e+00 5.73196173e-01 1.50064915e-01
1.08445477e+00 -9.75229740e-01 8.27493548e-01 1.80753559e-01
-1.51521671e+00 -2.97240347e-01 -6.60214961e-01 2.76190937e-01
-2.65724421e-01 3.36181432e-01 -6.04894578e-01 3.11228096e-01
7.09048152e-01 5.45623064e-01 -5.33294678e-01 9.30802345e-01
-1.63913276e-02 1.77838907e-01 -2.51163125e-01 3.13164234e-01
5.31422347e-02 -1.17001958e-01 3.58672261e-01 1.17165530e+00
1.63232401e-01 1.71762705e-01 3.28384563e-02 9.18481708e-01
-4.41019654e-01 3.39720815e-01 -6.53894424e-01 3.40405017e-01
3.76709253e-01 1.98385382e+00 -7.53754199e-01 -3.04351142e-03
-6.01854503e-01 6.50480866e-01 1.89139754e-01 3.12556118e-01
-7.86211431e-01 -1.77876621e-01 7.13279963e-01 1.03742979e-01
3.64242643e-01 1.84040129e-01 3.48669499e-01 -7.88056135e-01
1.46661609e-01 -9.38827693e-01 3.28077406e-01 -3.55841935e-01
-1.07908583e+00 7.00325251e-01 -2.00819030e-01 -1.12410831e+00
-9.25662667e-02 -9.87731397e-01 -7.71169186e-01 8.24634075e-01
-1.83951080e+00 -1.49497807e+00 -5.21075130e-01 9.07560885e-01
6.06696546e-01 -5.56017756e-01 6.84258938e-01 6.53162181e-01
-1.05909562e+00 1.12606609e+00 -1.95508108e-01 6.25840425e-01
9.07345414e-01 -4.63262826e-01 1.57774296e-02 8.57670665e-01
7.70104155e-02 7.15832353e-01 2.63042867e-01 -2.42746040e-01
-1.60141695e+00 -1.32031488e+00 4.53160226e-01 -1.98043570e-01
2.41138220e-01 -5.41133404e-01 -6.64249420e-01 4.38200444e-01
-2.59026170e-01 9.82637107e-01 4.85852689e-01 -2.33940169e-01
-4.98066872e-01 -4.85684425e-01 -1.60316288e+00 3.42792809e-01
1.14563787e+00 -3.77350807e-01 -7.68214166e-02 4.89395827e-01
1.02746505e-02 -3.19836468e-01 -6.68210983e-01 6.74610972e-01
8.77986908e-01 -1.13984978e+00 1.22902727e+00 -4.32494909e-01
6.42973483e-02 -3.85848314e-01 -1.58707738e-01 -6.61971688e-01
-1.52167693e-01 -5.91733098e-01 -2.10754260e-01 1.48107445e+00
-1.27427176e-01 -5.21878719e-01 8.88225317e-01 2.03820229e-01
2.03576043e-01 -9.91664290e-01 -1.26806319e+00 -5.97096086e-01
-2.57662773e-01 -2.22924531e-01 5.94855428e-01 6.33914948e-01
-5.55061340e-01 -1.74689703e-02 -4.25358564e-01 1.09015062e-01
8.76140356e-01 8.05239826e-02 8.11309636e-01 -1.30279481e+00
1.85382545e-01 -6.01388454e-01 -6.98086381e-01 -8.10977042e-01
3.93591970e-01 -5.76047063e-01 -1.12332448e-01 -9.13495541e-01
4.16890025e-01 -3.09811324e-01 -2.43718296e-01 5.77994049e-01
-6.82160705e-02 8.19747150e-01 1.07494615e-01 -4.49619703e-02
-5.57407737e-01 4.71423894e-01 1.21985018e+00 -8.37435131e-04
9.19006243e-02 5.90781234e-02 -7.18343854e-01 8.85293305e-01
4.47667569e-01 -1.27685606e-01 -9.46092233e-02 1.46128526e-02
-2.54769862e-01 -1.19493129e-02 5.76233983e-01 -8.34956169e-01
3.13530564e-01 3.89369428e-02 1.00306034e+00 -5.86567044e-01
5.15868843e-01 -8.21582437e-01 -1.47880375e-01 3.84655207e-01
2.23893568e-01 -2.11413503e-01 4.26595956e-01 4.08333093e-01
-2.68408924e-01 2.42674965e-02 1.32134831e+00 -1.26175091e-01
-9.20818269e-01 6.77658856e-01 1.26538798e-01 -2.44139165e-01
1.28687310e+00 -5.45117497e-01 -2.25208133e-01 -5.27546890e-02
-4.81085479e-01 -5.51098697e-02 1.60276935e-01 6.75793290e-01
8.35084975e-01 -1.47605813e+00 -1.00614107e+00 6.87575877e-01
1.87008053e-01 -1.13133982e-01 3.16937029e-01 1.01827574e+00
-2.14711040e-01 4.91820335e-01 -3.48227233e-01 -6.57636464e-01
-1.84486187e+00 4.65220034e-01 4.80406255e-01 2.13438347e-01
-3.12263638e-01 1.15126491e+00 5.26577115e-01 -3.53927314e-01
5.59500217e-01 1.01988345e-01 -4.69637066e-01 2.32159287e-01
9.61257696e-01 2.55968034e-01 6.80814981e-02 -1.28993940e+00
-5.95539510e-01 1.04921603e+00 -1.54224202e-01 6.01823390e-01
1.18926334e+00 3.44840921e-02 -4.26447153e-01 -3.41261119e-01
1.16544688e+00 -7.62653798e-02 -1.12344325e+00 -3.81940424e-01
-3.58167738e-01 -7.09730089e-01 1.33783698e-01 -6.03318214e-01
-1.54288876e+00 9.84368980e-01 1.06548834e+00 -3.69013906e-01
1.26819193e+00 -3.34182307e-02 3.78740013e-01 2.05691412e-01
5.07341087e-01 -9.90904629e-01 1.59786493e-01 2.29894340e-01
1.21379077e+00 -1.60834599e+00 1.57373726e-01 -6.88968837e-01
-1.93927705e-01 1.31928539e+00 9.06772256e-01 1.78424641e-01
7.80472040e-01 2.39654154e-01 -1.36093087e-02 -1.98626593e-01
-4.30909753e-01 -2.56560475e-01 5.21153092e-01 5.67151010e-01
5.77854514e-01 -1.96547642e-01 2.82611419e-02 3.62488836e-01
3.02073777e-01 -4.20910381e-02 -5.68190329e-02 5.67336440e-01
-5.28513908e-01 -8.45731080e-01 -8.59274507e-01 2.88613170e-01
-7.77033508e-01 9.07258317e-02 -3.60180348e-01 1.05947399e+00
3.93066883e-01 8.48250389e-01 -8.15860331e-02 -9.00106784e-03
8.67247023e-03 -3.61383962e-03 6.90292895e-01 -4.81722653e-01
-5.08148789e-01 3.68685395e-01 -2.94263363e-01 -7.13702679e-01
-4.81550485e-01 -5.65003753e-01 -9.95380104e-01 -1.77555040e-01
-5.77645302e-01 -3.13558847e-01 6.35254741e-01 7.39621639e-01
1.46916077e-01 1.88596696e-01 7.97798455e-01 -1.14744210e+00
-5.97599864e-01 -1.17506170e+00 -8.00962210e-01 3.58311087e-01
2.46423855e-01 -1.08071899e+00 -1.88892737e-01 -2.92027175e-01] | [13.384683609008789, 0.6343944668769836] |
78a1f210-2389-4af2-9c32-0c1aabb22f3a | in-context-analogical-reasoning-with-pre | 2305.17626 | null | https://arxiv.org/abs/2305.17626v2 | https://arxiv.org/pdf/2305.17626v2.pdf | In-Context Analogical Reasoning with Pre-Trained Language Models | Analogical reasoning is a fundamental capacity of human cognition that allows us to reason abstractly about novel situations by relating them to past experiences. While it is thought to be essential for robust reasoning in AI systems, conventional approaches require significant training and/or hard-coding of domain knowledge to be applied to benchmark tasks. Inspired by cognitive science research that has found connections between human language and analogy-making, we explore the use of intuitive language-based abstractions to support analogy in AI systems. Specifically, we apply large pre-trained language models (PLMs) to visual Raven's Progressive Matrices (RPM), a common relational reasoning test. By simply encoding the perceptual features of the problem into language form, we find that PLMs exhibit a striking capacity for zero-shot relational reasoning, exceeding human performance and nearing supervised vision-based methods. We explore different encodings that vary the level of abstraction over task features, finding that higher-level abstractions further strengthen PLMs' analogical reasoning. Our detailed analysis reveals insights on the role of model complexity, in-context learning, and prior knowledge in solving RPM tasks. | ['Joyce Chai', 'Richard L. Lewis', 'Shane Storks', 'Xiaoyang Hu'] | 2023-05-28 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [ 3.48627061e-01 2.18081623e-01 2.45451123e-01 -2.86495656e-01
5.94829035e-04 -6.12859011e-01 9.87583280e-01 7.15845406e-01
-4.49456185e-01 1.86586782e-01 4.46131676e-01 -8.17854643e-01
-5.26944101e-01 -8.88472974e-01 -5.67836821e-01 3.89038287e-02
-1.75538972e-01 6.49652123e-01 2.10901618e-01 -6.95264041e-01
8.01671445e-01 7.69208252e-01 -1.43343723e+00 7.70573497e-01
9.27990317e-01 7.03462720e-01 3.67565662e-01 6.57766521e-01
-3.39758098e-01 1.46413350e+00 -3.80370229e-01 -6.15103364e-01
2.73695976e-01 -6.07194662e-01 -1.15485883e+00 -3.56830239e-01
4.83131826e-01 -4.21281874e-01 -5.26595950e-01 8.16744268e-01
-8.97190273e-02 5.62208891e-01 8.52777183e-01 -9.71961141e-01
-1.67728233e+00 5.57310224e-01 -2.41232350e-01 7.53961146e-01
8.51157248e-01 4.77385014e-01 1.08299601e+00 -7.37490118e-01
4.78412896e-01 1.63874960e+00 7.85699487e-01 3.24130237e-01
-1.53158164e+00 -3.49947393e-01 1.79251000e-01 7.62995899e-01
-1.09090519e+00 -1.58934787e-01 4.90859509e-01 -7.19357669e-01
1.25455356e+00 2.02168137e-01 1.03849578e+00 6.71840191e-01
4.00124252e-01 4.25332546e-01 1.46438253e+00 -6.98581278e-01
4.46931481e-01 -9.10656154e-02 3.53982568e-01 7.38635659e-01
3.35625976e-01 3.53956789e-01 -5.77803254e-01 -8.68216306e-02
1.14066410e+00 2.43794709e-01 -1.11853234e-01 -2.05754802e-01
-1.35193312e+00 6.44618034e-01 1.00106573e+00 3.79188269e-01
-5.67845464e-01 3.56206894e-02 2.93110222e-01 5.77175796e-01
-4.56177294e-01 1.39025557e+00 2.38381401e-02 3.09436560e-01
-5.62187850e-01 3.58540505e-01 5.66781819e-01 6.31066442e-01
5.57031631e-01 8.28727633e-02 -4.98849571e-01 7.19022632e-01
1.55404031e-01 2.11465791e-01 7.19965458e-01 -1.34058154e+00
1.96452320e-01 6.73542857e-01 -1.50755674e-01 -1.09997487e+00
-3.20470423e-01 -3.76016468e-01 -3.10255259e-01 3.29537868e-01
3.54975700e-01 5.41092038e-01 -6.41601622e-01 1.78505647e+00
-1.63260296e-01 -2.66602159e-01 2.74205625e-01 8.71368706e-01
6.26700461e-01 5.20897210e-01 5.13571799e-01 6.37500659e-02
1.48513472e+00 -9.10789549e-01 -3.54836076e-01 -7.55561233e-01
5.11987269e-01 -3.78172219e-01 1.69971192e+00 3.91978770e-01
-1.35769165e+00 -7.88271606e-01 -1.11166930e+00 -6.06366515e-01
-5.83439827e-01 -5.05969763e-01 9.43056107e-01 3.18523079e-01
-1.17746520e+00 8.39904666e-01 -1.87685862e-01 -6.39838457e-01
5.87331474e-01 -1.28256530e-01 -3.61378878e-01 -4.85942811e-01
-1.09328437e+00 1.56744504e+00 5.57839751e-01 -8.53841901e-02
-7.20308721e-01 -7.48628557e-01 -7.75308549e-01 3.41567427e-01
1.73598975e-01 -1.04248714e+00 1.33444452e+00 -7.42120504e-01
-1.13508427e+00 1.02277935e+00 5.36608584e-02 -5.42853892e-01
5.75641878e-02 -1.21957794e-01 -3.23509485e-01 3.99749070e-01
7.99956098e-02 7.69823849e-01 4.93301809e-01 -1.11277580e+00
-3.66447344e-02 -1.70539141e-01 5.40509403e-01 3.11562598e-01
-2.26495311e-01 -1.99233428e-01 2.23721564e-01 -7.96925902e-01
2.94937819e-01 -7.91484535e-01 -1.20511167e-01 5.05179405e-01
1.86551526e-01 -3.03756654e-01 2.88152825e-02 -8.32515419e-01
7.87629068e-01 -2.17143822e+00 2.36248598e-01 1.65392309e-01
4.49141473e-01 8.25943053e-02 -4.00975376e-01 6.06613159e-01
-1.28791720e-01 -5.05770370e-02 -1.39089808e-01 2.62304723e-01
5.24265230e-01 3.62671733e-01 -8.03177178e-01 -1.13837905e-01
1.22066073e-01 1.54267538e+00 -1.12580884e+00 -3.66371870e-01
1.93418846e-01 -5.88721596e-02 -9.17314649e-01 2.62660116e-01
-2.70628929e-01 2.46509779e-02 1.95196778e-01 2.27806836e-01
3.00470412e-01 -4.10985768e-01 3.01904082e-01 -9.01068449e-02
2.26716712e-01 5.55309355e-01 -5.92352986e-01 1.77332759e+00
-4.97721970e-01 8.94128561e-01 -7.21976280e-01 -9.04592454e-01
7.91748643e-01 6.82742661e-03 -4.89136338e-01 -1.16277027e+00
-7.03560337e-02 -1.61870599e-01 7.37979054e-01 -5.06854057e-01
4.19667035e-01 -6.55376673e-01 1.61476746e-01 5.68640769e-01
1.35231921e-02 -7.46126473e-01 2.19048157e-01 4.19011742e-01
1.01797557e+00 4.89473417e-02 7.35803783e-01 -4.58501965e-01
3.41054082e-01 1.95076704e-01 2.85318643e-02 1.11260688e+00
-4.59501632e-02 9.65550691e-02 3.68366539e-01 -5.73094487e-01
-1.02257907e+00 -1.83098590e+00 7.93584138e-02 1.36292791e+00
-5.93102276e-02 -3.36549371e-01 -4.12829101e-01 1.06485404e-01
-2.01826151e-02 1.45965779e+00 -8.35056245e-01 -7.40137160e-01
-2.82263845e-01 -2.17269525e-01 4.32316393e-01 8.52735043e-01
4.70077455e-01 -1.47972238e+00 -9.93712425e-01 1.85752381e-02
1.92052662e-01 -1.07708776e+00 1.27430171e-01 3.89092788e-03
-1.01146603e+00 -8.90410185e-01 -3.23732585e-01 -7.30354607e-01
5.69430351e-01 3.66440982e-01 1.46389961e+00 4.31446820e-01
-5.21589458e-01 9.04711306e-01 -1.61935583e-01 -4.06707436e-01
-5.97186744e-01 -6.25107825e-01 -3.04333046e-02 -8.59105825e-01
5.00435472e-01 -7.74906397e-01 -5.19805014e-01 -2.65203826e-02
-9.51131105e-01 2.77838618e-01 9.71978128e-01 7.17205346e-01
1.70082018e-01 -1.70318127e-01 4.19185787e-01 -3.77939284e-01
1.18598735e+00 -5.65518498e-01 -1.75728813e-01 4.45856273e-01
-3.70057970e-01 2.47860476e-01 6.00559354e-01 -5.96099854e-01
-1.08779395e+00 -5.99114060e-01 3.47142994e-01 -3.82952988e-01
-1.21864334e-01 5.96029878e-01 3.73739481e-01 -1.26739293e-01
1.15330696e+00 4.42936301e-01 -1.05439112e-01 -6.66719526e-02
7.58914292e-01 -4.97810952e-02 7.67795444e-01 -1.23186171e+00
8.17468047e-01 3.08882922e-01 9.54990759e-02 -8.12334299e-01
-1.06061327e+00 1.49112597e-01 -5.16157985e-01 1.06105790e-01
1.02500010e+00 -7.69197106e-01 -1.01472044e+00 -2.80161858e-01
-1.13277805e+00 -4.66526389e-01 -7.42575467e-01 4.52469051e-01
-8.07333052e-01 3.67373973e-01 -8.04320574e-01 -4.55697536e-01
4.84821647e-02 -7.93247044e-01 3.60186279e-01 4.62299995e-02
-9.84516263e-01 -1.01395094e+00 2.52764914e-02 4.01632667e-01
5.24348557e-01 -1.25518024e-01 1.70300388e+00 -6.73748195e-01
-4.95476007e-01 1.63655072e-01 -5.24795234e-01 9.40860361e-02
-1.51535705e-01 -5.80600560e-01 -8.81947994e-01 1.00088738e-01
2.19180465e-01 -6.43698514e-01 7.16970146e-01 -9.32099819e-02
1.33167624e+00 -3.35810810e-01 -4.08643438e-03 3.01517248e-01
1.13697314e+00 2.77325690e-01 8.63967180e-01 4.22720373e-01
4.12965864e-01 6.98089063e-01 2.62551844e-01 1.20802045e-01
6.15737200e-01 2.71240711e-01 -1.83658272e-01 2.38050640e-01
-1.43985584e-01 -5.29867411e-01 1.13974683e-01 4.60506082e-01
-1.44037083e-01 4.11053479e-01 -1.32284856e+00 4.45169896e-01
-1.40348041e+00 -1.33751225e+00 2.50696212e-01 1.84891665e+00
1.04615033e+00 3.23036849e-01 -2.62787789e-01 1.43261492e-01
1.61175892e-01 -1.11174971e-01 -6.04430079e-01 -9.58939970e-01
8.36293101e-02 4.89931941e-01 -2.50474066e-01 3.73306125e-01
-2.21004486e-01 1.03299689e+00 7.23055077e+00 4.39338952e-01
-5.70021391e-01 -2.00709552e-01 1.55037597e-01 3.41981687e-02
-4.67998952e-01 8.84415284e-02 1.36386424e-01 -2.37250134e-01
8.65284860e-01 -4.63560462e-01 1.02029729e+00 5.73419333e-01
-3.20728153e-01 -1.72964826e-01 -1.85729539e+00 8.84744644e-01
1.26183212e-01 -1.43543780e+00 7.95377314e-01 -3.12821716e-01
5.25815129e-01 -3.89939696e-01 2.75866270e-01 6.44317210e-01
7.02461302e-01 -1.50387430e+00 7.62838304e-01 7.97990263e-01
4.01811570e-01 -2.46948272e-01 1.90633088e-02 2.99667269e-01
-8.86032641e-01 -5.11551917e-01 -6.32467330e-01 -9.70689356e-01
-2.03815296e-01 -6.40771911e-02 -8.83408546e-01 1.57512482e-02
6.47067845e-01 4.95251477e-01 -8.85807812e-01 6.63844466e-01
-4.51761782e-01 3.11690830e-02 1.44405037e-01 2.63945088e-02
2.95846686e-02 3.96430399e-03 1.23354048e-01 9.55966353e-01
1.24484695e-01 7.03951955e-01 2.55568009e-02 1.30875373e+00
1.77836344e-01 -1.79943621e-01 -7.97970116e-01 -3.34190786e-01
6.73191726e-01 7.32868671e-01 -6.82109177e-01 -5.16362965e-01
-4.90141332e-01 1.00516593e+00 7.14080513e-01 4.30526465e-01
-4.38308507e-01 -1.11080319e-01 6.60321474e-01 2.07081676e-01
9.46650207e-02 -6.22377455e-01 -6.20843649e-01 -9.05937195e-01
-1.70610383e-01 -1.00475669e+00 4.05134618e-01 -1.59220219e+00
-1.54589140e+00 3.32594544e-01 1.35366529e-01 -7.49689698e-01
-2.99352854e-01 -1.07394695e+00 -8.07360232e-01 1.03184450e+00
-1.09884953e+00 -9.20694172e-01 -4.29853290e-01 6.57271504e-01
5.07535636e-01 -2.95074165e-01 1.06479561e+00 -5.79752564e-01
1.97086871e-01 3.09041739e-01 -3.65114808e-01 1.16465144e-01
3.95326167e-01 -1.13276303e+00 8.35660517e-01 4.66535360e-01
3.86751056e-01 1.38264084e+00 7.63399243e-01 -4.46833313e-01
-1.36045098e+00 -3.09610337e-01 5.41826904e-01 -8.25491190e-01
9.98215556e-01 -2.61181444e-01 -1.28185225e+00 1.04219460e+00
3.58177394e-01 -1.68033972e-01 1.02306592e+00 1.71252191e-01
-1.12778509e+00 2.20626935e-01 -1.08969307e+00 1.21284664e+00
1.42672789e+00 -1.12806010e+00 -1.99417710e+00 2.45469019e-01
9.39528346e-01 -3.41381638e-05 -8.43813717e-01 -1.15546875e-01
6.49017930e-01 -1.01535952e+00 1.46507633e+00 -1.05180037e+00
9.13294852e-01 -2.58483857e-01 -4.21452135e-01 -1.42345464e+00
-1.05900478e+00 -4.34567630e-02 6.42577186e-03 5.52985609e-01
1.71513427e-02 -7.03630984e-01 5.25548272e-02 7.74572730e-01
-4.65628728e-02 -3.34871471e-01 -5.43085515e-01 -9.46910739e-01
4.67499524e-01 -3.82596791e-01 5.34697175e-01 1.20988345e+00
5.30688345e-01 4.68825072e-01 5.16194999e-01 3.01120784e-02
5.03213346e-01 1.03403367e-01 4.08829004e-01 -1.61652529e+00
-5.57810247e-01 -7.85557151e-01 -4.17426258e-01 -6.71883523e-01
3.59651357e-01 -1.33915603e+00 -2.71207631e-01 -1.74432659e+00
2.64131963e-01 -1.24891602e-01 -3.49739164e-01 3.89698178e-01
-7.13327974e-02 6.15800731e-03 6.84616208e-01 2.96090215e-01
-4.07255799e-01 2.86612362e-01 1.35213375e+00 -5.83128817e-02
4.82818559e-02 -8.71962309e-01 -1.21498096e+00 9.39621150e-01
6.79122210e-01 -4.66914549e-02 -8.56244028e-01 -6.49960220e-01
6.22990966e-01 3.99534106e-02 1.05734563e+00 -1.31027520e+00
3.11606318e-01 -5.58181822e-01 9.19997156e-01 8.93008113e-02
3.70371729e-01 -7.60717452e-01 -1.71133429e-01 6.94613576e-01
-7.67413795e-01 5.64714909e-01 7.54660428e-01 5.28648674e-01
1.10066459e-01 -2.26565838e-01 6.59998536e-01 -5.23569107e-01
-1.16488993e+00 -3.36307555e-01 -5.05539477e-01 3.80331010e-01
7.66996384e-01 -3.61095488e-01 -5.68520069e-01 -1.96215838e-01
-8.22621286e-01 7.30730593e-02 6.26011610e-01 4.72680658e-01
8.35771859e-01 -1.31171954e+00 -5.48223317e-01 -6.92910841e-03
4.49195445e-01 -5.30779839e-01 2.86224842e-01 3.81002009e-01
-6.06077492e-01 3.91993970e-01 -9.36035454e-01 -2.39987805e-01
-6.27798915e-01 1.17395592e+00 2.83101350e-01 1.10849090e-01
-7.82758057e-01 7.61213183e-01 6.40850902e-01 -2.48066321e-01
-2.87228674e-01 -5.06556809e-01 -1.58845007e-01 -2.90235043e-01
6.95565283e-01 2.39005536e-01 -3.63111377e-01 -4.74149585e-02
-2.11919457e-01 5.06663382e-01 -2.45666772e-01 -3.35323811e-01
1.07079661e+00 1.41431034e-01 -3.93053383e-01 7.36253917e-01
6.10345244e-01 -3.73042554e-01 -9.85909581e-01 -4.66946900e-01
2.35732123e-01 -4.64844465e-01 -3.59627366e-01 -1.06038892e+00
-1.45869395e-02 1.23185122e+00 2.48455212e-01 -8.21132511e-02
9.45394874e-01 1.67458862e-01 1.88911662e-01 1.17207074e+00
2.80606210e-01 -9.06736553e-01 6.31594777e-01 6.94089711e-01
1.46646333e+00 -9.97363210e-01 2.93144554e-01 -8.17523524e-02
-6.66786134e-01 1.04311109e+00 7.28757560e-01 -3.05321544e-01
1.64074585e-01 -1.36818558e-01 -3.67570460e-01 -2.24106252e-01
-8.62279236e-01 2.14830171e-02 4.78074640e-01 8.49857986e-01
4.44489807e-01 -1.16815247e-01 4.34149690e-02 6.10880911e-01
-7.94636667e-01 5.85517623e-02 4.04978454e-01 9.71208036e-01
-5.79093993e-01 -3.34656835e-01 -5.22478580e-01 5.01282632e-01
2.90531665e-01 -6.21135533e-01 -4.00376827e-01 6.96529269e-01
-7.94719681e-02 5.30115485e-01 4.52198684e-01 3.20820161e-03
4.10835683e-01 4.30635989e-01 1.07176614e+00 -6.80542290e-01
-5.33561528e-01 -1.09807754e+00 -1.94910094e-01 -5.91143668e-01
5.36051430e-02 -5.52446485e-01 -1.48534763e+00 -5.85913777e-01
6.62882090e-01 -6.36869892e-02 1.11955032e-03 1.20770597e+00
1.77173033e-01 6.52391434e-01 -3.88917238e-01 -7.09368169e-01
-6.30811155e-01 -7.23316073e-01 -3.18591833e-01 7.58618534e-01
2.07125872e-01 -8.02292168e-01 -1.82679713e-01 9.91659090e-02] | [10.596675872802734, 2.3239083290100098] |
5ce1c916-8c9b-401a-a4e3-f765fafd099e | subpixel-heatmap-regression-for-facial | 2111.02360 | null | https://arxiv.org/abs/2111.02360v1 | https://arxiv.org/pdf/2111.02360v1.pdf | Subpixel Heatmap Regression for Facial Landmark Localization | Deep Learning models based on heatmap regression have revolutionized the task of facial landmark localization with existing models working robustly under large poses, non-uniform illumination and shadows, occlusions and self-occlusions, low resolution and blur. However, despite their wide adoption, heatmap regression approaches suffer from discretization-induced errors related to both the heatmap encoding and decoding process. In this work we show that these errors have a surprisingly large negative impact on facial alignment accuracy. To alleviate this problem, we propose a new approach for the heatmap encoding and decoding process by leveraging the underlying continuous distribution. To take full advantage of the newly proposed encoding-decoding mechanism, we also introduce a Siamese-based training that enforces heatmap consistency across various geometric image transformations. Our approach offers noticeable gains across multiple datasets setting a new state-of-the-art result in facial landmark localization. Code alongside the pretrained models will be made available at https://www.adrianbulat.com/face-alignment | ['Georgios Tzimiropoulos', 'Enrique Sanchez', 'Adrian Bulat'] | 2021-11-03 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [-2.77684648e-02 -7.41635486e-02 -4.20820341e-02 -8.12845528e-01
-1.07020867e+00 -3.77409875e-01 7.00542808e-01 -1.87164932e-01
-1.80003822e-01 4.37005788e-01 6.88018352e-02 2.20423594e-01
2.02314377e-01 -5.05113006e-01 -8.09453070e-01 -6.49149179e-01
-6.94040209e-02 3.89241070e-01 -1.86908215e-01 3.73123027e-02
2.91289777e-01 6.11893594e-01 -1.47048366e+00 -2.17031613e-02
7.09559977e-01 1.07062685e+00 -3.58187675e-01 3.98984373e-01
1.46209802e-02 2.24379569e-01 -2.18528777e-01 -5.43853045e-01
5.61092138e-01 -2.81719625e-01 -5.53607821e-01 -8.30766782e-02
1.22886717e+00 -3.40419024e-01 -2.64559567e-01 1.00441253e+00
7.20414400e-01 2.02273596e-02 5.76906562e-01 -1.48552299e+00
-6.03977203e-01 -3.85065116e-02 -9.96303856e-01 -2.53871471e-01
2.85011113e-01 -2.77934391e-02 7.60840535e-01 -1.16252136e+00
6.88290238e-01 1.23589885e+00 1.06693995e+00 5.90916932e-01
-1.45214653e+00 -9.89662170e-01 -9.06542614e-02 1.98966026e-01
-1.69547772e+00 -8.94799948e-01 8.92193437e-01 -2.58454204e-01
6.33747518e-01 2.44419873e-01 5.78791082e-01 9.92586374e-01
1.45358309e-01 5.65641999e-01 1.21099794e+00 -4.72859919e-01
1.40767440e-01 -3.81260319e-03 -5.18788874e-01 1.03096044e+00
8.62302666e-04 -1.13966815e-01 -6.81947291e-01 -2.75974184e-01
7.97508597e-01 2.88092364e-02 -4.88088559e-03 -8.29813600e-01
-5.73444426e-01 7.61611521e-01 6.23339295e-01 1.43217100e-02
-8.18238109e-02 3.89232039e-01 1.32101461e-01 2.15112362e-02
8.09398651e-01 2.27813169e-01 -4.23669130e-01 -1.04413971e-01
-1.36902976e+00 2.21452951e-01 6.15074933e-01 8.82529020e-01
1.14376211e+00 -1.50484547e-01 -4.11969274e-02 1.11475134e+00
4.43013132e-01 5.42181134e-01 2.58771479e-01 -9.98079300e-01
2.73603410e-01 4.33334202e-01 -8.51444751e-02 -1.41148674e+00
-4.94890094e-01 -1.86758593e-01 -6.13738537e-01 3.51525962e-01
5.85290074e-01 9.90865193e-03 -1.13142073e+00 1.93071222e+00
6.33443236e-01 4.29604471e-01 -4.94161904e-01 8.56917799e-01
5.80956221e-01 2.73007751e-01 6.44648308e-03 1.38148591e-01
1.16139615e+00 -1.04029799e+00 -6.73145711e-01 -2.28910726e-02
4.11140144e-01 -9.53075409e-01 9.72367167e-01 1.28401607e-01
-1.17335534e+00 -1.51888162e-01 -1.00451040e+00 -4.76578861e-01
-4.24541265e-01 1.96066156e-01 6.67976081e-01 7.92916179e-01
-1.38444448e+00 5.22152543e-01 -1.06661630e+00 -7.32986152e-01
1.05046117e+00 5.58438361e-01 -6.77643239e-01 -1.76580027e-01
-5.86202085e-01 8.86234343e-01 -4.57658410e-01 2.25777447e-01
-7.17620850e-01 -8.77166212e-01 -9.90991294e-01 -1.23687319e-01
1.30998805e-01 -4.66864407e-01 9.95950818e-01 -9.49457407e-01
-1.66158748e+00 1.05094326e+00 -3.96194905e-01 -1.20671503e-01
8.01081300e-01 -3.59007418e-01 -8.57813880e-02 -1.75456591e-02
5.29412627e-02 1.02201200e+00 1.02058005e+00 -1.06685710e+00
-1.52748331e-01 -6.16659403e-01 -4.57599491e-01 3.16678779e-03
-4.07624602e-01 1.00712292e-01 -7.29238391e-01 -5.67024589e-01
1.14617109e-01 -1.05984128e+00 -5.74012920e-02 4.38414514e-01
-3.77153397e-01 4.37553450e-02 8.04545045e-01 -6.55602872e-01
9.77434278e-01 -2.11117578e+00 -9.59234089e-02 3.38791251e-01
4.94063087e-02 1.45802408e-01 -3.11455578e-01 1.33944958e-01
-4.93996870e-03 2.85121761e-02 -3.92170697e-01 -9.13829923e-01
1.37472376e-01 5.59116788e-02 -5.53164557e-02 8.34196806e-01
3.43239158e-01 9.23204362e-01 -5.96663117e-01 -6.05371416e-01
1.10474028e-01 1.04349780e+00 -7.08901167e-01 3.26787867e-02
6.11009784e-02 5.20937979e-01 1.82874296e-02 9.99712110e-01
1.11926198e+00 -4.13936824e-02 9.64776846e-04 -2.31370434e-01
4.58706021e-02 2.85624303e-02 -9.92058516e-01 2.20740485e+00
-4.05560315e-01 7.95733452e-01 2.08865017e-01 -6.90105557e-01
8.94853711e-01 9.50200483e-02 6.47624552e-01 -7.63266325e-01
1.41935080e-01 2.51218855e-01 -3.25360715e-01 -1.72472849e-01
3.97420377e-01 -5.46677560e-02 3.30974132e-01 4.18151706e-01
1.81527302e-01 -7.07511753e-02 -1.48834884e-01 -6.31876616e-03
8.71014059e-01 5.58251977e-01 2.78654657e-02 -3.00685823e-01
2.51637518e-01 -3.33058655e-01 5.27890623e-01 3.27986717e-01
-3.80154938e-01 1.05878711e+00 4.81462896e-01 -7.06422091e-01
-1.17628622e+00 -8.38196099e-01 -4.49149936e-01 1.10714364e+00
-8.33811238e-02 -5.77224255e-01 -1.05035508e+00 -6.81582749e-01
2.26328999e-01 1.04815260e-01 -1.08363140e+00 -6.66079717e-03
-6.33579552e-01 -9.06461298e-01 7.24013627e-01 4.95365590e-01
4.56788719e-01 -5.55170953e-01 -4.15103167e-01 -1.32335439e-01
-2.11279299e-02 -1.12901700e+00 -6.11947656e-01 -2.00015143e-01
-6.15695596e-01 -1.01471102e+00 -8.48954141e-01 -5.39945841e-01
1.04891348e+00 -1.51113756e-02 9.30902362e-01 8.29430372e-02
-7.52310157e-01 4.31080461e-01 1.25461757e-01 -2.92421371e-01
-2.93939281e-02 9.04603899e-02 -3.96867469e-02 2.22026750e-01
4.12732780e-01 -6.19042337e-01 -8.43315423e-01 4.83596444e-01
-5.19440234e-01 1.80735886e-02 3.49964678e-01 7.25627899e-01
5.15621722e-01 -4.82578069e-01 7.49021322e-02 -7.58562624e-01
2.03324899e-01 -2.71767348e-01 -6.90239787e-01 1.91567346e-01
-6.82863355e-01 2.84394287e-02 1.59758464e-01 -2.51013637e-01
-9.03448105e-01 2.41475731e-01 -2.03362361e-01 -3.64657998e-01
-1.09087400e-01 -8.08383748e-02 -1.15627222e-01 -6.62596464e-01
4.99322087e-01 -1.39917940e-01 2.72067547e-01 -5.32390893e-01
4.74848479e-01 3.13850522e-01 4.32999283e-01 -5.53314924e-01
9.74819779e-01 7.19478488e-01 3.78636211e-01 -6.98008478e-01
-5.22574961e-01 -1.45082757e-01 -9.02726054e-01 -7.92530626e-02
7.19978273e-01 -8.19609880e-01 -5.27307749e-01 4.71198887e-01
-9.80372369e-01 -4.42390144e-01 1.11287780e-01 1.30994484e-01
-5.18029511e-01 2.95923859e-01 -3.94589126e-01 -6.79645598e-01
-2.81903088e-01 -1.08439469e+00 1.28254116e+00 3.55495244e-01
-2.16173425e-01 -9.19709563e-01 2.55100310e-01 1.27269402e-01
7.86530554e-01 5.73312342e-01 5.85570097e-01 -2.42589056e-01
-5.80423236e-01 -3.25567544e-01 -4.48881686e-01 3.81822996e-02
1.63368538e-01 2.40518823e-01 -1.26925099e+00 -3.10862690e-01
-4.51100022e-01 -3.78598958e-01 7.23200560e-01 3.92617494e-01
1.14611328e+00 -3.68378431e-01 -2.47623056e-01 1.17128420e+00
1.34753394e+00 -3.85556996e-01 7.33288467e-01 4.24313098e-01
8.91113341e-01 6.89891398e-01 4.29489940e-01 3.77754211e-01
5.41670740e-01 1.07616591e+00 4.38463509e-01 -4.71616089e-01
-3.48705798e-01 -4.12641644e-01 4.38831970e-02 3.14261973e-01
-1.38244808e-01 2.77960181e-01 -9.54885423e-01 3.42443168e-01
-1.68874311e+00 -7.45339692e-01 1.21671565e-01 2.25529313e+00
9.28593278e-01 -4.53841507e-01 1.47988960e-01 -2.39983633e-01
3.64218086e-01 2.03597099e-01 -4.06313688e-01 -3.38268161e-01
-2.92220842e-02 4.09424484e-01 5.87846696e-01 6.27936125e-01
-1.25904858e+00 1.23661399e+00 6.47756004e+00 7.74557233e-01
-1.42125618e+00 2.90508807e-01 7.11855650e-01 -3.16632777e-01
-1.65869415e-01 -1.19055070e-01 -7.57140696e-01 4.46324646e-01
5.82931101e-01 3.40364248e-01 4.84689862e-01 8.00934792e-01
-5.13486490e-02 -9.03290436e-02 -1.00372052e+00 1.19311249e+00
4.20499176e-01 -1.09603965e+00 -2.61580467e-01 1.98410586e-01
9.01716650e-01 5.23122884e-02 4.25113380e-01 -7.02223331e-02
8.63193199e-02 -1.30385530e+00 5.49287558e-01 3.93632561e-01
1.13927603e+00 -5.66314995e-01 3.97581249e-01 -4.06893104e-01
-1.09301043e+00 2.08319217e-01 -3.73723805e-01 1.69599086e-01
-1.35983720e-01 3.82726699e-01 -8.91328514e-01 2.74672151e-01
6.40960813e-01 6.49072230e-01 -9.00932550e-01 1.09068286e+00
-2.72479832e-01 2.91280448e-01 -5.62884271e-01 3.37768763e-01
3.23046409e-02 -1.02737077e-01 1.90896273e-01 1.35054088e+00
3.91618520e-01 -2.87545592e-01 -1.79119810e-01 8.30930829e-01
-1.90155327e-01 2.31353387e-01 -4.65659499e-01 2.72048682e-01
3.71843070e-01 1.50706506e+00 -7.57556200e-01 2.45990474e-02
-3.06661129e-01 1.08560991e+00 5.97664714e-01 5.57894409e-01
-9.20688093e-01 -2.74110973e-01 9.16723371e-01 3.08361858e-01
1.89861774e-01 -4.66915429e-01 -6.12651467e-01 -1.01514065e+00
8.87066051e-02 -5.51838756e-01 1.63850769e-01 -4.25731361e-01
-9.39133942e-01 4.80863273e-01 -7.90240616e-02 -8.55010629e-01
-1.23307638e-01 -5.35040975e-01 -5.97293913e-01 7.22075701e-01
-1.78191900e+00 -1.48599315e+00 -6.02329791e-01 5.32346964e-01
2.29421616e-01 -7.75113031e-02 9.70164239e-01 5.05953550e-01
-6.68111086e-01 1.04265547e+00 2.48933639e-02 1.49617895e-01
1.24562836e+00 -1.19446850e+00 5.01364410e-01 7.09838867e-01
3.54578823e-01 6.48390472e-01 5.49680650e-01 -3.51634532e-01
-1.42032421e+00 -9.31137204e-01 8.57379615e-01 -6.17131114e-01
4.04857278e-01 -7.83803225e-01 -8.22939754e-01 7.36841321e-01
9.99044850e-02 4.30563271e-01 7.41779625e-01 2.74920017e-01
-7.16213822e-01 -4.16260511e-01 -1.39332175e+00 5.38241446e-01
1.04718375e+00 -5.46712041e-01 2.83224899e-02 3.31858814e-01
9.69233736e-03 -6.19266868e-01 -6.88055098e-01 4.89568263e-01
1.03092873e+00 -1.01869988e+00 9.11907613e-01 -3.17926347e-01
3.33138466e-01 -2.19067901e-01 -8.10786188e-02 -1.09394729e+00
-1.80942982e-01 -8.81337583e-01 -6.29583970e-02 1.25747812e+00
3.14193845e-01 -6.97982490e-01 9.04007316e-01 9.18942809e-01
1.68577492e-01 -9.43948746e-01 -1.14033699e+00 -4.65260386e-01
6.00767955e-02 -1.76495448e-01 6.69267237e-01 1.00735974e+00
-5.97924143e-02 -1.94529310e-01 -4.67858255e-01 8.58472213e-02
7.91560531e-01 -1.35692835e-01 1.17452049e+00 -9.88372505e-01
1.25473469e-01 -4.92515057e-01 -6.51214242e-01 -8.30231726e-01
1.75309628e-01 -7.48025894e-01 1.11672275e-01 -1.15705836e+00
1.91550955e-01 -7.73081541e-01 -1.26429915e-01 8.65714312e-01
-2.26791911e-02 1.04712427e+00 1.20370202e-01 2.81799495e-01
-5.61889410e-01 5.37322521e-01 9.15496230e-01 1.12720959e-01
8.21413472e-03 -3.47005248e-01 -4.77741182e-01 7.10874856e-01
8.66974175e-01 -4.63497341e-01 -1.01542035e-02 -7.28713095e-01
1.22730665e-01 -6.47084594e-01 3.18550408e-01 -9.32269335e-01
4.38448727e-01 1.12721644e-01 7.24618316e-01 -1.39600933e-01
6.82015121e-01 -7.64030993e-01 1.23488553e-01 1.64218694e-01
-2.80261248e-01 1.22651830e-01 4.19461995e-01 3.47551823e-01
-2.57022865e-02 3.10624182e-01 1.02874577e+00 2.11187214e-01
-2.94119000e-01 5.35017967e-01 2.68826902e-01 -6.17716648e-02
1.00126183e+00 -2.15857774e-01 -2.76301205e-01 -4.52169746e-01
-5.05295157e-01 -7.35325413e-03 8.48281860e-01 4.18413579e-01
3.92742962e-01 -1.52173591e+00 -7.03761756e-01 6.69647872e-01
-2.41040085e-02 -2.04159006e-01 7.71698430e-02 1.29002500e+00
-6.48481250e-01 2.69243062e-01 -3.72920185e-01 -6.43565595e-01
-1.40038824e+00 6.69872612e-02 3.41135770e-01 1.14371412e-01
-4.64241624e-01 1.11963201e+00 1.02085061e-01 -3.78846765e-01
3.89862806e-01 -9.14549977e-02 2.63679147e-01 -6.43927157e-02
5.15321255e-01 3.73864591e-01 1.94440633e-01 -9.09807920e-01
-5.70967555e-01 1.15369749e+00 -1.25768989e-01 -1.60486147e-01
1.24576557e+00 -9.50071067e-02 -1.80390805e-01 6.23688437e-02
1.46102953e+00 2.41023600e-01 -1.58826005e+00 -1.39735946e-02
-8.03370625e-02 -9.46217299e-01 -9.20016915e-02 -7.42147207e-01
-1.19861150e+00 9.62375224e-01 7.95714021e-01 -2.09145367e-01
1.01956391e+00 -1.12533316e-01 7.66863167e-01 -2.08629936e-01
2.93010712e-01 -9.99005258e-01 4.82478142e-02 2.39976808e-01
8.56533051e-01 -1.44340026e+00 1.53292164e-01 -4.07085955e-01
-3.26317281e-01 1.09933543e+00 5.02400160e-01 -1.46506608e-01
6.41114771e-01 4.38465118e-01 3.38256150e-01 -1.21000782e-01
-3.42473924e-01 5.44654019e-03 4.47715968e-01 7.45393395e-01
6.51791811e-01 -1.52133003e-01 -2.98194382e-02 -7.53308274e-03
-3.38211864e-01 -1.10497586e-01 -2.68765595e-02 7.64231801e-01
-3.21155675e-02 -1.20188606e+00 -4.67272729e-01 2.15839464e-02
-5.35741091e-01 -7.52213895e-02 -3.62855464e-01 6.89613640e-01
4.12611403e-02 5.65942824e-01 1.59361333e-01 -1.55812636e-01
4.11508009e-02 2.07230538e-01 8.35641921e-01 -2.41020903e-01
-1.89405918e-01 1.25889212e-01 -2.62401372e-01 -1.02405810e+00
-3.11603338e-01 -7.10074544e-01 -1.04864585e+00 -4.45147365e-01
-1.39253378e-01 -1.41093716e-01 1.18161917e+00 6.10302985e-01
7.39017606e-01 4.51787151e-02 5.36066115e-01 -1.10679734e+00
-3.35842192e-01 -9.90940392e-01 -3.65597218e-01 3.99372905e-01
3.61196965e-01 -7.89319396e-01 -2.86654681e-01 -4.11639288e-02] | [13.444022178649902, 0.37481269240379333] |
06a17d19-9c18-4f53-bbef-95616ce7c0a0 | hifisinger-towards-high-fidelity-neural | 2009.01776 | null | https://arxiv.org/abs/2009.01776v1 | https://arxiv.org/pdf/2009.01776v1.pdf | HiFiSinger: Towards High-Fidelity Neural Singing Voice Synthesis | High-fidelity singing voices usually require higher sampling rate (e.g., 48kHz) to convey expression and emotion. However, higher sampling rate causes the wider frequency band and longer waveform sequences and throws challenges for singing voice synthesis (SVS) in both frequency and time domains. Conventional SVS systems that adopt small sampling rate cannot well address the above challenges. In this paper, we develop HiFiSinger, an SVS system towards high-fidelity singing voice. HiFiSinger consists of a FastSpeech based acoustic model and a Parallel WaveGAN based vocoder to ensure fast training and inference and also high voice quality. To tackle the difficulty of singing modeling caused by high sampling rate (wider frequency band and longer waveform), we introduce multi-scale adversarial training in both the acoustic model and vocoder to improve singing modeling. Specifically, 1) To handle the larger range of frequencies caused by higher sampling rate, we propose a novel sub-frequency GAN (SF-GAN) on mel-spectrogram generation, which splits the full 80-dimensional mel-frequency into multiple sub-bands and models each sub-band with a separate discriminator. 2) To model longer waveform sequences caused by higher sampling rate, we propose a multi-length GAN (ML-GAN) for waveform generation to model different lengths of waveform sequences with separate discriminators. 3) We also introduce several additional designs and findings in HiFiSinger that are crucial for high-fidelity voices, such as adding F0 (pitch) and V/UV (voiced/unvoiced flag) as acoustic features, choosing an appropriate window/hop size for mel-spectrogram, and increasing the receptive field in vocoder for long vowel modeling. Experiment results show that HiFiSinger synthesizes high-fidelity singing voices with much higher quality: 0.32/0.44 MOS gain over 48kHz/24kHz baseline and 0.83 MOS gain over previous SVS systems. | ['Tie-Yan Liu', 'Xu Tan', 'Tao Qin', 'Jian Luan', 'Jiawei Chen'] | 2020-09-03 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [ 1.84529901e-01 -1.40343262e-02 1.11068867e-01 1.11776911e-01
-9.86194968e-01 -7.37704992e-01 1.96343854e-01 -7.37059116e-01
1.81818575e-01 6.45085156e-01 4.89217550e-01 -1.94418028e-01
3.63744497e-01 -6.80806220e-01 -5.86688697e-01 -7.05875754e-01
1.24546647e-01 -2.67559677e-01 -1.56611856e-02 -3.42474490e-01
-4.45673943e-01 2.14400917e-01 -1.60136616e+00 2.83209562e-01
8.29524696e-01 7.83403754e-01 3.82462919e-01 1.18557405e+00
1.13937318e-01 2.21362695e-01 -1.22708523e+00 -1.38809532e-01
4.05731052e-01 -1.08323479e+00 -3.29531014e-01 -1.46805450e-01
3.13050121e-01 -3.98909032e-01 -3.56482267e-01 8.59359503e-01
9.50226128e-01 2.23991320e-01 6.43292665e-01 -1.00556970e+00
-5.64396739e-01 6.86266482e-01 -1.72246501e-01 3.71683575e-02
2.71207541e-01 5.65331936e-01 9.78919446e-01 -5.81075788e-01
1.42613173e-01 1.38152814e+00 7.45040953e-01 9.45729613e-01
-1.04586709e+00 -1.13055420e+00 -4.51590002e-01 -8.51993412e-02
-1.17179513e+00 -5.39836407e-01 1.07843876e+00 -2.26854131e-01
7.27775812e-01 7.46075928e-01 7.46411383e-01 1.24888742e+00
-7.87425339e-02 4.04663175e-01 9.80894327e-01 -4.19941366e-01
-3.93568911e-02 -2.04121582e-02 -5.50258875e-01 1.33370668e-01
-4.45104271e-01 4.81063426e-01 -3.32751423e-01 1.18258884e-02
1.02996159e+00 -4.40597296e-01 -5.33997118e-01 6.99141145e-01
-1.00452983e+00 7.66019464e-01 1.33657917e-01 4.18923706e-01
-2.64409751e-01 3.40307474e-01 3.87991637e-01 5.19789338e-01
3.84015590e-02 6.84997380e-01 -2.01743484e-01 -3.16209018e-01
-1.08497036e+00 3.19453597e-01 6.67275012e-01 6.69866920e-01
1.89660132e-01 1.15646505e+00 -6.32790327e-01 1.25002897e+00
2.80745953e-01 8.42919350e-01 7.83176422e-01 -9.62624550e-01
4.52252328e-01 -2.95789808e-01 -1.66277885e-01 -6.70447767e-01
-2.60619428e-02 -5.58354020e-01 -9.49637949e-01 9.75624323e-02
1.72332153e-01 -5.15360534e-01 -9.15426731e-01 1.92707992e+00
2.75256068e-01 4.48925585e-01 2.27086127e-01 1.00893927e+00
8.55449855e-01 1.30313754e+00 -2.89626271e-01 -4.72388923e-01
1.34396124e+00 -1.11508524e+00 -1.07339525e+00 6.89107180e-02
8.16970244e-02 -1.06818569e+00 1.52190983e+00 2.66769618e-01
-1.26593828e+00 -1.14008379e+00 -1.14191425e+00 3.43319140e-02
6.54985383e-02 1.94910675e-01 -1.13153026e-01 9.66338456e-01
-9.09572363e-01 6.09300494e-01 -3.20712209e-01 3.33336651e-01
-1.65414840e-01 2.55381584e-01 3.70174274e-02 4.52798963e-01
-1.61982310e+00 3.87323588e-01 1.13509230e-01 -8.59070197e-02
-1.05806577e+00 -1.00877953e+00 -8.96920919e-01 1.33709088e-01
-1.10103741e-01 -5.09434402e-01 1.36242735e+00 -9.05130029e-01
-2.15171814e+00 4.56960648e-02 3.80930752e-02 -3.50087315e-01
1.95537850e-01 -9.66775268e-02 -9.49891031e-01 -1.16478028e-02
-3.53563040e-01 4.43424582e-01 1.43322551e+00 -1.15426886e+00
-2.81474739e-01 1.91885203e-01 -2.75466770e-01 2.40826353e-01
-5.11635244e-01 -3.40102911e-02 1.53981477e-01 -1.25659931e+00
-3.69679958e-01 -7.85901129e-01 2.23867781e-03 -5.51145434e-01
-4.26990151e-01 2.97977850e-02 1.06477296e+00 -9.40804124e-01
1.50790203e+00 -2.29443622e+00 1.01184115e-01 -1.58329099e-01
-7.37070665e-02 8.57001960e-01 -3.09042901e-01 4.67217654e-01
-1.27090469e-01 1.89847946e-01 -3.15389663e-01 -2.56014198e-01
-3.16115320e-02 1.23522751e-01 -5.38322151e-01 3.58877108e-02
3.30245018e-01 7.31631398e-01 -7.43942499e-01 -1.92420468e-01
2.11888164e-01 8.31723332e-01 -7.61533380e-01 6.32390738e-01
-3.95406559e-02 7.45842814e-01 -1.30558267e-01 3.66719544e-01
5.68273783e-01 7.25318134e-01 -2.71366984e-01 -4.51816171e-01
-1.09384559e-01 4.95203406e-01 -1.31780899e+00 1.31345367e+00
-8.28653336e-01 4.02375191e-01 4.35595930e-01 -5.47209263e-01
1.31763637e+00 7.80465543e-01 6.29488677e-02 -4.36778843e-01
1.02387734e-01 3.67790192e-01 3.51052910e-01 -5.27360559e-01
4.16165918e-01 -7.91985095e-01 8.49143788e-03 8.32079723e-02
3.04041713e-01 -8.17646801e-01 -3.48554611e-01 -4.35745746e-01
7.82550931e-01 -1.10505708e-01 -1.08718358e-01 -3.68873365e-02
6.86948478e-01 -7.63517857e-01 6.91288292e-01 2.61101007e-01
-1.65214688e-02 8.93241048e-01 1.20998763e-01 1.21410035e-01
-1.28664935e+00 -9.31589305e-01 -3.97452898e-02 8.37331891e-01
-2.82405823e-01 -4.79613692e-01 -9.67692971e-01 -1.22931100e-01
-1.20216273e-01 7.22647905e-01 -1.57392025e-01 -4.13895071e-01
-8.37424159e-01 -2.21860498e-01 1.25180507e+00 4.56772715e-01
3.26621890e-01 -1.32702363e+00 -1.31490976e-01 4.43031818e-01
-5.16218804e-02 -8.79786789e-01 -1.13459253e+00 -5.70637956e-02
-5.98450840e-01 -3.59279424e-01 -1.01091170e+00 -8.49596918e-01
-5.11982404e-02 -2.26210326e-01 7.34268248e-01 -2.01555699e-01
-1.27863914e-01 1.35765849e-02 -5.33780515e-01 -4.81574237e-01
-1.00925267e+00 -8.21155161e-02 3.97319108e-01 8.57218131e-02
-5.29615998e-01 -8.36322606e-01 -6.33102298e-01 3.15751880e-01
-1.04249930e+00 -1.32387191e-01 3.30078572e-01 1.07758856e+00
4.77406114e-01 1.66001320e-01 1.25533259e+00 -4.70432729e-01
9.40614641e-01 -3.64994675e-01 -3.17735255e-01 -2.10157096e-01
-1.82228938e-01 -2.82824099e-01 1.41805196e+00 -9.18663800e-01
-9.36657131e-01 -3.86601061e-01 -8.49258661e-01 -7.21424639e-01
-3.99895236e-02 3.35123427e-02 -3.35296482e-01 2.82674670e-01
6.96802020e-01 3.01339060e-01 1.59078673e-01 -5.79385221e-01
3.60600233e-01 1.15807188e+00 7.18123078e-01 -4.14215237e-01
1.16795230e+00 -2.00564981e-01 -3.05971742e-01 -1.01281261e+00
-3.49029571e-01 -1.14978798e-01 1.68087929e-02 -2.81273127e-02
7.71774054e-01 -9.25250053e-01 -6.12019897e-01 6.30255103e-01
-1.00979877e+00 -4.83651102e-01 -6.69406056e-01 7.56232560e-01
-5.80871522e-01 2.53537923e-01 -9.46852505e-01 -1.21680355e+00
-8.54765117e-01 -1.11507952e+00 8.54840875e-01 4.46909547e-01
-2.73307413e-01 -6.65691972e-01 2.42334660e-02 4.87229735e-01
8.49589407e-01 4.44111824e-01 6.65082157e-01 -2.10089177e-01
4.79245521e-02 4.49911579e-02 4.33544636e-01 9.69659686e-01
3.83567303e-01 -3.27945761e-02 -1.26337802e+00 -3.24289590e-01
3.82074058e-01 -2.77648062e-01 5.25184393e-01 3.32334906e-01
9.71177101e-01 -6.67758882e-01 3.64837795e-01 7.89957106e-01
9.91052628e-01 6.02132797e-01 6.92186892e-01 -4.27228361e-01
9.32415009e-01 4.83875990e-01 6.12799525e-01 2.98845470e-01
-7.34128132e-02 6.82463288e-01 1.73090339e-01 -2.49277964e-01
-8.52210283e-01 -5.53424358e-01 8.07447076e-01 1.63197565e+00
-3.08326185e-02 -2.99430937e-01 -2.17926875e-01 6.00031316e-01
-1.01018178e+00 -8.97706091e-01 -7.40339011e-02 2.17775083e+00
1.24994600e+00 -1.20995268e-01 4.45483059e-01 6.54861569e-01
8.62647176e-01 4.39944446e-01 -5.78164339e-01 -8.77840817e-01
-9.49794948e-02 7.92497218e-01 1.39378890e-01 5.84491670e-01
-5.87088466e-01 8.31946313e-01 5.64219809e+00 1.42076051e+00
-1.44199169e+00 1.54551953e-01 4.24020708e-01 -2.01523155e-01
-6.78683937e-01 -2.30034471e-01 -7.72925377e-01 6.78030849e-01
1.28359091e+00 -2.94347517e-02 9.43578422e-01 5.30483305e-01
3.34783018e-01 6.73889637e-01 -7.09046543e-01 9.14595366e-01
-7.53329769e-02 -1.06333029e+00 3.26718017e-02 -4.00023237e-02
7.03297615e-01 -4.07005489e-01 2.74224043e-01 4.86379027e-01
-2.34942272e-01 -1.42137086e+00 9.04315889e-01 7.02424347e-02
1.59534097e+00 -8.78857255e-01 4.56173778e-01 3.44426364e-01
-1.38263929e+00 -9.07804593e-02 -1.87246531e-01 1.47479447e-02
2.35850647e-01 3.89515549e-01 -9.66801345e-01 4.39878583e-01
3.39534312e-01 2.61814117e-01 2.13332638e-01 6.71192050e-01
-1.26424655e-01 1.24823928e+00 -2.75337607e-01 6.84936866e-02
6.90950975e-02 -5.08949682e-02 8.57371509e-01 1.29080176e+00
6.34300470e-01 6.43039346e-02 -2.07273751e-01 9.50027525e-01
-1.61433339e-01 1.97207555e-02 -4.55344707e-01 -2.49065697e-01
7.72806227e-01 1.11886823e+00 7.94268996e-02 -4.76974063e-02
-5.69493920e-02 8.16305518e-01 -4.30177927e-01 3.53307247e-01
-1.01165223e+00 -7.78180838e-01 8.92203987e-01 2.50548393e-01
3.25685382e-01 -1.70121729e-01 7.20550120e-02 -7.39098608e-01
-1.48719445e-01 -1.25561523e+00 -1.66829359e-02 -5.74489236e-01
-1.02132928e+00 8.83077860e-01 -4.75133657e-01 -1.30659533e+00
-4.75292414e-01 -2.34673947e-01 -1.01657057e+00 1.20430017e+00
-1.60565388e+00 -1.14551294e+00 1.84546784e-02 5.71449578e-01
9.71807539e-01 -1.66541770e-01 9.90686059e-01 4.44156498e-01
-3.66183043e-01 9.23227966e-01 3.03562032e-03 5.00938017e-03
7.40405738e-01 -1.22086430e+00 6.84566557e-01 6.69190526e-01
5.39518008e-03 3.25325608e-01 7.14253128e-01 -3.10552180e-01
-1.53757143e+00 -1.30175960e+00 5.33342421e-01 -5.68353981e-02
3.02032441e-01 -5.47158837e-01 -9.65031564e-01 3.96677926e-02
2.16660053e-01 -1.11223377e-01 7.69430280e-01 -4.99154270e-01
-2.42407322e-01 -3.43083680e-01 -1.40327704e+00 4.86761779e-01
7.15415597e-01 -7.54699707e-01 -4.86772537e-01 -9.69919041e-02
1.26104772e+00 -4.80137408e-01 -1.17892849e+00 5.55800617e-01
4.73262668e-01 -7.50274420e-01 9.30616260e-01 -1.62453055e-01
4.60923702e-01 -5.02281308e-01 -1.99840352e-01 -1.65613413e+00
-2.99331456e-01 -1.30876911e+00 -1.90801829e-01 1.72778761e+00
3.07140827e-01 -5.26389420e-01 2.50664800e-01 -2.07556352e-01
-4.28962529e-01 -6.99568510e-01 -8.81335914e-01 -9.20970142e-01
2.75659084e-01 -3.87375981e-01 8.83826375e-01 8.61103714e-01
-1.20977327e-01 4.62064505e-01 -8.16918075e-01 1.31884754e-01
2.38234922e-01 -6.15023673e-02 7.03766048e-01 -6.16219640e-01
-6.68978810e-01 -3.01967561e-01 1.42939552e-03 -9.66784716e-01
-1.38810873e-01 -5.52033126e-01 1.82164460e-01 -1.02739811e+00
-6.12525105e-01 -3.91673714e-01 -2.07010239e-01 1.82493731e-01
-2.89587587e-01 4.28685516e-01 5.21999776e-01 -1.01363003e-01
3.82953793e-01 8.28313828e-01 1.77258730e+00 4.89850082e-02
-3.32643896e-01 1.43370181e-01 -3.73759955e-01 5.77823877e-01
7.95374751e-01 -2.88254857e-01 -5.68264425e-01 -1.15954384e-01
-4.84279364e-01 6.75250173e-01 9.44604874e-02 -1.17981577e+00
-7.22715631e-02 3.33392322e-02 1.97706431e-01 -2.34852284e-01
7.43524551e-01 -5.35052955e-01 4.08994615e-01 5.15451908e-01
-2.24678382e-01 -3.30030799e-01 3.26440483e-01 2.85566270e-01
-4.45477724e-01 -1.08170301e-01 1.02145493e+00 5.37827089e-02
3.22597325e-02 8.22101906e-02 -3.01634640e-01 1.19777821e-01
6.54211462e-01 -1.36232555e-01 -7.31131509e-02 -6.88702881e-01
-4.78011668e-01 -1.65463075e-01 1.40559345e-01 5.44902742e-01
5.62632620e-01 -1.50430429e+00 -9.47105646e-01 6.54713154e-01
-3.88764143e-01 2.32841186e-02 6.05079234e-01 4.37058747e-01
-2.65962541e-01 1.68938801e-01 -4.06838730e-02 -2.27763981e-01
-1.11878228e+00 1.95894122e-01 4.33114856e-01 -1.98151786e-02
-4.52326357e-01 9.60659087e-01 2.66981184e-01 -4.16396081e-01
1.22917786e-01 -3.23980778e-01 -3.18221182e-01 -4.76832576e-02
4.16637570e-01 3.93369615e-01 -8.05583969e-02 -7.64995635e-01
-2.11969810e-03 6.38854682e-01 5.19766986e-01 -2.04080477e-01
1.10996175e+00 7.57719278e-02 2.42366552e-01 5.14492869e-01
1.30989361e+00 6.31582916e-01 -1.33557880e+00 2.08515808e-01
-7.22615302e-01 -3.95354778e-01 1.10321008e-01 -7.38088548e-01
-1.08276153e+00 9.31586564e-01 4.35556799e-01 5.18011689e-01
1.33558106e+00 -3.80729079e-01 1.65086412e+00 -5.38288176e-01
-2.98714098e-02 -8.99654686e-01 1.13778979e-01 4.86619681e-01
1.14632905e+00 -4.79990810e-01 -6.71373188e-01 -3.57133567e-01
-8.83132577e-01 1.05446148e+00 5.03483593e-01 -2.65160114e-01
3.58383983e-01 5.01124680e-01 3.15768778e-01 3.56367081e-01
-5.57957351e-01 6.99908435e-02 4.96115148e-01 6.38545513e-01
5.04145205e-01 2.29683086e-01 -2.14030132e-01 9.02221978e-01
-9.76884246e-01 -4.18790221e-01 4.05463248e-01 2.13396832e-01
-2.79421806e-01 -1.32986653e+00 -7.03987539e-01 3.68135005e-01
-8.47814381e-01 -2.97395557e-01 -9.92126837e-02 1.22052751e-01
3.28323603e-01 1.24468398e+00 3.28673162e-02 -8.67743492e-01
5.03222585e-01 2.20322132e-01 2.85752565e-01 -5.16747355e-01
-1.07360590e+00 6.03296816e-01 1.50254235e-01 -1.08638898e-01
3.35406214e-02 -2.90109068e-01 -1.28411996e+00 -1.05069265e-01
-7.07903862e-01 3.78684640e-01 7.23336756e-01 5.48058212e-01
3.13720793e-01 1.08281851e+00 1.22863126e+00 -7.04716027e-01
-8.02157819e-01 -1.21314692e+00 -8.38526070e-01 3.55421036e-01
7.34013200e-01 -1.06357537e-01 -6.75957203e-01 1.05482869e-01] | [15.507224082946777, 6.187005519866943] |
401b8f8d-8acd-4d54-92f0-2c69240777b2 | double-matching-under-complementary | 2301.10230 | null | https://arxiv.org/abs/2301.10230v2 | https://arxiv.org/pdf/2301.10230v2.pdf | Double Matching Under Complementary Preferences | In this paper, we propose a new algorithm for addressing the problem of matching markets with complementary preferences, where agents' preferences are unknown a priori and must be learned from data. The presence of complementary preferences can lead to instability in the matching process, making this problem challenging to solve. To overcome this challenge, we formulate the problem as a bandit learning framework and propose the Multi-agent Multi-type Thompson Sampling (MMTS) algorithm. The algorithm combines the strengths of Thompson Sampling for exploration with a double matching technique to achieve a stable matching outcome. Our theoretical analysis demonstrates the effectiveness of MMTS as it is able to achieve stability at every matching step, satisfies the incentive-compatibility property, and has a sublinear Bayesian regret over time. Our approach provides a useful method for addressing complementary preferences in real-world scenarios. | ['Xiaowu Dai', 'Guang Cheng', 'Yuantong Li'] | 2023-01-24 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 1.88098233e-02 1.33358404e-01 -9.37618732e-01 -1.54094785e-01
-1.11789167e+00 -7.75012255e-01 3.30919206e-01 -2.54094880e-02
-3.85866523e-01 8.98574173e-01 -3.51077989e-02 -6.54576421e-01
-7.23692834e-01 -8.01884115e-01 -7.71312475e-01 -6.82466269e-01
4.70687635e-02 8.11298251e-01 -6.59652352e-02 6.62237331e-02
2.44302377e-01 1.89531550e-01 -1.28843522e+00 1.39623731e-01
1.17665982e+00 1.29456258e+00 1.59613296e-01 3.02278429e-01
-7.71306083e-02 5.86968541e-01 7.65582100e-02 -5.90403378e-01
9.70653951e-01 -3.22049946e-01 -7.96002924e-01 9.15485807e-03
-1.61894634e-01 -5.90200484e-01 1.42135188e-01 1.21658766e+00
3.29161376e-01 2.09171206e-01 5.57219446e-01 -1.41917801e+00
-1.66956604e-01 9.04105246e-01 -8.25624943e-01 -4.97490279e-02
2.08106697e-01 -1.14791781e-01 1.51910460e+00 -2.83794850e-01
5.62038481e-01 1.35097349e+00 3.91256303e-01 4.22276825e-01
-1.41289091e+00 -5.99236906e-01 4.86186653e-01 4.17945087e-02
-7.38947332e-01 -2.01723099e-01 7.96620905e-01 -2.87230521e-01
3.61489117e-01 4.21740979e-01 8.25246036e-01 8.45110953e-01
-9.60523710e-02 1.36374569e+00 1.45361280e+00 -6.50282919e-01
6.73935771e-01 2.32896283e-01 3.91820036e-02 2.04399988e-01
5.33565044e-01 7.59537756e-01 -5.14177322e-01 -6.63080394e-01
8.32373083e-01 5.27481511e-02 -2.07211152e-02 -1.06138706e+00
-6.42765999e-01 1.34081006e+00 2.96878546e-01 -9.74466950e-02
-6.96156740e-01 1.07386991e-01 -2.24608649e-02 7.51362443e-01
5.83662808e-01 2.53515303e-01 -4.88693744e-01 -3.19643058e-02
-7.79704809e-01 6.67245150e-01 8.48015845e-01 8.11393201e-01
5.63630760e-01 -4.18373853e-01 -1.89764798e-01 5.84975481e-01
6.27239168e-01 4.80554432e-01 2.92229533e-01 -1.24167633e+00
7.13469326e-01 3.22108179e-01 9.28732932e-01 -5.59173107e-01
-2.17677906e-01 -5.03058732e-01 -2.58782893e-01 4.52783555e-01
7.42623746e-01 -3.92389685e-01 -4.65760946e-01 2.00037622e+00
6.38455868e-01 -2.11012095e-01 1.00337870e-01 8.77814770e-01
2.61757541e-02 5.53712964e-01 -2.02506810e-01 -6.03725731e-01
1.10261071e+00 -9.85054553e-01 -6.62827671e-01 -2.76897222e-01
4.05896246e-01 -5.16733587e-01 5.79304576e-01 4.14760828e-01
-1.39042079e+00 1.33092329e-01 -8.32822800e-01 7.09236145e-01
4.60782498e-02 -5.22071481e-01 1.10547554e+00 9.68850195e-01
-7.01105833e-01 5.36463559e-01 -4.57711041e-01 -5.43306358e-02
3.75497490e-01 5.38648248e-01 3.70297760e-01 2.11156886e-02
-1.00132585e+00 1.02210546e+00 1.61182836e-01 -1.00103870e-01
-5.75870574e-01 -6.94498837e-01 -4.64684457e-01 1.61899850e-01
8.57277632e-01 -8.02175701e-01 1.83859956e+00 -1.28203666e+00
-1.72093463e+00 4.24301893e-01 -1.18200660e-01 -5.09579182e-01
1.14968467e+00 -9.60191786e-02 6.41173422e-02 -9.85770375e-02
2.33492672e-01 5.79819344e-02 6.64070666e-01 -1.25035858e+00
-1.07416654e+00 -3.83661568e-01 3.09857219e-01 3.25036108e-01
-2.76177246e-02 -4.45620827e-02 1.32431462e-02 -7.32115328e-01
5.09294383e-02 -9.84064043e-01 -6.97640777e-01 -3.76792163e-01
6.88523427e-02 -2.32993782e-01 -2.33128271e-03 -8.83925334e-02
9.44874108e-01 -1.76692033e+00 3.18833329e-02 6.48308694e-01
-2.52611399e-01 -2.53606826e-01 -1.83409810e-01 4.52058345e-01
1.74509391e-01 -2.00497508e-02 -9.60401073e-02 -1.15787588e-01
5.00313163e-01 3.78230624e-02 -4.50309157e-01 5.02387702e-01
-4.04080451e-01 9.82087076e-01 -9.46384668e-01 -1.99084640e-01
-9.81081054e-02 -5.10119796e-01 -7.14205265e-01 1.60713404e-01
-3.72867227e-01 1.51810646e-02 -7.80428410e-01 7.01931477e-01
8.18907976e-01 -1.98709965e-01 7.28005767e-01 5.19939363e-01
-9.57660899e-02 1.12161897e-01 -1.56150961e+00 1.24919391e+00
-3.01295698e-01 -2.98392940e-02 4.94542629e-01 -1.14069116e+00
5.01313448e-01 2.36738831e-01 6.60671353e-01 -7.75926709e-01
2.85475105e-01 5.38304865e-01 -2.00602293e-01 -2.96352953e-01
1.70768112e-01 -4.55050081e-01 -1.06282659e-01 1.10671854e+00
-3.04504693e-01 9.41031799e-03 1.92041740e-01 -8.22219178e-02
6.35746837e-01 1.99376568e-02 3.83184195e-01 -5.78926742e-01
-5.25660515e-02 6.41917437e-02 9.51855779e-01 1.34079325e+00
-1.00592397e-01 7.92720541e-03 5.33149242e-01 -4.41524416e-01
-7.43755162e-01 -1.02018893e+00 6.31088018e-02 9.28438663e-01
3.11822385e-01 2.74311423e-01 -5.11300564e-01 -7.84078777e-01
6.36133313e-01 5.80744267e-01 -7.53700197e-01 1.03901930e-01
-1.44523501e-01 -1.04229045e+00 -3.85876536e-01 4.74161953e-01
4.38804358e-01 -6.38708293e-01 -7.82484651e-01 4.22974497e-01
-3.04596633e-01 -5.45025945e-01 -5.30253053e-01 1.06535278e-01
-9.29831624e-01 -1.17782593e+00 -7.53508031e-01 -4.66809273e-01
5.27710438e-01 4.03645724e-01 8.82226586e-01 -3.13605338e-01
1.70779407e-01 4.57974076e-01 -1.80888131e-01 -6.12429857e-01
-1.41144559e-01 -2.20424868e-02 3.86291817e-02 3.53794158e-01
3.09946835e-01 -3.26408237e-01 -7.73523748e-01 3.38244528e-01
-7.05634832e-01 -2.11535677e-01 5.91469705e-01 1.07551885e+00
3.10680747e-01 1.28130928e-01 8.68691146e-01 -8.85940373e-01
9.74704444e-01 -5.86059213e-01 -1.49007368e+00 5.08748651e-01
-9.52516675e-01 1.00834697e-01 1.04158126e-01 -6.84693217e-01
-1.22602022e+00 2.52986550e-01 5.38015366e-01 -7.40305260e-02
4.39479053e-01 7.25448132e-01 -5.75482026e-02 -2.29745567e-01
2.29512319e-01 -1.95514590e-01 2.87523508e-01 -6.11979008e-01
3.89466524e-01 6.23441696e-01 -5.98526001e-02 -9.10323739e-01
7.40488231e-01 5.28970540e-01 -5.26345475e-03 3.61525267e-02
-7.13216364e-01 -3.78034353e-01 -1.07050218e-01 -2.23269269e-01
2.57086515e-01 -8.05593312e-01 -1.12998426e+00 1.72453508e-01
-8.16254795e-01 -4.20076907e-01 -1.88613802e-01 6.66373372e-01
-1.02757764e+00 2.63295799e-01 -3.97409409e-01 -1.59375811e+00
-1.28112555e-01 -1.06401026e+00 2.80782938e-01 3.49674076e-01
-2.41592735e-01 -9.41325188e-01 2.59614646e-01 5.27241230e-01
3.03658932e-01 2.52002705e-04 7.38922596e-01 -5.55742681e-01
-1.02388096e+00 -6.79853857e-02 2.33881716e-02 -4.86255467e-01
-1.71071652e-03 -4.03140157e-01 -6.17554367e-01 -6.85374737e-01
8.63535181e-02 -4.67233956e-01 9.60662842e-01 8.54836226e-01
5.48339128e-01 -5.42834103e-01 -3.51142913e-01 1.71806663e-01
1.37031996e+00 5.42887330e-01 -2.03833207e-02 9.09664452e-01
-4.31439221e-01 8.19061279e-01 1.00203347e+00 8.40829611e-01
4.57171351e-01 7.64764249e-01 5.25510550e-01 6.12604432e-02
8.18812191e-01 -4.12517816e-01 1.21393621e-01 -6.45786449e-02
5.58640808e-02 -1.88353229e-02 -4.02058512e-01 6.84527099e-01
-2.61026263e+00 -1.08793306e+00 3.09596449e-01 2.64127350e+00
7.80955315e-01 1.00533934e-02 7.17784286e-01 -7.50007853e-02
6.52323902e-01 -1.98874578e-01 -8.10788691e-01 -5.39299250e-01
-4.91585620e-02 -8.98692459e-02 7.48257875e-01 6.64326906e-01
-1.00450766e+00 6.68164074e-01 7.48022604e+00 7.99068749e-01
-5.57699144e-01 2.12927938e-01 7.65128613e-01 -3.79016280e-01
-7.79792666e-01 2.48706773e-01 -5.52501798e-01 4.91692036e-01
3.40998024e-01 -4.83293474e-01 7.52538264e-01 7.66982913e-01
1.76481113e-01 -3.18416446e-01 -1.02681136e+00 7.41781175e-01
-5.42752743e-01 -1.41542912e+00 -3.31044376e-01 3.31123441e-01
1.05516016e+00 -1.76139176e-01 2.22902879e-01 9.65982080e-02
1.22767019e+00 -6.91202521e-01 8.56351554e-01 1.69897318e-01
3.54039788e-01 -1.05277741e+00 5.16736150e-01 5.19066155e-01
-8.62822592e-01 -7.14282036e-01 -2.92229056e-01 -3.06319237e-01
3.21903795e-01 6.02314949e-01 -3.14043969e-01 7.55218148e-01
5.35454273e-01 1.85585260e-01 2.97213823e-01 1.46080458e+00
-2.51360051e-02 5.10079600e-02 -6.11693442e-01 -3.28916371e-01
4.86788094e-01 -6.90167725e-01 3.53342891e-01 4.68281120e-01
3.99793863e-01 2.15879008e-01 3.37337554e-01 8.56012762e-01
-8.83199945e-02 1.37121037e-01 -4.12311107e-01 1.51459470e-01
6.68816507e-01 7.54776359e-01 -6.76647663e-01 -2.44069040e-01
-3.75790358e-01 5.65742016e-01 2.49633417e-01 5.00399232e-01
-6.08439028e-01 1.28788322e-01 6.64889336e-01 -3.52112383e-01
4.28981692e-01 3.02038431e-01 -4.92432594e-01 -1.25468123e+00
2.58594781e-01 -8.86143327e-01 9.86685336e-01 -7.70532340e-02
-1.22579420e+00 -3.22279893e-02 8.02953467e-02 -1.11171758e+00
-5.13006330e-01 -3.51177871e-01 -5.39165020e-01 6.12609446e-01
-1.76816976e+00 -8.18573356e-01 4.72985327e-01 5.30471921e-01
4.36425716e-01 -6.62607178e-02 4.99652445e-01 -4.47644927e-02
-4.12801802e-01 5.07346690e-01 6.76877201e-01 -4.55742896e-01
2.94053376e-01 -1.41195333e+00 -1.24759693e-02 5.96841156e-01
-7.92585835e-02 4.30028051e-01 6.90359771e-01 -6.66606724e-01
-1.55710340e+00 -5.46189606e-01 6.26990259e-01 -1.60093829e-01
8.33565116e-01 -5.73421344e-02 -7.16937557e-02 6.63230896e-01
2.46242195e-01 -5.43324232e-01 7.54905879e-01 4.86466974e-01
-2.71834046e-01 -4.35517371e-01 -1.42427266e+00 6.05409026e-01
7.65049040e-01 -2.15285450e-01 -6.13550365e-01 2.85953254e-01
2.81181991e-01 -1.67287678e-01 -6.24876618e-01 3.42352390e-01
8.52696598e-01 -9.54245150e-01 9.12901282e-01 -6.49370849e-01
-1.22493178e-01 7.26165175e-02 -1.05623223e-01 -1.46648622e+00
-4.06043738e-01 -1.18435192e+00 3.25648226e-02 1.00834572e+00
6.89234912e-01 -1.11413181e+00 9.31614518e-01 1.11380446e+00
6.78627908e-01 -4.91241634e-01 -1.32954061e+00 -1.05491984e+00
4.31840092e-01 -2.04167798e-01 8.76948714e-01 8.56262863e-01
5.73257267e-01 -1.67042583e-01 -7.02758372e-01 -1.29087716e-01
1.28193116e+00 1.03430164e+00 4.87551302e-01 -1.41991198e+00
-5.74195862e-01 -7.80095637e-01 5.85983455e-01 -1.20584416e+00
1.25348628e-01 -4.87410039e-01 -2.62683213e-01 -1.18578601e+00
6.67611718e-01 -8.41098666e-01 -3.92515630e-01 7.04175457e-02
6.79086298e-02 -1.35359287e-01 4.50296015e-01 9.90594625e-02
-5.10152996e-01 4.23066229e-01 9.77854729e-01 -3.37992311e-01
-2.74307102e-01 7.30108440e-01 -1.03871536e+00 3.72545570e-01
8.68270516e-01 -3.94559562e-01 -3.53386849e-01 -2.52894521e-01
5.07604301e-01 5.80638230e-01 -1.84486397e-02 4.64518368e-02
3.45206946e-01 -7.00127661e-01 9.60260034e-02 -6.58465803e-01
2.13240951e-01 -9.96892154e-01 4.43731457e-01 8.83383512e-01
-4.57885087e-01 -5.75626269e-03 -2.20268250e-01 7.53104687e-01
7.18090758e-02 -5.04486859e-01 6.83065414e-01 -4.15907204e-01
-1.11607209e-01 2.94216782e-01 -6.08927548e-01 -1.63833529e-01
1.10645890e+00 -3.82426046e-02 4.36342843e-02 -7.39839852e-01
-5.72390795e-01 8.69964063e-01 5.25553882e-01 2.75093406e-01
2.82814503e-01 -1.53401208e+00 -5.45948088e-01 2.84195580e-02
-1.08859293e-01 -4.81137693e-01 9.36408788e-02 7.21931279e-01
9.28923562e-02 4.66561466e-01 -3.44362035e-02 -1.42901689e-01
-1.03460312e+00 7.74997294e-01 4.07098413e-01 -5.46258330e-01
-8.21613297e-02 6.41023815e-01 5.37903123e-02 -3.16713601e-01
3.88581574e-01 -1.83575660e-01 5.19538037e-02 4.34889466e-01
4.88495916e-01 4.43158239e-01 -2.73300081e-01 1.47279343e-02
-5.40901273e-02 5.36343873e-01 -7.43702799e-02 -5.52227974e-01
1.37878692e+00 -4.11471874e-01 -1.46653391e-02 1.63624212e-01
5.15595615e-01 -8.49425942e-02 -1.30283773e+00 -5.75896859e-01
4.92122501e-01 -9.14180160e-01 1.03490002e-01 -8.84121060e-01
-1.06083572e+00 1.49501026e-01 3.03883851e-01 6.32611275e-01
7.91758358e-01 -1.64075479e-01 4.70738739e-01 1.89977869e-01
7.66136110e-01 -1.52103126e+00 -2.12077364e-01 -3.38899307e-02
4.90301996e-01 -1.37215078e+00 -1.82547256e-01 -2.11840108e-01
-6.15545630e-01 8.08063388e-01 1.48087144e-01 6.20684922e-02
6.97865009e-01 1.06464038e-02 8.74919742e-02 2.34581411e-01
-9.43452597e-01 -4.14199740e-01 -6.11191895e-03 3.33306611e-01
-2.72437543e-01 3.72959286e-01 -6.72829986e-01 7.08058655e-01
7.34234974e-02 -6.26074970e-02 4.09329832e-01 9.60555732e-01
-4.14790332e-01 -1.53563702e+00 -6.20187402e-01 5.37026942e-01
-5.80852270e-01 2.12820858e-01 -4.70943093e-01 5.76802909e-01
-6.05877161e-01 1.19758844e+00 -1.32656366e-01 2.34838963e-01
2.21559361e-01 -1.92709938e-01 6.92288756e-01 7.31607601e-02
-4.48925257e-01 7.32097745e-01 1.58930108e-01 -5.70380449e-01
-7.08804548e-01 -1.04984891e+00 -5.15016139e-01 -4.23389286e-01
-6.03905439e-01 5.98471940e-01 2.52444476e-01 9.35310006e-01
1.54009447e-01 -2.09409803e-01 1.13549471e+00 -5.61859846e-01
-1.34305298e+00 -5.47760546e-01 -9.59419429e-01 3.31042290e-01
4.78558093e-01 -9.53763127e-01 -4.69954163e-01 -7.87214875e-01] | [4.47898530960083, 3.245450019836426] |
70545602-f99d-44b2-ad68-179d8981017a | temporal-hallucinating-for-action-recognition | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Wang_Temporal_Hallucinating_for_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Wang_Temporal_Hallucinating_for_CVPR_2018_paper.pdf | Temporal Hallucinating for Action Recognition With Few Still Images | Action recognition in still images has been recently promoted by deep learning. However, the success of these deep models heavily depends on huge amount of training images for various action categories, which may not be available in practice. Alternatively, humans can classify new action categories after seeing few images, since we may not only compare appearance similarities between images on hand, but also attempt to recall importance motion cues from relevant action videos in our memory. To mimic this capacity, we propose a novel Hybrid Video Memory (HVM) machine, which can hallucinate temporal features of still images from video memory, in order to boost action recognition with few still images. First, we design a temporal memory module consisting of temporal hallucinating and predicting. Temporal hallucinating can generate temporal features of still images in an unsupervised manner. Hence, it can be flexibly used in realistic scenarios, where image and video categories may not be consistent. Temporal predicting can effectively infer action categories for query image, by integrating temporal features of training images and videos within a domain-adaptation manner. Second, we design a spatial memory module for spatial predicting. As spatial and temporal features are complementary to represent different actions, we apply spatial-temporal prediction fusion to further boost performance. Finally, we design a video selection module to select strongly-relevant videos as memory. In this case, we can balance the number of images and videos to reduce prediction bias as well as preserve computation efficiency. To show the effectiveness, we conduct extensive experiments on three challenging data sets, where our HVM outperforms a number of recent approaches by temporal hallucinating from video memory. | ['Yu Qiao', 'Yali Wang', 'Lei Zhou'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['action-recognition-in-still-images'] | ['computer-vision'] | [ 2.60649204e-01 -3.36059034e-01 -3.43712032e-01 -3.13128531e-01
-3.99327546e-01 -3.01266789e-01 6.34701431e-01 -2.81173855e-01
-4.20199066e-01 5.56173503e-01 3.15753520e-01 2.26057425e-01
6.64226487e-02 -7.97066569e-01 -8.61373603e-01 -6.97025478e-01
-2.97820829e-02 -1.03427611e-01 7.14088082e-01 2.79512286e-01
3.94702762e-01 2.73630500e-01 -1.70708096e+00 7.92932332e-01
5.95713198e-01 1.21602130e+00 7.19673038e-01 3.32191378e-01
7.58503526e-02 1.26748037e+00 -2.90321976e-01 6.63586855e-02
8.57733712e-02 -5.15114546e-01 -7.60955989e-01 5.73017359e-01
2.53043711e-01 -7.63734162e-01 -8.98174465e-01 8.60022962e-01
2.82471418e-01 6.04739606e-01 4.37534899e-01 -1.19730771e+00
-6.33755624e-01 2.08595872e-01 -4.73875791e-01 3.71321827e-01
5.37684083e-01 5.72029352e-01 5.49846828e-01 -9.48677897e-01
7.35130847e-01 1.33056450e+00 1.17868714e-01 5.88216603e-01
-8.68369222e-01 -7.12145329e-01 4.13924754e-01 9.94800091e-01
-1.39640021e+00 -5.49601555e-01 7.22689331e-01 -3.25610399e-01
8.63652527e-01 1.76584437e-01 8.36206973e-01 1.03815830e+00
2.04201564e-01 1.26012516e+00 8.65652025e-01 5.57835288e-02
2.26734087e-01 -9.99125838e-02 -2.90010124e-01 5.93672395e-01
-4.50527370e-01 1.48945868e-01 -8.40622067e-01 1.54767886e-01
8.56371224e-01 6.91221774e-01 -5.92831135e-01 -4.32695776e-01
-1.54664457e+00 5.93834162e-01 5.52132845e-01 3.78818870e-01
-6.14687741e-01 1.34938449e-01 2.81157434e-01 3.17233235e-01
1.58420965e-01 1.04230173e-01 -2.89983362e-01 -1.67539611e-01
-8.74364018e-01 -1.60209518e-02 1.59049720e-01 8.82841647e-01
8.16629112e-01 -1.36912510e-01 -4.33949322e-01 8.15499365e-01
2.22789973e-01 4.58663762e-01 9.39301908e-01 -1.27122569e+00
4.01427269e-01 6.21248484e-01 9.79811177e-02 -1.23585510e+00
-1.57907948e-01 2.06362307e-01 -1.02554119e+00 -2.82098770e-01
6.88124597e-02 3.40401649e-01 -1.00249076e+00 1.59885561e+00
1.34373233e-01 4.69263226e-01 -7.41890259e-03 1.19114292e+00
5.98064303e-01 8.70892823e-01 1.45351872e-01 -6.79084897e-01
1.04958296e+00 -1.08943009e+00 -6.25172377e-01 -3.45876396e-01
5.40999115e-01 -6.28100753e-01 9.19919550e-01 3.17787588e-01
-1.11914980e+00 -9.78404701e-01 -7.53779888e-01 7.78784007e-02
-6.98481500e-02 9.55107361e-02 4.26821679e-01 -9.75877494e-02
-8.19731891e-01 7.54399776e-01 -9.05036390e-01 -4.44361180e-01
4.90974367e-01 2.91023225e-01 -5.28248608e-01 -5.47425926e-01
-1.18323994e+00 5.66619515e-01 6.69252813e-01 1.89372003e-01
-1.12702858e+00 -2.08618641e-01 -8.42296004e-01 -9.95154306e-02
4.46243823e-01 -5.44570327e-01 1.02467799e+00 -1.28258395e+00
-1.24020171e+00 4.72808421e-01 -3.88487846e-01 -4.12411571e-01
3.21539611e-01 -1.19793132e-01 -6.23686969e-01 6.24347389e-01
8.46163332e-02 1.05624008e+00 1.03852236e+00 -9.09660757e-01
-7.65910327e-01 -2.60813862e-01 1.11926638e-01 2.22383797e-01
-5.62514365e-01 -2.45788381e-01 -7.04567552e-01 -6.11939132e-01
3.06172162e-01 -9.30052936e-01 -2.75815666e-01 1.20943032e-01
6.25395682e-03 -7.24066496e-02 1.01389134e+00 -6.45815551e-01
1.20439470e+00 -2.40260625e+00 2.44950637e-01 1.62087975e-03
1.36385085e-02 2.77826607e-01 -2.91812211e-01 2.81680018e-01
1.27648428e-01 -8.43549594e-02 -4.21782136e-02 1.52215054e-02
-3.10505390e-01 3.71915668e-01 -4.39858168e-01 2.38089368e-01
1.68291003e-01 9.71359253e-01 -9.39989567e-01 -7.13293672e-01
3.46676946e-01 3.11088860e-01 -5.43392658e-01 4.61271763e-01
-2.44432896e-01 6.00107610e-01 -5.38740337e-01 6.30423784e-01
6.26815200e-01 -4.73713040e-01 1.95717663e-01 -3.84029746e-01
-4.21080645e-03 -4.59826626e-02 -1.07099831e+00 1.80731928e+00
-3.95063639e-01 4.94810581e-01 -3.94500345e-01 -1.23880076e+00
7.08565891e-01 3.00174773e-01 5.10262132e-01 -1.07603967e+00
-1.66418150e-01 5.53951487e-02 -1.98673993e-01 -8.14146042e-01
3.29402566e-01 -6.78525269e-02 2.76375651e-01 4.33692008e-01
8.25735629e-02 2.84423709e-01 1.86271191e-01 1.84195310e-01
1.16846967e+00 2.42199808e-01 2.24528238e-01 3.13891947e-01
6.23068392e-01 5.18411724e-03 7.98685670e-01 5.02308786e-01
-3.57442588e-01 7.31386662e-01 9.40660536e-02 -6.14990890e-01
-6.92063630e-01 -1.02920306e+00 1.56807169e-01 1.07219315e+00
6.62599206e-01 -3.76882792e-01 -2.04716668e-01 -7.17444718e-01
-2.83118039e-01 4.46744323e-01 -5.53073764e-01 -4.78759348e-01
-7.24746585e-01 -4.04319525e-01 -1.98089853e-02 8.22074831e-01
9.19356644e-01 -1.45076454e+00 -6.78492129e-01 2.37297207e-01
-4.59875673e-01 -1.06550765e+00 -6.86910510e-01 -1.75987631e-01
-9.91805434e-01 -1.10496473e+00 -8.16440225e-01 -8.70063365e-01
5.61034203e-01 8.77657831e-01 7.70297468e-01 1.68788597e-01
-2.07639043e-03 5.15692532e-01 -6.40034258e-01 3.43335837e-01
-4.01876606e-02 -4.40245390e-01 1.15935080e-01 4.04668659e-01
4.43411499e-01 -6.55160367e-01 -9.68106627e-01 6.29707873e-01
-1.23622477e+00 3.48773122e-01 9.68575120e-01 9.24707472e-01
6.46843672e-01 1.01814054e-01 3.42144072e-01 -3.72607470e-01
-2.04235986e-02 -5.45166373e-01 -8.20804685e-02 3.73215765e-01
-1.52958438e-01 -9.86725911e-02 6.44569576e-01 -8.34651232e-01
-1.11853004e+00 2.97991991e-01 2.40537301e-01 -9.88039017e-01
-1.00702681e-01 4.97777104e-01 -2.68656939e-01 1.04523368e-01
2.50173092e-01 8.66940677e-01 8.27071965e-02 -2.73274392e-01
2.46482253e-01 5.58671474e-01 3.25143814e-01 -1.95276052e-01
4.49381322e-01 4.95426744e-01 -2.19804004e-01 -5.60624003e-01
-6.74721777e-01 -4.95518148e-01 -8.48434746e-01 -3.76308024e-01
9.49541211e-01 -1.00875819e+00 -5.43484449e-01 4.09035444e-01
-1.08843064e+00 -3.04213703e-01 6.18932471e-02 7.16317594e-01
-6.81121826e-01 5.43656230e-01 -6.66515231e-01 -4.67925251e-01
1.70969516e-01 -1.13201630e+00 9.66494203e-01 2.70796359e-01
-5.43923005e-02 -6.58450961e-01 -2.91623175e-01 3.91523778e-01
2.08814994e-01 -1.25224605e-01 6.32751584e-01 -3.19453269e-01
-1.13946688e+00 6.60985336e-02 -3.55602443e-01 3.06447119e-01
6.72844350e-02 -2.81826049e-01 -7.14351952e-01 -3.29414278e-01
6.16479991e-03 -5.33652902e-01 1.11675286e+00 2.82799184e-01
1.44238448e+00 -4.19239074e-01 -5.41070342e-01 3.15341890e-01
1.02673995e+00 5.70592642e-01 7.84562290e-01 2.95305461e-01
6.81655765e-01 6.71195924e-01 1.15271401e+00 4.79822606e-01
2.65159547e-01 8.40820670e-01 2.71455020e-01 2.17784420e-01
-1.66175179e-02 -3.39558035e-01 6.76099122e-01 9.14400995e-01
-1.78391159e-01 -7.19502270e-02 -6.20077133e-01 5.02732217e-01
-2.21070981e+00 -1.36942828e+00 3.33359510e-01 2.16733623e+00
5.86275280e-01 -9.24443901e-02 5.20970561e-02 -6.82348013e-02
7.08789170e-01 3.85245383e-01 -7.22548544e-01 2.88992912e-01
1.61707420e-02 -2.45072901e-01 2.20495183e-02 -3.02669443e-02
-1.13378751e+00 7.63409495e-01 5.28189754e+00 9.92868125e-01
-1.22151411e+00 1.41261831e-01 6.95080280e-01 -3.44787776e-01
-1.51920035e-01 9.99413803e-02 -3.42609435e-01 6.34366512e-01
6.29895031e-01 -1.84434518e-01 2.92483985e-01 8.48866522e-01
2.63564408e-01 -1.14032514e-01 -1.14896786e+00 1.20433795e+00
2.38335475e-01 -1.22539508e+00 3.67124200e-01 -1.16695967e-02
6.10013545e-01 -2.58764595e-01 7.08340555e-02 6.02639675e-01
-1.74906343e-01 -9.78054762e-01 5.46229303e-01 8.25663507e-01
4.34457541e-01 -5.58225632e-01 5.03150880e-01 4.96454000e-01
-1.35846376e+00 -4.92785037e-01 -6.78292811e-01 -9.50565115e-02
2.26006076e-01 2.54470348e-01 -3.84650767e-01 4.05620694e-01
8.52783203e-01 1.27514386e+00 -6.73985124e-01 9.96638179e-01
1.17924817e-01 2.75372595e-01 1.57709625e-02 1.48878559e-01
3.01355034e-01 -4.01481129e-02 1.30017042e-01 8.06444764e-01
5.46674430e-01 6.36794925e-01 4.86189246e-01 5.19833744e-01
1.31116018e-01 1.03447564e-01 -6.37517095e-01 -1.21369071e-01
4.88290757e-01 9.73735690e-01 -5.00576079e-01 -5.82585573e-01
-7.01634467e-01 1.35274541e+00 2.03370959e-01 4.74972516e-01
-8.47577751e-01 7.68883973e-02 3.42037499e-01 1.85462236e-01
3.95083338e-01 -2.05617458e-01 4.79243666e-01 -1.40401661e+00
1.39027804e-01 -8.31270814e-01 5.31010270e-01 -1.07188380e+00
-9.57827866e-01 4.94801372e-01 -3.78882363e-02 -1.92384171e+00
-2.86197305e-01 -3.02193612e-01 -4.48098660e-01 2.75774986e-01
-1.22089529e+00 -1.09010744e+00 -5.18245041e-01 8.73127699e-01
9.13655341e-01 -2.29368493e-01 4.26653504e-01 3.71101707e-01
-4.12745148e-01 1.60606593e-01 8.80299893e-04 -3.76740992e-02
8.70803237e-01 -6.73965216e-01 -5.74735254e-02 7.89903939e-01
3.31790864e-01 5.79647064e-01 2.56111324e-01 -5.94456971e-01
-1.50241721e+00 -1.43925369e+00 5.54655373e-01 -1.81551993e-01
4.41492707e-01 1.32660180e-01 -1.20978796e+00 5.74179769e-01
9.26075205e-02 2.60841489e-01 5.72508931e-01 -3.36721539e-01
-1.37276396e-01 -1.77321240e-01 -6.52951777e-01 6.58178568e-01
1.29857683e+00 -6.03840649e-01 -5.96049488e-01 4.39578563e-01
6.32829130e-01 1.06073068e-02 -7.65803039e-01 6.39397562e-01
6.01114154e-01 -1.22024763e+00 9.43776071e-01 -4.11716193e-01
6.10239565e-01 -5.11214137e-01 -3.42342675e-01 -1.09171736e+00
-4.43300217e-01 -1.77537113e-01 -3.30666602e-01 1.17283940e+00
-6.55208379e-02 -1.88084424e-01 7.08020747e-01 5.02968967e-01
-5.42443097e-02 -7.82975376e-01 -8.11961532e-01 -8.94835651e-01
-3.17924559e-01 -2.97588617e-01 4.37588781e-01 8.10155630e-01
-1.56180735e-03 6.75034896e-02 -8.06364715e-01 3.75017114e-02
1.86948746e-01 5.92956901e-01 7.22903013e-01 -6.68676078e-01
-5.38627982e-01 -3.37927818e-01 -6.46897972e-01 -1.49197447e+00
1.72802612e-01 -5.94795346e-01 1.19960822e-01 -1.26898777e+00
5.87123096e-01 4.95636426e-02 -6.12996936e-01 5.27085602e-01
-2.30986774e-01 4.09199595e-01 3.58952314e-01 7.49094605e-01
-1.12434614e+00 9.20765817e-01 1.49328065e+00 -3.81046206e-01
-1.58643872e-01 -2.14613095e-01 -2.08744124e-01 7.35914111e-01
4.38029081e-01 -1.92139938e-01 -6.14960611e-01 -3.64966184e-01
-3.63222003e-01 5.47631800e-01 5.90538442e-01 -1.22009206e+00
3.37727576e-01 -5.44157326e-01 8.28844726e-01 -5.69739163e-01
5.14138103e-01 -6.82763040e-01 2.25310370e-01 4.78269279e-01
-3.31059009e-01 -8.25874954e-02 -1.39962941e-01 8.67879450e-01
-7.13154972e-01 1.22561231e-01 7.38494754e-01 -3.95879954e-01
-1.58480227e+00 6.67109191e-01 -4.08131093e-01 -3.36927772e-01
1.32290614e+00 -3.75371635e-01 -2.65714645e-01 -4.82313901e-01
-7.99152851e-01 3.87718379e-01 5.78961849e-01 6.46066308e-01
1.09153569e+00 -1.69281423e+00 -2.24105641e-01 2.95148402e-01
3.57408404e-01 -3.15334916e-01 9.13140178e-01 1.09342349e+00
-2.19002232e-01 3.38723183e-01 -4.87919927e-01 -7.15945363e-01
-1.24582529e+00 1.17326188e+00 2.15291250e-02 1.85836613e-01
-6.18514895e-01 6.35622263e-01 7.18125224e-01 4.71402816e-02
1.27028480e-01 -6.57380596e-02 -3.62556040e-01 1.79245044e-02
8.77723813e-01 2.77933627e-02 -4.23915684e-01 -7.42471039e-01
-3.83544773e-01 5.57405412e-01 -1.83363140e-01 9.55909118e-03
1.06249917e+00 -4.34508801e-01 9.28060114e-02 4.74847674e-01
1.26787817e+00 -4.40014333e-01 -1.51727855e+00 -4.29641038e-01
-1.59019828e-01 -8.99710298e-01 -1.06930308e-01 -4.32757020e-01
-1.23779047e+00 1.11572373e+00 6.94693625e-01 -1.20765492e-01
1.62208569e+00 -1.32547960e-01 1.08210838e+00 4.39892173e-01
4.46917087e-01 -1.12853372e+00 7.09145010e-01 1.97019234e-01
8.23079169e-01 -1.44927573e+00 -6.79787546e-02 -2.22367108e-01
-1.01562178e+00 1.13659298e+00 9.57423389e-01 5.59534021e-02
6.03003561e-01 -4.38179255e-01 -8.18593055e-02 3.73234041e-02
-1.06019866e+00 -2.70128101e-01 3.74022007e-01 4.91252005e-01
1.47105142e-01 -1.55090138e-01 -1.87837496e-01 6.08804166e-01
4.79200542e-01 1.37372866e-01 2.75783807e-01 9.03752744e-01
-6.01431966e-01 -1.04770744e+00 -1.86648741e-01 4.38304216e-01
-6.56095967e-02 6.93419203e-02 -1.97757706e-01 5.26275635e-01
1.93466783e-01 9.04771626e-01 1.16039649e-01 -8.27539027e-01
1.11068591e-01 -1.06101818e-01 4.56956059e-01 -4.64035720e-01
6.65995479e-02 1.92204505e-01 -3.44991297e-01 -9.46825325e-01
-7.99633265e-01 -6.71289146e-01 -1.16618443e+00 -2.12141171e-01
1.28394703e-03 1.22155279e-01 -1.25387251e-01 1.11308706e+00
4.65283096e-01 2.59251386e-01 7.32485294e-01 -9.93984222e-01
-2.66180843e-01 -9.19390976e-01 -7.47693002e-01 6.69377446e-01
1.12327658e-01 -9.63786364e-01 -1.35579839e-01 4.22084123e-01] | [8.637910842895508, 0.5901516079902649] |
03082a8c-55ed-4c57-9f43-ae96c572c5fa | unsupervised-joint-training-of-bilingual-word | null | null | https://aclanthology.org/P19-1312 | https://aclanthology.org/P19-1312.pdf | Unsupervised Joint Training of Bilingual Word Embeddings | State-of-the-art methods for unsupervised bilingual word embeddings (BWE) train a mapping function that maps pre-trained monolingual word embeddings into a bilingual space. Despite its remarkable results, unsupervised mapping is also well-known to be limited by the original dissimilarity between the word embedding spaces to be mapped. In this work, we propose a new approach that trains unsupervised BWE jointly on synthetic parallel data generated through unsupervised machine translation. We demonstrate that existing algorithms that jointly train BWE are very robust to noisy training data and show that unsupervised BWE jointly trained significantly outperform unsupervised mapped BWE in several cross-lingual NLP tasks. | ['Atsushi Fujita', 'Benjamin Marie'] | 2019-07-01 | null | null | null | acl-2019-7 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [-1.04743056e-01 -2.12273989e-02 -5.83250999e-01 -4.32308614e-01
-9.41112280e-01 -6.48577511e-01 9.29806113e-01 3.26105654e-02
-8.21071446e-01 8.86268735e-01 6.37870967e-01 -6.44954801e-01
2.78562486e-01 -6.50393307e-01 -8.35927248e-01 -3.74986112e-01
1.83365479e-01 1.00521243e+00 -2.09840983e-01 -4.97138798e-01
-1.10775620e-01 2.93045282e-01 -8.45860481e-01 -2.07833812e-01
9.44845319e-01 -1.92976356e-01 1.00342333e-01 5.44075429e-01
-3.38822037e-01 4.87757097e-06 -1.91324219e-01 -4.57701236e-01
3.22677702e-01 -6.73068464e-01 -7.51274049e-01 -1.53482571e-01
2.99825013e-01 5.04846051e-02 -5.81982672e-01 1.23132145e+00
5.14369428e-01 -3.76317464e-02 6.48743987e-01 -1.04481816e+00
-1.41761816e+00 8.17368925e-01 -1.36555165e-01 4.34305608e-01
2.45763466e-01 4.41883355e-02 1.43828857e+00 -1.49016082e+00
1.00324476e+00 1.05141282e+00 6.86691403e-01 4.48986828e-01
-1.97254491e+00 -5.67932367e-01 -2.44090274e-01 3.33919793e-01
-1.49348974e+00 -4.63843554e-01 4.87412840e-01 -5.12799740e-01
1.41517603e+00 -7.49469772e-02 5.28207064e-01 1.47391069e+00
1.94191068e-01 4.59427506e-01 1.46003354e+00 -9.19216752e-01
-1.59320489e-01 4.42604363e-01 4.26591709e-02 4.43946451e-01
2.80839771e-01 4.97146368e-01 -7.04132378e-01 1.15645481e-02
6.20985687e-01 -3.65818143e-01 -9.37561914e-02 -5.96117496e-01
-1.81167364e+00 1.11881161e+00 3.16478163e-01 6.98051512e-01
-8.68401751e-02 7.28986934e-02 4.06296581e-01 7.47312069e-01
6.54260397e-01 9.12289619e-01 -5.69387853e-01 -1.87728003e-01
-8.15182388e-01 -1.37292027e-01 5.72772384e-01 1.02088332e+00
1.02942574e+00 5.03068008e-02 1.14845909e-01 9.99612153e-01
2.20867097e-01 5.30378401e-01 1.06340134e+00 -7.10132122e-02
3.44887793e-01 1.33193687e-01 -3.77914459e-02 -6.67394042e-01
-2.20727082e-02 -1.94972932e-01 -3.80345285e-01 -2.88322389e-01
1.30912721e-01 1.28544867e-01 -7.49694645e-01 1.88637936e+00
4.56062437e-04 6.29632100e-02 5.07289648e-01 9.83872414e-01
6.14490747e-01 7.77845204e-01 -1.28774475e-02 -2.53840163e-03
1.05017900e+00 -1.26360178e+00 -1.06882048e+00 -3.70178252e-01
6.84423804e-01 -1.09244955e+00 1.27489984e+00 -1.24035813e-01
-7.41050065e-01 -7.18996227e-01 -1.33119726e+00 -2.07035810e-01
-8.18037987e-01 2.34980863e-02 5.86332202e-01 6.57701313e-01
-1.01330614e+00 3.14166456e-01 -8.34199846e-01 -8.86464000e-01
-3.55174835e-03 3.90142560e-01 -9.44174170e-01 -1.44990161e-01
-1.43834138e+00 1.45039058e+00 6.73829138e-01 -6.94581419e-02
-8.39509785e-01 -6.52224183e-01 -1.17658055e+00 -2.29742602e-01
-1.18749313e-01 -2.77238786e-01 9.14036751e-01 -8.47453713e-01
-1.57608175e+00 1.13561571e+00 -1.85091391e-01 -3.83656591e-01
1.98482975e-01 -2.44387090e-01 -7.67580092e-01 -2.16579586e-01
3.54083627e-01 7.96774566e-01 5.76431274e-01 -1.17906642e+00
-1.47167519e-01 -3.65524068e-02 -3.36963028e-01 9.92595404e-02
-7.66929686e-01 1.94406167e-01 -1.37167484e-01 -1.01447964e+00
3.70806903e-02 -7.64246941e-01 -3.01911086e-01 -3.82723331e-01
-2.99631774e-01 -2.79049188e-01 5.41704237e-01 -6.39338374e-01
9.77034688e-01 -1.96129072e+00 6.55374765e-01 1.42909866e-02
-9.82389748e-02 1.97824955e-01 -5.81351042e-01 9.42887604e-01
-2.89737046e-01 1.40888512e-01 -4.29002881e-01 -2.76134998e-01
3.46645743e-01 8.19701314e-01 -5.16521215e-01 6.82417393e-01
4.70537096e-01 1.14668655e+00 -1.34628272e+00 -4.50926453e-01
2.49111265e-01 3.10834706e-01 -4.88093436e-01 4.39961106e-01
2.30663210e-01 5.64929187e-01 1.31612092e-01 3.90592903e-01
5.43631077e-01 2.31937617e-01 7.62995660e-01 -1.65145293e-01
-2.16991514e-01 5.85665107e-01 -6.53929293e-01 2.25371480e+00
-7.11234450e-01 1.03534329e+00 -5.19336164e-01 -1.45939577e+00
9.73085523e-01 4.76100594e-01 1.77919760e-01 -6.81237400e-01
-2.44196393e-02 7.96394408e-01 5.32265641e-02 -4.99374688e-01
6.40315354e-01 -6.54507339e-01 -3.21153373e-01 7.66026855e-01
9.90742922e-01 -3.61578792e-01 3.14685255e-01 -5.96577581e-03
6.91575885e-01 3.53721201e-01 3.85361403e-01 -7.52695799e-01
4.24693346e-01 1.82111323e-01 5.26101708e-01 3.71924192e-01
-1.29973479e-02 5.55101812e-01 1.13085940e-01 -6.09411120e-01
-1.43639636e+00 -1.62867939e+00 -4.45092052e-01 1.12897444e+00
1.11628145e-01 -5.28417289e-01 -5.12517452e-01 -9.66901958e-01
-1.15219690e-01 7.64840424e-01 -6.35121346e-01 -5.62678389e-02
-7.72274792e-01 -8.63251925e-01 7.13246226e-01 5.99938691e-01
-1.75749153e-01 -9.77950692e-01 -4.18349765e-02 4.39398378e-01
-1.03553444e-01 -1.38929129e+00 -8.41874301e-01 3.94765854e-01
-7.39147246e-01 -8.26892197e-01 -6.31435096e-01 -1.44670212e+00
1.00664902e+00 -7.13181682e-03 1.35202360e+00 -2.69542307e-01
-3.25309545e-01 1.17770128e-01 -5.14909804e-01 -9.94780660e-02
-7.42283106e-01 3.13400745e-01 7.96436369e-01 -1.23953201e-01
1.07268286e+00 -6.21779680e-01 2.78666187e-02 1.94961518e-01
-1.08306205e+00 -1.31751839e-02 4.78519231e-01 1.39533114e+00
7.62763679e-01 -6.58758402e-01 4.95144337e-01 -9.22610462e-01
8.48445058e-01 -3.67502749e-01 -4.41203117e-01 4.05111879e-01
-7.15435982e-01 5.87425172e-01 5.27602077e-01 -5.70321739e-01
-5.04557788e-01 -2.78128237e-01 -3.11825424e-01 -1.93830326e-01
1.51472492e-02 5.68987429e-01 7.52138644e-02 7.16721192e-02
6.70977056e-01 3.87278587e-01 -2.12230340e-01 -4.90779728e-01
9.85947609e-01 8.63580406e-01 5.09093344e-01 -7.49312282e-01
1.20936060e+00 8.88688043e-02 -5.28010249e-01 -7.25224316e-01
-5.50802290e-01 -3.98103833e-01 -1.17078710e+00 2.01188490e-01
1.06721556e+00 -9.20161843e-01 3.24055731e-01 -1.56300768e-01
-1.13299310e+00 -2.35467926e-01 -5.24441898e-01 1.03331709e+00
-4.86911923e-01 2.11921930e-01 -4.25010622e-01 -1.56822667e-01
-1.06940493e-01 -1.26390779e+00 7.57049084e-01 -3.23631763e-01
-5.54530978e-01 -1.51530218e+00 8.75429571e-01 -1.14174642e-01
4.61789787e-01 -9.62212831e-02 8.80071640e-01 -8.33005309e-01
-2.58781314e-02 -8.96785557e-02 -1.65200725e-01 5.86265147e-01
5.39575815e-01 -2.66156375e-01 -6.72506988e-01 -6.18518949e-01
-1.99599475e-01 -3.92155021e-01 6.84630334e-01 -1.72417209e-01
4.18267131e-01 -1.81419477e-01 -2.74225831e-01 8.59486878e-01
1.64038348e+00 -3.00754637e-01 3.07255566e-01 5.19343197e-01
7.23345160e-01 3.99242908e-01 3.52911204e-01 -3.93672973e-01
1.34552762e-01 8.13116729e-01 -1.32660091e-01 -3.87573004e-01
-1.28791332e-01 -5.81569552e-01 6.57352388e-01 2.11572289e+00
9.19799283e-02 3.43336046e-01 -1.00094664e+00 1.14174974e+00
-1.64549232e+00 -4.95671988e-01 1.39133692e-01 1.99563849e+00
1.26113033e+00 -6.25723833e-03 -1.91840008e-01 -1.25789776e-01
5.67430437e-01 3.02859008e-01 7.24157766e-02 -9.20760632e-01
-4.28234845e-01 1.12998652e+00 8.35686147e-01 7.92932272e-01
-8.85551214e-01 1.45154262e+00 6.90074301e+00 5.64492702e-01
-9.71615195e-01 1.02553272e+00 -1.55599281e-01 2.20442072e-01
-8.05781603e-01 9.82361510e-02 -3.78882796e-01 1.53271779e-01
9.58017945e-01 -2.68306047e-01 5.61901152e-01 3.83403987e-01
-2.36566454e-01 4.36593682e-01 -1.43312633e+00 8.83556128e-01
3.07380378e-01 -1.20056343e+00 1.67818516e-01 -4.69818991e-03
1.32601953e+00 4.33112174e-01 1.13073602e-01 4.05453891e-01
5.65057099e-01 -1.29340768e+00 4.50385422e-01 -2.64092926e-02
1.23804808e+00 -6.15227878e-01 8.53802621e-01 3.75514403e-02
-1.06265116e+00 3.95685673e-01 -6.83690548e-01 1.63535118e-01
2.55950719e-01 2.62254477e-01 -7.54383683e-01 6.93266094e-01
3.00896764e-01 8.88034463e-01 -3.59516948e-01 4.06717956e-01
-7.95995712e-01 5.48546970e-01 -9.25515294e-02 1.13761440e-01
6.19911075e-01 -6.02311134e-01 5.67207038e-01 1.58313167e+00
4.74151164e-01 -4.51500803e-01 2.04574876e-02 8.28657508e-01
-2.97512710e-01 4.90042180e-01 -8.90321314e-01 -6.68945134e-01
2.52495378e-01 9.32595313e-01 -4.52039987e-01 -2.81127989e-01
-6.12529755e-01 1.24657786e+00 6.95949912e-01 4.43232715e-01
-7.41088092e-01 -5.20966351e-01 1.08595502e+00 -1.81317315e-01
4.46981817e-01 -7.55365729e-01 -2.73067858e-02 -1.64075625e+00
-1.15842775e-01 -9.07475293e-01 -4.01674695e-02 -3.33703190e-01
-1.68495381e+00 8.57367039e-01 -2.07435876e-01 -1.34947968e+00
-1.36621609e-01 -1.11850834e+00 -4.79756057e-01 1.01335645e+00
-1.62096190e+00 -1.08137178e+00 3.88350636e-01 2.73338497e-01
7.23642826e-01 -5.79080939e-01 1.58926952e+00 3.73617530e-01
-3.65580797e-01 8.36147070e-01 4.67562139e-01 3.87334049e-01
1.06342936e+00 -1.46133280e+00 8.64110947e-01 9.53660548e-01
1.13356280e+00 9.46877360e-01 7.66376078e-01 -3.81096452e-01
-1.49555719e+00 -9.33900654e-01 1.62146115e+00 -6.26671493e-01
1.19872141e+00 -7.28085876e-01 -6.20784163e-01 1.01846290e+00
7.74833858e-01 4.04555112e-01 1.02323043e+00 2.99176216e-01
-8.28235388e-01 8.10731389e-03 -6.61162615e-01 6.25665128e-01
9.84324217e-01 -1.26920211e+00 -1.29200613e+00 6.07783318e-01
6.33522272e-01 -1.52575597e-01 -1.04312181e+00 2.87801355e-01
4.92957890e-01 -5.43495953e-01 8.92908931e-01 -1.15285695e+00
4.51078504e-01 -2.39243224e-01 -3.71875137e-01 -1.96453786e+00
-4.75966744e-02 -5.08274555e-01 2.63096690e-01 9.25676465e-01
7.51353741e-01 -8.46781909e-01 2.08948821e-01 -5.00492215e-01
-7.33061060e-02 -5.08151710e-01 -1.14116430e+00 -1.11654055e+00
5.32266438e-01 -2.85928160e-01 5.66536844e-01 1.46436691e+00
4.09640193e-01 4.12182420e-01 -3.93456727e-01 4.11535874e-02
5.20841777e-01 -3.56144793e-02 6.36024177e-01 -6.75116718e-01
-2.49893591e-01 -4.04202461e-01 -5.90711653e-01 -1.09515154e+00
5.70088506e-01 -1.50290966e+00 3.04633796e-01 -1.56204045e+00
1.05628021e-01 -2.52281785e-01 -7.18901515e-01 1.78812712e-01
-2.63897330e-01 7.44739056e-01 -2.17827663e-01 2.37315640e-01
-1.02208614e-01 6.88984215e-01 9.29044425e-01 -3.31094295e-01
3.90368029e-02 -8.80590737e-01 -1.31585643e-01 2.45616674e-01
7.95325458e-01 -8.93982530e-01 -1.84202269e-01 -1.07181323e+00
2.30874151e-01 -5.79827368e-01 -1.99625969e-01 -5.07136941e-01
-7.22535849e-02 -1.37179466e-02 -6.42030761e-02 -1.25183135e-01
9.29310266e-03 -6.81196868e-01 -3.06548886e-02 3.58712822e-01
-3.47887278e-01 4.77965474e-01 1.61011264e-01 3.02044898e-01
-5.96128523e-01 -1.50429398e-01 4.51220214e-01 6.37624934e-02
-5.96477270e-01 1.49386168e-01 -4.11470264e-01 2.32428953e-01
8.01710069e-01 6.32993355e-02 1.91922225e-02 1.81059584e-01
-7.81925082e-01 -8.40855837e-02 5.23995757e-01 8.62334430e-01
3.97250682e-01 -2.05422306e+00 -9.51134980e-01 7.01279283e-01
6.17510796e-01 -6.81126297e-01 -4.42836255e-01 9.36312020e-01
-7.35425651e-01 5.47802806e-01 -4.74765211e-01 -4.19462025e-01
-7.64604628e-01 7.25210547e-01 -5.09105623e-02 -2.91735411e-01
-5.09485543e-01 9.03036594e-01 -2.12341055e-01 -1.27307665e+00
-2.00474247e-01 -2.37142280e-01 2.15567276e-01 4.09070663e-02
1.15300454e-02 -9.98984948e-02 -1.14454711e-02 -1.18682289e+00
-3.55394036e-01 7.30747819e-01 1.13259383e-01 -5.87151825e-01
1.36749661e+00 1.10525198e-01 -3.73327583e-01 6.90881848e-01
1.50555468e+00 2.93456614e-01 -4.58816946e-01 -5.85812271e-01
2.96818674e-01 -5.13438940e-01 -1.69600964e-01 -1.76792115e-01
-6.87747538e-01 1.02419865e+00 5.01498044e-01 7.53611093e-03
5.87573230e-01 -5.00548258e-03 9.33802247e-01 3.25396657e-01
3.84590000e-01 -1.33058035e+00 -1.35320157e-01 5.80353141e-01
5.86755753e-01 -1.37026119e+00 -5.24883159e-02 -5.29242344e-02
-3.05801839e-01 1.33649504e+00 2.84512073e-01 -4.31598216e-01
6.92511261e-01 2.68104196e-01 5.67735732e-01 1.85694285e-02
-2.89088845e-01 -3.00899476e-01 4.69062358e-01 6.44808650e-01
6.11345232e-01 3.34923834e-01 -8.51264358e-01 3.49321604e-01
-4.94992793e-01 -3.25590819e-01 2.04793915e-01 6.79440439e-01
1.33014128e-01 -1.89436626e+00 -1.92642093e-01 -1.97468419e-02
-3.43318135e-01 -4.65470701e-01 -2.57068783e-01 8.98793340e-01
2.18068421e-01 5.37155867e-01 1.34249479e-01 -4.33108956e-01
1.82420686e-01 4.19430345e-01 7.57796645e-01 -9.79991436e-01
-4.44005668e-01 -1.57968223e-01 1.10170327e-01 -3.25711429e-01
-6.98699176e-01 -3.07753056e-01 -8.94990265e-01 2.86447983e-02
-2.85089940e-01 5.85952044e-01 7.51291394e-01 1.11301517e+00
-1.08161278e-01 2.96256065e-01 6.00905359e-01 -7.51368642e-01
-5.88385642e-01 -1.32793200e+00 -3.21526587e-01 6.52427852e-01
1.66998744e-01 -5.73006868e-01 -2.56882459e-01 2.02044532e-01] | [11.058248519897461, 10.039112091064453] |
41727ab7-689b-48dd-8935-4ab0be6c2fb9 | ghost-free-multi-exposure-image-fusion | null | null | https://www.sciencedirect.com/science/article/pii/S1047320319301750 | https://www.sciencedirect.com/science/article/pii/S1047320319301750 | Ghost-free multi exposure image fusion technique using dense SIFT descriptor and guided filter | A ghost-free multi-exposure image fusion technique using the dense SIFT descriptor and the guided filter is proposed in this paper. The results suggest that the presented scheme produces high-quality images using ordinary cameras and that too without the ghosting artifact. To do so, the dense SIFT descriptor is used to extract the local contrast information from source images. Whereas, for the dynamic scenes, the histogram equalization and median filtering are used to calculate the color dissimilarity feature. Three weighting terms: local contrast, brightness, and color dissimilarity feature are used to estimate the initial weights. The estimated initial weights contain discontinuities. Therefore, the guided filter is used to remove the noise and discontinuity in initial weights. Finally, the fusion is performed using a pyramid decomposition method. Experimental results prove the superiority of the proposed technique over existing state-of-the-art methods in terms of both subjective and objective evaluation. | ['Muhammad Imran', 'Naila Hayat'] | 2019-07-01 | null | null | null | journal-2019-7 | ['multi-exposure-image-fusion'] | ['computer-vision'] | [ 2.88171947e-01 -8.89066637e-01 1.96904898e-01 -1.60607129e-01
-3.31182986e-01 -2.11399104e-02 4.17641729e-01 2.00829003e-02
-5.77016175e-01 5.39023995e-01 3.59079614e-02 4.05051112e-01
-2.22538814e-01 -6.70479178e-01 -2.05225393e-01 -1.10449886e+00
1.02949198e-02 -5.73637605e-01 4.96476710e-01 -1.91407919e-01
6.32989109e-01 3.16069067e-01 -1.94700503e+00 5.39111942e-02
1.13621557e+00 1.30803084e+00 3.57966334e-01 3.22815329e-01
6.79847077e-02 7.48160124e-01 -4.20444280e-01 -1.22416183e-01
4.68778551e-01 -5.02840459e-01 -7.84420818e-02 5.07263541e-01
4.14059073e-01 -7.94917226e-01 -1.88321233e-01 1.47949183e+00
4.46788311e-01 4.66862917e-01 4.00779933e-01 -1.04092205e+00
-4.99469876e-01 -2.14135543e-01 -1.06300163e+00 2.31067568e-01
5.35456777e-01 1.07160792e-01 3.62980992e-01 -1.10048461e+00
4.05422688e-01 1.16465926e+00 3.91204268e-01 -2.47901082e-01
-9.85908568e-01 -6.03569031e-01 -1.58674300e-01 5.77609003e-01
-1.53599083e+00 -4.45140719e-01 1.08873141e+00 -1.20956987e-01
4.24710661e-01 2.17233121e-01 6.68171883e-01 1.58364505e-01
7.03240871e-01 2.26266965e-01 1.52672088e+00 -6.92887962e-01
7.34882057e-02 1.25576004e-01 5.29052652e-02 8.17097008e-01
4.36774701e-01 3.00273776e-01 -4.38401729e-01 -2.15808079e-01
7.38455117e-01 4.87702519e-01 -4.51280326e-01 -1.82400346e-01
-1.14370930e+00 5.32066941e-01 2.70451695e-01 5.43499708e-01
-6.06777430e-01 -1.59288198e-01 2.52813607e-01 2.45576411e-01
1.91957489e-01 -1.79136634e-01 1.37951434e-01 1.58768460e-01
-1.05262041e+00 -1.05537819e-02 3.66571933e-01 6.81282401e-01
9.38475728e-01 1.58374593e-01 -1.04311280e-01 6.60474360e-01
5.41972339e-01 5.75368226e-01 4.49895501e-01 -1.02998567e+00
9.70485359e-02 5.03013253e-01 4.10000056e-01 -1.44898558e+00
7.94489235e-02 -3.91677320e-02 -9.62350070e-01 7.95612395e-01
1.94498748e-01 2.16039211e-01 -7.51419485e-01 1.17310202e+00
3.99508029e-01 2.21860051e-01 2.09711552e-01 1.09394515e+00
6.11519456e-01 8.47131968e-01 -1.52036041e-01 -5.61344802e-01
1.37087262e+00 -7.45453238e-01 -1.29582381e+00 1.92039937e-01
-2.12752476e-01 -1.36629677e+00 7.04366684e-01 6.58219695e-01
-1.32944942e+00 -1.13396156e+00 -1.38897419e+00 1.37289628e-01
-1.40445650e-01 2.85263926e-01 3.77383024e-01 6.33283138e-01
-8.36898267e-01 4.69982237e-01 -6.36130929e-01 -1.20940790e-01
-1.30187124e-01 2.35293984e-01 -1.99602276e-01 -3.13000441e-01
-1.07832348e+00 8.77829552e-01 4.08610612e-01 2.01492652e-01
-2.97639698e-01 -2.75813490e-01 -7.14228749e-01 1.38569608e-01
-1.65557563e-02 -4.26403254e-01 5.46237111e-01 -1.10977435e+00
-1.53436470e+00 3.78764808e-01 -1.81171775e-01 6.04370348e-02
2.48344213e-01 5.58076799e-02 -6.32182300e-01 6.75409496e-01
-2.48157717e-02 1.15482867e-01 1.13631701e+00 -1.22935688e+00
-9.62926626e-01 -1.45915046e-01 -2.57927448e-01 5.47155678e-01
-1.29885077e-01 -3.21780145e-02 -5.07029355e-01 -5.77293336e-01
4.22304481e-01 -4.25089240e-01 -3.11484709e-02 1.55583605e-01
4.60897982e-02 2.59327680e-01 1.00669265e+00 -8.51901114e-01
1.31365132e+00 -2.60401487e+00 -1.00717112e-01 4.89121675e-01
1.05007365e-02 3.49610187e-02 1.62149027e-01 3.90308589e-01
1.71614885e-01 -6.43769324e-01 -1.77291542e-01 3.74695659e-02
-2.95803785e-01 -2.60255754e-01 1.39637724e-01 8.82651865e-01
-1.99943632e-01 -2.50336882e-02 -6.30654812e-01 -7.59366095e-01
8.54143322e-01 7.08940208e-01 -2.31419355e-01 9.22958702e-02
6.05719626e-01 2.28239805e-01 -2.87472457e-01 6.76870823e-01
1.43333995e+00 2.88948625e-01 -5.48095666e-02 -7.36423016e-01
-5.86868703e-01 -4.98191118e-01 -1.81352711e+00 1.47322166e+00
-2.04272017e-01 2.34446064e-01 3.34663540e-01 -5.93581200e-01
9.99578595e-01 1.59930542e-01 5.89412928e-01 -8.13048303e-01
5.35113752e-01 3.18231702e-01 -7.93897659e-02 -4.39301521e-01
7.15837836e-01 -1.11948429e-02 2.75809646e-01 1.37097379e-02
-1.62979364e-01 -1.66197792e-01 3.79064053e-01 8.76079872e-02
5.75446725e-01 1.39841333e-01 3.66816670e-01 -3.71480376e-01
1.15583265e+00 -2.41515934e-01 7.10636497e-01 2.95657903e-01
-2.79503822e-01 3.20812225e-01 -2.94521481e-01 -3.11610758e-01
-7.94817388e-01 -1.05412531e+00 -1.27728328e-01 3.88049096e-01
9.82917786e-01 -5.55109717e-02 -5.93552411e-01 6.68571368e-02
-1.33584857e-01 2.83010840e-01 -1.64348647e-01 -2.58573771e-01
-2.62625694e-01 -4.82251674e-01 -1.60587743e-01 1.69350635e-02
1.28720963e+00 -6.58260405e-01 -8.28994513e-01 3.41497242e-01
-2.80719906e-01 -9.00744677e-01 -6.14260495e-01 -4.29272681e-01
-8.16607594e-01 -1.02879035e+00 -8.07035446e-01 -9.43559647e-01
7.26395488e-01 8.80340219e-01 3.16043735e-01 3.95857483e-01
-3.59120369e-01 3.05524707e-01 -3.49713832e-01 8.73599648e-02
-5.21767773e-02 -8.66610050e-01 1.59913732e-03 4.43992794e-01
3.21150988e-01 -3.58378381e-01 -1.02638876e+00 3.12183499e-01
-1.11850882e+00 5.51065721e-04 7.68333972e-01 8.79899442e-01
6.32590413e-01 8.13722670e-01 6.36345968e-02 -1.47411719e-01
6.21243834e-01 1.30006075e-01 -8.54596317e-01 1.99329183e-01
-6.54964328e-01 -1.97647005e-01 6.69224918e-01 -4.51969802e-01
-1.55858231e+00 4.84484509e-02 1.98227912e-01 -1.67119950e-01
1.43168628e-01 1.96794406e-01 -6.05969802e-02 -6.66966796e-01
-1.36489142e-03 5.60001612e-01 1.79976642e-01 -2.71082669e-01
2.03317210e-01 6.42334521e-01 6.59486234e-01 -1.71411201e-01
8.34206343e-01 8.17291439e-01 1.32424816e-01 -8.93808067e-01
-4.44958359e-02 -6.78038895e-01 -2.85861611e-01 -4.79371876e-01
1.00772965e+00 -1.00540197e+00 -9.30664480e-01 9.52732384e-01
-1.04737043e+00 4.60739166e-01 8.14587697e-02 9.89928961e-01
-4.45194364e-01 8.90051067e-01 -6.83762491e-01 -8.51296425e-01
-3.91495615e-01 -1.36347342e+00 6.77620292e-01 7.79515743e-01
4.00202364e-01 -7.66307533e-01 -1.67158410e-01 4.61059855e-03
6.25675738e-01 1.94044858e-01 4.12296563e-01 3.31533313e-01
-8.16283464e-01 -3.61195952e-02 -3.76196921e-01 5.01067519e-01
3.69789630e-01 1.74936563e-01 -6.20437741e-01 -3.59391123e-01
4.51550245e-01 9.06825885e-02 6.44641280e-01 5.13272047e-01
7.46363342e-01 2.83137299e-02 -3.45045328e-02 7.87747741e-01
2.03819036e+00 5.98492086e-01 9.65696871e-01 4.58344072e-01
2.34189406e-01 3.08312118e-01 1.08400035e+00 5.15726268e-01
1.19013861e-01 4.75777477e-01 2.19494417e-01 -5.07093132e-01
-1.47757709e-01 1.49949893e-01 3.37447643e-01 8.98435593e-01
-2.21718282e-01 -2.35144962e-02 -1.11017279e-01 3.57744515e-01
-1.56201649e+00 -1.16753817e+00 -3.70671600e-01 2.42506623e+00
6.02034032e-01 -1.32437916e-02 -4.63175684e-01 4.16596621e-01
9.95012522e-01 2.24511057e-01 -6.93991035e-02 -4.57081407e-01
-2.88509011e-01 1.26193807e-01 7.74517119e-01 6.08148515e-01
-1.07572138e+00 3.69749814e-01 5.40169096e+00 9.13509071e-01
-9.73587632e-01 -2.48257853e-02 1.81974649e-01 4.84286875e-01
-8.53736922e-02 2.00014353e-01 -4.16343004e-01 7.51630127e-01
2.58379936e-01 -1.24688655e-01 5.15974522e-01 4.84463990e-01
3.72858107e-01 -9.64101374e-01 -2.44333014e-01 1.28922713e+00
2.92537332e-01 -8.31923306e-01 -1.20213881e-01 8.18139408e-03
7.40013480e-01 -6.89113855e-01 1.47678226e-01 -3.14807713e-01
-1.30220428e-01 -2.06118554e-01 6.50084436e-01 9.96984601e-01
5.51942050e-01 -9.80341256e-01 9.12213862e-01 2.28734757e-03
-1.54985332e+00 -4.77793515e-02 -4.87834275e-01 -5.42113464e-03
3.13871413e-01 7.04921603e-01 1.04802936e-01 8.48877311e-01
5.74378729e-01 5.46784222e-01 -4.89700496e-01 1.36214793e+00
1.31979324e-02 -9.23694968e-02 -2.93512702e-01 2.92240113e-01
2.47654513e-01 -8.13639104e-01 3.65077764e-01 9.10530806e-01
6.03564918e-01 2.89930582e-01 2.03217626e-01 4.26384956e-01
3.43768686e-01 2.94452399e-01 -2.78016329e-01 5.16021311e-01
3.76878709e-01 1.28011513e+00 -6.93031132e-01 -5.49565375e-01
-7.06532776e-01 1.37314773e+00 -5.58173358e-01 4.54891711e-01
-8.07225704e-01 -9.52779353e-01 2.92014420e-01 -1.37904420e-01
5.02568007e-01 -2.06472069e-01 3.33372271e-03 -1.06281340e+00
-6.62503690e-02 -7.38623798e-01 2.32544675e-01 -9.55711365e-01
-9.63259399e-01 2.74998486e-01 7.27509409e-02 -1.67954075e+00
1.38590530e-01 -2.92770207e-01 -7.54551232e-01 1.03074324e+00
-1.74049854e+00 -7.38965690e-01 -6.02533340e-01 8.36548448e-01
3.54548991e-01 -5.98497353e-02 3.30860943e-01 4.89866614e-01
-1.90682828e-01 1.36959448e-01 5.83606899e-01 -2.92051524e-01
8.58932674e-01 -7.67858624e-01 -4.87439007e-01 1.20010924e+00
-5.86465120e-01 4.82203096e-01 7.22934186e-01 -6.88557267e-01
-1.45951843e+00 -4.21853989e-01 6.35591924e-01 6.06651843e-01
3.24611574e-01 1.82850614e-01 -8.48346591e-01 4.46809679e-02
6.05609775e-01 4.44995835e-02 4.05388981e-01 -8.06082845e-01
9.53970701e-02 -4.85118508e-01 -1.61283755e+00 2.57540405e-01
2.92861432e-01 -2.48905227e-01 -4.43068415e-01 -2.06693605e-01
2.27829322e-01 -2.27065340e-01 -1.14416718e+00 1.85877278e-01
8.33671153e-01 -1.45750821e+00 1.07190967e+00 7.02320039e-01
-1.53115019e-02 -9.55338359e-01 -3.22723836e-01 -1.02164960e+00
-5.22328436e-01 -3.33701938e-01 3.82150024e-01 1.36699080e+00
-2.99750298e-01 -8.08359444e-01 4.95155752e-02 2.98812449e-01
1.10205509e-01 -1.93134457e-01 -7.59983540e-01 -5.04845202e-01
-9.77357864e-01 3.59778404e-01 3.77951294e-01 6.48296058e-01
-1.25253364e-01 -5.23914918e-02 -4.87725854e-01 2.89084494e-01
1.34292316e+00 1.06960982e-01 5.38012862e-01 -1.00552988e+00
-4.24353667e-02 -9.28659961e-02 -6.86234117e-01 -7.12290168e-01
-3.59569371e-01 -1.23504318e-01 -8.15391243e-02 -1.42454422e+00
1.77959189e-01 2.75149215e-02 -5.02964139e-01 -2.94343531e-01
-3.29703480e-01 4.33616221e-01 3.18415761e-01 3.22944611e-01
-4.03281480e-01 5.59172213e-01 1.34686744e+00 -1.08252577e-01
-1.20052859e-01 -2.23468691e-01 -3.58815342e-01 6.66006267e-01
6.57438397e-01 -1.96373284e-01 -3.08527142e-01 -1.69026732e-01
-3.27296495e-01 1.38275936e-01 4.04752284e-01 -1.30019248e+00
4.04213518e-01 -1.66311875e-01 7.07617760e-01 -7.92602956e-01
5.39856732e-01 -1.28121305e+00 2.21358269e-01 6.98331118e-01
2.74145871e-01 1.09702080e-01 -1.11646019e-02 5.81303000e-01
-5.78106165e-01 -2.05466926e-01 1.09092939e+00 -4.59940359e-02
-1.01528943e+00 -8.34575072e-02 -4.13531482e-01 -6.19012833e-01
1.17489672e+00 -7.84253418e-01 -2.87358880e-01 -3.12574506e-01
-3.79188925e-01 -1.38928548e-01 7.78938711e-01 -1.14069454e-01
1.03943443e+00 -1.43614614e+00 -5.27245939e-01 6.07221782e-01
-2.60168463e-02 -5.97474456e-01 7.59400606e-01 1.18032086e+00
-8.24092329e-01 -1.41983390e-01 -6.53458834e-01 -4.27973539e-01
-1.41558981e+00 6.42210901e-01 -1.58223193e-02 -5.73370494e-02
-4.69268709e-01 2.41444126e-01 -5.91351576e-02 5.16334474e-01
1.45729452e-01 -2.87892252e-01 -2.91914999e-01 -9.37150866e-02
8.93011212e-01 6.86363935e-01 -1.83757231e-01 -8.79818797e-01
-4.27944094e-01 1.00922537e+00 9.88546833e-02 -2.43194446e-01
1.06301618e+00 -7.19766438e-01 -3.86501700e-01 7.70198926e-02
1.25070584e+00 3.48293871e-01 -1.19103014e+00 -1.08787835e-01
-4.48575944e-01 -1.10002053e+00 4.47878122e-01 -4.60735351e-01
-1.04410648e+00 5.80861449e-01 1.29596722e+00 1.13487579e-01
1.78752148e+00 -7.67352164e-01 7.76895404e-01 -6.18301630e-02
3.20970237e-01 -1.26793480e+00 -1.05927326e-01 -5.14419414e-02
4.52297777e-01 -1.09753203e+00 5.34810185e-01 -4.42851156e-01
-4.81789321e-01 1.19561303e+00 3.93449247e-01 -4.50144470e-01
4.91356254e-01 1.92947343e-01 -7.73758292e-02 8.18897635e-02
-2.49338835e-01 -3.88088197e-01 1.39344245e-01 5.27362525e-01
2.32449129e-01 -4.12124127e-01 -9.41424429e-01 -1.03303075e-01
4.72415894e-01 1.69857726e-01 6.27315640e-01 9.60379839e-01
-8.70047212e-01 -8.49319637e-01 -9.97132063e-01 1.51843503e-01
-4.27432448e-01 2.02451106e-02 3.15813273e-01 6.46657944e-01
3.17566574e-01 1.30866027e+00 6.70999335e-03 -3.57607633e-01
3.43777657e-01 -3.68560374e-01 6.46474183e-01 2.11778745e-01
-4.23698008e-01 4.92588848e-01 -3.01626921e-01 -5.88318765e-01
-9.70598638e-01 -4.33599323e-01 -1.17812622e+00 -3.98906916e-01
-4.38383549e-01 2.78296769e-01 8.05974185e-01 4.02616560e-01
-8.45911205e-02 4.13532466e-01 9.89650428e-01 -1.03268242e+00
-2.87936330e-01 -7.11648226e-01 -9.10686970e-01 5.78166127e-01
3.39341164e-01 -8.29879224e-01 -7.47504056e-01 2.67185569e-01] | [10.919923782348633, -2.446815252304077] |
d8cd8ac6-1128-48d1-916d-541dc34635fe | uncertainty-based-class-activation-maps-for | 2002.10309 | null | https://arxiv.org/abs/2002.10309v1 | https://arxiv.org/pdf/2002.10309v1.pdf | Uncertainty based Class Activation Maps for Visual Question Answering | Understanding and explaining deep learning models is an imperative task. Towards this, we propose a method that obtains gradient-based certainty estimates that also provide visual attention maps. Particularly, we solve for visual question answering task. We incorporate modern probabilistic deep learning methods that we further improve by using the gradients for these estimates. These have two-fold benefits: a) improvement in obtaining the certainty estimates that correlate better with misclassified samples and b) improved attention maps that provide state-of-the-art results in terms of correlation with human attention regions. The improved attention maps result in consistent improvement for various methods for visual question answering. Therefore, the proposed technique can be thought of as a recipe for obtaining improved certainty estimates and explanations for deep learning models. We provide detailed empirical analysis for the visual question answering task on all standard benchmarks and comparison with state of the art methods. | ['Vinay P. Namboodiri', 'Mayank Lunayach', 'Badri N. Patro'] | 2020-01-23 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-3.00491333e-01 4.28886205e-01 -1.87290132e-01 -5.87156475e-01
-1.11500978e+00 -5.55459142e-01 7.87589788e-01 2.58614123e-01
-1.87559009e-01 6.58586323e-01 2.50198096e-01 -4.79683518e-01
-1.19336382e-01 -7.00516939e-01 -9.96824443e-01 -3.50599885e-01
1.76656321e-01 5.04035652e-01 2.60938019e-01 1.63505480e-01
5.81674695e-01 2.09459886e-01 -1.55736744e+00 7.12686002e-01
9.21653688e-01 1.29671574e+00 8.62422511e-02 8.42046261e-01
-6.56968534e-01 1.31582463e+00 -5.74209452e-01 -8.80969942e-01
-2.89992422e-01 -2.77760655e-01 -1.06911421e+00 -2.67228842e-01
9.50081587e-01 -4.18483526e-01 -2.05672383e-01 1.11923945e+00
2.04659346e-02 6.68877289e-02 1.08356249e+00 -1.49035680e+00
-1.43895602e+00 4.24968511e-01 -7.94948220e-01 3.94037396e-01
2.66117781e-01 1.41742736e-01 1.36665738e+00 -1.24608505e+00
2.63624430e-01 1.63723636e+00 4.68129545e-01 6.36615157e-01
-8.86721373e-01 -3.01580667e-01 3.77339512e-01 1.10569513e+00
-1.01530433e+00 -7.77026266e-02 6.38730645e-01 -4.82036293e-01
9.01532054e-01 3.01212728e-01 2.64737785e-01 1.07245028e+00
1.87106133e-01 1.20105481e+00 1.09589565e+00 -3.86257410e-01
3.87661397e-01 4.58111495e-01 5.33784866e-01 1.07904935e+00
2.83925444e-01 8.14342052e-02 -6.78951323e-01 -1.01285325e-02
5.20616472e-01 -4.06719632e-02 -3.11854631e-01 -5.50473392e-01
-9.79138911e-01 1.03251779e+00 1.10329592e+00 -1.54830381e-01
-3.87130439e-01 7.38094926e-01 1.45431072e-01 -1.06506206e-01
4.12675530e-01 4.02534962e-01 -3.42526555e-01 1.10484652e-01
-8.17188025e-01 3.42862666e-01 5.34867883e-01 8.71438801e-01
8.33321869e-01 3.92077267e-02 -7.24914372e-01 6.10038877e-01
6.78077281e-01 3.82853657e-01 6.95390552e-02 -1.28055620e+00
3.87469441e-01 6.33589625e-01 2.55744278e-01 -1.11002970e+00
-3.12278420e-01 -4.64485914e-01 -7.35257328e-01 4.79057431e-01
7.36663461e-01 3.12866032e-01 -1.27147222e+00 1.74786425e+00
2.80264139e-01 8.85474533e-02 4.61473409e-03 1.03528535e+00
1.23540664e+00 7.91035831e-01 4.30493772e-01 3.94685358e-01
1.84058142e+00 -1.38388181e+00 -7.47450829e-01 -4.58556771e-01
1.27858698e-01 -3.77390563e-01 1.51602435e+00 2.37878829e-01
-7.90005028e-01 -7.22997189e-01 -1.04178977e+00 -5.37761033e-01
-4.76103038e-01 2.94974893e-01 6.68217719e-01 6.29920125e-01
-1.02916348e+00 3.33184212e-01 -6.04648173e-01 -5.84719703e-02
8.62689912e-01 -2.23242924e-01 -6.18459173e-02 -7.65213221e-02
-1.04660571e+00 1.11832035e+00 1.49682730e-01 6.12281896e-02
-1.32206142e+00 -8.15181553e-01 -9.70759690e-01 5.61536491e-01
2.86916673e-01 -1.00393987e+00 1.27858937e+00 -6.60635352e-01
-9.77940559e-01 9.61219192e-01 -5.94734073e-01 -5.84732771e-01
5.31539261e-01 -4.38994050e-01 2.32649401e-01 3.84275526e-01
1.11067921e-01 1.03191352e+00 1.08897638e+00 -1.74294639e+00
-6.62850976e-01 -2.74254501e-01 5.03907442e-01 -5.75775802e-02
-9.49238688e-02 -4.73897368e-01 -4.02875423e-01 -3.91765088e-01
-1.30738020e-01 -4.44738030e-01 8.70071072e-03 6.59989357e-01
-3.88197362e-01 -5.88942826e-01 5.52987337e-01 -8.79851401e-01
7.54092932e-01 -1.73917377e+00 1.11085311e-01 -1.93455964e-01
7.14919329e-01 -3.19822691e-02 -1.39946818e-01 -1.16006799e-01
9.76948738e-02 3.49318564e-01 -1.66938633e-01 -5.76086104e-01
3.55381995e-01 1.15229644e-01 -7.33768582e-01 1.67269200e-01
6.98496819e-01 1.39677525e+00 -9.11693275e-01 -6.85807645e-01
2.79520780e-01 3.75214875e-01 -3.12719107e-01 4.38968688e-01
-5.05855381e-01 1.40092373e-02 -2.49882296e-01 7.38672078e-01
6.78969622e-01 -7.36105859e-01 -2.97590911e-01 -3.26072782e-01
3.18270922e-01 1.41366675e-01 -7.67380953e-01 1.32198799e+00
-2.79253423e-01 1.05141723e+00 -3.40401441e-01 -8.58853400e-01
9.03238177e-01 2.34404542e-02 -4.98959184e-01 -5.24200499e-01
4.90141734e-02 -1.98572591e-01 -3.00974160e-01 -5.85649014e-01
5.86853266e-01 -1.62141964e-01 3.26464117e-01 3.34243238e-01
3.49618345e-01 -1.04030110e-01 -8.12860131e-02 5.35478950e-01
6.59412146e-01 1.58269510e-01 3.33161592e-01 -3.45610529e-01
4.68883753e-01 -8.00063834e-03 2.72888243e-02 1.06717837e+00
-6.08064294e-01 5.95033050e-01 9.61024106e-01 -5.09505451e-01
-1.14854217e+00 -1.45641577e+00 1.37114853e-01 1.06870234e+00
3.16640317e-01 -2.34331846e-01 -7.28840768e-01 -1.03834128e+00
1.70680821e-01 1.08200026e+00 -1.28977787e+00 -1.67235419e-01
-6.94779605e-02 -3.77816498e-01 3.67384195e-01 1.14541972e+00
4.26365703e-01 -1.01616144e+00 -4.92793411e-01 -3.93562049e-01
-2.44048327e-01 -1.03706074e+00 -9.46358815e-02 7.34019876e-02
-7.98694015e-01 -1.17966115e+00 -1.05023098e+00 -7.25569963e-01
6.36774480e-01 2.53373712e-01 1.56227612e+00 3.02348942e-01
-7.85090104e-02 6.01585805e-01 -6.52834848e-02 -6.99858606e-01
-1.64838314e-01 -1.21009998e-01 -4.53463137e-01 -2.48013780e-01
3.97989899e-01 5.47347218e-03 -8.31343412e-01 1.40241861e-01
-7.68693686e-01 -1.89312790e-02 4.33357030e-01 9.17099774e-01
5.65447152e-01 -6.91823006e-01 8.62050772e-01 -6.88388407e-01
7.09581316e-01 -6.27931714e-01 -5.87668002e-01 7.86966383e-01
-6.56742632e-01 7.55381227e-01 4.28464472e-01 -2.26527467e-01
-1.27435493e+00 -2.52006650e-01 -1.38877764e-01 -4.46848363e-01
-1.21404514e-01 1.89326659e-01 2.23367661e-01 9.81083438e-02
8.84407461e-01 -2.10746288e-01 -4.82847512e-01 -3.20030421e-01
9.95154619e-01 3.36404800e-01 6.25111759e-01 -6.75897717e-01
5.94515622e-01 6.16472960e-01 -1.76947877e-01 -2.15262339e-01
-1.40648890e+00 -2.94095337e-01 -2.14800388e-01 -3.84613216e-01
1.09560978e+00 -6.45848989e-01 -1.18891883e+00 7.59754106e-02
-1.50469184e+00 -1.67829152e-02 -6.69471025e-02 1.30840763e-01
-4.60496336e-01 5.31521261e-01 -4.97847825e-01 -1.27700770e+00
-4.30947781e-01 -1.07618642e+00 1.18709517e+00 5.55284262e-01
-1.36992782e-01 -1.10147488e+00 6.88400939e-02 5.35493851e-01
4.91225064e-01 1.07563131e-01 1.40838611e+00 -4.21379924e-01
-9.47731495e-01 2.55181789e-01 -9.63980556e-01 2.27117926e-01
-3.99162859e-01 -1.82454791e-02 -1.56095529e+00 1.44365624e-01
-6.82379007e-02 -6.41875744e-01 1.40643954e+00 5.51644742e-01
1.56256783e+00 -1.29861012e-01 -2.96367019e-01 3.41905743e-01
1.44769144e+00 -3.73523116e-01 7.25439131e-01 3.42404038e-01
7.80507445e-01 8.01075280e-01 4.87413317e-01 1.71926647e-01
8.05907309e-01 2.74346352e-01 9.49564278e-01 -1.26597658e-01
-3.09486389e-01 -4.97901529e-01 -1.50323763e-01 2.10762873e-01
3.08087058e-02 -2.15975732e-01 -1.08489704e+00 1.00637341e+00
-2.11226082e+00 -9.07005191e-01 -3.69829178e-01 1.76982665e+00
5.74320078e-01 -1.87030230e-02 -1.21068925e-01 -8.46401155e-02
7.58070588e-01 2.52996951e-01 -8.06582749e-01 -3.94438297e-01
1.49915159e-01 2.31456324e-01 7.05089793e-03 7.24047661e-01
-8.68569314e-01 8.07913125e-01 6.87944746e+00 8.98425817e-01
-5.75793743e-01 1.65519476e-01 1.09941769e+00 2.64670312e-01
-7.52042770e-01 -5.46034612e-02 -7.19659090e-01 -1.61829181e-02
6.83968186e-01 -2.00361572e-02 2.81786293e-01 1.19521952e+00
-3.19236636e-01 -2.88570732e-01 -1.36651146e+00 1.26745522e+00
2.82575697e-01 -1.66941404e+00 2.87769467e-01 -4.22644138e-01
9.00571108e-01 -2.86127716e-01 3.42916399e-01 5.38286507e-01
2.22146645e-01 -1.31870520e+00 1.01437426e+00 7.95539796e-01
5.13520062e-01 -7.46936202e-01 7.60416150e-01 2.21812129e-01
-1.02879882e+00 3.55281495e-02 -7.19242275e-01 6.81935549e-02
1.01098754e-01 6.22285128e-01 -7.99774706e-01 3.34057897e-01
1.03675473e+00 4.03478265e-01 -8.30511630e-01 1.04382157e+00
-8.59240711e-01 6.34031475e-01 1.04261689e-01 -3.85782391e-01
3.10943335e-01 3.29079598e-01 1.74680561e-01 1.09175789e+00
5.90237603e-02 -2.37971827e-01 -5.85037947e-01 1.70537198e+00
-3.46658289e-01 -2.37056270e-01 -2.48520464e-01 1.77105606e-01
3.06677252e-01 1.31201363e+00 -5.41884959e-01 -7.10050464e-01
-2.97354877e-01 8.54133725e-01 1.02697933e+00 5.52732289e-01
-1.31625605e+00 -2.51672983e-01 6.95467889e-01 -2.69987881e-01
4.86408949e-01 -1.84448153e-01 -7.00932145e-01 -1.19560981e+00
5.09079266e-03 -7.30784893e-01 4.16124314e-01 -1.30630589e+00
-1.71422744e+00 7.44458616e-01 -3.07071693e-02 -6.07496560e-01
-1.13379344e-01 -1.18577099e+00 -5.23135126e-01 9.03242469e-01
-2.07792234e+00 -1.01308823e+00 -7.32138991e-01 4.44964081e-01
5.47174394e-01 9.40443724e-02 6.53889239e-01 -1.37870371e-01
-1.40512571e-01 6.19857728e-01 -2.25624740e-01 -1.26203001e-01
6.00243807e-01 -1.70281541e+00 6.77926183e-01 6.77452683e-01
6.23667836e-01 5.47836900e-01 6.03117526e-01 -2.58263081e-01
-8.63856435e-01 -7.92560935e-01 7.93043494e-01 -1.10192692e+00
5.96218050e-01 -3.44372958e-01 -1.12233961e+00 6.81963146e-01
4.40630585e-01 -1.78911150e-01 4.14273500e-01 3.59904796e-01
-8.02264035e-01 1.58926725e-01 -1.10417974e+00 5.65359950e-01
6.55121982e-01 -9.02997255e-01 -1.01718378e+00 2.11376607e-01
7.57218480e-01 -3.98329377e-01 -3.55068803e-01 2.51773298e-01
5.68486571e-01 -1.13405013e+00 1.13487506e+00 -9.32756543e-01
8.75977099e-01 -3.99852633e-01 -9.16358978e-02 -1.08771479e+00
-2.87483871e-01 -2.55600978e-02 -5.99528790e-01 9.48664725e-01
6.81835473e-01 -3.30521166e-01 8.36977363e-01 8.17635655e-01
5.06894700e-02 -8.87806237e-01 -8.70960891e-01 -3.78185064e-01
1.25567034e-01 -6.16170645e-01 4.58344817e-01 7.17788160e-01
-1.39246419e-01 3.52335900e-01 -4.57149059e-01 2.64771551e-01
8.61000657e-01 3.81679647e-02 4.85343903e-01 -8.81709337e-01
-1.78514138e-01 -4.30826455e-01 -2.47619286e-01 -1.49296522e+00
2.10547119e-01 -6.62148952e-01 2.35895827e-01 -2.05526114e+00
7.13636994e-01 1.01782948e-01 -3.16414416e-01 5.54124415e-01
-8.40672791e-01 1.56000331e-01 3.36760372e-01 -7.16639906e-02
-8.42674732e-01 7.32694447e-01 1.08697414e+00 -3.88637185e-01
4.93902594e-01 -3.67466182e-01 -9.48909760e-01 8.21028173e-01
6.88590109e-01 -5.74032962e-01 -2.85047591e-01 -7.68190205e-01
5.12676179e-01 -2.39117563e-01 1.00905776e+00 -8.63127291e-01
6.42045066e-02 2.69786026e-02 7.00865209e-01 -7.75692821e-01
3.97482961e-01 -5.49810469e-01 -6.69723749e-01 3.55087668e-01
-7.07353175e-01 1.07142337e-01 3.59484702e-01 9.78667319e-01
-2.66342431e-01 -3.39955419e-01 7.41533399e-01 -1.16105184e-01
-9.12960708e-01 5.55313751e-02 -1.29379585e-01 2.43465930e-01
6.86530232e-01 3.01834382e-02 -9.31914926e-01 -8.53472471e-01
-6.74249113e-01 5.08168638e-01 7.22302007e-04 4.98199850e-01
8.55802238e-01 -1.45682752e+00 -7.49612093e-01 -4.33149576e-01
3.57115686e-01 -3.15070003e-01 3.46046001e-01 4.31317389e-01
-4.72754210e-01 6.20587230e-01 -1.36134714e-01 -8.49845290e-01
-9.05596077e-01 7.85877049e-01 4.08998638e-01 -3.42283458e-01
-1.11048780e-01 1.22683477e+00 7.26010859e-01 -3.46317619e-01
5.58873355e-01 -7.89844096e-01 -3.63831848e-01 -2.46210545e-01
7.82639503e-01 3.57324690e-01 -2.83652455e-01 1.15729561e-02
-5.51662743e-01 4.99337405e-01 1.37188673e-01 -1.55895660e-02
9.74375069e-01 -1.33093297e-01 1.94326088e-01 2.82497168e-01
9.92557883e-01 -2.39969179e-01 -1.52293217e+00 -3.44377458e-02
3.22649404e-02 -5.88349938e-01 -4.12367918e-02 -1.08497095e+00
-1.00074208e+00 1.49326575e+00 7.15827703e-01 3.19727153e-01
6.76151395e-01 6.39228046e-01 3.27455282e-01 4.63066697e-01
-1.22956894e-01 -5.96182883e-01 5.27798831e-01 4.10222292e-01
1.28562939e+00 -1.65393078e+00 -1.49729401e-01 -1.41026869e-01
-9.27435040e-01 1.16670692e+00 8.58880937e-01 -1.31525293e-01
3.69351596e-01 -1.87191844e-01 1.89410493e-01 -5.27301550e-01
-9.57958817e-01 -3.48482609e-01 8.41911077e-01 8.68629873e-01
4.97915834e-01 -1.19562391e-02 -6.58503100e-02 7.63612926e-01
8.46785530e-02 -2.86330789e-01 1.51340827e-01 1.40539244e-01
-7.36199021e-01 -4.78348374e-01 -4.26874042e-01 2.60035783e-01
-1.04108386e-01 -3.62218350e-01 -4.82737780e-01 7.51961470e-01
-2.64788896e-01 9.88717377e-01 -5.93080977e-03 1.09029897e-02
1.92529738e-01 1.43872410e-01 5.46033502e-01 -2.82058924e-01
-1.44045472e-01 -7.41258383e-01 5.99648356e-02 -5.31977892e-01
-2.20869318e-01 3.98867487e-05 -1.31211913e+00 -2.36920342e-01
-4.38073874e-01 4.03600466e-03 7.39151597e-01 1.10424268e+00
3.60911518e-01 6.60917819e-01 -6.89054579e-02 -6.52071297e-01
-6.86829567e-01 -9.09475625e-01 -1.10434450e-01 4.52308953e-01
3.75537813e-01 -8.12819839e-01 -4.90112901e-01 2.20553894e-02] | [10.819918632507324, 1.8040192127227783] |
36e46cfb-c4c0-43a3-a904-594754917979 | adversarial-consistent-learning-on-partial | 2009.09289 | null | https://arxiv.org/abs/2009.09289v1 | https://arxiv.org/pdf/2009.09289v1.pdf | Adversarial Consistent Learning on Partial Domain Adaptation of PlantCLEF 2020 Challenge | Domain adaptation is one of the most crucial techniques to mitigate the domain shift problem, which exists when transferring knowledge from an abundant labeled sourced domain to a target domain with few or no labels. Partial domain adaptation addresses the scenario when target categories are only a subset of source categories. In this paper, to enable the efficient representation of cross-domain plant images, we first extract deep features from pre-trained models and then develop adversarial consistent learning ($ACL$) in a unified deep architecture for partial domain adaptation. It consists of source domain classification loss, adversarial learning loss, and feature consistency loss. Adversarial learning loss can maintain domain-invariant features between the source and target domains. Moreover, feature consistency loss can preserve the fine-grained feature transition between two domains. We also find the shared categories of two domains via down-weighting the irrelevant categories in the source domain. Experimental results demonstrate that training features from NASNetLarge model with proposed $ACL$ architecture yields promising results on the PlantCLEF 2020 Challenge. | ['Youshan Zhang', 'Brian D. Davison'] | 2020-09-19 | null | null | null | null | ['partial-domain-adaptation'] | ['methodology'] | [ 4.36919212e-01 7.91239887e-02 -3.14276904e-01 -6.22454047e-01
-6.99643672e-01 -9.92335618e-01 4.06609863e-01 -1.22626813e-03
9.43615474e-03 8.87136042e-01 -1.47482559e-01 1.21355906e-01
-3.11081239e-04 -8.18427622e-01 -1.06316245e+00 -5.47144771e-01
1.76985085e-01 4.93504196e-01 3.22252333e-01 -3.11849654e-01
-1.74001068e-01 4.43268836e-01 -1.07898784e+00 5.75222671e-01
1.00157368e+00 1.22764587e+00 2.96740651e-01 1.61320865e-01
-3.73422831e-01 7.15499818e-01 -7.40554929e-01 -2.95323104e-01
4.94485736e-01 -4.17677552e-01 -1.01723874e+00 -4.13572602e-02
6.21082664e-01 -3.96876067e-01 -3.09054226e-01 1.29807484e+00
3.35607886e-01 -2.07264517e-02 8.12354982e-01 -1.57983196e+00
-1.47011459e+00 3.68027180e-01 -6.88221991e-01 -4.13417770e-03
1.16299674e-01 1.26571149e-01 6.32562101e-01 -7.65815973e-01
8.27027857e-01 1.26078618e+00 7.15643227e-01 6.78319216e-01
-1.28918099e+00 -9.55629408e-01 4.01384860e-01 2.64869422e-01
-1.37797940e+00 -1.19614162e-01 8.86684656e-01 -4.55734253e-01
6.07574165e-01 -3.27408344e-01 -1.19677968e-01 1.33281827e+00
1.79100350e-01 5.43493152e-01 9.07539368e-01 -2.79274762e-01
1.90380126e-01 3.11698467e-01 5.73985726e-02 3.50764334e-01
1.49320573e-01 1.62558407e-01 -2.82677054e-01 -2.68431753e-01
6.33092999e-01 2.25721002e-02 -3.28013599e-01 -1.05679154e+00
-1.06834662e+00 7.99983680e-01 8.46933722e-01 2.25527972e-01
-2.61529028e-01 -3.16001862e-01 8.01970601e-01 7.04096258e-01
3.37139189e-01 3.46392363e-01 -1.09554911e+00 6.06638432e-01
-4.03232098e-01 2.08084762e-01 5.21810830e-01 1.58093727e+00
7.73747087e-01 1.23314504e-02 -1.80034161e-01 9.88615930e-01
1.37111470e-01 7.08472192e-01 6.33653998e-01 -7.73668706e-01
4.22846496e-01 7.30834365e-01 1.53874233e-01 -7.91207969e-01
5.47187105e-02 -2.85978377e-01 -9.56656396e-01 2.77901918e-01
5.44236779e-01 -2.52257027e-02 -1.07438195e+00 2.21895576e+00
4.55127507e-01 1.19560130e-01 3.84628624e-01 6.73427165e-01
8.87569726e-01 5.18333077e-01 5.54309726e-01 1.58221588e-01
1.18725383e+00 -8.95333588e-01 -5.22476554e-01 -5.50757289e-01
3.60830635e-01 -6.90221906e-01 1.22020602e+00 -1.62648279e-02
-8.19005907e-01 -9.80189562e-01 -1.17161834e+00 -2.02616423e-01
-8.13914835e-01 5.21889105e-02 3.77938211e-01 2.40337461e-01
-4.72966254e-01 6.44196928e-01 -3.82450700e-01 -5.86401999e-01
8.11154306e-01 2.42988646e-01 -7.77625203e-01 -5.08137286e-01
-1.56655061e+00 8.40411603e-01 7.52639055e-01 -4.18272167e-01
-1.16686368e+00 -1.06645203e+00 -1.01404476e+00 1.31821379e-01
2.17244044e-01 -4.51181263e-01 1.13816631e+00 -1.41798806e+00
-1.18348646e+00 1.17421079e+00 1.70009822e-01 -5.36173701e-01
1.84558362e-01 -7.93787614e-02 -7.80791581e-01 -3.31248902e-02
5.07202625e-01 7.50536025e-01 9.53018308e-01 -1.14787400e+00
-8.00691187e-01 -5.05398214e-01 -1.17893130e-01 7.77743682e-02
-3.92744064e-01 -3.55546474e-01 7.17502087e-02 -6.53038621e-01
2.88715884e-02 -8.25791895e-01 2.00011030e-01 4.51120228e-01
-2.40646511e-01 -1.79749519e-01 1.18415654e+00 -8.13176155e-01
5.21091580e-01 -2.41174865e+00 2.65380125e-02 -1.32230029e-01
-1.09538406e-01 4.73484486e-01 -5.61854661e-01 -3.96485254e-02
-5.96000731e-01 -1.81121022e-01 -5.54217935e-01 3.73986840e-01
9.31716114e-02 2.59859532e-01 -6.15999222e-01 2.81193554e-01
6.45438492e-01 7.56494641e-01 -9.53315496e-01 -2.46709824e-01
1.27982587e-01 2.76955128e-01 -2.28452072e-01 3.30011427e-01
-3.07000965e-01 4.96346623e-01 -4.40114677e-01 8.16345215e-01
1.38935113e+00 -1.21674590e-01 1.10394925e-01 -3.38751435e-01
4.67515886e-01 -1.86024737e-02 -9.28963423e-01 2.16663790e+00
-2.24555582e-01 8.66933763e-02 2.04178736e-01 -1.20108461e+00
1.18054843e+00 1.22482456e-01 3.62129956e-01 -6.82825685e-01
-1.21389948e-01 2.88752675e-01 -1.26705676e-01 1.02622628e-01
9.82738435e-02 -2.03809008e-01 -5.34422815e-01 -6.48075417e-02
6.92644715e-01 -2.99550086e-01 -1.12237565e-01 -7.01847598e-02
1.07608008e+00 3.89943182e-01 5.27099073e-01 -3.20525825e-01
6.82748199e-01 2.97151059e-01 9.10062909e-01 4.53262389e-01
-8.16063046e-01 5.96591592e-01 4.25078362e-01 -3.81525069e-01
-1.15385377e+00 -1.45809352e+00 -5.34565970e-02 1.32545006e+00
2.85181940e-01 3.35992664e-01 -4.95830685e-01 -1.42564952e+00
3.98453206e-01 7.15464652e-01 -6.79311872e-01 -9.53368127e-01
-3.90750527e-01 -2.84685552e-01 5.99167228e-01 6.63776398e-01
7.34205902e-01 -1.06663013e+00 1.38747916e-01 1.81609780e-01
-1.28813252e-01 -9.65724945e-01 -6.07399225e-01 7.62739778e-01
-7.03024268e-01 -1.18554318e+00 -8.11275601e-01 -1.37857258e+00
5.92136562e-01 2.41525829e-01 1.42455614e+00 -5.39462924e-01
2.63770148e-02 -2.58115120e-02 -4.06948209e-01 -5.59642673e-01
-5.18202722e-01 1.99245244e-01 -5.72243668e-02 -4.40418243e-01
8.97537529e-01 -4.99152958e-01 -3.02226752e-01 3.69922698e-01
-9.48724627e-01 -6.23437583e-01 4.70501900e-01 1.21224296e+00
9.01070178e-01 1.95302784e-01 9.62167621e-01 -1.13800681e+00
2.92030394e-01 -8.01893651e-01 -3.90834659e-01 3.72925490e-01
-3.87482941e-01 5.18814288e-02 1.10107255e+00 -5.42923391e-01
-1.35046947e+00 3.92705411e-01 1.30872950e-01 -6.03355050e-01
-5.05447090e-01 9.62684155e-02 -8.93164277e-01 1.10020839e-01
1.02360773e+00 1.81524992e-01 -1.42462030e-01 -3.88596982e-01
5.94989777e-01 4.34641331e-01 9.09992516e-01 -6.51187778e-01
1.04996848e+00 5.02722204e-01 -2.57794529e-01 -2.29851350e-01
-9.69798148e-01 -3.19250315e-01 -1.01018906e+00 6.16013765e-01
6.31676018e-01 -1.16727102e+00 2.24422850e-02 5.23632348e-01
-1.10865355e+00 -1.73595428e-01 -8.74916971e-01 1.36069208e-01
-6.96277738e-01 2.83519536e-01 -3.22138816e-01 1.21539064e-01
-2.05207199e-01 -6.28831267e-01 9.75759208e-01 3.76419812e-01
4.38064002e-02 -9.01405036e-01 1.47873506e-01 -1.97978899e-01
2.31333554e-01 5.45348525e-01 1.16522038e+00 -1.10070527e+00
-7.06703886e-02 -9.72286463e-02 -4.72910523e-01 7.24293411e-01
6.01366222e-01 -3.80368352e-01 -1.03292811e+00 -5.77656448e-01
-5.53329736e-02 -6.17083311e-01 7.62608767e-01 2.64450908e-01
1.14778757e+00 1.24673843e-01 -5.49251616e-01 7.22784460e-01
1.43342030e+00 3.78927082e-01 4.81891930e-01 2.82311976e-01
5.84499598e-01 5.78650773e-01 1.00874650e+00 2.21689999e-01
1.24044694e-01 2.69938976e-01 5.19902527e-01 -2.16854542e-01
-3.86602402e-01 -4.53664273e-01 2.27570370e-01 1.68789864e-01
8.16098571e-01 -1.61646947e-01 -6.74424708e-01 8.23519349e-01
-1.57752049e+00 -9.30073798e-01 4.76815552e-01 2.10725999e+00
9.05717731e-01 9.40399319e-02 -1.10186666e-01 -1.52447179e-01
1.15512645e+00 -1.52952701e-01 -1.29806149e+00 -3.83392334e-01
-2.17700109e-01 2.96766549e-01 5.38389206e-01 -5.42557277e-02
-1.60765457e+00 1.10771680e+00 5.76431322e+00 8.78331840e-01
-9.99862015e-01 1.70272738e-01 4.18297619e-01 2.71111518e-01
-8.97648484e-02 -1.85214847e-01 -6.45508051e-01 5.14962018e-01
7.94013381e-01 -3.03531915e-01 8.43527466e-02 1.48603880e+00
-4.59111035e-01 5.20761788e-01 -1.27527988e+00 5.05454481e-01
-2.07007647e-01 -7.88876116e-01 1.31260872e-01 -1.45694092e-01
8.47321630e-01 -1.50932446e-01 1.94464833e-01 8.68336976e-01
8.23387742e-01 -8.09108078e-01 4.56065387e-01 9.19571444e-02
1.18824112e+00 -9.47732329e-01 6.41591132e-01 2.51460433e-01
-1.38329649e+00 -1.53536066e-01 -1.05516350e+00 3.84379119e-01
-2.20735371e-01 4.53893542e-01 -9.12147224e-01 7.05374479e-01
9.76210654e-01 8.08719516e-01 -4.15605038e-01 5.62332213e-01
-1.20813623e-01 1.57769576e-01 -8.55711699e-02 7.71292627e-01
7.60787651e-02 1.21185981e-01 2.83387125e-01 9.22465682e-01
3.48470420e-01 -2.64081299e-01 3.03590238e-01 1.03888500e+00
-3.20671886e-01 -2.59165227e-01 -1.01338601e+00 -3.93692106e-02
8.21658432e-01 1.01778471e+00 -2.76076257e-01 -2.97863215e-01
-4.49128896e-01 1.50429940e+00 5.05955756e-01 7.30422884e-02
-9.57554460e-01 -8.10938537e-01 1.09259808e+00 -1.18815459e-01
5.11120141e-01 2.81322032e-01 -2.24711388e-01 -1.05033469e+00
-1.40803516e-01 -9.56114650e-01 7.14187086e-01 -6.05288625e-01
-2.00431323e+00 5.54432571e-01 -1.73395932e-01 -1.41299045e+00
-1.10760681e-01 -6.55624509e-01 -4.99118209e-01 1.27802527e+00
-1.83553755e+00 -1.64579213e+00 -4.35985386e-01 1.04968476e+00
6.15493715e-01 -5.31008124e-01 1.15759397e+00 3.62195164e-01
-2.72471219e-01 9.03827071e-01 5.47940493e-01 1.53632075e-01
1.38403249e+00 -1.35443854e+00 5.74573994e-01 7.93520987e-01
-4.50909257e-01 4.64937001e-01 3.32767934e-01 -7.27155983e-01
-6.37776613e-01 -1.71983743e+00 7.49774337e-01 -3.51526111e-01
4.02617097e-01 -2.33057976e-01 -1.34336066e+00 8.42628598e-01
8.19669515e-02 4.72332150e-01 7.13515699e-01 -2.46700764e-01
-8.97778451e-01 -2.70930469e-01 -1.79313684e+00 6.16863891e-02
1.06928146e+00 -6.80810988e-01 -8.12069833e-01 3.99035394e-01
1.07285285e+00 -3.45669806e-01 -1.07118928e+00 5.53127229e-01
3.20146382e-01 -4.93344367e-01 1.28391206e+00 -9.91991758e-01
4.09841269e-01 -4.22171950e-01 -3.82802367e-01 -1.67631257e+00
-7.85372198e-01 -7.34282564e-03 9.50679779e-02 1.65605879e+00
1.37299627e-01 -3.11190903e-01 7.78358042e-01 3.72356385e-01
-1.63488358e-01 1.20567963e-01 -6.21140540e-01 -1.06508315e+00
9.31381226e-01 3.97084743e-01 1.09933352e+00 1.33347845e+00
-4.08611089e-01 2.19793290e-01 -1.16777346e-01 2.51643538e-01
6.43242538e-01 2.46527344e-01 5.69680452e-01 -1.40559125e+00
4.17271517e-02 -1.19481869e-02 -1.60352945e-01 -7.71226346e-01
6.43368185e-01 -1.11506522e+00 1.22001164e-01 -1.16523039e+00
1.19382314e-01 -4.55847681e-01 -7.67589152e-01 6.85894012e-01
-5.95854148e-02 -1.67111978e-02 7.67013207e-02 8.50426257e-02
-3.70205045e-01 5.73464751e-01 1.31897342e+00 -4.79686439e-01
-5.83160948e-03 -2.81932473e-01 -1.10850346e+00 7.22763062e-01
9.20009494e-01 -6.53459609e-01 -6.52888238e-01 -5.45563340e-01
-6.32493854e-01 -4.37262803e-01 2.92021275e-01 -9.60340202e-01
-1.32671461e-01 -4.16822433e-01 1.00119710e+00 -4.57513809e-01
7.11243413e-03 -1.31161177e+00 2.86072842e-03 4.74949330e-01
-4.30729419e-01 -3.95952076e-01 4.52903897e-01 7.02786565e-01
-3.23817044e-01 -1.41743734e-01 1.46374273e+00 -2.70019025e-01
-1.40760946e+00 3.29185426e-01 2.58158267e-01 2.03302845e-01
1.39577246e+00 -1.24080405e-01 -4.82843250e-01 2.23400116e-01
-8.69763315e-01 2.22428188e-01 5.37966073e-01 5.74155092e-01
3.82175237e-01 -1.78319943e+00 -7.17254043e-01 3.96357328e-01
5.41936457e-01 1.97568357e-01 3.98956358e-01 -1.48958191e-01
-2.86517411e-01 4.19278592e-01 -1.00609398e+00 -5.50716519e-01
-9.10837471e-01 1.05503643e+00 3.99017036e-01 -4.43154484e-01
-1.62120104e-01 1.08860195e+00 7.73069382e-01 -1.19501090e+00
4.15310115e-02 -1.52431577e-01 -4.82517108e-02 -1.72506899e-01
4.70206738e-01 1.59750357e-01 1.28556594e-01 -5.77154577e-01
-6.17381752e-01 3.75147313e-01 -3.96725714e-01 5.49386561e-01
1.16691184e+00 -1.88364238e-01 1.59216434e-01 -1.82753727e-02
1.32934165e+00 -5.24286628e-01 -1.64499807e+00 -6.31620407e-01
-6.54126033e-02 -3.85347247e-01 -3.00583869e-01 -1.29281616e+00
-9.87084389e-01 8.09513509e-01 1.09301579e+00 -1.29343852e-01
1.47591770e+00 -4.56046239e-02 8.64223182e-01 2.98865974e-01
2.92741507e-01 -1.16090512e+00 8.93720686e-02 7.08024621e-01
9.98092711e-01 -1.42888641e+00 -3.55026603e-01 -4.46971595e-01
-8.40430856e-01 7.07596719e-01 1.21235585e+00 -4.81918275e-01
6.63974047e-01 4.62336168e-02 -3.59716155e-02 1.80030093e-01
-5.21361053e-01 -1.61124170e-01 9.57457125e-02 1.57322288e+00
3.48748207e-01 3.54915522e-02 4.84936982e-02 1.08351648e+00
1.97337404e-01 5.52028194e-02 7.43939951e-02 9.54838395e-01
-4.41609174e-01 -1.30892181e+00 -3.17077786e-01 2.23072425e-01
-2.97904015e-01 2.57140268e-02 -7.01626837e-01 9.15788591e-01
5.44378042e-01 7.05820441e-01 4.92574833e-02 -2.41485536e-01
8.75084937e-01 3.71773064e-01 3.62758696e-01 -6.67275488e-01
-4.98903781e-01 -3.66391540e-01 -4.78424668e-01 -4.57322240e-01
-6.24928735e-02 -3.68657529e-01 -1.21237421e+00 -3.48447293e-01
-1.97179645e-01 -7.42602795e-02 3.55892032e-01 5.03631711e-01
6.96593344e-01 6.91330671e-01 5.99190116e-01 -3.16680759e-01
-1.13749552e+00 -8.98993075e-01 -9.50547755e-01 9.73912179e-01
2.85544127e-01 -8.13469052e-01 -1.35346860e-01 3.79156917e-01] | [10.355196952819824, 3.1032400131225586] |
a2c97185-e551-4e72-bb2c-e6578ad0ca45 | xer-an-explainable-model-for-entity | null | null | https://aclanthology.org/2021.trustnlp-1.5 | https://aclanthology.org/2021.trustnlp-1.5.pdf | xER: An Explainable Model for Entity Resolution using an Efficient Solution for the Clique Partitioning Problem | In this paper, we propose a global, self- explainable solution to solve a prominent NLP problem: Entity Resolution (ER). We formu- late ER as a graph partitioning problem. Every mention of a real-world entity is represented by a node in the graph, and the pairwise sim- ilarity scores between the mentions are used to associate these nodes to exactly one clique, which represents a real-world entity in the ER domain. In this paper, we use Clique Partition- ing Problem (CPP), which is an Integer Pro- gram (IP) to formulate ER as a graph partition- ing problem and then highlight the explainable nature of this method. Since CPP is NP-Hard, we introduce an efficient solution procedure, the xER algorithm, to solve CPP as a combi- nation of finding maximal cliques in the graph and then performing generalized set packing using a novel formulation. We discuss the advantages of using xER over the traditional methods and provide the computational exper- iments and results of applying this method to ER data sets. | ['Wen-mei Hwu', 'JinJun Xiong', 'Rakesh Nagi', 'Samhita Vadrevu'] | null | null | null | null | naacl-trustnlp-2021-6 | ['graph-partitioning', 'entity-resolution'] | ['graphs', 'natural-language-processing'] | [ 1.58212125e-01 9.70002353e-01 -3.32351893e-01 -2.93942750e-01
-6.78363979e-01 -7.85482407e-01 -1.20691448e-01 4.72295702e-01
-3.58418263e-02 1.26062167e+00 -3.03703725e-01 -2.69030660e-01
-5.25409520e-01 -1.03378296e+00 -7.81979501e-01 -1.91826150e-01
-4.55652297e-01 1.35720825e+00 3.87066305e-02 2.03703297e-03
-1.46395966e-01 5.39129078e-01 -1.12429357e+00 8.88608173e-02
1.24307394e+00 2.40187436e-01 2.45951727e-01 2.42243126e-01
-2.84403801e-01 4.02620375e-01 -7.29019284e-01 -6.60957277e-01
1.85968518e-01 -1.82605147e-01 -1.30361319e+00 2.91849196e-01
-1.15714908e-01 4.42202330e-01 -2.24588960e-01 1.08223939e+00
-6.04580343e-02 2.90614009e-01 7.07549751e-01 -1.81696665e+00
-5.42336762e-01 1.07356441e+00 -1.03015196e+00 -9.32927709e-03
8.08753610e-01 -9.55588996e-01 1.57174492e+00 -6.69538736e-01
1.16282320e+00 1.18207681e+00 4.50792849e-01 3.66918832e-01
-1.40755391e+00 -5.10607183e-01 2.66630322e-01 2.86636442e-01
-1.79691696e+00 1.66359231e-01 5.96979260e-01 -8.22463706e-02
1.10424757e+00 8.13495934e-01 5.78435659e-01 9.71465334e-02
5.12078144e-02 8.12260628e-01 8.56742740e-01 -4.57501471e-01
3.27700704e-01 4.95275892e-02 5.10694504e-01 5.36364913e-01
1.06278515e+00 -5.77705681e-01 -3.05411071e-01 -4.21236962e-01
5.84450722e-01 -2.83670634e-01 -2.67180920e-01 -4.04917300e-01
-8.93183589e-01 8.49864900e-01 4.96469617e-01 2.72810251e-01
-2.06519350e-01 1.00142039e-01 -1.00311078e-01 1.58494841e-02
3.82130951e-01 8.59791338e-01 -7.68151402e-01 5.90778410e-01
-7.99428463e-01 3.79983872e-01 1.32491028e+00 1.29455626e+00
8.95253658e-01 -5.30089021e-01 3.35755765e-01 4.25302148e-01
4.89715517e-01 1.45894885e-01 -3.15970957e-01 -7.49822915e-01
6.84182107e-01 9.11553681e-01 2.63880700e-01 -1.43005240e+00
-5.71069896e-01 -1.42191518e-02 -7.21791267e-01 -4.50415522e-01
1.56813040e-01 -5.46307899e-02 -8.82982314e-01 1.76860106e+00
6.51300669e-01 3.46405983e-01 2.81527549e-01 7.20478594e-01
1.17273843e+00 8.43923032e-01 1.93668678e-01 -8.03906024e-01
1.64872003e+00 -8.20581377e-01 -1.04981339e+00 -1.66388720e-01
7.29538679e-01 -2.45463535e-01 1.17599517e-01 2.21901566e-01
-1.04744828e+00 -2.20025796e-03 -9.44032729e-01 -1.84756905e-01
-5.03859103e-01 -1.91961184e-01 9.48320031e-01 4.13662106e-01
-8.97437692e-01 4.99096066e-01 -5.83747208e-01 -3.16119939e-01
2.58775149e-02 9.96759832e-01 -8.12541425e-01 -2.09195450e-01
-1.24126554e+00 7.90187478e-01 9.10588324e-01 -2.40906086e-02
5.87595105e-02 -4.75442320e-01 -1.06416559e+00 4.29358512e-01
8.83284748e-01 -7.45797098e-01 7.67314553e-01 -6.36681378e-01
-6.37654841e-01 1.12178791e+00 -6.51404858e-01 -3.74229372e-01
-1.17349185e-01 3.61742496e-01 -5.85307002e-01 4.78581935e-02
2.80393034e-01 3.90535593e-01 5.73032759e-02 -1.65742028e+00
-5.31438172e-01 -4.69289362e-01 3.30251187e-01 2.51966029e-01
2.08552480e-01 2.69380510e-01 -6.69996679e-01 -3.66158396e-01
6.94680095e-01 -1.10414672e+00 -6.30694211e-01 -9.23696399e-01
-1.00084865e+00 -5.67658663e-01 4.14714217e-01 -5.45430720e-01
1.75013351e+00 -1.79100525e+00 4.34157193e-01 8.41354668e-01
9.15320575e-01 -4.83615957e-02 6.47563487e-02 7.14148939e-01
-5.87428033e-01 6.30046964e-01 -3.96512449e-01 -2.58176029e-01
1.89362213e-01 5.95163822e-01 -1.14084125e-01 3.88685495e-01
-4.33504675e-03 9.05609012e-01 -8.91495943e-01 -6.63827837e-01
-1.93836138e-01 -1.32713869e-01 -3.96812230e-01 7.61350021e-02
-4.46771473e-01 6.74944893e-02 -5.56211829e-01 4.42167103e-01
1.12548423e+00 -5.38424969e-01 1.01256979e+00 1.44682061e-02
8.38569552e-02 4.25906628e-02 -1.85686278e+00 1.15704739e+00
2.07734734e-01 3.19255382e-01 1.05058879e-01 -9.93317187e-01
1.00491261e+00 2.17729628e-01 6.25571489e-01 -4.73833531e-02
-6.65405616e-02 3.51104945e-01 -1.63892373e-01 -3.98170352e-01
8.76620114e-01 2.14533240e-01 -3.68111581e-01 2.92475879e-01
-2.74670035e-01 3.48513350e-02 6.01628602e-01 7.16448843e-01
1.27615893e+00 -3.92322570e-01 9.58385766e-01 -3.56850713e-01
3.68939728e-01 4.14281994e-01 1.02070653e+00 6.22090459e-01
3.12855154e-01 4.72392559e-01 8.33110571e-01 -4.11708802e-01
-7.44917691e-01 -8.06165993e-01 -5.64769618e-02 5.45098424e-01
5.70958853e-01 -9.93331134e-01 -7.36562550e-01 -6.88079119e-01
2.32240781e-01 6.17918313e-01 -5.80777287e-01 4.86187518e-01
-6.38553143e-01 -8.39005291e-01 -3.95798497e-02 2.55955070e-01
6.85700253e-02 -9.00498569e-01 3.88654880e-03 3.33827943e-01
-5.84244728e-01 -1.28866470e+00 -1.69799596e-01 4.27324712e-01
-5.72015584e-01 -1.28248823e+00 5.80960000e-03 -1.04423690e+00
1.13111758e+00 2.82647580e-01 1.26231432e+00 3.58024389e-01
-2.14426279e-01 -1.27057256e-02 -3.60330641e-01 -2.22322062e-01
-1.65876113e-02 2.68797189e-01 3.46807763e-02 -4.12099928e-01
5.60936034e-01 -3.66691649e-01 -3.09093352e-02 3.79320681e-01
-7.30373204e-01 2.35575438e-01 2.27863774e-01 4.40242022e-01
1.17631924e+00 5.70984840e-01 5.42198062e-01 -1.68892038e+00
6.26678348e-01 -8.70874345e-01 -5.92284858e-01 7.58972704e-01
-5.11021852e-01 -3.78032662e-02 3.06976169e-01 -1.20786279e-01
-6.73211098e-01 4.83462632e-01 1.76606447e-01 -7.39022493e-02
1.26921814e-02 1.11661363e+00 -6.43932581e-01 1.06802389e-01
1.21308282e-01 -2.94637829e-01 -6.66647494e-01 -2.93051988e-01
4.85541910e-01 5.19208550e-01 4.12613302e-01 -6.28728867e-01
9.63908136e-01 2.30840281e-01 3.03616971e-01 -4.63487267e-01
-8.74203801e-01 -8.61489534e-01 -5.57578504e-01 1.44789115e-01
8.85188758e-01 -7.90547490e-01 -1.00017369e+00 -3.55586261e-01
-1.37994087e+00 2.71438628e-01 -2.07898200e-01 1.95835978e-01
-4.04566079e-01 5.12306273e-01 -4.38458443e-01 -9.44126725e-01
-7.60783479e-02 -8.31058741e-01 8.25147092e-01 3.87215495e-01
-3.00811887e-01 -7.68137217e-01 1.83804035e-01 5.37702918e-01
-3.90380383e-01 6.94366157e-01 9.60089743e-01 -1.23768067e+00
-8.31685424e-01 -6.82076290e-02 -5.28476775e-01 -7.08991528e-01
-1.04571514e-01 -8.56091231e-02 -3.74451250e-01 -3.68399382e-01
-3.83716136e-01 2.70400673e-01 4.75123435e-01 2.34307930e-01
1.19043696e+00 -4.66123462e-01 -1.00187409e+00 3.35474908e-01
1.65156484e+00 1.49485737e-01 7.89178133e-01 3.27692598e-01
6.22259438e-01 6.49127781e-01 1.00014830e+00 2.30122775e-01
7.80955017e-01 8.01496327e-01 2.93436915e-01 -3.85035276e-01
4.43545550e-01 -2.92015940e-01 -4.10019338e-01 7.18418241e-01
-4.84832734e-01 -8.43888581e-01 -8.61913860e-01 4.32236463e-01
-2.28561425e+00 -8.46585333e-01 -9.30509150e-01 1.90589988e+00
7.92707682e-01 -3.79390836e-01 7.84567297e-02 -2.17838287e-02
1.08867168e+00 -2.37939015e-01 -2.39101216e-01 -4.06225830e-01
-3.40864062e-01 2.10936368e-01 5.35377681e-01 6.28964841e-01
-1.10397899e+00 1.15419531e+00 6.40392494e+00 7.42427289e-01
-1.77357540e-01 -7.40719661e-02 4.57917184e-01 5.16852200e-01
-5.05071580e-01 3.41659218e-01 -8.16501439e-01 5.28427437e-02
7.96625674e-01 -6.77711010e-01 5.00118852e-01 7.89501309e-01
-9.34939831e-02 -2.65131146e-01 -1.04312134e+00 6.85055256e-01
-3.55039239e-02 -1.14889300e+00 -2.99832989e-02 3.78385484e-01
9.30312634e-01 -4.98862237e-01 -4.64086771e-01 1.25691429e-01
5.47362149e-01 -1.26072097e+00 9.55465659e-02 1.18362106e-01
8.34648252e-01 -1.03415060e+00 7.41527498e-01 3.30890268e-01
-1.68038642e+00 7.09530488e-02 -7.05459476e-01 1.81399509e-01
3.43481749e-01 6.75015390e-01 -1.04396951e+00 1.27122593e+00
5.24683952e-01 4.06097770e-01 -9.59321763e-03 1.15409899e+00
-3.73948783e-01 2.13404387e-01 -4.96169984e-01 1.57719567e-01
-3.71208072e-01 -4.65826690e-01 7.31986642e-01 1.17533386e+00
1.40443668e-01 1.04847884e+00 3.05407315e-01 8.49580228e-01
-5.51523209e-01 2.54714280e-01 -3.39355648e-01 -1.34270385e-01
8.91306043e-01 1.45111191e+00 -1.13447475e+00 -1.88591003e-01
-1.57782003e-01 8.02601933e-01 7.38615274e-01 3.83433551e-01
-9.85092461e-01 -5.20839810e-01 3.97025794e-01 5.62454090e-02
3.17485362e-01 -1.27033755e-01 -2.63586909e-01 -1.18865430e+00
-1.05969593e-01 -7.99448669e-01 7.68406868e-01 -8.35195720e-01
-1.12115395e+00 6.43616021e-01 2.69485831e-01 -7.20090330e-01
6.00499138e-02 -2.95313388e-01 -3.74981850e-01 6.97457731e-01
-1.31918526e+00 -8.93036783e-01 -7.63218701e-02 4.60378587e-01
1.39435334e-03 5.03106296e-01 9.15525615e-01 9.70095694e-02
-6.33875251e-01 4.67178613e-01 -1.77865967e-01 -2.70397030e-02
2.67262250e-01 -1.67032659e+00 2.88219810e-01 8.27569008e-01
2.72616595e-01 1.04561722e+00 1.03026855e+00 -1.12166488e+00
-1.44100332e+00 -1.04916561e+00 1.62716138e+00 -2.66355842e-01
5.46633661e-01 -4.69671309e-01 -9.33764696e-01 1.17537975e+00
5.83229810e-02 -6.54206127e-02 9.17267382e-01 5.77758670e-01
-1.07240662e-01 3.12020034e-01 -1.49267840e+00 3.67959201e-01
1.25951874e+00 -5.13757346e-03 -9.79983985e-01 6.62036002e-01
1.06508732e+00 -7.74142325e-01 -1.22121131e+00 3.67098153e-01
-1.52439368e-03 -3.14253360e-01 9.63439941e-01 -8.04473162e-01
1.53621942e-01 -6.46777630e-01 -1.60721899e-03 -1.05617869e+00
-4.95211571e-01 -8.38717699e-01 -3.44632477e-01 1.26438642e+00
8.25826108e-01 -6.69319808e-01 9.31421816e-01 1.01382339e+00
2.03438103e-01 -8.98629725e-01 -1.06608391e+00 -6.35910749e-01
-3.19601029e-01 -4.60747592e-02 9.29683030e-01 1.45305932e+00
6.17391884e-01 5.02135754e-01 -2.51453757e-01 8.38605821e-01
7.26260185e-01 6.64803207e-01 6.88101888e-01 -1.68816662e+00
-2.67380953e-01 3.15064043e-01 -3.05936992e-01 -8.65397871e-01
3.87676686e-01 -1.05852854e+00 -1.18991747e-01 -1.99865294e+00
6.10963285e-01 -7.39064753e-01 1.60594851e-01 7.05300629e-01
-2.73671359e-01 -2.88967807e-02 3.79453033e-01 2.27157757e-01
-1.09558451e+00 -1.43132340e-02 1.01927376e+00 1.39122158e-02
-5.52177787e-01 -6.93978593e-02 -8.84481668e-01 4.83139813e-01
4.64796960e-01 -8.67323458e-01 -1.10065401e-01 -2.00554729e-02
6.65072143e-01 5.78162432e-01 -2.79213756e-01 -3.30391407e-01
4.33020413e-01 -4.37640786e-01 2.24688947e-02 -9.16895509e-01
1.55482009e-01 -1.06148267e+00 1.01845014e+00 2.84350097e-01
-9.43376347e-02 1.29504323e-01 1.02409124e-01 7.09810197e-01
-1.99930936e-01 -3.98945540e-01 2.95054168e-01 4.44693565e-02
-7.11372256e-01 3.64494592e-01 -3.33640389e-02 1.04903683e-01
1.48262620e+00 -3.35759342e-01 -3.70753229e-01 -1.94721352e-02
-1.20331478e+00 7.31786132e-01 4.07659113e-01 -5.08699194e-02
5.48042834e-01 -8.41637731e-01 -8.25769484e-01 -3.04873854e-01
4.44331095e-02 3.29307616e-01 1.15119249e-01 4.19504493e-01
-5.04269242e-01 4.03817564e-01 3.50379407e-01 -1.18537799e-01
-1.73762000e+00 8.77798080e-01 -2.29591988e-02 -7.79636621e-01
-4.18128997e-01 7.32390821e-01 2.68956870e-01 -5.62545419e-01
-1.56271070e-01 9.68317315e-02 -6.23120666e-01 -1.89250976e-01
1.50628671e-01 3.21768254e-01 -2.28207424e-01 -7.35475898e-01
-6.11105919e-01 2.83447742e-01 -1.62996188e-01 3.29580873e-01
1.61534715e+00 -2.35662371e-01 -8.33884120e-01 -2.66121072e-03
7.20074594e-01 2.90752560e-01 -2.01112673e-01 -8.94044936e-02
2.87199289e-01 -3.50641727e-01 -4.38799173e-01 -6.61154330e-01
-9.81324494e-01 -2.03673571e-01 -3.62840146e-01 6.72871590e-01
1.02144337e+00 6.46724343e-01 6.37356043e-01 3.90176356e-01
6.39341891e-01 -9.21935618e-01 -7.16262817e-01 4.06454146e-01
7.12877870e-01 -9.00412738e-01 3.91786903e-01 -1.59865749e+00
-5.81418753e-01 9.39743996e-01 6.35141790e-01 -9.37365666e-02
4.79838461e-01 2.72911280e-01 -6.65404320e-01 -4.49575394e-01
-6.05444789e-01 -3.11622798e-01 2.95125127e-01 4.05330420e-01
9.21591818e-02 5.13744891e-01 -1.01688218e+00 1.04055333e+00
-3.30958039e-01 -1.75892606e-01 7.00138450e-01 6.94713116e-01
-4.81685817e-01 -1.27331960e+00 -3.53686750e-01 3.88297200e-01
-5.87280154e-01 3.26616243e-02 -8.96869898e-01 1.06356275e+00
2.76875585e-01 1.19334316e+00 -5.97800352e-02 -2.91317046e-01
2.22352847e-01 -5.63111246e-01 3.66520792e-01 -9.27165747e-01
-3.44756901e-01 -1.60796911e-01 5.92829168e-01 -3.49667460e-01
-5.51804483e-01 -4.84790206e-01 -1.93911612e+00 -5.66351950e-01
-9.24021840e-01 7.94827402e-01 4.50984031e-01 1.05285060e+00
2.45498627e-01 3.71889412e-01 5.37846744e-01 -3.02776873e-01
5.85423820e-02 -3.38146627e-01 -1.13314128e+00 2.53373027e-01
-2.22933382e-01 -5.89110970e-01 -5.16501606e-01 -3.11077327e-01] | [7.0887885093688965, 5.337768077850342] |
1d755f68-9143-447e-82f7-1cb278c04412 | ssd-single-shot-multibox-detector | 1512.02325 | null | http://arxiv.org/abs/1512.02325v5 | http://arxiv.org/pdf/1512.02325v5.pdf | SSD: Single Shot MultiBox Detector | We present a method for detecting objects in images using a single deep
neural network. Our approach, named SSD, discretizes the output space of
bounding boxes into a set of default boxes over different aspect ratios and
scales per feature map location. At prediction time, the network generates
scores for the presence of each object category in each default box and
produces adjustments to the box to better match the object shape. Additionally,
the network combines predictions from multiple feature maps with different
resolutions to naturally handle objects of various sizes. Our SSD model is
simple relative to methods that require object proposals because it completely
eliminates proposal generation and subsequent pixel or feature resampling stage
and encapsulates all computation in a single network. This makes SSD easy to
train and straightforward to integrate into systems that require a detection
component. Experimental results on the PASCAL VOC, MS COCO, and ILSVRC datasets
confirm that SSD has comparable accuracy to methods that utilize an additional
object proposal step and is much faster, while providing a unified framework
for both training and inference. Compared to other single stage methods, SSD
has much better accuracy, even with a smaller input image size. For $300\times
300$ input, SSD achieves 72.1% mAP on VOC2007 test at 58 FPS on a Nvidia Titan
X and for $500\times 500$ input, SSD achieves 75.1% mAP, outperforming a
comparable state of the art Faster R-CNN model. Code is available at
https://github.com/weiliu89/caffe/tree/ssd . | ['Cheng-Yang Fu', 'Christian Szegedy', 'Wei Liu', 'Scott Reed', 'Dumitru Erhan', 'Dragomir Anguelov', 'Alexander C. Berg'] | 2015-12-08 | null | null | null | null | ['lidar-semantic-segmentation', 'surgical-tool-detection'] | ['computer-vision', 'computer-vision'] | [ 1.08210770e-02 -2.54739136e-01 -3.28592658e-02 -5.93914211e-01
-6.78719401e-01 -6.15865171e-01 1.96285963e-01 -2.09802866e-01
-6.70670331e-01 2.34567493e-01 -3.82237405e-01 -3.76337171e-01
5.60735464e-01 -8.02847028e-01 -9.76746798e-01 -3.13789189e-01
2.38942549e-01 2.28738949e-01 9.35607731e-01 1.61014661e-01
6.64022788e-02 7.83065856e-01 -1.80044901e+00 4.97844428e-01
3.68435651e-01 1.33823729e+00 4.03834581e-01 9.34220076e-01
1.50330231e-01 6.44397199e-01 -8.00271392e-01 -3.06775600e-01
6.12120032e-01 2.39226595e-01 -4.85574365e-01 -8.34391564e-02
1.20513201e+00 -7.67158449e-01 -4.66078639e-01 9.64213133e-01
5.12257218e-01 1.39590397e-01 3.08624566e-01 -1.12568235e+00
-5.98741353e-01 2.83275485e-01 -7.08734214e-01 3.11079025e-01
-7.50651285e-02 4.48443860e-01 8.97224605e-01 -1.32084239e+00
4.77588207e-01 1.15601754e+00 7.97430158e-01 5.42942226e-01
-1.32252693e+00 -8.81761909e-01 -3.81836034e-02 -6.45613596e-02
-1.55372989e+00 -3.27601254e-01 1.75253361e-01 -2.51913011e-01
1.41127312e+00 1.90751851e-01 4.79753345e-01 5.75943291e-01
1.63390920e-01 9.12602365e-01 8.13477755e-01 5.86954597e-03
3.98070440e-02 1.69729903e-01 1.27195418e-01 7.36598969e-01
3.05116296e-01 4.16756608e-02 -3.60648632e-01 1.99012578e-01
1.06658816e+00 2.45420948e-01 -4.03844677e-02 -3.90292078e-01
-1.12850714e+00 7.53770709e-01 9.64655936e-01 -3.14671427e-01
-1.16761826e-01 5.74427605e-01 4.30476665e-01 -1.25428259e-01
1.56729385e-01 3.02384704e-01 -5.69291353e-01 1.37196735e-01
-1.00284290e+00 3.25093657e-01 5.08327425e-01 1.16296303e+00
6.68411851e-01 2.05811471e-01 -1.54563129e-01 6.86920345e-01
1.06589295e-01 6.03957474e-01 2.42220357e-01 -1.22382247e+00
4.29946810e-01 5.86220920e-01 1.93423659e-01 -7.90647566e-01
-4.20807868e-01 -5.17163157e-01 -4.90739971e-01 7.23694324e-01
4.68505859e-01 -4.85166684e-02 -1.06662428e+00 1.33359325e+00
3.83032203e-01 -8.34999532e-02 -2.77098209e-01 9.86402810e-01
1.14776289e+00 6.98536456e-01 1.56248081e-02 5.85450888e-01
1.43774414e+00 -1.20194852e+00 -1.15161613e-01 -4.50680226e-01
4.04316276e-01 -7.67139256e-01 1.18826866e+00 5.39789677e-01
-1.14350009e+00 -1.13264453e+00 -1.27736545e+00 -5.99523783e-01
-6.04479074e-01 7.48664856e-01 7.40095079e-01 6.07238352e-01
-1.35249269e+00 6.60955369e-01 -9.07668829e-01 -2.32772291e-01
7.11403668e-01 6.89396799e-01 -1.05652176e-01 7.40675852e-02
-5.93346059e-01 7.48319387e-01 4.11714494e-01 4.68835644e-02
-6.96883500e-01 -7.19625473e-01 -9.28124249e-01 1.57325685e-01
2.26874918e-01 -3.41589242e-01 1.47396159e+00 -8.93950939e-01
-1.16989386e+00 7.63260007e-01 -1.42229483e-01 -6.90294027e-01
5.98528564e-01 -3.68919253e-01 -2.10528687e-01 -2.26024613e-02
9.07431915e-02 1.56263947e+00 8.74008238e-01 -7.74953008e-01
-8.68008077e-01 -2.17657954e-01 9.87640470e-02 1.59187689e-01
-1.47871375e-02 1.82579339e-01 -8.67505133e-01 -4.13118809e-01
-8.69251862e-02 -9.04255509e-01 -1.88939169e-01 7.30768800e-01
-2.76490748e-01 -1.96786016e-01 9.76612210e-01 -3.81486684e-01
8.89661431e-01 -2.22847509e+00 -4.66118127e-01 -7.74821043e-02
2.13117316e-01 5.06566644e-01 -2.27562860e-01 -2.31646046e-01
1.66171119e-02 -1.82921998e-02 9.33948681e-02 -5.21316588e-01
-1.43827498e-01 -9.25351009e-02 -3.39601606e-01 4.58136618e-01
4.30288792e-01 8.60949337e-01 -5.42495489e-01 -3.45326006e-01
5.28670847e-01 7.63823986e-01 -7.20890641e-01 -8.86715725e-02
-1.17945589e-01 -2.52185643e-01 -4.96949954e-03 8.23720634e-01
9.35042500e-01 -4.95764285e-01 -1.34492308e-01 -3.93609494e-01
-2.93261141e-01 3.43916178e-01 -1.43652368e+00 1.39978242e+00
-2.32114017e-01 9.31018353e-01 -5.36247343e-03 -4.07532901e-01
1.10723567e+00 -2.56642580e-01 4.56881784e-02 -5.20579398e-01
2.75587082e-01 2.06757426e-01 -4.17495668e-02 1.33146018e-01
8.90555561e-01 5.22952437e-01 7.58613497e-02 2.62522340e-01
8.37714747e-02 -3.70332479e-01 3.10091287e-01 1.00113124e-01
1.16146624e+00 3.56470346e-01 2.32692972e-01 3.94736324e-03
2.22135961e-01 7.62952343e-02 4.25111532e-01 9.12519276e-01
-4.66300011e-01 8.83327603e-01 2.34464109e-01 -8.43887269e-01
-1.18634486e+00 -1.33701396e+00 -5.15379667e-01 1.29802907e+00
2.78178960e-01 -2.08183259e-01 -6.00337982e-01 -7.13955998e-01
2.87601829e-01 5.00810146e-01 -5.66369236e-01 1.70300975e-01
-6.32744133e-01 -6.33186519e-01 4.81986821e-01 1.23465562e+00
6.74360573e-01 -1.22374713e+00 -1.13667190e+00 5.45700230e-02
3.12274277e-01 -1.18557870e+00 -5.56698024e-01 4.24867272e-01
-8.64980876e-01 -1.03217685e+00 -3.76711220e-01 -8.37319732e-01
7.62744248e-01 4.43882555e-01 1.09131515e+00 -3.36612365e-03
-7.41237938e-01 -5.27123101e-02 -6.12844154e-02 -6.07882738e-01
3.39295827e-02 -3.56981680e-02 -9.83796716e-02 -6.06995523e-01
7.01753199e-01 1.35819316e-01 -8.80100846e-01 5.05135953e-01
-7.20928669e-01 1.57826379e-01 5.57333291e-01 7.26247489e-01
8.57324064e-01 -2.59769917e-01 2.65876532e-01 -5.64084351e-01
1.41135724e-02 -1.71979107e-02 -1.05308437e+00 -1.60682037e-01
-2.72315085e-01 -2.04784364e-01 6.23885393e-01 -4.29396182e-01
-8.72670650e-01 6.21426523e-01 -8.39192793e-02 -5.08689046e-01
-2.91576385e-01 -3.64764869e-01 2.79608071e-01 -1.92705840e-01
8.48540664e-01 1.01085767e-01 3.53976004e-02 -3.23801458e-01
4.12352294e-01 6.23049676e-01 6.81159079e-01 -4.49494869e-02
7.58640885e-01 6.50797307e-01 -3.58590513e-01 -4.86077547e-01
-6.42277718e-01 -5.62815905e-01 -7.36742675e-01 -9.28625837e-02
8.37125719e-01 -1.22432697e+00 -7.90150583e-01 4.14486825e-01
-1.22667885e+00 -5.95907509e-01 -4.18903977e-01 3.83390784e-01
-2.99390435e-01 -1.88950151e-01 -7.59639025e-01 -4.31554586e-01
-5.29608369e-01 -1.21404624e+00 1.27905238e+00 5.18642366e-01
-2.36426353e-01 -3.66536975e-01 -3.33669752e-01 2.50459075e-01
4.42859769e-01 -1.18671969e-01 1.99502870e-01 -4.88822997e-01
-8.66072476e-01 -3.86750549e-01 -8.34447503e-01 5.60417593e-01
-9.86533388e-02 3.31325889e-01 -1.17408538e+00 -3.93986017e-01
-4.28384811e-01 -4.97254491e-01 1.03488934e+00 5.18224835e-01
1.60658109e+00 7.64484853e-02 -3.16514462e-01 7.49152541e-01
1.51152098e+00 2.46362507e-01 5.95124125e-01 2.84500927e-01
5.95042944e-01 2.22398236e-01 7.12755620e-01 3.46791953e-01
2.90363163e-01 5.90672433e-01 5.30992687e-01 -4.99401748e-01
-3.18378091e-01 6.25370964e-02 3.33534122e-01 -7.18393875e-03
1.10530108e-01 -1.16304353e-01 -8.49116802e-01 3.84570271e-01
-1.41183197e+00 -7.97698081e-01 -1.53478742e-01 2.22103882e+00
6.86235011e-01 4.18380111e-01 1.85298026e-01 -2.90125281e-01
7.36290574e-01 1.25459194e-01 -8.02529037e-01 -5.24634540e-01
-7.11531052e-03 4.37061161e-01 9.29050446e-01 2.73260057e-01
-1.39183140e+00 1.10806894e+00 6.29802847e+00 6.32125914e-01
-1.28321326e+00 -5.06793745e-02 8.70736301e-01 -5.19290149e-01
3.93229127e-01 -4.29582506e-01 -1.37689114e+00 3.34416777e-01
6.84162974e-01 3.62467170e-01 2.71408081e-01 1.42213225e+00
7.86815062e-02 -3.10691983e-01 -1.20134211e+00 8.88394594e-01
9.45871025e-02 -1.41025984e+00 -2.90011525e-01 -1.45833924e-01
6.33598864e-01 6.90065444e-01 2.75771320e-01 3.46808583e-01
4.78600442e-01 -1.02251792e+00 9.81462479e-01 4.95387875e-02
1.05960274e+00 -6.49310827e-01 7.83233047e-01 4.99479324e-02
-1.39733732e+00 -1.55398622e-01 -7.41835535e-01 -4.73575033e-02
-2.41951063e-01 1.65213093e-01 -1.18697655e+00 -2.51094013e-01
1.27020204e+00 4.18492168e-01 -9.95107651e-01 1.10248816e+00
-1.58849180e-01 2.11750016e-01 -5.28032184e-01 -2.04779401e-01
6.56636059e-02 3.55071157e-01 1.54036760e-01 1.44355714e+00
2.63667494e-01 -8.59085321e-02 4.29506935e-02 1.09765899e+00
-2.18290776e-01 -1.07679762e-01 -3.98564994e-01 4.71143782e-01
7.13839531e-01 1.42800391e+00 -9.72630918e-01 -5.55156946e-01
-5.33180654e-01 9.48991954e-01 4.35612798e-01 1.25918701e-01
-1.24277604e+00 -6.94311321e-01 6.31497920e-01 1.25649676e-01
9.28528786e-01 -2.79820234e-01 -5.07016242e-01 -8.53505671e-01
1.26930013e-01 -5.88843286e-01 1.92401916e-01 -1.03220201e+00
-8.42923760e-01 6.90425098e-01 -2.66383618e-01 -1.27240503e+00
1.93253960e-02 -1.22941422e+00 -5.58525324e-01 7.60152221e-01
-1.32706630e+00 -9.59996521e-01 -6.16086483e-01 3.28568488e-01
7.78202236e-01 1.80844933e-01 6.10720336e-01 1.91326171e-01
-7.53239334e-01 6.96591020e-01 -3.63093726e-02 3.91233623e-01
7.16873288e-01 -1.41008413e+00 9.76621568e-01 8.31140101e-01
8.13063011e-02 6.48319840e-01 3.43247294e-01 -5.65523982e-01
-1.19441330e+00 -1.44706178e+00 5.25553823e-01 -6.40965104e-01
4.45621371e-01 -5.21754265e-01 -7.90021777e-01 6.67088151e-01
3.19045223e-02 6.18070006e-01 3.49372268e-01 -1.60240516e-01
-5.87952554e-01 -3.62493545e-01 -1.18174434e+00 4.91922498e-01
9.17567909e-01 -2.98593879e-01 -1.91721320e-01 2.76352316e-01
6.53801858e-01 -8.79497886e-01 -6.54292285e-01 4.62384880e-01
7.22389996e-01 -1.06303549e+00 1.09700787e+00 -1.31435692e-01
2.64016390e-01 -6.66651249e-01 -1.77583069e-01 -7.31789351e-01
-4.20282781e-01 -1.16475746e-01 -1.14171989e-01 7.43979096e-01
5.71588278e-01 -5.17322898e-01 1.00214601e+00 6.62311196e-01
-1.68067664e-01 -8.92604768e-01 -7.11913764e-01 -7.34155357e-01
-1.46240309e-01 -5.74070573e-01 5.35559833e-01 2.62937814e-01
-5.93254089e-01 -1.78506374e-02 7.40535930e-02 2.38570064e-01
5.05846560e-01 1.08968690e-01 1.05457914e+00 -8.59329760e-01
-3.23621362e-01 -5.47967672e-01 -4.70831096e-01 -1.20915484e+00
-3.08205664e-01 -5.33128321e-01 2.65752137e-01 -1.35207701e+00
7.29654506e-02 -5.71125388e-01 -1.73972607e-01 8.57322812e-01
-3.50411027e-03 9.84489262e-01 3.35609913e-01 1.47746801e-01
-7.16737032e-01 1.21933483e-01 8.82789314e-01 -6.82409331e-02
-2.19160214e-01 -3.72905470e-02 -5.36200404e-01 9.25991952e-01
9.20375526e-01 -3.03898245e-01 -8.29810575e-02 -5.63187361e-01
-5.87565191e-02 -4.81398225e-01 6.50596678e-01 -1.47191501e+00
2.36361295e-01 1.24398813e-01 1.10497355e+00 -1.06582010e+00
5.41542292e-01 -5.74488103e-01 -1.00200571e-01 5.95197380e-01
-2.32888117e-01 1.64567217e-01 6.18901432e-01 2.42337599e-01
8.12784359e-02 -1.35260797e-03 1.07251000e+00 5.74810430e-04
-1.01969182e+00 2.85249084e-01 -1.29755482e-01 -2.50973761e-01
1.13636804e+00 -6.02646828e-01 -5.76030254e-01 1.03666104e-01
-4.44110334e-01 1.51237145e-01 5.54341853e-01 6.17524266e-01
7.46996701e-01 -1.11776733e+00 -4.10127729e-01 3.77178967e-01
-3.97210009e-03 3.51033658e-01 2.19629019e-01 4.77644354e-01
-9.27156568e-01 4.55021292e-01 -1.72664553e-01 -8.56837153e-01
-1.38067508e+00 3.25535268e-01 5.19227564e-01 1.62937269e-01
-5.77752173e-01 1.25803578e+00 4.29301202e-01 -4.14163619e-01
4.69067752e-01 -6.69667125e-01 6.48432970e-03 -2.32345000e-01
7.95019746e-01 4.04889137e-01 1.83762059e-01 -5.30153632e-01
-4.04893190e-01 5.08988082e-01 -2.81793088e-01 1.22690864e-01
1.05330646e+00 2.29579642e-01 2.92999655e-01 5.34416176e-02
1.13004565e+00 -1.53975859e-01 -1.81488371e+00 -1.59813032e-01
-4.86496717e-01 -5.40183663e-01 9.63459164e-02 -8.43206048e-01
-1.15921366e+00 8.45496058e-01 8.68000388e-01 -9.71529931e-02
9.20654297e-01 6.59865811e-02 5.21935046e-01 5.29166043e-01
2.31190547e-01 -1.13796377e+00 1.99879661e-01 5.71966469e-01
8.63761544e-01 -1.39232850e+00 2.09952399e-01 -3.97822320e-01
-5.62464774e-01 1.20204556e+00 1.32669365e+00 -4.27867025e-01
3.40990514e-01 6.55904233e-01 4.37900759e-02 -1.19092658e-01
-6.09915137e-01 -3.76218855e-02 4.29337680e-01 4.35902894e-01
4.07432377e-01 1.76282123e-01 2.47667357e-01 2.84695268e-01
-3.66494954e-01 -1.65321573e-03 5.18445969e-01 8.65331292e-01
-6.80764019e-01 -5.40713370e-01 -4.09453273e-01 6.64299548e-01
-4.66495603e-01 -2.16470450e-01 -9.12295431e-02 9.43303406e-01
3.82389963e-01 6.02892578e-01 7.06998289e-01 -3.76603633e-01
2.76599586e-01 -2.80160725e-01 3.00092101e-01 -8.20725977e-01
-6.86532438e-01 -6.46448806e-02 -1.25185102e-01 -1.02263761e+00
-2.97892350e-03 -6.96166575e-01 -1.47490573e+00 -2.57738501e-01
-5.86494207e-01 -4.88565087e-01 8.58503938e-01 3.53615820e-01
5.26212394e-01 5.05441010e-01 3.67169231e-01 -1.20658720e+00
-5.72626054e-01 -7.89497256e-01 -3.19791883e-01 -7.04142302e-02
1.73616007e-01 -5.12589395e-01 -2.05289572e-01 1.93057433e-02] | [8.813780784606934, -0.060054875910282135] |
46a1f327-8eef-4165-8f71-79cef3d1f752 | a-comprehensive-survey-on-multi-hop-machine | 2212.04070 | null | https://arxiv.org/abs/2212.04070v1 | https://arxiv.org/pdf/2212.04070v1.pdf | A Comprehensive Survey on Multi-hop Machine Reading Comprehension Datasets and Metrics | Multi-hop Machine reading comprehension is a challenging task with aim of answering a question based on disjoint pieces of information across the different passages. The evaluation metrics and datasets are a vital part of multi-hop MRC because it is not possible to train and evaluate models without them, also, the proposed challenges by datasets often are an important motivation for improving the existing models. Due to increasing attention to this field, it is necessary and worth reviewing them in detail. This study aims to present a comprehensive survey on recent advances in multi-hop MRC evaluation metrics and datasets. In this regard, first, the multi-hop MRC problem definition will be presented, then the evaluation metrics based on their multi-hop aspect will be investigated. Also, 15 multi-hop datasets have been reviewed in detail from 2017 to 2022, and a comprehensive analysis has been prepared at the end. Finally, open issues in this field have been discussed. | ['Ahmad Baraani', 'Reza Ramezani', 'Azade Mohammadi'] | 2022-12-08 | null | null | null | null | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 3.16360146e-01 1.82886839e-01 -2.52423257e-01 -2.67352253e-01
-1.35058415e+00 -5.83154321e-01 2.11930469e-01 7.33909667e-01
-6.48897052e-01 1.06008041e+00 5.26493847e-01 -3.38825017e-01
-7.28178442e-01 -5.71700871e-01 -6.54344440e-01 -3.39237034e-01
-1.28902672e-02 8.44313875e-02 3.34773451e-01 -6.14363551e-01
6.93679214e-01 -8.09202641e-02 -1.93733764e+00 5.53785443e-01
1.38033724e+00 7.53131032e-01 5.04258335e-01 1.25237799e+00
-7.80203715e-02 9.59772825e-01 -9.87205029e-01 -4.84958619e-01
-4.51118290e-01 -8.20014775e-01 -1.44306910e+00 -4.90780771e-01
6.89081252e-01 -1.08952403e-01 -1.72893807e-01 6.33048236e-01
8.55405092e-01 4.97230560e-01 5.23753941e-01 -9.21989620e-01
-7.14571595e-01 6.86714113e-01 -1.76198676e-01 5.62274814e-01
1.15790701e+00 -3.72476190e-01 1.34056973e+00 -6.49526715e-01
3.66964668e-01 8.39139223e-01 4.78249043e-01 4.71349567e-01
-3.45602244e-01 -7.16726994e-03 3.99445236e-01 1.13447797e+00
-9.76756811e-01 -4.27963436e-02 6.20260894e-01 -3.43663655e-02
8.43905747e-01 8.21719587e-01 4.98232275e-01 1.37663293e+00
1.94261074e-02 1.15626216e+00 1.16972578e+00 -8.41319799e-01
-1.50469750e-01 -1.28843829e-01 7.97708571e-01 3.99489194e-01
-3.49470936e-02 -5.09022832e-01 -7.26595521e-01 2.36030951e-01
-2.42227793e-01 -6.10559523e-01 -7.68115342e-01 3.16583961e-01
-1.22609317e+00 7.43175864e-01 1.18615106e-01 6.55778289e-01
4.49856892e-02 -4.58734393e-01 1.99551478e-01 6.58518672e-01
7.99550414e-02 7.38775194e-01 -6.67412102e-01 -5.18983066e-01
-8.55530024e-01 4.68510985e-01 9.64610815e-01 1.08238232e+00
8.89743939e-02 -6.07977688e-01 -6.92302823e-01 9.68230069e-01
2.30355258e-03 3.40369076e-01 4.74568635e-01 -8.61787021e-01
9.48306203e-01 4.24244165e-01 -1.25574335e-01 -9.32119012e-01
-7.72709429e-01 -6.59920990e-01 -8.80904675e-01 -5.51216841e-01
4.69057113e-01 -1.37263894e-01 -2.71526933e-01 1.45443869e+00
8.57174918e-02 -1.16865315e-01 1.77228704e-01 6.75744772e-01
1.64458799e+00 5.88817954e-01 -1.20274700e-01 -3.63585562e-01
1.54463315e+00 -1.42387581e+00 -1.12183249e+00 2.50937700e-01
6.81301236e-01 -9.86250758e-01 1.39771807e+00 4.02120560e-01
-1.36679590e+00 -7.45189607e-01 -1.11622238e+00 -6.04400694e-01
-5.28798878e-01 1.99496727e-02 2.78255455e-02 4.43179160e-01
-8.27593148e-01 3.17567080e-01 -1.76809445e-01 -3.09241951e-01
-6.27991837e-03 -2.59486049e-01 -4.59929220e-02 -3.57270688e-01
-1.80073273e+00 1.10391569e+00 4.23744559e-01 4.90690172e-02
-6.16576314e-01 -8.09110045e-01 -5.91173351e-01 -1.36339515e-01
1.76389560e-01 -6.20578468e-01 1.61458707e+00 -2.41875887e-01
-1.42867494e+00 5.66794217e-01 -2.46617541e-01 -2.85065353e-01
6.29315794e-01 -5.08156598e-01 -7.29744494e-01 5.49480200e-01
4.66492623e-02 3.14551681e-01 3.17710668e-01 -8.85401368e-01
-9.97931004e-01 4.47602645e-02 5.96868873e-01 5.91718137e-01
-3.00260067e-01 1.24740005e-01 -4.58855957e-01 -4.95127827e-01
-2.69234687e-01 -3.05641204e-01 2.98274606e-01 -7.94505358e-01
-5.89219034e-01 -6.06771588e-01 4.78795677e-01 -8.78965497e-01
1.95609689e+00 -1.59825289e+00 3.18567634e-01 -4.12804604e-01
3.45578283e-01 2.21983716e-01 -2.07413837e-01 9.12490129e-01
3.00763935e-01 1.80035815e-01 -4.16500598e-01 -1.04715213e-01
-6.36524484e-02 -1.63258612e-01 -3.30888405e-02 2.13097543e-01
-3.18792500e-02 8.90176475e-01 -9.78413701e-01 -5.79929471e-01
5.09870355e-04 2.89192915e-01 -1.06816985e-01 2.56882906e-01
-2.95598060e-01 6.16735756e-01 -5.42737067e-01 6.86082244e-01
5.83309352e-01 -3.51405054e-01 -3.51084411e-01 1.95525177e-02
-2.71801412e-01 5.27516961e-01 -7.57483423e-01 1.85736585e+00
-4.46269214e-01 7.83238411e-01 -4.15508837e-01 -1.03295052e+00
4.68961716e-01 5.03148973e-01 2.72106230e-01 -1.04114115e+00
2.73335278e-02 3.08442593e-01 8.10733587e-02 -9.24455464e-01
7.61357307e-01 4.52236682e-01 -1.27044603e-01 3.66932958e-01
-1.42463803e-01 8.43933523e-02 7.79754162e-01 2.22390890e-01
1.16518009e+00 -1.67549029e-01 2.39153042e-01 -8.61912444e-02
1.22345543e+00 1.53501958e-01 3.39420065e-02 9.29554880e-01
-2.26251692e-01 8.80981803e-01 3.46564531e-01 4.67389785e-02
-6.69178247e-01 -9.13525403e-01 -4.64828938e-01 1.14486599e+00
1.95642382e-01 -2.94411331e-01 -9.44603860e-01 -8.66725385e-01
-4.22050983e-01 9.34381604e-01 -7.54291594e-01 -8.79922807e-02
-7.20229506e-01 -7.74328113e-01 7.26133704e-01 2.81505287e-01
9.53261197e-01 -9.99412239e-01 -6.50298715e-01 1.59276977e-01
-1.23323190e+00 -1.23025298e+00 -9.85500216e-02 -1.49133950e-01
-6.41433239e-01 -1.40485811e+00 -1.07673311e+00 -1.15452635e+00
2.02853680e-01 6.26878798e-01 1.55339885e+00 4.39680308e-01
-9.75619480e-02 8.16666722e-01 -1.25225532e+00 -7.00654745e-01
-1.07016981e-01 8.07279527e-01 -4.00884598e-01 -4.17588770e-01
5.20580292e-01 1.29762208e-02 -6.92196131e-01 3.88718754e-01
-7.17387438e-01 1.51261585e-02 6.19798660e-01 5.53150892e-01
5.87835193e-01 -1.63647011e-01 1.21511829e+00 -6.87148392e-01
1.07433426e+00 -7.25467503e-01 -2.16728151e-02 8.71061087e-01
-3.84982973e-01 -4.01596278e-01 6.02378786e-01 -1.80861443e-01
-1.04628527e+00 -8.93839717e-01 -5.06466806e-01 5.77137232e-01
-7.18054771e-02 7.33056545e-01 -9.90926623e-02 6.45948425e-02
6.39503777e-01 1.07771754e-01 -3.90945673e-01 -8.05471361e-01
4.63316023e-01 8.86806428e-01 2.56779701e-01 -5.26120305e-01
4.61102784e-01 -1.43996388e-01 -2.60290772e-01 -1.19892550e+00
-1.70770085e+00 -6.37849867e-01 -6.83447003e-01 -5.31603992e-01
1.14265835e+00 -5.89732587e-01 -4.90682393e-01 5.66252768e-01
-1.17099988e+00 -4.99828681e-02 -1.33088395e-01 6.76035821e-01
-5.65800905e-01 4.13565487e-01 -6.08093739e-01 -4.43303376e-01
-2.74841517e-01 -1.05689418e+00 6.70095801e-01 5.69526911e-01
-1.26104429e-01 -1.23508835e+00 4.05290872e-02 1.23928630e+00
2.92613417e-01 2.27912530e-01 9.94880974e-01 -5.75543284e-01
-3.78691435e-01 -1.02363944e-01 3.23663023e-03 4.48928416e-01
-1.92427374e-02 -1.67264730e-01 -1.02315187e+00 -3.10801983e-01
1.61474690e-01 -6.84905469e-01 6.75106943e-01 3.19993138e-01
1.53068209e+00 1.99770540e-01 9.82527062e-02 1.31576493e-01
1.30045557e+00 -1.22020185e-01 5.43628156e-01 6.79820240e-01
4.41367477e-01 9.45971191e-01 1.03788352e+00 6.91467002e-02
9.14453924e-01 2.87063420e-01 3.89000028e-01 2.79098541e-01
-4.35307771e-01 -1.41983569e-01 1.37876377e-01 1.95997381e+00
-2.22655594e-01 -7.32798636e-01 -8.92939210e-01 5.62360704e-01
-1.48370588e+00 -8.53248715e-01 -1.01368058e+00 1.90937829e+00
7.59409487e-01 -2.37633005e-01 -2.22196709e-03 5.43069184e-01
6.24832809e-01 2.81792909e-01 -2.73464650e-01 -4.04467553e-01
-7.09657311e-01 1.60452023e-01 -7.75361434e-02 5.46818614e-01
-9.16335821e-01 4.59457397e-01 6.83739185e+00 8.26849520e-01
-4.57768053e-01 1.46611184e-01 4.07074481e-01 7.30663836e-02
-3.13251585e-01 -3.51701409e-01 -9.09878314e-01 2.96522409e-01
1.11199486e+00 -2.29077503e-01 1.77681372e-02 5.66642731e-03
7.88629055e-02 -3.92065048e-01 -8.63979697e-01 6.71384156e-01
5.19248545e-01 -9.36449051e-01 1.09190740e-01 -3.98187369e-01
8.03760588e-01 -6.27890751e-02 1.71069041e-01 5.11425972e-01
-2.30972245e-01 -1.11048079e+00 3.79821569e-01 7.98695505e-01
5.71152031e-01 -8.01374793e-01 9.94604051e-01 6.95155621e-01
-1.03184414e+00 -1.17788963e-01 -3.88444394e-01 -1.31587580e-01
3.09713960e-01 4.44478393e-01 -8.65941197e-02 1.23464763e+00
8.54304492e-01 5.52073777e-01 -7.80382574e-01 1.32936215e+00
-5.57067931e-01 9.03720319e-01 2.17501417e-01 -4.42346305e-01
3.29884499e-01 1.14454538e-01 4.23381746e-01 1.25877082e+00
4.12904322e-01 2.25774571e-01 -1.98555410e-01 8.96228105e-02
-1.68433920e-01 5.67803204e-01 -2.00020254e-01 2.09388286e-01
3.59729379e-01 1.01398933e+00 -5.09409830e-02 -1.27837643e-01
-9.20163870e-01 6.36144578e-01 5.66899538e-01 2.45865032e-01
-9.55527127e-01 -8.42434764e-01 1.74734533e-01 -1.52586177e-01
-2.31273100e-01 -1.07303172e-01 -1.94879755e-01 -1.03219569e+00
3.37875187e-01 -1.27308977e+00 7.22219765e-01 -6.52908504e-01
-1.27844846e+00 4.74655986e-01 1.22576945e-01 -1.42299294e+00
1.18501686e-01 -3.15200448e-01 -3.54946107e-01 8.64968777e-01
-2.00709891e+00 -6.31608844e-01 -7.92765200e-01 4.43670779e-01
9.64636385e-01 -5.63715771e-02 8.40913832e-01 4.61777061e-01
-6.49552822e-01 9.33610916e-01 3.50738168e-01 -7.42008612e-02
7.93755770e-01 -1.38233936e+00 2.34912543e-04 6.80182815e-01
3.35217081e-02 3.53631467e-01 5.95196545e-01 -2.37445071e-01
-1.25668454e+00 -8.20892334e-01 1.48527241e+00 -7.38298297e-01
5.52983046e-01 1.06215194e-01 -1.24153757e+00 3.31111282e-01
7.95607030e-01 -6.67252660e-01 1.02713454e+00 1.15434349e-01
7.57347122e-02 5.01163378e-02 -9.96385753e-01 5.35657883e-01
1.10377228e+00 -3.64482075e-01 -9.67051387e-01 5.92460930e-01
1.07017112e+00 -5.38174689e-01 -1.35459876e+00 4.81378168e-01
2.62786716e-01 -9.69925344e-01 8.16790760e-01 -6.39190137e-01
6.32020354e-01 -5.48462085e-02 -2.20918357e-01 -1.27732277e+00
-1.89248681e-01 -3.88801932e-01 -3.05501550e-01 1.16371012e+00
7.08606839e-01 -4.97184873e-01 4.54513222e-01 -8.39253739e-02
-4.40837890e-01 -1.05909681e+00 -1.00991476e+00 -5.94040275e-01
6.55208707e-01 -5.37432492e-01 5.69139957e-01 8.14187527e-01
1.54396575e-02 5.84995568e-01 -2.30107427e-01 9.35311019e-02
5.74993610e-01 -2.23789420e-02 4.19064730e-01 -9.96725321e-01
-5.45465425e-02 -6.27538979e-01 4.62714434e-02 -1.59059346e+00
-9.14730281e-02 -8.90072703e-01 -2.12223068e-01 -2.09305263e+00
1.54829279e-01 -1.26727805e-01 -4.31932896e-01 -2.77574480e-01
-6.66823506e-01 -8.69122967e-02 9.40915719e-02 9.26229954e-02
-1.02344811e+00 6.16115093e-01 1.96980953e+00 -1.58407196e-01
2.58387417e-01 2.80372530e-01 -9.26488340e-01 4.44986939e-01
1.38459897e+00 -1.34941876e-01 -7.98204482e-01 -8.72355342e-01
3.64199340e-01 1.02672018e-01 -4.63440567e-02 -1.18661368e+00
3.85425329e-01 4.81503606e-02 1.02290533e-01 -1.05900884e+00
2.38529563e-01 -4.23522621e-01 -7.54744589e-01 2.11039320e-01
-8.48900735e-01 5.02924502e-01 -9.57163423e-02 5.87106586e-01
-6.35053754e-01 -6.86667681e-01 6.83885813e-01 -1.19591191e-01
-6.28936052e-01 -1.73295513e-02 -3.84540796e-01 8.36558342e-01
1.05707169e+00 4.82302085e-02 -5.83597243e-01 -4.50513154e-01
-5.19327939e-01 8.15474987e-01 -2.30123818e-01 8.59347939e-01
6.38180196e-01 -1.05052626e+00 -1.22367251e+00 -4.67195034e-01
4.42320704e-01 3.98373976e-02 8.53212595e-01 1.06230581e+00
-5.71532726e-01 8.91186774e-01 -5.31819277e-02 -5.02239466e-01
-1.44441319e+00 3.35945249e-01 1.03389531e-01 -4.97443646e-01
-5.09022713e-01 8.21480393e-01 -3.62183452e-01 -5.77325463e-01
6.11892641e-01 -2.68009692e-01 -1.25908422e+00 3.28613222e-01
9.11091387e-01 1.00288832e+00 5.05761564e-01 -6.22423351e-01
2.76808580e-03 8.28919113e-01 -1.45950958e-01 2.64272951e-02
8.93659294e-01 -5.97820044e-01 -1.30716369e-01 7.80729294e-01
1.29529190e+00 -1.36094525e-01 -5.34395814e-01 -2.11569130e-01
1.33448765e-01 -1.04979627e-01 -2.71881014e-01 -1.27327240e+00
-4.60178107e-01 1.06976116e+00 4.01262194e-01 4.32779670e-01
1.28534484e+00 -6.75291717e-02 1.07526636e+00 5.22107363e-01
3.40949476e-01 -1.56826770e+00 3.01905274e-01 8.96076739e-01
1.22067332e+00 -1.29814684e+00 -1.43708602e-01 -4.56725359e-01
-4.31388825e-01 1.03471518e+00 6.42243505e-01 4.30628181e-01
4.67408627e-01 -3.13617140e-01 1.33697495e-01 -8.19360614e-02
-6.71624780e-01 -2.68883109e-01 7.12868869e-01 6.40496492e-01
7.73432076e-01 -4.46207784e-02 -9.37040687e-01 5.93195021e-01
-6.50777638e-01 -3.44974697e-01 6.44819617e-01 1.04771781e+00
-7.19854593e-01 -1.26249707e+00 -2.53445685e-01 5.53632736e-01
-6.41476274e-01 -5.20451814e-02 -3.43682647e-01 9.50632274e-01
-1.39305159e-01 1.84548211e+00 -4.75793034e-01 -2.97024816e-01
7.11203396e-01 -9.23262611e-02 5.05321383e-01 -4.52345073e-01
-7.48068213e-01 -8.34712267e-01 3.30714345e-01 -2.42264420e-01
-5.23529530e-01 -6.12922788e-01 -9.93066967e-01 -3.39795679e-01
-2.36873806e-01 5.33364475e-01 6.30144954e-01 9.67538536e-01
5.31188287e-02 8.98953021e-01 4.55953270e-01 1.88220277e-01
-6.44459665e-01 -1.25514138e+00 -2.50343323e-01 9.05590355e-02
4.89493281e-01 -4.73120868e-01 -5.62166452e-01 -2.84871310e-01] | [11.364263534545898, 8.117300033569336] |
84c05116-9fd3-4b11-81eb-72e3a7b80e5f | instance-segmentation-based-6d-pose | null | null | https://www.sciencedirect.com/science/article/pii/S0736584523000170 | https://www.sciencedirect.com/science/article/pii/S0736584523000170 | Instance segmentation based 6D pose estimation of industrial objects using point clouds for robotic bin-picking | 3D object pose estimation for robotic grasping and manipulation is a crucial task in the manufacturing industry.
In cluttered and occluded scenes, the 6D pose estimation of the low-textured or textureless industrial object is a
challenging problem due to the lack of color information. Thus, point cloud that is hardly affected by the lighting
conditions is gaining popularity as an alternative solution for pose estimation. This article proposes a deep
learning-based pose estimation using point cloud as input, which consists of instance segmentation and instance
point cloud pose estimation. The instance segmentation divides the scene point cloud into multiple instance point
clouds, and each instance point cloud pose is accurately predicted by fusing the depth and normal feature maps.
In order to reduce the time consumption of the dataset acquisition and annotation, a physically-simulated engine
is constructed to generate the synthetic dataset. Finally, several experiments are conducted on the public, synthetic and real datasets to verify the effectiveness of the pose estimation network. The experimental results show
that the point cloud based pose estimation network can effectively and robustly predict the poses of objects in
cluttered and occluded scenes. | ['Han Ding', 'Shaofei Li', 'Chungang Zhuang'] | 2023-08-10 | null | null | null | robotics-and-computer-integrated | ['6d-pose-estimation-1', 'robotic-grasping'] | ['computer-vision', 'robots'] | [-0.03628322 -0.24296111 0.19885643 -0.3988318 -0.34179145 -0.29685462
0.25174007 -0.04430562 -0.01561185 0.18019919 -0.6746481 0.05891504
-0.32839248 -0.79985505 -0.8830565 -0.7314285 0.03896917 1.0126407
0.24635424 0.033198 0.3229553 0.9352481 -1.5696867 0.04656348
0.70864886 1.3899217 0.796736 0.01313994 -0.35995927 0.03921491
-0.60969424 -0.04693819 0.7745028 0.33773533 -0.291528 0.55364865
0.3239515 -0.7487105 -0.21648549 1.0517681 0.20420796 -0.10593852
0.39137158 -1.432168 -0.047814 -0.01824328 -0.72688395 -0.65925837
0.05258427 0.17428607 0.21670942 -0.92606497 0.6127312 1.6199865
0.21014382 0.25332338 -0.5958573 -0.8123903 0.43847102 0.34991768
-1.355399 0.10595257 1.1190836 -0.38026825 0.36600617 0.18576843
0.7791703 0.99523216 0.30162686 0.8417125 1.055764 -0.17033915
0.13593765 -0.07043538 -0.23508954 0.4839259 0.4575643 -0.04955169
-0.18922469 -0.01562715 1.3022434 0.5502161 -0.06183443 -1.0234601
-1.5210682 0.2970832 0.65008336 -0.39119694 -0.4200332 0.084405
0.22566697 0.06203628 0.5681744 0.28707147 -0.7843907 -0.03134237
-0.27601647 0.28572902 0.65057796 1.564032 0.676675 -0.11358563
0.2384425 0.58450145 0.6323466 0.8842435 -0.24432613 -1.1058997
0.57406116 0.98385054 0.52088886 -1.2312087 -0.3093922 -0.33034882
-0.61043006 0.4202748 0.35663265 0.13610907 -1.0723301 0.97169566
0.7677697 0.0563475 -0.360898 1.3202003 0.76081383 0.5919016
-0.33212668 -0.04581979 1.159336 -0.69958013 -0.55426013 -0.16117106
0.14072087 -1.0287737 0.92171395 0.6849301 -0.6988976 -0.60034126
-0.987214 -0.03853312 -0.14483808 0.5470188 0.74580383 0.03223322
-0.11216183 0.42939436 -1.0560452 -0.06659208 0.5156263 0.460809
-0.33532605 -0.6120947 -0.6003135 0.90724057 0.6864718 0.7225378
-0.8902393 -0.41298434 -0.58698565 -0.1640449 0.6124174 -0.44177133
0.95135957 -0.5351059 -1.4496082 0.7016917 0.35454503 0.22526816
0.8027975 -0.49424824 0.13140348 0.06096063 -0.08409435 0.5465564
0.8701754 -1.670166 -0.4457446 -0.8476153 0.1640348 0.3883515
0.19319454 -0.35730842 -0.794105 -0.13101543 0.91183126 -0.9587986
-0.17638692 0.42762667 -0.50886893 -0.24792692 1.3958973 -0.7277692
0.02084652 -2.166015 0.13113044 0.31130457 -0.0417767 -0.14735568
-0.06830629 0.2086877 0.15340847 -0.45451286 0.15233943 -0.05757607
0.06506501 0.3025244 -0.07463588 0.54397976 0.13397436 0.64846915
-0.7556648 -0.6464331 0.6181797 0.47467586 -0.2826603 0.34628344
-0.6572364 0.6817953 -0.91044897 0.9925322 1.1646222 0.05281387
0.03581326 -0.59129465 -0.1875305 -0.26333913 -1.4243591 2.0742202
-0.4540107 0.19946732 0.28274488 -0.56655246 1.4438648 0.10092125
0.6860847 -0.47456098 0.34998575 0.41659954 -0.13224451 -0.67584383
0.03138743 0.26000652 -0.03206726 -0.14245598 -0.35361993 -0.8048209
-0.17788093 -0.2719844 0.61966926 0.7190743 -0.36041415 0.06598368
0.19737987 0.34137943 0.67947346 0.14194651 0.18644443 0.5338219
0.27401185 -0.6742524 -1.267439 -1.1078879 -0.21669579 0.2571137
0.9805931 0.21481948 -0.37781623 -0.42551062 0.36993796 0.32315513
-0.0882396 -0.0909581 -0.6663599 -0.42283165 -0.52723604 0.44572794
0.68026555 -1.00221 -0.7116878 -0.03296262 -0.1243017 -1.1335119
0.2773106 -0.10649159 -0.96248376 -1.2693467 -0.5458048 -0.8405854
0.9626473 0.37507105 0.73046595 0.10733355 -0.4200869 0.09752008
-0.47568065 -0.9027246 -0.04836561 -0.09380198 0.10179254 -0.12868817
0.32182205 -0.17880557 -0.80681777 0.56881297 -0.6059953 0.33591455
0.9218431 0.46880847 0.8475825 0.18988791 0.15858375 -0.36651498
0.11992025 -0.18266849 -1.0946476 0.02782382 0.07832849 -0.48904043
0.10895621 -0.5064516 -1.0218676 0.5804792 0.32505912 -0.94429666
-0.228197 0.2599457 -0.58150345 -0.11673454 0.07479902 -0.02371751
0.30221683 -0.8825625 0.13989991 0.84681165 0.2887699 -0.63572294
1.0737364 0.46707225 0.27051586 -0.74847376 -0.60248804 -0.30195728
-0.9099383 -0.60328805 0.8331432 -0.89676195 -1.1734316 0.5304893
-1.5961931 -0.07748137 0.06061622 0.7606653 -0.5860712 0.3296653
-0.5129234 -0.65858316 -0.14773324 -1.386063 1.5360466 0.05833682
0.3850824 -0.19921985 -0.6190843 0.54847974 -0.1262342 0.6707725
0.87075114 -0.13121821 -1.4345251 -0.6680166 -0.33263353 0.1860293
0.1472239 0.08405237 -0.63763 -0.32524726 0.27056843 -0.14539021
0.08963218 0.20321032 1.5768273 0.03073487 -0.5026877 0.46518058
1.4115489 0.35278556 0.3476831 0.44077995 0.95766115 0.6646159
1.5633773 0.5271064 -0.0246053 0.7262544 1.4119825 -0.1219938
0.2758812 -0.04183083 -0.22318147 0.7786488 -0.19124487 0.07423463
-0.9230518 0.10078461 -1.8054191 -0.41075507 -0.5075831 2.206346
0.4144331 0.22025563 -0.3333463 0.10307725 0.7277034 -0.35403842
-0.866078 0.15125822 0.37350625 -0.1318131 0.43475744 0.03449346
-0.8871092 0.8300211 4.6137986 0.64460546 -1.1177212 -0.21507977
0.20757408 0.05494374 0.1007229 -0.20161185 -0.48869014 0.46556002
-0.07993077 0.4251918 0.2872311 1.331439 0.20718105 -0.08987775
-1.1685927 1.1689043 -0.25816193 -0.9724189 0.03884741 0.0835603
0.51558137 -0.03724005 -0.01773059 0.11482213 -0.02779466 -0.69969505
0.8692076 0.7057756 0.70615894 -0.75727797 0.98317355 0.79320157
-0.9696493 -0.172009 -0.7739378 -0.06740658 0.06908546 0.6633929
-0.94653887 0.70423734 0.8580823 0.4405084 -0.29964384 1.3383135
-0.10498525 -0.1104274 -0.41124967 -0.17878714 0.01302653 -0.51892334
0.4995183 0.38681814 0.33178696 0.02188366 0.7338252 1.056082
0.30896786 -0.08305493 -0.44356078 0.3240711 0.59137255 1.3823886
-0.80823356 0.04127761 -0.13478419 0.7791381 -0.04296361 0.2725955
-0.718248 -0.37558597 0.4131135 0.15509175 0.1324056 -0.79739535
-0.1445763 -0.8922782 0.39284655 -0.6136282 -0.40946323 -1.1937557
-1.1137637 0.29591087 0.32271665 -1.7092358 0.27034053 -1.0502789
-0.14540309 0.9554981 -1.3382077 -1.2305493 -0.98791784 0.28140286
0.84928864 0.04827137 0.5924736 0.12226187 -0.44041583 -0.2357392
0.01617971 -0.03176508 0.30404434 -0.86650527 0.00594269 0.16430397
-0.46077415 0.4617453 0.6115404 -0.7747801 -2.040636 -1.072651
-0.06989994 -0.4210427 0.19437483 -0.7565125 -0.76576495 0.57488644
-0.41355765 0.2261882 -0.24851054 -0.33124453 0.18495275 -0.36324263
-1.2288868 0.4644966 0.94182765 0.06366744 -0.35099903 0.9163205
0.7686214 -1.081609 -0.9319046 0.830348 0.555762 -0.51121604
1.0706524 -0.25604606 0.4629351 -0.53056115 -0.04600256 -1.0077045
-0.16675755 -0.03489095 0.03160882 0.9709531 -0.07794648 -0.4428423
1.176628 0.41770244 -0.32270715 -0.8165509 -0.88984054 -0.7675191
-0.22323366 -0.17216758 0.93586814 0.721971 -0.6053615 -0.01887982
-0.02673065 0.57520604 0.7676737 0.6280656 1.2756542 -1.6992527
0.21012135 0.19650048 -0.64124477 -1.1823223 0.08577061 -0.6055422
0.39049298 -1.796924 -0.01167066 -0.8921907 0.13334186 0.18012717
0.2558797 -0.07955211 0.2209003 0.41930947 -0.29736245 0.6554783
1.9143726 -0.28867587 0.04229018 0.34121498 0.2622929 0.7761583
0.7313748 -0.23760527 -0.47680268 -0.76997477 0.1189217 0.32837516
0.50505984 -0.9725955 -0.0928819 -0.48060316 0.74676204 -1.2649535
0.8974927 -1.624483 0.19426358 0.6464902 0.16458106 -0.08656542
0.02455503 0.5748556 0.10798175 -0.19199471 0.442678 -0.5223632
-0.65925896 0.8123502 0.4655833 -0.6842392 1.3580898 -0.47850525
-0.10144414 0.07905322 -0.4539373 0.3894384 0.6144535 0.6848082
0.96259874 -1.3364772 -0.6295098 0.20955236 0.18123217 0.97281504
0.2526356 0.5868208 -1.1337606 0.15605746 -0.34948605 -1.1749142
-1.1288193 0.6072826 0.1940543 0.43133482 -0.6639448 0.6584355
0.11562476 -0.7802424 0.27497166 -0.5615905 0.07010434 -0.556535
-0.09666117 0.49043906 0.11143644 -0.41924414 -0.22871575 0.75795615
-0.07970486 0.3470613 1.3925487 0.04199601 -0.3628459 0.3485246
1.1126705 -0.38913068 -1.563831 0.04822161 -0.25133756 -0.9030077
-0.13165353 -0.6106718 -1.182068 1.0565035 0.6860367 -0.09649172
0.7012299 -0.05297491 0.71944433 0.5657997 0.8508743 -1.0182797
0.16795263 0.4102666 1.2028873 -1.2731115 0.14314236 -1.0641413
-0.22093967 1.4136401 1.0875903 -0.27867827 0.44434083 0.04965761
0.1639206 -0.43390843 -0.17974983 0.37363335 0.12600991 0.6205103
-0.24811555 0.06310588 0.01487725 0.15348819 -0.21562718 -0.18606697
0.08814563 1.2820878 -0.5386685 -0.7974522 -0.6918463 0.39130732
0.01808313 0.62878436 -0.24832748 0.97025293 0.39970127 0.7009774
0.28343964 -0.30433854 0.529229 -0.3799339 0.6872697 -0.731238
-0.09616891 0.08776974 -0.2573329 -0.6438052 -0.11464655 -0.36463824
-1.4532095 0.08808666 -0.65187377 -0.17837268 1.4124739 0.7757926
0.13072117 0.4802235 0.8664028 -1.4114075 -0.74652094 -0.90560746
-0.6112504 0.4190745 -0.08640123 -1.0488732 -0.11615795 -0.28606927] | [7.314042568206787, -2.509852170944214] |
59968885-289c-484e-a04e-4f92f22de5cf | chatting-makes-perfect-chat-based-image | 2305.20062 | null | https://arxiv.org/abs/2305.20062v1 | https://arxiv.org/pdf/2305.20062v1.pdf | Chatting Makes Perfect -- Chat-based Image Retrieval | Chats emerge as an effective user-friendly approach for information retrieval, and are successfully employed in many domains, such as customer service, healthcare, and finance. However, existing image retrieval approaches typically address the case of a single query-to-image round, and the use of chats for image retrieval has been mostly overlooked. In this work, we introduce ChatIR: a chat-based image retrieval system that engages in a conversation with the user to elicit information, in addition to an initial query, in order to clarify the user's search intent. Motivated by the capabilities of today's foundation models, we leverage Large Language Models to generate follow-up questions to an initial image description. These questions form a dialog with the user in order to retrieve the desired image from a large corpus. In this study, we explore the capabilities of such a system tested on a large dataset and reveal that engaging in a dialog yields significant gains in image retrieval. We start by building an evaluation pipeline from an existing manually generated dataset and explore different modules and training strategies for ChatIR. Our comparison includes strong baselines derived from related applications trained with Reinforcement Learning. Our system is capable of retrieving the target image from a pool of 50K images with over 78% success rate after 5 dialogue rounds, compared to 75% when questions are asked by humans, and 64% for a single shot text-to-image retrieval. Extensive evaluations reveal the strong capabilities and examine the limitations of CharIR under different settings. | ['Dani Lischinski', 'Nir Darshan', 'Rami Ben-Ari', 'Matan Levy'] | 2023-05-31 | null | null | null | null | ['chat-based-image-retrieval', 'information-retrieval'] | ['computer-vision', 'natural-language-processing'] | [ 3.36302191e-01 -8.96096155e-02 -7.99925253e-02 -3.29076707e-01
-1.50816298e+00 -7.65345573e-01 7.08160639e-01 2.56510191e-02
-6.74328208e-01 4.36012834e-01 2.94256121e-01 -2.74647385e-01
6.24929145e-02 -3.28456312e-01 -3.25945616e-01 -5.19690990e-01
2.90583283e-01 7.35867918e-01 2.94947267e-01 -3.56900781e-01
3.99406999e-01 2.53107786e-01 -1.30475843e+00 6.70951843e-01
3.83876592e-01 7.83449709e-01 6.48061156e-01 8.14667284e-01
-1.31374881e-01 1.16631556e+00 -8.93913448e-01 -3.36096317e-01
4.02015820e-02 -3.82925361e-01 -1.39623356e+00 2.07690880e-01
2.13135302e-01 -1.07527030e+00 -2.94805527e-01 4.64415550e-01
8.33378494e-01 1.24012977e-01 5.97620249e-01 -1.12104321e+00
-7.48157620e-01 2.25216210e-01 -1.47104025e-01 2.21668750e-01
7.39416599e-01 3.77293468e-01 1.02081823e+00 -8.10353875e-01
9.99210358e-01 1.09812677e+00 1.39233917e-01 7.98708439e-01
-1.00339496e+00 -4.40328598e-01 -2.63665766e-01 2.19514579e-01
-1.32297981e+00 -5.86063087e-01 2.66203254e-01 -1.97016403e-01
1.02432334e+00 2.43421897e-01 3.68907034e-01 1.02839291e+00
-3.53374667e-02 1.10520399e+00 7.91265786e-01 -5.79603195e-01
6.29740953e-02 3.81191015e-01 -4.05133218e-02 3.58632565e-01
-5.53442717e-01 -1.98770672e-01 -4.54236269e-01 -1.41530886e-01
6.85000479e-01 4.01376225e-02 -6.84272200e-02 -1.64117664e-02
-1.17847335e+00 1.00479889e+00 4.99423295e-01 2.69391268e-01
-2.85534143e-01 1.74611196e-01 4.61036175e-01 5.03192306e-01
2.98404068e-01 6.07413411e-01 -5.12251444e-02 -2.19552875e-01
-7.66888976e-01 2.53334910e-01 9.25000906e-01 1.05527425e+00
6.66066051e-01 -5.20180404e-01 -5.53366184e-01 1.06171989e+00
8.89571458e-02 4.20923680e-01 4.33904976e-01 -1.27016521e+00
1.53712288e-01 3.62299323e-01 2.99067676e-01 -5.35004854e-01
-3.64271700e-02 1.20519526e-01 -3.09069902e-01 -2.99784895e-02
2.54324853e-01 -2.75537036e-02 -8.08519602e-01 1.36029088e+00
2.40240589e-01 -2.78303444e-01 3.31945628e-01 1.02422404e+00
1.12788117e+00 5.91205060e-01 1.62918776e-01 4.64599952e-02
1.71422791e+00 -1.09106648e+00 -5.24916470e-01 -1.69073239e-01
6.15424156e-01 -1.06373429e+00 1.21594346e+00 1.29710943e-01
-1.18709171e+00 -3.76876086e-01 -4.67407435e-01 -3.15411121e-01
-2.51991719e-01 2.08072424e-01 1.84344411e-01 2.45846078e-01
-1.46651328e+00 1.62968367e-01 -3.39916646e-01 -7.32474089e-01
1.41147241e-01 2.42764637e-01 -2.47771099e-01 -3.14028770e-01
-1.08941257e+00 9.21574473e-01 4.87082824e-03 -1.99610621e-01
-9.98668730e-01 -5.13330638e-01 -5.28861761e-01 3.46803218e-02
6.42459810e-01 -6.35588109e-01 1.78345418e+00 -7.91878700e-01
-1.43252885e+00 1.11046994e+00 -1.52759150e-01 -4.80669439e-01
5.15794516e-01 -1.92076072e-01 9.95353460e-02 9.47839081e-01
2.59206086e-01 1.26369989e+00 7.72861421e-01 -1.07366312e+00
-5.24890244e-01 4.17039394e-02 5.93489766e-01 4.78117079e-01
-1.67598426e-01 4.35246646e-01 -1.09158301e+00 -1.85394764e-01
-5.40968955e-01 -1.11164141e+00 -1.57710850e-01 -4.64289412e-02
-2.02233225e-01 -3.34431767e-01 7.98652172e-01 -5.50280571e-01
9.26249325e-01 -2.19302964e+00 -2.64179319e-01 -1.00847319e-01
1.00229539e-01 1.86598644e-01 -4.44068491e-01 9.87695873e-01
3.05558175e-01 2.03391239e-01 1.25207528e-01 -4.00496691e-01
-1.65117934e-01 5.95411658e-02 -3.53636742e-01 -7.34262243e-02
3.73966157e-01 1.29209018e+00 -9.48125184e-01 -9.66772854e-01
2.02666670e-01 4.43725586e-01 -4.62683678e-01 6.04247332e-01
-2.37131163e-01 2.88489014e-01 -5.90146542e-01 5.33634841e-01
1.19171897e-02 -7.76579201e-01 8.82765129e-02 -4.16608229e-02
2.49122843e-01 4.43246998e-02 -4.82431382e-01 1.71322048e+00
-5.06876230e-01 6.75994039e-01 4.70046885e-02 -7.49201715e-01
6.08153701e-01 5.74673593e-01 5.62343001e-01 -9.89055097e-01
9.32331681e-02 1.31672984e-02 -1.72700733e-01 -7.47949064e-01
7.88782358e-01 1.67412892e-01 -1.58605695e-01 9.58152533e-01
1.39579177e-02 -2.96143055e-01 3.71454060e-01 7.01436520e-01
1.38164604e+00 -1.51946411e-01 4.02587391e-02 2.03133240e-01
3.77727181e-01 4.66974437e-01 -2.62488574e-01 1.20665014e+00
-2.25538626e-01 9.46574330e-01 2.85627693e-01 -3.33984613e-01
-9.62018371e-01 -8.19303751e-01 1.95860282e-01 1.22481072e+00
4.04102266e-01 -2.90600389e-01 -7.57000625e-01 -7.28175998e-01
-3.83102596e-01 4.10070866e-01 -3.95670563e-01 4.84309457e-02
-3.96496356e-01 -3.39170158e-01 5.50366104e-01 2.60192126e-01
6.08861029e-01 -1.69276142e+00 -9.43318963e-01 1.03602387e-01
-5.51102996e-01 -1.41417503e+00 -8.17181408e-01 -2.25676626e-01
-5.09862304e-01 -1.09981716e+00 -1.10129857e+00 -9.20977890e-01
5.52823424e-01 6.49873495e-01 1.36703944e+00 6.41851366e-01
-8.35204065e-01 1.07720566e+00 -6.38781786e-01 -2.79376566e-01
-6.74936771e-01 1.92217290e-01 -5.22943854e-01 -4.30066049e-01
1.85527042e-01 4.58134897e-02 -1.09472692e+00 6.36875808e-01
-1.28385329e+00 1.54425846e-02 8.38093221e-01 9.09237921e-01
2.26766601e-01 -6.63496673e-01 7.66402721e-01 -7.33453393e-01
9.81636703e-01 -4.61373627e-01 -3.60518157e-01 4.73522156e-01
-3.87096733e-01 -9.99346897e-02 1.49292916e-01 -5.61468959e-01
-9.72797632e-01 1.19918495e-01 -1.82575554e-01 -4.76405829e-01
-1.98718607e-01 4.77323532e-01 5.22207320e-01 1.08825611e-02
6.36031747e-01 4.16666776e-01 3.83240491e-01 -1.77823186e-01
3.75975966e-01 9.78671014e-01 3.98608625e-01 -4.62708235e-01
4.36394811e-01 3.89283091e-01 -6.83349133e-01 -8.88381243e-01
-5.84361255e-01 -1.01465619e+00 -3.59138638e-01 -3.31602395e-01
8.90652597e-01 -9.84012008e-01 -9.54310417e-01 1.23871811e-01
-1.19696903e+00 -5.06057262e-01 -1.49481729e-01 1.88991338e-01
-5.80340803e-01 2.67315179e-01 -8.18708837e-01 -8.52293491e-01
-8.06089997e-01 -1.40715528e+00 1.59158742e+00 2.41197020e-01
-2.87561715e-01 -7.29831278e-01 -6.23270757e-02 7.84301043e-01
6.28845334e-01 -4.95094866e-01 6.71693981e-01 -8.91022742e-01
-8.56903374e-01 -2.70135552e-01 -4.73604530e-01 6.50203899e-02
1.81799848e-03 -3.56318533e-01 -8.85159731e-01 -2.70153761e-01
-4.42278177e-01 -1.00068676e+00 4.59990144e-01 -6.66757524e-02
8.35418522e-01 -1.54684529e-01 -2.38784790e-01 -1.50948003e-01
1.24815786e+00 3.26471508e-01 8.63654613e-01 3.10565799e-01
-6.02786057e-03 6.65480256e-01 6.75097287e-01 1.92455545e-01
4.17826325e-01 7.16380775e-01 1.95292845e-01 -1.15305506e-01
-1.07776135e-01 -8.13859478e-02 1.85234189e-01 6.00549221e-01
2.34479561e-01 -3.08540136e-01 -6.45238519e-01 7.09410489e-01
-1.75261199e+00 -9.30913687e-01 4.87739801e-01 1.87027681e+00
9.33256626e-01 -2.21343115e-01 1.33168325e-01 -5.15921891e-01
4.38800246e-01 -2.34819879e-03 -5.64948559e-01 -1.69125840e-01
3.06773722e-01 2.05454662e-01 1.23410247e-01 4.67725992e-01
-7.85028458e-01 1.23499227e+00 6.86267662e+00 8.18714321e-01
-1.18590295e+00 3.32488678e-02 7.91931748e-01 -1.39441743e-01
-9.21164602e-02 1.77200455e-02 -5.72378695e-01 1.22695617e-01
8.35738659e-01 -1.66130915e-01 4.69511718e-01 7.21794307e-01
2.02241048e-01 -4.50405687e-01 -1.11002076e+00 1.00750959e+00
3.26587290e-01 -1.44743431e+00 1.32831812e-01 -6.07622415e-02
3.25586677e-01 2.01052383e-01 -1.16286211e-01 4.79890019e-01
3.16633552e-01 -1.09456539e+00 3.28497827e-01 3.61841857e-01
9.00366187e-01 -2.40943700e-01 6.29442811e-01 5.26206791e-01
-8.41201425e-01 1.76907003e-01 -1.92999482e-01 1.76079661e-01
2.41652280e-01 -2.01327622e-01 -1.58030665e+00 3.03079814e-01
8.23222280e-01 1.86089054e-01 -6.08575046e-01 8.77151668e-01
-2.83013587e-03 2.61754930e-01 -2.17417359e-01 -3.87706488e-01
3.19708318e-01 -2.26087533e-02 1.28673077e-01 1.23663819e+00
2.06743926e-03 3.30464989e-01 3.33874434e-01 7.37472832e-01
-3.76610368e-01 3.19995314e-01 -6.47477269e-01 -6.13292791e-02
4.18833613e-01 1.47393906e+00 -8.64917278e-01 -5.61222196e-01
-3.57216477e-01 1.33154464e+00 2.98951715e-01 5.62114894e-01
-5.06146789e-01 -4.62722868e-01 2.41957992e-01 -2.38840673e-02
3.02708387e-01 8.48935395e-02 4.63889629e-01 -9.12453175e-01
6.48754761e-02 -1.26017797e+00 3.83483827e-01 -1.39494884e+00
-1.11073697e+00 6.93416953e-01 1.16900742e-01 -1.04389501e+00
-7.17587531e-01 -1.55517787e-01 -5.92275381e-01 8.06986511e-01
-1.63148916e+00 -1.06038451e+00 -3.92048806e-01 6.92107260e-01
1.07013166e+00 5.02767488e-02 1.01191139e+00 2.61549503e-01
-1.31930068e-01 4.16263729e-01 -5.89852855e-02 5.15260637e-01
1.03469026e+00 -9.75294113e-01 2.35459000e-01 2.05554724e-01
3.29882652e-01 7.32420266e-01 5.45773864e-01 -4.64759588e-01
-1.50563002e+00 -5.62096000e-01 7.02118516e-01 -4.95069653e-01
5.11979342e-01 -2.93483466e-01 -7.62916625e-01 3.75134766e-01
7.58637786e-01 -1.25871539e-01 4.56184834e-01 -3.42212707e-01
-6.35779798e-02 5.21661006e-02 -1.03362405e+00 7.25017548e-01
5.60904682e-01 -8.48417461e-01 -5.62085092e-01 5.68043768e-01
7.42208898e-01 -3.93460572e-01 -8.05777311e-01 1.05055831e-02
8.11857522e-01 -7.25388765e-01 1.18095720e+00 -5.97945273e-01
7.13279188e-01 -6.01676386e-03 1.32472023e-01 -8.24320078e-01
2.32126325e-01 -7.39036620e-01 5.05080938e-01 9.78897274e-01
3.97937149e-01 -2.23427683e-01 7.84353256e-01 9.35467958e-01
2.24500641e-01 -6.61173582e-01 -6.65097952e-01 -3.48688036e-01
-1.88617483e-01 -2.20615312e-01 2.23317936e-01 4.85788196e-01
1.63444340e-01 7.06081390e-01 -3.53526533e-01 -3.72662321e-02
2.96585798e-01 2.44752824e-01 1.06260312e+00 -7.54229665e-01
-2.17680350e-01 -1.47286043e-01 -4.88957912e-02 -1.35093379e+00
-1.58672780e-02 -6.36524498e-01 3.90747428e-01 -1.74797976e+00
6.72881603e-01 -3.83082032e-01 5.96580580e-02 5.54145098e-01
-1.58864141e-01 5.82171321e-01 4.35433030e-01 6.90310359e-01
-1.12930632e+00 2.10861281e-01 1.09194565e+00 -2.33560145e-01
-2.08253235e-01 -1.51893407e-01 -7.08675325e-01 3.79006743e-01
6.27834618e-01 -2.44197518e-01 -5.23152351e-01 -3.86341393e-01
4.89467569e-02 4.87059683e-01 3.55224013e-01 -5.56495845e-01
5.17373621e-01 6.39676228e-02 1.54414266e-01 -5.49146414e-01
5.88608682e-01 -7.38995910e-01 -1.67122573e-01 3.30442727e-01
-6.73793733e-01 2.52785146e-01 3.87084261e-02 3.64425123e-01
-3.53753150e-01 -4.91493285e-01 4.02475655e-01 -4.86907452e-01
-8.15076530e-01 5.08270375e-02 -6.08331919e-01 2.13271290e-01
9.40852702e-01 -1.83049124e-02 -4.35368419e-01 -1.08285153e+00
-6.24266207e-01 5.64197361e-01 3.09923559e-01 4.82875794e-01
8.64908397e-01 -9.13053572e-01 -6.72279179e-01 -2.41043136e-01
3.75779122e-01 -1.87315881e-01 1.45669311e-01 6.74131632e-01
-5.83272994e-01 4.79410231e-01 6.85535744e-02 -8.53677928e-01
-1.63275719e+00 4.51556385e-01 6.14045113e-02 -4.84010190e-01
-6.89185143e-01 3.62327725e-01 2.91017622e-01 -2.28565395e-01
3.24560434e-01 4.30257879e-02 -1.68908253e-01 3.86239626e-02
7.33337581e-01 -1.62887543e-01 3.17494869e-02 -3.73609990e-01
-1.92742392e-01 3.95747304e-01 -4.29169625e-01 -5.28250635e-01
1.20616651e+00 -4.57117587e-01 1.57212749e-01 2.23315001e-01
1.36460316e+00 -2.91158944e-01 -8.54508758e-01 -3.53909314e-01
1.14577310e-02 -3.21758658e-01 -3.16159487e-01 -9.69313085e-01
-8.31423998e-01 6.89348459e-01 5.18189192e-01 2.74442852e-01
1.00214410e+00 4.80055243e-01 9.00326788e-01 8.91416848e-01
3.70472133e-01 -1.02514017e+00 7.71110713e-01 4.14453387e-01
9.05706167e-01 -1.60890758e+00 -9.12874490e-02 -1.29172295e-01
-1.05862784e+00 9.86873686e-01 4.13003296e-01 -2.61595044e-02
3.02111894e-01 -4.01102528e-02 6.02357924e-01 -3.54604453e-01
-1.03033888e+00 -4.17402655e-01 1.08946159e-01 3.22738439e-01
4.51064557e-01 -4.72370923e-01 -2.62545832e-02 3.17429639e-02
9.66040790e-02 1.97307810e-01 4.37374890e-01 1.22613454e+00
-2.48436838e-01 -1.10764313e+00 -2.79291660e-01 3.74574989e-01
-6.73503697e-01 -3.76632184e-01 -5.78784943e-01 8.44728470e-01
-7.33516037e-01 1.21513677e+00 3.67819183e-02 -3.79586704e-02
3.15639317e-01 1.73659444e-01 2.07284078e-01 -7.55372584e-01
-9.33761299e-01 2.26630658e-01 1.75432742e-01 -6.87197685e-01
-5.29753864e-01 -4.15996999e-01 -1.04319370e+00 1.34949937e-01
-3.46050560e-01 3.35582227e-01 7.15079188e-01 1.07677317e+00
5.53893030e-01 1.71828479e-01 4.68692690e-01 -7.56559551e-01
-5.23434937e-01 -9.84050632e-01 -4.85084578e-02 6.39229953e-01
2.80120313e-01 -1.11703873e-01 -1.52884781e-01 3.57272625e-01] | [10.920382499694824, 1.3651435375213623] |
60dfa1e0-8c84-459f-a9e1-60831f677219 | construction-of-english-resume-corpus-and | 2208.03219 | null | https://arxiv.org/abs/2208.03219v2 | https://arxiv.org/pdf/2208.03219v2.pdf | Construction of English Resume Corpus and Test with Pre-trained Language Models | Information extraction(IE) has always been one of the essential tasks of NLP. Moreover, one of the most critical application scenarios of information extraction is the information extraction of resumes. Constructed text is obtained by classifying each part of the resume. It is convenient to store these texts for later search and analysis. Furthermore, the constructed resume data can also be used in the AI resume screening system. Significantly reduce the labor cost of HR. This study aims to transform the information extraction task of resumes into a simple sentence classification task. Based on the English resume dataset produced by the prior study. The classification rules are improved to create a larger and more fine-grained classification dataset of resumes. This corpus is also used to test some current mainstream Pre-training language models (PLMs) performance.Furthermore, in order to explore the relationship between the number of training samples and the correctness rate of the resume dataset, we also performed comparison experiments with training sets of different train set sizes.The final multiple experimental results show that the resume dataset with improved annotation rules and increased sample size of the dataset improves the accuracy of the original resume dataset. | ['Tatsunori Mori', 'Chengguang Gan'] | 2022-08-05 | null | null | null | null | ['sentence-classification'] | ['natural-language-processing'] | [ 4.38176244e-01 -7.33154938e-02 -3.30332577e-01 -3.48580390e-01
-5.88953257e-01 -5.88231504e-01 5.24514437e-01 5.15442371e-01
-4.54714715e-01 1.01030672e+00 1.96678728e-01 -4.48489130e-01
-2.28638723e-01 -6.44371688e-01 -3.16801101e-01 -3.05245489e-01
1.10323250e-01 4.46204066e-01 2.06466794e-01 -3.33191961e-01
6.15814209e-01 5.20392716e-01 -1.80198956e+00 4.37665522e-01
1.23975420e+00 1.06783760e+00 5.04486978e-01 6.29204988e-01
-4.46937174e-01 9.32952762e-01 -1.07645524e+00 -2.44445264e-01
1.68793917e-01 -5.77586651e-01 -1.00690830e+00 9.18885320e-02
-6.29185289e-02 -7.67347142e-02 -1.45398438e-01 7.57197618e-01
1.75157219e-01 3.99787575e-01 5.92917979e-01 -1.12026381e+00
7.14866072e-02 6.87191904e-01 -7.96745569e-02 4.09613818e-01
4.74280030e-01 -5.12273490e-01 6.39861345e-01 -6.73800230e-01
7.32301593e-01 7.90513635e-01 2.94327497e-01 2.51620263e-01
-5.55207312e-01 -9.23401058e-01 -2.96315342e-01 3.46056759e-01
-1.20892358e+00 -5.79173565e-01 7.44678915e-01 -2.89711446e-01
8.58371556e-01 4.41310048e-01 4.29877698e-01 6.75668776e-01
5.22903383e-01 8.86180043e-01 1.04107404e+00 -9.00599360e-01
-2.58623928e-01 5.98398328e-01 5.93359470e-01 6.54531300e-01
2.23036155e-01 -4.21439916e-01 -7.94361055e-01 1.32885680e-01
2.45627612e-01 -1.61650762e-01 -3.45898449e-01 3.05813223e-01
-1.02127278e+00 7.52634704e-01 -1.84798166e-01 6.55243039e-01
-2.14557081e-01 -7.81832278e-01 7.39198864e-01 2.66053230e-01
4.33932781e-01 6.14181995e-01 -6.43765450e-01 -4.71768856e-01
-9.75594997e-01 -3.49879637e-02 1.15629756e+00 1.20617115e+00
4.04445887e-01 -2.23164052e-01 -2.46697381e-01 8.84470940e-01
-2.05500871e-01 1.59324154e-01 5.68186641e-01 -4.10560310e-01
9.32696283e-01 1.14696586e+00 4.55646217e-02 -9.65438366e-01
-5.44486463e-01 -4.87576157e-01 -9.58682716e-01 -5.27756333e-01
3.00069690e-01 -9.69530344e-02 -7.06119239e-01 1.06259716e+00
1.78304404e-01 -5.12448132e-01 2.14914918e-01 4.96333092e-01
9.33920383e-01 1.02440512e+00 -4.91790799e-03 -6.33176863e-01
1.45439327e+00 -8.90836120e-01 -1.13194776e+00 -3.00906032e-01
6.85382307e-01 -9.73383486e-01 8.74738812e-01 6.06320024e-01
-5.45222044e-01 -7.17797935e-01 -1.13265562e+00 -1.81836203e-01
-4.62906331e-01 8.75794113e-01 8.97874296e-01 3.14941913e-01
-3.23244631e-01 4.88012999e-01 -4.45639908e-01 -3.78220588e-01
-4.41544801e-02 4.41827804e-01 -4.76627052e-01 1.33204341e-01
-1.30909967e+00 8.92031133e-01 9.94966090e-01 1.26277298e-01
-3.17395866e-01 -1.02531917e-01 -7.41152167e-01 1.10655442e-01
5.53962827e-01 4.34734039e-02 1.10003221e+00 -6.90401733e-01
-1.17176688e+00 9.29397047e-01 -1.83842063e-01 -2.77324349e-01
1.28246285e-02 -2.72592962e-01 -6.50735915e-01 1.69155061e-01
4.29567099e-01 1.25222364e-02 4.17150050e-01 -9.91786659e-01
-9.40520704e-01 -2.99515873e-01 -2.79898375e-01 1.66815192e-01
-3.59094262e-01 4.73940045e-01 -6.06353581e-01 -2.57998139e-01
-5.59316315e-02 -9.63096499e-01 3.57884049e-01 -1.04970789e+00
-4.38676745e-01 -5.43026984e-01 6.40947461e-01 -1.20353532e+00
1.90811551e+00 -2.15757990e+00 -5.97858913e-02 1.18382990e-01
2.78966911e-02 3.79791260e-01 3.98107588e-01 7.55827010e-01
9.04306173e-02 2.10820571e-01 -1.86805194e-03 1.34590762e-02
-2.40283623e-01 5.70105128e-02 -2.83396959e-01 -1.16295461e-02
-2.09078402e-03 3.57399583e-01 -6.05029464e-01 -1.14910316e+00
-3.50136198e-02 -2.87077874e-01 -2.66727656e-02 5.75792074e-01
8.70525390e-02 3.56065333e-01 -6.32436693e-01 5.68060935e-01
2.41602361e-01 8.01408589e-02 1.42472342e-01 -1.46772161e-01
-3.26133847e-01 5.33732414e-01 -9.04642522e-01 1.50199282e+00
-4.89041030e-01 7.85886705e-01 -6.99415728e-02 -1.01187837e+00
1.23618078e+00 2.96032608e-01 2.73261130e-01 -5.86069345e-01
2.27614835e-01 2.09178656e-01 1.26044318e-01 -8.81012499e-01
1.02106178e+00 -6.10325448e-02 -5.57680368e-01 3.40673439e-02
5.48551418e-02 -2.02358544e-01 7.91233897e-01 3.92116874e-01
1.00955236e+00 4.04311493e-02 5.96347332e-01 -2.05764979e-01
6.55479252e-01 4.99787807e-01 8.47671807e-01 5.08658588e-01
-1.14564290e-02 4.57211286e-02 5.34888387e-01 -2.61107087e-01
-7.54794419e-01 -4.53926355e-01 -4.73389417e-01 9.48880851e-01
-1.15011707e-02 -7.44000912e-01 -6.29995584e-01 -8.20474148e-01
-3.97338361e-01 9.79840696e-01 -2.03152090e-01 -1.36020277e-02
-4.55439240e-01 -7.56775200e-01 5.87805271e-01 3.27233076e-01
8.15911889e-01 -1.31542766e+00 -4.40119535e-01 1.62465468e-01
-8.18059087e-01 -1.09350383e+00 -1.14561111e-01 5.87551892e-01
-7.17818439e-01 -1.18717325e+00 2.21023142e-01 -1.09712029e+00
4.15039927e-01 -1.13246530e-01 9.26371932e-01 2.96889961e-01
4.39619385e-02 -1.30942836e-01 -7.88796663e-01 -6.43357337e-01
-9.20848370e-01 6.77222311e-01 1.14721149e-01 -3.24677765e-01
5.34904420e-01 -4.54157405e-02 -2.82455217e-02 2.08083332e-01
-8.85513842e-01 4.78953093e-01 6.61429942e-01 6.59731328e-01
3.11797023e-01 4.21101272e-01 1.09472966e+00 -1.15483093e+00
1.08080602e+00 -3.43441099e-01 -2.98479527e-01 5.71163774e-01
-7.43654668e-01 9.44427624e-02 9.23942149e-01 -2.19353020e-01
-1.30276155e+00 -1.56551912e-01 -1.98566496e-01 1.99143365e-01
-2.11282372e-01 7.14263320e-01 -2.67063648e-01 3.77240628e-01
3.89394522e-01 4.04557168e-01 -2.65236318e-01 -5.98353624e-01
-3.65649074e-01 1.60242808e+00 5.87604582e-01 -4.50365812e-01
3.74273837e-01 -4.27866369e-01 -1.42118018e-02 -9.84324455e-01
-9.85830128e-01 -7.69449592e-01 -7.70417690e-01 -4.60382074e-01
6.76885128e-01 -4.35660034e-01 -7.07306623e-01 3.07933837e-01
-9.96885061e-01 -2.05732882e-02 1.85071919e-02 5.39883673e-01
-2.25367486e-01 3.83162171e-01 -7.17107356e-01 -7.83086121e-01
-6.74229920e-01 -8.47102165e-01 9.31825042e-01 3.29181731e-01
-2.31577531e-01 -5.77892721e-01 -2.02818021e-01 8.11850727e-01
-8.81784111e-02 1.04963511e-01 1.20986736e+00 -1.24183810e+00
-2.71828622e-01 -6.01275563e-01 6.85971836e-03 4.28097993e-01
2.12804094e-01 -1.68924208e-03 -5.05337715e-01 8.94477591e-03
3.76875699e-01 -4.01298970e-01 5.84036171e-01 -2.90265739e-01
9.59042072e-01 -2.87788689e-01 -3.47763181e-01 2.45420814e-01
1.08536828e+00 6.43071890e-01 6.57841563e-01 6.17547691e-01
3.62100929e-01 7.38794208e-01 1.43967354e+00 4.12660986e-01
1.08427405e-01 3.43887717e-01 -3.07968885e-01 5.08918643e-01
6.43145591e-02 -3.31330329e-01 4.14448380e-01 1.60095167e+00
4.34437059e-02 -4.32164639e-01 -1.00930607e+00 3.85155976e-01
-1.58855033e+00 -7.81553090e-01 -2.79863983e-01 2.11757255e+00
1.17285943e+00 1.98483273e-01 -2.05264211e-01 6.42214060e-01
5.71183980e-01 -2.27296188e-01 4.00204919e-02 -5.26207447e-01
3.37322533e-01 1.60451204e-01 2.91852146e-01 2.25314602e-01
-1.10580015e+00 1.01206100e+00 5.23290205e+00 1.21904945e+00
-8.10548306e-01 -2.34215781e-01 4.03137922e-01 3.19661230e-01
2.76877820e-01 9.51497182e-02 -1.42085791e+00 5.14039278e-01
1.20075309e+00 -2.40179256e-01 2.49178380e-01 6.40556157e-01
1.40790448e-01 -7.02043712e-01 -1.05634081e+00 7.96922326e-01
2.62844533e-01 -1.09478188e+00 -1.12844713e-01 -7.68136978e-02
3.23737770e-01 -2.86981672e-01 -1.00183821e+00 8.80372047e-01
-3.96233648e-01 -7.54361033e-01 4.82342899e-01 6.99067950e-01
7.53203034e-01 -8.01568329e-01 1.16401625e+00 1.09072924e+00
-1.03288317e+00 -5.47781251e-02 -2.57883877e-01 -1.09412961e-01
-1.32770136e-01 4.11077976e-01 -1.16681886e+00 7.85180330e-01
5.32499373e-01 3.06611866e-01 -9.79247987e-01 8.44293535e-01
-2.27017403e-01 7.93746173e-01 -4.37178403e-01 -4.36548531e-01
-4.97299694e-02 -3.95430863e-01 3.01516503e-01 1.27666605e+00
6.53498748e-04 1.21477410e-01 3.52417499e-01 3.95021915e-01
-2.31373712e-01 5.25321424e-01 -5.27814388e-01 -3.44222218e-01
5.33396065e-01 1.30705357e+00 -9.36644614e-01 -4.10387069e-01
-2.29560152e-01 7.10590184e-01 3.23735505e-01 8.00600126e-02
-4.70638007e-01 -1.08926702e+00 -2.43550032e-01 1.90626889e-01
-2.31207579e-01 -2.72104740e-01 -1.26092732e-01 -1.15776956e+00
1.10651359e-01 -1.01102901e+00 2.94583261e-01 -5.08055925e-01
-8.46035480e-01 7.98334658e-01 3.39547187e-01 -1.16931593e+00
-4.64062780e-01 -3.85975957e-01 -3.71921420e-01 7.24485695e-01
-9.89840508e-01 -9.14465308e-01 -3.46461266e-01 3.29821020e-01
8.84199202e-01 -1.95398360e-01 1.01150703e+00 3.72670889e-01
-9.12464797e-01 2.77628392e-01 1.40684858e-01 5.08682847e-01
7.29695916e-01 -1.04285431e+00 -1.77043349e-01 9.79032457e-01
8.66185781e-03 7.20415831e-01 4.51217324e-01 -9.95356858e-01
-1.36886024e+00 -8.05565774e-01 1.45431376e+00 -3.43643099e-01
4.49654132e-01 -5.57855546e-01 -8.39242339e-01 7.57020414e-01
1.47372976e-01 -7.40282893e-01 7.04967678e-01 1.93122894e-01
5.45198083e-01 -1.90046087e-01 -7.93736160e-01 2.28799433e-01
5.83313882e-01 -4.50204432e-01 -9.19219136e-01 3.13836366e-01
5.01833498e-01 -4.38388288e-01 -1.03174174e+00 4.67067778e-01
2.67090261e-01 -5.86405218e-01 3.31972957e-01 -4.24050242e-01
4.03979480e-01 -6.27440289e-02 1.73763812e-01 -9.17250514e-01
6.23960979e-02 -3.71897906e-01 8.85942578e-02 1.86519992e+00
4.89701122e-01 -2.25801229e-01 4.48819965e-01 6.94689095e-01
-2.42076710e-01 -7.51100719e-01 -5.92391789e-01 -6.01450682e-01
-3.86322647e-01 -4.71676320e-01 2.42276803e-01 7.07992435e-01
4.80545104e-01 1.03767967e+00 -3.35346431e-01 -2.90854633e-01
5.44954687e-02 4.89190966e-01 9.25944030e-01 -1.21677101e+00
1.95236176e-01 1.45172998e-01 -1.80894241e-01 -1.00268805e+00
2.38253251e-01 -1.11427855e+00 2.85390049e-01 -1.77922153e+00
4.01288748e-01 -5.70660830e-01 1.57254040e-01 4.70354974e-01
-2.56890446e-01 -3.35103661e-01 2.54625231e-01 4.68852162e-01
-6.28510058e-01 4.56048489e-01 1.10160458e+00 2.95476824e-01
-5.09377480e-01 5.03138661e-01 -4.22046989e-01 5.97757101e-01
9.54694629e-01 -7.93009341e-01 -3.16302270e-01 2.58143216e-01
2.55028218e-01 5.38236082e-01 -3.06682348e-01 -9.39645112e-01
3.50324899e-01 -1.65782407e-01 3.19133818e-01 -1.02182758e+00
2.79433429e-02 -7.15667725e-01 -1.48106545e-01 3.89424525e-02
-3.00352335e-01 -5.78408735e-03 3.48301232e-01 2.35062167e-01
-4.89394039e-01 -1.03737879e+00 2.22236112e-01 5.49414009e-02
-6.87354326e-01 -2.09645405e-01 -6.37652576e-01 1.81838647e-01
8.55978072e-01 -1.49116114e-01 -1.17161475e-01 -1.49766982e-01
-3.49533349e-01 2.66741991e-01 1.75659865e-01 5.19166887e-01
4.69253600e-01 -6.87540650e-01 -5.18569827e-01 9.03455690e-02
8.17352682e-02 2.99015567e-02 -2.15062439e-01 7.49560237e-01
-8.42966616e-01 7.32482791e-01 -2.04195991e-01 -9.15205181e-02
-1.54520440e+00 5.47382534e-01 -2.12053522e-01 -6.82872593e-01
-5.87398827e-01 2.50367701e-01 -2.51672298e-01 -3.35978001e-01
3.25412393e-01 -4.34993297e-01 -7.28174388e-01 2.12291956e-01
3.19985211e-01 2.89693743e-01 2.21043840e-01 -4.22666848e-01
-3.69683743e-01 -4.28853408e-02 -1.76016212e-01 1.60535008e-01
1.25710094e+00 3.71195115e-02 -4.43367213e-01 5.52762568e-01
9.02511060e-01 6.37608767e-01 -4.47697520e-01 9.45197791e-02
4.96522874e-01 -2.33154938e-01 -3.11639328e-02 -9.41289842e-01
-5.80526233e-01 5.43077528e-01 8.87482911e-02 3.42984527e-01
1.37144446e+00 -1.60800129e-01 8.28782201e-01 6.26087010e-01
4.74792749e-01 -1.25808239e+00 -4.41422284e-01 5.31796753e-01
8.67262602e-01 -1.28868365e+00 2.72663951e-01 -6.11027598e-01
-7.49471366e-01 1.35252428e+00 5.33969402e-01 3.45171779e-01
3.52918208e-01 3.21532875e-01 -2.85541594e-01 -2.04040930e-01
-5.09899855e-01 -7.99543560e-02 3.54143679e-01 3.08672283e-02
5.21027923e-01 4.67567891e-02 -9.84382451e-01 1.37429106e+00
-4.20758218e-01 1.70772493e-01 4.80225474e-01 1.06110597e+00
-5.83656430e-01 -1.32987726e+00 -1.84888572e-01 9.01359677e-01
-8.32228065e-01 -1.11031517e-01 -7.48347759e-01 1.04693878e+00
1.68515801e-01 1.38008225e+00 -4.15261209e-01 -6.64908826e-01
2.48917893e-01 5.54781139e-01 2.27683946e-01 -7.29390562e-01
-1.06031847e+00 -3.04440916e-01 7.57496834e-01 -4.66397265e-04
-3.50720286e-01 -4.45187509e-01 -1.51015663e+00 -2.19010189e-01
-8.76555443e-01 1.00390756e+00 8.30535591e-01 1.13425303e+00
1.94573939e-01 5.15322864e-01 5.05395353e-01 -3.21221709e-01
-4.00218219e-01 -1.49551976e+00 -5.71720958e-01 4.04482037e-01
-1.76974177e-01 -3.75959367e-01 -3.52450490e-01 2.48197436e-01] | [9.787022590637207, 8.624054908752441] |
635ba76e-d967-4476-b3ee-1585f0a6594c | faster-dan-multi-target-queries-with-document | 2301.10593 | null | https://arxiv.org/abs/2301.10593v1 | https://arxiv.org/pdf/2301.10593v1.pdf | Faster DAN: Multi-target Queries with Document Positional Encoding for End-to-end Handwritten Document Recognition | Recent advances in handwritten text recognition enabled to recognize whole documents in an end-to-end way: the Document Attention Network (DAN) recognizes the characters one after the other through an attention-based prediction process until reaching the end of the document. However, this autoregressive process leads to inference that cannot benefit from any parallelization optimization. In this paper, we propose Faster DAN, a two-step strategy to speed up the recognition process at prediction time: the model predicts the first character of each text line in the document, and then completes all the text lines in parallel through multi-target queries and a specific document positional encoding scheme. Faster DAN reaches competitive results compared to standard DAN, while being at least 4 times faster on whole single-page and double-page images of the RIMES 2009, READ 2016 and MAURDOR datasets. Source code and trained model weights are available at https://github.com/FactoDeepLearning/FasterDAN. | ['Thierry Paquet', 'Clément Chatelain', 'Denis Coquenet'] | 2023-01-25 | null | null | null | null | ['handwritten-document-recognition'] | ['computer-vision'] | [ 2.57712990e-01 -2.85332024e-01 -2.33504087e-01 -4.13612664e-01
-9.58655596e-01 -8.28448832e-01 7.54005492e-01 -2.19288602e-01
-3.72649103e-01 3.41003031e-01 2.45468408e-01 -3.68335068e-01
4.57949974e-02 -4.36872452e-01 -7.78239369e-01 -6.70043468e-01
5.60209513e-01 1.05102682e+00 1.16853036e-01 1.59461752e-01
7.96374977e-01 7.63726294e-01 -1.21761537e+00 7.98499227e-01
4.19221818e-01 9.84282494e-01 2.98903406e-01 1.49260914e+00
-3.78412664e-01 1.04760921e+00 -4.38751757e-01 -7.38679588e-01
5.60863055e-02 -4.00940478e-01 -9.33333278e-01 2.71852434e-01
8.13063443e-01 -5.90259075e-01 -5.60276031e-01 6.53828323e-01
5.03424108e-01 2.92114373e-02 6.92568779e-01 -6.75223768e-01
-9.09524620e-01 8.33966672e-01 -7.04935372e-01 1.38137013e-01
1.72258094e-01 -4.72358949e-02 1.21052945e+00 -1.50880075e+00
5.77749491e-01 8.64239395e-01 5.54988503e-01 5.93538463e-01
-7.15917706e-01 -2.32540205e-01 4.04832140e-02 3.95314753e-01
-1.03469336e+00 -4.86019403e-01 3.89034927e-01 -4.43507791e-01
1.50523388e+00 4.15042967e-01 3.95171732e-01 1.26908708e+00
1.23500988e-01 1.44098794e+00 5.13813615e-01 -7.14838684e-01
-1.46078825e-01 -3.78272772e-01 6.09780371e-01 7.85786510e-01
-3.84598561e-02 -3.09013277e-01 -7.25482047e-01 3.02150190e-01
5.49781859e-01 3.97886075e-02 -2.38728836e-01 6.04376663e-03
-9.91720319e-01 5.61453462e-01 1.22590125e-01 3.02087724e-01
-4.82534200e-01 8.03163797e-02 4.24576759e-01 5.73836314e-03
1.42798260e-01 2.16665357e-01 -3.66760015e-01 -7.24242210e-01
-1.36845148e+00 3.11527774e-02 1.04795325e+00 7.97026575e-01
3.91816854e-01 -1.43826974e-03 -4.69801098e-01 9.33629513e-01
9.30285603e-02 6.03167593e-01 6.05529070e-01 -4.69461918e-01
7.88319051e-01 3.49290937e-01 -4.63181026e-02 -7.24437654e-01
-2.26377904e-01 -3.73705149e-01 -9.16740179e-01 7.24708661e-02
5.88119805e-01 -9.10773426e-02 -1.23569489e+00 6.50582671e-01
-1.52839646e-01 -3.91641915e-01 -1.31978139e-01 1.01458967e+00
2.95876950e-01 1.01046753e+00 -4.55108345e-01 3.57913166e-01
1.23039234e+00 -1.83196747e+00 -4.45528120e-01 -3.98251384e-01
5.62313437e-01 -1.00617623e+00 1.06078255e+00 7.75062501e-01
-1.25726569e+00 -5.81368446e-01 -1.07883835e+00 -5.58752596e-01
-2.28631273e-01 8.26815546e-01 3.51702154e-01 4.32598621e-01
-8.86188090e-01 4.84684020e-01 -9.72480536e-01 -1.47837460e-01
6.38173401e-01 2.83577770e-01 -1.69558629e-01 -3.37292284e-01
-3.98976922e-01 6.53500795e-01 2.14381337e-01 4.93415356e-01
-8.09967816e-01 -4.64683890e-01 -3.41145843e-01 4.81141567e-01
3.49548757e-01 -1.81047708e-01 1.36859095e+00 -1.20030797e+00
-1.75805628e+00 7.65010953e-01 -4.94574010e-01 -5.51155031e-01
8.99368048e-01 -7.99082756e-01 -1.39557093e-01 4.77065854e-02
-3.51052284e-01 5.56633174e-01 8.63870740e-01 -8.36379945e-01
-6.86912596e-01 -4.39551383e-01 -5.79177022e-01 2.03588545e-01
-1.85545608e-01 5.45422286e-02 -1.26706707e+00 -5.47516465e-01
-4.33835834e-02 -8.08035970e-01 2.17003107e-01 -7.48659298e-02
-8.11883032e-01 -1.64972514e-01 6.71559870e-01 -1.14909768e+00
9.58232224e-01 -2.17751575e+00 3.21719497e-01 2.11192548e-01
-2.57735304e-03 4.29678112e-01 -4.30754393e-01 5.19199431e-01
1.97489951e-02 -6.15748577e-02 -9.98457894e-02 -6.07078731e-01
1.66487083e-01 -2.32379548e-02 -8.36368859e-01 2.92469621e-01
2.07859963e-01 1.10140634e+00 -3.78843755e-01 -1.30682245e-01
-1.33968160e-01 5.04204392e-01 -1.72686607e-01 1.29900038e-01
-3.87813151e-01 -1.54836535e-01 -1.72856182e-01 6.15095139e-01
4.88618404e-01 -5.56240439e-01 1.62752599e-01 -9.80511121e-03
-1.65761039e-01 2.92709827e-01 -8.22730303e-01 1.39641082e+00
-2.14092672e-01 1.23003447e+00 -3.85022163e-01 -8.15208077e-01
1.05120051e+00 -5.68605447e-03 -3.55821662e-02 -8.18035543e-01
-4.80598956e-02 2.05872938e-01 -7.66448528e-02 -2.54551142e-01
8.27660322e-01 6.67229056e-01 3.26468021e-01 7.12988675e-01
-2.38853153e-02 2.71768928e-01 4.56236899e-01 1.29718259e-01
9.15603042e-01 1.85645208e-01 -2.36618191e-01 1.25761643e-01
5.81883967e-01 4.87294234e-02 -8.59000087e-02 1.07703006e+00
3.85883421e-01 8.88990760e-01 6.85506165e-01 -6.80407882e-01
-1.50040162e+00 -7.37910628e-01 1.66743264e-01 1.34734023e+00
-5.28592348e-01 -2.46865824e-01 -7.24121332e-01 -7.31925607e-01
2.58817449e-02 7.63414502e-01 -7.38592565e-01 2.67776012e-01
-6.56764030e-01 -3.23876828e-01 9.00952637e-01 8.97116125e-01
5.10565221e-01 -1.19917297e+00 -5.10843456e-01 1.94448352e-01
9.28362906e-02 -7.95447767e-01 -6.68414474e-01 4.11113709e-01
-6.60596728e-01 -9.06973958e-01 -1.19267786e+00 -7.81337380e-01
7.52795160e-01 -1.45308793e-01 8.46683383e-01 2.09834173e-01
-4.45066422e-01 1.38952017e-01 -3.42746884e-01 -1.46207422e-01
-1.39439806e-01 5.36279082e-01 -4.59128499e-01 6.36083707e-02
3.94959033e-01 1.95175558e-01 -3.42740983e-01 1.29353642e-01
-6.31141484e-01 3.75900537e-01 8.84662211e-01 1.15521991e+00
6.07695580e-01 -4.55038220e-01 -9.27623957e-02 -8.63099158e-01
5.72055995e-01 -3.20478603e-02 -8.19832206e-01 6.79266930e-01
-5.84524512e-01 2.51187861e-01 6.96340084e-01 -3.85679930e-01
-9.69004333e-01 2.51718283e-01 -2.18432844e-01 -2.22580656e-01
-8.71082842e-02 5.15926123e-01 2.34598771e-01 3.79359722e-01
3.61129373e-01 9.56889451e-01 -2.18461007e-01 -6.89493835e-01
2.23378330e-01 8.36133003e-01 6.69236541e-01 -3.85077417e-01
3.05921286e-01 1.88270614e-01 -2.77366549e-01 -7.76347280e-01
-8.71017218e-01 -3.21635395e-01 -8.75256240e-01 4.50968444e-02
6.38736427e-01 -5.51258802e-01 -8.04941654e-01 9.60861027e-01
-1.38995802e+00 -8.15079808e-01 -2.62222365e-02 2.21677274e-01
-3.89262319e-01 4.40565020e-01 -1.09573746e+00 -7.35844076e-01
-7.84559488e-01 -9.25758123e-01 1.04180217e+00 1.81777954e-01
-2.37115681e-01 -6.48169935e-01 -1.78036815e-03 3.90724480e-01
2.67984211e-01 -6.25110269e-01 8.66154015e-01 -9.03808594e-01
-6.60608947e-01 -5.64667463e-01 -5.03140211e-01 4.03596967e-01
-2.62324929e-01 5.91084421e-01 -7.51274407e-01 -2.39270836e-01
-5.47903836e-01 -3.64938915e-01 1.15330637e+00 1.88395530e-01
1.33877897e+00 -3.20813388e-01 -1.29431963e-01 6.86716974e-01
1.36485195e+00 3.22135240e-01 9.69547153e-01 4.84109968e-01
8.74473810e-01 2.33497664e-01 2.93694973e-01 6.05634272e-01
9.67274904e-02 5.40184736e-01 1.64638751e-03 1.10412657e-01
-3.78272593e-01 -1.01838477e-01 4.72755611e-01 7.75192440e-01
-5.58854192e-02 -8.62353981e-01 -1.38915372e+00 4.13946658e-01
-1.61608160e+00 -9.82993245e-01 -3.05025667e-01 1.99818587e+00
7.11168468e-01 2.06498653e-01 -2.10900202e-01 1.42566115e-01
5.10198414e-01 1.28067032e-01 -6.48475349e-01 -6.35594130e-01
-5.00099361e-01 1.53206140e-01 6.56744778e-01 6.51134908e-01
-7.82364964e-01 1.15911913e+00 6.36818981e+00 8.31231713e-01
-1.45023263e+00 -3.00147682e-01 9.68147814e-01 -2.29510427e-01
3.75540219e-02 -2.16554299e-01 -1.33398628e+00 3.77500445e-01
9.92193758e-01 8.59431848e-02 6.15997493e-01 8.49803507e-01
-2.77369201e-01 -1.16160877e-01 -1.07074320e+00 9.43156064e-01
4.67004567e-01 -1.65385008e+00 2.95285940e-01 -3.24918367e-02
7.46922016e-01 3.57942820e-01 2.50468910e-01 1.71284359e-02
2.19617665e-01 -1.04800689e+00 8.84288609e-01 9.56482232e-01
7.16782093e-01 -6.46255791e-01 6.60059154e-01 2.47660205e-01
-6.52970433e-01 -1.03665054e-01 -4.98257518e-01 1.46083429e-01
-7.28732049e-02 3.33493710e-01 -9.26010966e-01 3.53786826e-01
5.59498429e-01 6.92970991e-01 -7.93167293e-01 1.05107641e+00
-4.70379412e-01 9.36340630e-01 -4.77895364e-02 -4.23025042e-01
6.54790699e-01 -2.57774573e-02 8.59068856e-02 1.53587139e+00
5.31832814e-01 -1.99347153e-01 -3.25431138e-01 5.28578401e-01
-2.70682067e-01 8.53877962e-02 -2.99089178e-02 -4.97035950e-01
1.54524460e-01 1.00028837e+00 -7.37561643e-01 -6.48030221e-01
-3.81915390e-01 1.75375617e+00 6.07877076e-01 4.07938570e-01
-7.18307674e-01 -6.25926495e-01 -5.25062345e-02 -1.89898998e-01
1.01863062e+00 -4.04271752e-01 -4.43920553e-01 -9.39913511e-01
1.77576452e-01 -1.09252191e+00 2.44814590e-01 -1.13617551e+00
-9.90766048e-01 7.22263634e-01 -9.39526975e-01 -6.12182617e-01
-2.67866731e-01 -1.05659366e+00 -4.79580432e-01 1.02550471e+00
-1.04868114e+00 -1.10472417e+00 -3.27322245e-01 2.44041428e-01
1.00264454e+00 -2.93392122e-01 8.32873046e-01 6.82708770e-02
-8.04931402e-01 8.40317130e-01 7.82469451e-01 6.47642136e-01
6.00714147e-01 -1.18228114e+00 8.85173738e-01 9.42252398e-01
5.82435429e-01 2.76862770e-01 3.09675843e-01 -6.59284711e-01
-1.57961357e+00 -6.29929900e-01 1.11436296e+00 -4.64312553e-01
6.10338867e-01 -3.93257499e-01 -1.04390967e+00 9.13277805e-01
4.63098884e-01 -3.20266187e-01 5.28427184e-01 1.16390422e-01
-5.92098057e-01 4.92652245e-02 -3.32694471e-01 5.72945178e-01
5.68553329e-01 -4.38951015e-01 -2.90807754e-01 3.96728367e-01
5.82639314e-02 -6.53941751e-01 -5.25942504e-01 -3.73085648e-01
9.71438468e-01 -8.16117883e-01 4.98691291e-01 -6.29740953e-01
9.96506274e-01 5.95422424e-02 -1.27044365e-01 -8.65804911e-01
-4.83137995e-01 -4.98532623e-01 -4.72263068e-01 1.00403690e+00
8.38054955e-01 -3.07143360e-01 1.07621121e+00 3.84774566e-01
-2.26737976e-01 -8.77417743e-01 -6.70656621e-01 -6.36871576e-01
3.49621661e-02 -4.62630600e-01 4.73374069e-01 4.67570990e-01
-2.76532859e-01 2.53886819e-01 -6.31303370e-01 1.10793367e-01
3.32077146e-01 2.30526283e-01 7.08933532e-01 -7.45813251e-01
-4.69982266e-01 -8.56590748e-01 -1.61816567e-01 -1.65534782e+00
-8.42390768e-03 -8.32335472e-01 2.28744894e-01 -1.53178704e+00
3.41192007e-01 1.57159820e-01 -3.74859534e-02 7.07221687e-01
1.89271644e-02 9.68012661e-02 4.16328043e-01 4.19444561e-01
-6.49427295e-01 3.44072908e-01 9.82551277e-01 -4.81677204e-01
2.54375041e-01 -1.66899726e-01 -2.70376354e-01 2.77165204e-01
6.87038362e-01 -2.25987330e-01 1.31790489e-01 -1.10584497e+00
2.72000968e-01 9.62981433e-02 1.46806225e-01 -1.05903506e+00
8.16790521e-01 2.68199652e-01 1.01937497e+00 -1.16405797e+00
3.00149024e-01 -4.70469743e-01 -2.74694920e-01 5.49338758e-01
-6.18279159e-01 2.08597198e-01 2.91412294e-01 2.84194261e-01
-1.94243386e-01 -4.89760965e-01 4.95998234e-01 2.00863659e-01
-7.50520825e-01 2.40649983e-01 -5.12443841e-01 -3.05686474e-01
6.27940297e-01 -3.86882007e-01 -6.11796677e-01 -4.19050664e-01
-5.87248981e-01 -3.52850836e-03 2.27501839e-01 4.89838332e-01
6.76515341e-01 -9.05632555e-01 -8.42718124e-01 2.79035926e-01
-1.63539171e-01 -1.57013699e-01 4.02656198e-01 7.58441210e-01
-9.85375583e-01 9.66194928e-01 -2.86770135e-01 -4.97103691e-01
-1.37790763e+00 3.30881149e-01 3.38761210e-01 -4.51852709e-01
-5.75766027e-01 1.14425457e+00 -2.95707017e-01 -7.81914815e-02
4.81147766e-01 -1.48068760e-02 1.05104744e-01 1.86245479e-02
9.09223795e-01 3.90692174e-01 1.78367272e-01 -4.02617246e-01
-2.14734390e-01 7.40549743e-01 -7.47198701e-01 -2.49005422e-01
1.35711086e+00 1.28980651e-01 -4.82358634e-02 4.49214578e-01
1.02106547e+00 5.93409203e-02 -1.60174620e+00 -1.48426786e-01
3.89881700e-01 -3.57036591e-01 -4.17538173e-02 -1.27019155e+00
-1.04263461e+00 1.03983450e+00 3.12919527e-01 -6.21969327e-02
8.20722044e-01 -2.20770597e-01 6.30102158e-01 9.07052100e-01
-2.61756241e-01 -1.22865105e+00 1.82490557e-01 1.01853538e+00
1.11333120e+00 -9.45301533e-01 -1.23528279e-02 9.32313874e-02
-9.27953064e-01 1.69019377e+00 4.74387646e-01 -1.12781949e-01
9.80836675e-02 4.17850345e-01 8.65884870e-02 7.32561061e-03
-1.00164282e+00 2.62340248e-01 5.07480323e-01 1.11533523e-01
6.70397520e-01 -1.87456682e-01 1.89422950e-01 5.03960371e-01
-1.09523468e-01 -9.68805552e-02 4.40727264e-01 7.65007019e-01
-3.37974429e-01 -1.03602242e+00 -2.53375918e-01 6.23080790e-01
-4.51518714e-01 -4.78133321e-01 -6.74379230e-01 5.75558543e-01
-6.67129874e-01 4.28642541e-01 5.49321413e-01 -1.42158955e-01
2.37010151e-01 4.32280958e-01 5.17008901e-01 -2.06259027e-01
-4.38982725e-01 1.07551090e-01 -1.65127777e-02 -3.89066756e-01
3.58509809e-01 -7.32529759e-01 -1.27441132e+00 -4.32075858e-01
-2.27155104e-01 -4.99737002e-02 8.31019223e-01 8.07791293e-01
6.16027415e-01 4.38204229e-01 4.26115930e-01 -8.15981567e-01
-7.56464779e-01 -1.18851376e+00 -3.10409337e-01 3.50431278e-02
2.27514908e-01 1.48876905e-01 -8.43078569e-02 3.50568831e-01] | [11.875056266784668, 2.4249794483184814] |
890ebf4f-3ee4-42c7-831e-f337b503fde6 | dynamic-virtual-network-embedding-algorithm | 2202.02140 | null | https://arxiv.org/abs/2202.02140v1 | https://arxiv.org/pdf/2202.02140v1.pdf | Dynamic Virtual Network Embedding Algorithm based on Graph Convolution Neural Network and Reinforcement Learning | Network virtualization (NV) is a technology with broad application prospects. Virtual network embedding (VNE) is the core orientation of VN, which aims to provide more flexible underlying physical resource allocation for user function requests. The classical VNE problem is usually solved by heuristic method, but this method often limits the flexibility of the algorithm and ignores the time limit. In addition, the partition autonomy of physical domain and the dynamic characteristics of virtual network request (VNR) also increase the difficulty of VNE. This paper proposed a new type of VNE algorithm, which applied reinforcement learning (RL) and graph neural network (GNN) theory to the algorithm, especially the combination of graph convolutional neural network (GCNN) and RL algorithm. Based on a self-defined fitness matrix and fitness value, we set up the objective function of the algorithm implementation, realized an efficient dynamic VNE algorithm, and effectively reduced the degree of resource fragmentation. Finally, we used comparison algorithms to evaluate the proposed method. Simulation experiments verified that the dynamic VNE algorithm based on RL and GCNN has good basic VNE characteristics. By changing the resource attributes of physical network and virtual network, it can be proved that the algorithm has good flexibility. | ['Lei Liu', 'Weishan Zhang', 'Neeraj Kumar', 'Chao Wang', 'Peiying Zhang'] | 2022-02-03 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-3.03720713e-01 -4.49837923e-01 -3.76194715e-01 3.46407950e-01
9.08434510e-01 -2.73327410e-01 1.56741254e-02 -5.45489788e-01
-4.26645845e-01 8.49247277e-01 -4.58293885e-01 -6.53000236e-01
-7.48091936e-01 -1.09611750e+00 -7.22403824e-03 -6.93004966e-01
-4.28218901e-01 3.85835111e-01 1.27089560e-01 -3.57056260e-01
2.26038992e-01 9.44441140e-01 -1.36662507e+00 -3.49094421e-01
7.53448069e-01 8.93288374e-01 4.89895076e-01 3.51939499e-01
-7.76333690e-01 6.36111379e-01 -9.59333479e-01 3.36100012e-01
2.09773704e-01 -4.03389335e-01 -6.60443723e-01 8.47714841e-02
-8.66750658e-01 -5.86139709e-02 -7.15387046e-01 9.28245246e-01
5.13459265e-01 3.98394763e-01 2.44457796e-01 -1.94886839e+00
-5.83192706e-01 6.54275894e-01 -4.43820834e-01 6.40782177e-01
2.30199285e-02 1.56382158e-01 5.15376449e-01 -2.90581107e-01
7.17901587e-01 1.20138979e+00 3.10196608e-01 5.91692924e-01
-8.62010181e-01 -8.35899830e-01 2.47437000e-01 5.19493520e-01
-1.51551032e+00 2.14220405e-01 8.97029698e-01 -7.05744177e-02
9.69652891e-01 3.69670331e-01 1.07916808e+00 5.11253715e-01
2.90712863e-01 3.91531199e-01 7.40785897e-01 -4.60239083e-01
1.57775968e-01 -2.55165361e-02 -2.67552942e-01 5.58382094e-01
4.04126197e-01 3.39155734e-01 4.50015128e-01 -2.13844329e-01
1.04073000e+00 1.89308986e-01 -4.49934423e-01 -6.85010850e-01
-9.73148704e-01 7.26086676e-01 4.47216183e-01 4.40427095e-01
-3.22658807e-01 1.75689068e-02 7.77811587e-01 4.34777051e-01
-2.59889573e-01 1.52957126e-01 -4.89105523e-01 -1.57216355e-01
-3.83662164e-01 -3.38956892e-01 7.36011863e-01 7.73305953e-01
5.58302224e-01 7.21167922e-01 1.62834123e-01 6.12197161e-01
1.31472036e-01 6.01993501e-01 5.40533781e-01 -6.92845345e-01
9.94738191e-02 6.41507804e-01 -3.01248431e-01 -1.26836944e+00
-7.47334421e-01 -4.06673431e-01 -1.13999677e+00 4.08794433e-01
-2.91106373e-01 -3.61108959e-01 -5.90704322e-01 1.73462534e+00
3.97274554e-01 5.30039847e-01 1.12888545e-01 7.87668765e-01
7.52532661e-01 9.18390155e-01 -1.90533668e-01 -7.33462811e-01
7.56391585e-01 -1.09896004e+00 -1.15624964e+00 5.98554850e-01
5.32520473e-01 -5.80427229e-01 8.45782578e-01 3.85907263e-01
-5.88599563e-01 -4.98424977e-01 -1.17209184e+00 9.08758759e-01
-5.15268862e-01 -3.24776858e-01 9.13992405e-01 9.41341579e-01
-1.11789036e+00 5.72501898e-01 -2.11947098e-01 -3.76448333e-01
1.56525001e-01 7.80921757e-01 -2.16170356e-01 1.57104835e-01
-1.45113516e+00 5.30018985e-01 9.52435017e-01 2.85861373e-01
-4.60640132e-01 -2.68518090e-01 -4.74531651e-01 3.63629997e-01
7.83247888e-01 -5.36854088e-01 6.36310875e-01 -1.04377091e+00
-1.63614213e+00 -7.03323167e-04 6.84024096e-01 3.69671779e-03
2.04524234e-01 7.57960677e-01 -1.32673228e+00 1.06249504e-01
-4.79436994e-01 6.90210611e-02 7.49449313e-01 -1.11899424e+00
-5.91808200e-01 2.14750215e-01 1.61785141e-01 1.67353034e-01
-4.08093721e-01 -7.45662302e-02 -1.68021575e-01 -1.19372740e-01
-1.76427484e-01 -6.97208345e-01 -2.26781830e-01 -3.67507279e-01
-2.71740742e-02 -1.09173439e-01 1.30709612e+00 -3.59883726e-01
1.67628336e+00 -2.15855336e+00 1.70898780e-01 8.45712245e-01
5.39797902e-01 7.72984803e-01 -2.46124983e-01 4.13107514e-01
-1.49521932e-01 3.25032830e-01 2.41079286e-01 8.90323699e-01
-1.45126954e-01 4.81848568e-01 1.49000421e-01 6.10037148e-02
-4.52547967e-01 5.74293256e-01 -8.51919353e-01 -5.89496136e-01
5.32543302e-01 3.48623931e-01 -4.78629261e-01 9.20373574e-02
-1.28797904e-01 1.58353850e-01 -4.17551309e-01 4.72186714e-01
9.56889808e-01 -2.42206141e-01 6.42162800e-01 -1.00287404e-02
3.11097950e-02 -5.26203811e-01 -1.73130667e+00 8.77908468e-01
-5.05661905e-01 4.01672125e-01 2.78790414e-01 -1.03188968e+00
1.11278772e+00 2.71208435e-01 8.21475387e-01 -9.20292079e-01
4.53940898e-01 2.25629732e-01 3.17557782e-01 -7.30069876e-01
3.28963965e-01 2.77946919e-01 3.00961554e-01 5.35912097e-01
-2.35868081e-01 4.11880583e-01 1.33187426e-02 7.48534948e-02
1.04586625e+00 -8.76726359e-02 4.04258490e-01 -2.42927969e-01
9.14591432e-01 -2.47424200e-01 6.30069792e-01 3.28080148e-01
-3.66758853e-01 -4.83382374e-01 5.28525710e-01 -5.73764145e-01
-1.15631843e+00 -9.16207194e-01 1.84489310e-01 6.51054144e-01
5.63761890e-01 -1.44707784e-02 -5.08498967e-01 -4.89867598e-01
-2.61973441e-02 4.91277546e-01 -2.58633405e-01 -4.88179445e-01
-4.69861329e-01 -4.92723286e-01 4.30080742e-01 -1.22588202e-01
6.26851201e-01 -1.51051366e+00 -2.84969360e-01 4.22472268e-01
1.88418016e-01 -9.78187621e-01 -3.43716890e-01 -1.67812690e-01
-8.32420588e-01 -1.10809612e+00 -1.97037011e-02 -1.06190026e+00
4.63734090e-01 7.30395854e-01 8.83099318e-01 7.98972428e-01
-3.28825593e-01 4.15518284e-01 -4.69900370e-01 2.11965144e-01
-4.60830390e-01 1.11474000e-01 4.26351368e-01 -1.37476847e-01
1.44375190e-01 -9.74522293e-01 -3.73296350e-01 3.80228609e-01
-1.01639891e+00 -8.84578079e-02 5.52260756e-01 1.02255893e+00
3.48288238e-01 7.21115828e-01 7.18101680e-01 -6.37031019e-01
9.46105897e-01 -5.13144433e-01 -8.51168931e-01 4.46739912e-01
-1.09869039e+00 -6.90693706e-02 1.13044405e+00 -6.01781785e-01
-7.14103580e-01 -6.89004004e-01 1.04309686e-01 -8.11215878e-01
3.44053090e-01 3.69306922e-01 -6.38628602e-01 -5.45971036e-01
1.05223894e-01 4.10203457e-01 3.76358956e-01 -5.81806488e-02
2.48774931e-01 8.56845498e-01 6.39881045e-02 -3.97937328e-01
7.06145644e-01 3.83755453e-02 3.99929583e-01 -6.74601436e-01
3.62192243e-01 7.49142170e-02 -1.35400994e-02 -5.99554181e-01
5.41020572e-01 -2.19344944e-01 -1.60527253e+00 1.34101436e-01
-8.55480373e-01 -2.82550544e-01 -5.73175848e-02 6.04423106e-01
-3.12846839e-01 5.67744017e-01 -5.41327953e-01 -7.92236865e-01
-3.43227416e-01 -1.06734753e+00 -1.89771608e-01 4.54210609e-01
5.82608461e-01 -1.24168313e+00 -1.48259938e-01 -1.89692155e-01
7.49033749e-01 3.38181138e-01 1.16297328e+00 -4.03086394e-01
-6.64557457e-01 1.96666002e-01 -4.42200333e-01 3.01061422e-01
1.40213564e-01 5.42509019e-01 -7.42853656e-02 -5.54524839e-01
-6.85702711e-02 1.77796587e-01 1.76237240e-01 2.21222594e-01
1.53088355e+00 -2.88116664e-01 -3.45995158e-01 9.07948434e-01
2.13502979e+00 8.37112308e-01 8.67434978e-01 7.62424767e-01
8.96754384e-01 1.73406154e-01 6.11294687e-01 6.88414514e-01
-2.90878229e-02 5.12589157e-01 9.15583193e-01 -1.58122018e-01
8.02382156e-02 -6.56254515e-02 -5.28401136e-02 1.30258000e+00
-2.90653259e-01 -6.07723176e-01 -6.59972250e-01 1.57451984e-02
-1.83602488e+00 -1.23758411e+00 1.03487730e-01 1.92146885e+00
-5.70617840e-02 4.09820408e-01 1.03720658e-01 3.30987483e-01
1.25818932e+00 9.95060429e-02 -5.27296722e-01 -1.02921283e+00
-1.54419720e-01 7.96529185e-03 5.58684945e-01 2.03323513e-01
-4.61338460e-01 8.21936369e-01 5.91787863e+00 1.39311731e+00
-1.38908482e+00 5.43794930e-02 6.03320673e-02 -1.71906073e-02
-3.09438080e-01 -9.25575048e-02 -1.47065207e-01 7.72009492e-01
8.58231723e-01 -4.34346080e-01 1.34391713e+00 8.32924962e-01
3.21309835e-01 3.56159776e-01 -3.64750743e-01 1.33254087e+00
-2.62330711e-01 -1.42661667e+00 2.74706334e-01 2.70213544e-01
5.38572550e-01 -2.31191739e-01 -1.16571397e-01 5.22181869e-01
2.10443810e-02 -9.98161674e-01 -5.04035912e-02 3.81227016e-01
9.62897718e-01 -1.29089880e+00 9.58284140e-01 2.31874853e-01
-1.53511798e+00 -4.79509741e-01 -6.27360523e-01 -1.83137596e-01
-2.75628846e-02 1.60628840e-01 -4.97735173e-01 9.85326648e-01
3.70540708e-01 1.98396966e-01 -2.02508748e-01 9.96219754e-01
1.56673461e-01 6.62732720e-02 -1.51979223e-01 -3.97078723e-01
3.07551175e-01 -6.34660542e-01 6.53857470e-01 7.30726063e-01
2.05857381e-01 1.40411839e-01 4.29733962e-01 5.60365915e-01
4.03978452e-02 5.32148540e-01 -5.95008492e-01 -3.72327358e-01
9.99768615e-01 1.41240454e+00 -9.28670466e-01 -1.00890905e-01
-3.88111562e-01 6.27658963e-01 3.69769871e-01 4.22551781e-01
-1.14831948e+00 -9.09101069e-01 4.93971765e-01 -8.78457874e-02
3.40185225e-01 -1.24569379e-01 1.75708279e-01 -7.66000211e-01
-1.50369942e-01 -9.07880902e-01 3.44613045e-01 -7.11635828e-01
-9.89919960e-01 8.57556760e-01 -3.29320461e-01 -1.40541255e+00
9.35128927e-02 -6.05529308e-01 -7.06314921e-01 7.02080309e-01
-1.36021614e+00 -6.67953253e-01 -4.00970250e-01 1.00362980e+00
-4.79871705e-02 -7.60119677e-01 9.08522964e-01 7.10595369e-01
-9.57578242e-01 6.85902894e-01 6.44250214e-01 -4.09809686e-02
6.00507110e-02 -6.29939973e-01 -2.70325661e-01 7.68051505e-01
-2.60428935e-01 2.91574121e-01 4.08477068e-01 -5.89458644e-01
-1.54405403e+00 -7.25477040e-01 2.07763344e-01 5.32470703e-01
5.84509015e-01 4.37692702e-02 -6.65515959e-01 3.67198169e-01
1.43943042e-01 5.97930625e-02 6.27792478e-01 -1.83263645e-01
1.01964191e-01 -2.12942883e-01 -1.34086537e+00 7.85282433e-01
1.19308817e+00 -1.84360504e-01 -1.72282290e-02 3.47829193e-01
1.36748397e+00 -1.68944687e-01 -7.84383357e-01 2.70105302e-01
2.73750603e-01 -8.67246151e-01 9.10644293e-01 -8.25167477e-01
-1.61623478e-01 -5.55511713e-01 -5.73456995e-02 -1.10111761e+00
-8.42593551e-01 -4.67283607e-01 -3.76364380e-01 1.07231057e+00
-7.67071033e-03 -1.07465255e+00 7.03502715e-01 -1.77250188e-02
2.37449557e-02 -9.58292425e-01 -8.09822917e-01 -1.28149390e+00
-5.54878473e-01 5.69237918e-02 1.41086257e+00 1.13167179e+00
-3.88167799e-02 2.08059728e-01 -4.39982951e-01 8.41261372e-02
3.04794610e-01 1.02679968e-01 5.99393547e-01 -1.30667257e+00
-4.08402562e-01 -8.11178446e-01 -7.33200788e-01 -5.21819234e-01
3.01092625e-01 -8.69286239e-01 -7.72540689e-01 -1.60901523e+00
-1.89352080e-01 -7.53134012e-01 -8.60059083e-01 1.27949275e-03
2.46193171e-01 -3.55518490e-01 2.24394724e-01 3.27208400e-01
-6.39288545e-01 5.95310748e-01 1.56479943e+00 5.15790954e-02
-6.18783832e-01 -7.49586448e-02 -9.97686312e-02 2.35203937e-01
1.23408723e+00 -1.94016472e-01 -9.46973324e-01 -9.98864025e-02
-1.01910513e-02 4.63032424e-01 -6.97545186e-02 -8.09316158e-01
1.40733436e-01 -6.30585194e-01 1.74456477e-01 -2.46559888e-01
-8.02446082e-02 -1.36260021e+00 5.54855406e-01 9.46146190e-01
3.28044564e-01 5.24984717e-01 1.29732549e-01 5.82617879e-01
7.09633231e-02 -3.01821101e-02 5.08295596e-01 1.58058137e-01
-1.16661024e+00 7.57330537e-01 -3.96330923e-01 -1.39437532e-02
1.29190147e+00 -6.42709613e-01 -1.37550905e-01 -7.84722790e-02
-6.67738020e-01 4.08419728e-01 3.94316703e-01 2.95276225e-01
8.35237563e-01 -1.59489954e+00 -1.70207888e-01 2.94592232e-01
-3.59865397e-01 -5.97033799e-01 5.58381021e-01 5.67191064e-01
-1.10060465e+00 1.82396635e-01 -9.46161270e-01 -2.55703539e-01
-1.21555555e+00 1.41297328e+00 5.16061783e-01 -3.75935018e-01
-6.39001250e-01 3.63256574e-01 -4.49939638e-01 -6.55057609e-01
2.34425858e-01 4.57056612e-01 -6.71112299e-01 -3.11952353e-01
2.34910265e-01 7.68753111e-01 -3.84291440e-01 -3.82766783e-01
-4.58643377e-01 2.69173294e-01 1.62247092e-01 6.28998876e-02
1.05110228e+00 -3.71445805e-01 -5.70633411e-01 -1.34500340e-01
1.06093371e+00 -2.24920034e-01 -4.92716104e-01 -2.00565651e-01
-2.90895104e-01 -8.19939077e-01 1.20233610e-01 -3.88067156e-01
-1.54671264e+00 3.78087312e-01 8.14389288e-01 5.57246029e-01
1.39714122e+00 -8.40522587e-01 9.55204606e-01 3.11063647e-01
7.33314931e-01 -1.35433292e+00 -4.93141543e-03 3.84123385e-01
8.43896195e-02 -7.28414893e-01 9.78274085e-03 -6.36366367e-01
-2.56310016e-01 1.38352561e+00 1.15431798e+00 -5.53722233e-02
7.99953818e-01 1.05037443e-01 -2.95379341e-01 -1.18852042e-01
-5.34777403e-01 -2.68799849e-02 -5.35427511e-01 9.08419728e-01
-3.82528096e-01 3.58802676e-01 -7.33752549e-01 3.84781659e-01
-3.57708298e-02 2.35370509e-02 8.00086021e-01 7.51265228e-01
-6.54183745e-01 -1.23205829e+00 -2.20451191e-01 4.25730884e-01
4.14175428e-02 5.84406480e-02 2.95993924e-01 1.10486794e+00
4.16724473e-01 9.13461924e-01 6.25243708e-02 -9.85747695e-01
1.27008617e-01 -3.75172973e-01 3.39626908e-01 -7.33006820e-02
-3.97092998e-01 -2.66877323e-01 -2.28698447e-01 -4.25558507e-01
-3.10329441e-02 7.20653906e-02 -1.60288596e+00 -1.18445218e+00
-4.21536565e-01 5.81912637e-01 6.15290344e-01 7.21101165e-01
4.39809352e-01 1.24448371e+00 1.18349743e+00 -2.49310270e-01
-2.26251319e-01 -3.48429561e-01 -1.08158374e+00 6.44923747e-02
-2.56055295e-01 -8.41939926e-01 -4.25088972e-01 -8.40766251e-01] | [5.858583927154541, 1.7233357429504395] |
c5e53df2-f62d-419b-ab6d-f5a16f110e35 | opinesum-entailment-based-self-training-for | 2212.10791 | null | https://arxiv.org/abs/2212.10791v1 | https://arxiv.org/pdf/2212.10791v1.pdf | OpineSum: Entailment-based self-training for abstractive opinion summarization | A typical product or place often has hundreds of reviews, and summarization of these texts is an important and challenging problem. Recent progress on abstractive summarization in domains such as news has been driven by supervised systems trained on hundreds of thousands of news articles paired with human-written summaries. However for opinion texts, such large scale datasets are rarely available. Unsupervised methods, self-training, and few-shot learning approaches bridge that gap. In this work, we present a novel self-training approach, OpineSum, for abstractive opinion summarization. The summaries in this approach are built using a novel application of textual entailment and capture the consensus of opinions across the various reviews for an item. This method can be used to obtain silver-standard summaries on a large scale and train both unsupervised and few-shot abstractive summarization systems. OpineSum achieves state-of-the-art performance in both settings. | ['Joshua Maynez', 'Annie Louis'] | 2022-12-21 | null | null | null | null | ['abstractive-text-summarization'] | ['natural-language-processing'] | [ 4.98839408e-01 2.86193430e-01 -4.96731102e-01 -3.14389974e-01
-1.37180138e+00 -4.67007101e-01 7.08082378e-01 1.03968596e+00
-1.49473444e-01 9.89273667e-01 9.79900897e-01 1.78725436e-01
1.69015333e-01 -5.91405213e-01 -4.35534179e-01 -3.67581934e-01
3.99245203e-01 6.54174984e-01 2.08440140e-01 -4.24842596e-01
9.04829144e-01 -1.39477968e-01 -1.63090372e+00 6.86875463e-01
1.42034686e+00 6.86880648e-01 -1.32338643e-01 9.59712982e-01
-5.30524492e-01 9.08750594e-01 -1.05532098e+00 -8.24775219e-01
-1.06646821e-01 -8.74371231e-01 -6.45833433e-01 2.68268257e-01
7.67997682e-01 -2.67428190e-01 2.05612168e-01 1.01061726e+00
8.56806695e-01 4.14032757e-01 9.14437056e-01 -7.78114498e-01
-6.16296470e-01 1.08267426e+00 -6.79642975e-01 3.65400642e-01
6.10654771e-01 -2.22037256e-01 1.55970955e+00 -8.86266589e-01
8.35414112e-01 1.04250383e+00 5.37303269e-01 3.60837638e-01
-1.01323521e+00 -6.11827448e-02 1.49797097e-01 6.04384765e-02
-5.97916067e-01 -6.99807405e-01 5.84645152e-01 -1.84301734e-01
1.27204394e+00 2.52614141e-01 7.48692393e-01 1.00733292e+00
5.40058672e-01 1.18204248e+00 4.29802299e-01 -5.25996804e-01
7.13770509e-01 9.28813219e-02 7.39493966e-01 2.45844796e-01
8.61201644e-01 -1.04651690e+00 -7.65726686e-01 -4.66279209e-01
-2.63963193e-01 1.84109047e-01 -1.53638616e-01 9.99054760e-02
-9.69144940e-01 1.06069553e+00 -2.05040589e-01 1.87069327e-01
-7.90125966e-01 -1.85117126e-01 1.02285564e+00 4.36188549e-01
1.28329337e+00 9.85500216e-01 -3.23458225e-01 -2.13077590e-01
-1.51556969e+00 5.27148247e-01 1.42419481e+00 8.34316492e-01
5.68159640e-01 -3.67725594e-03 -5.66739678e-01 8.62576663e-01
-3.25809777e-01 5.32101512e-01 8.54045689e-01 -7.75427282e-01
6.12752557e-01 6.99541211e-01 1.35073960e-01 -9.62970257e-01
-1.75841674e-01 -3.08532387e-01 -9.11444366e-01 -3.26648742e-01
-3.44262481e-01 -5.01311779e-01 -8.72351825e-01 9.14427400e-01
2.24667475e-01 -9.24967974e-02 5.02345979e-01 3.08457434e-01
1.27960503e+00 1.03824925e+00 -3.13809574e-01 -8.08226049e-01
1.23335910e+00 -1.33466160e+00 -1.07829523e+00 -4.14791703e-01
6.09588563e-01 -7.84364104e-01 6.77524865e-01 4.75717813e-01
-1.35767233e+00 -3.17380846e-01 -1.28099561e+00 -8.09099451e-02
-4.26134050e-01 6.95060939e-02 2.70091504e-01 1.33709654e-01
-7.58746445e-01 9.06018615e-01 -4.79425192e-01 -7.47901976e-01
4.65637177e-01 2.70010494e-02 -1.54902294e-01 -2.26318598e-01
-1.02907336e+00 1.01662207e+00 2.11070225e-01 -3.60166162e-01
-2.37890825e-01 -7.40709841e-01 -1.06817997e+00 1.51008666e-01
6.07325315e-01 -1.01277351e+00 1.68176365e+00 -9.06534433e-01
-1.48845255e+00 6.18483782e-01 -4.20545727e-01 -7.32145429e-01
7.50752389e-02 -6.61913276e-01 -3.64988357e-01 3.30701739e-01
5.31985998e-01 4.47372124e-02 8.61522794e-01 -9.07593608e-01
-7.12553799e-01 -3.75515461e-01 -3.29316646e-01 3.62893254e-01
-3.64444584e-01 1.60545126e-01 -1.09868944e-01 -5.36237359e-01
-4.81884152e-01 -4.46957022e-01 -4.65999514e-01 -8.63912106e-01
-7.29597688e-01 -5.03311574e-01 5.11529505e-01 -5.43612897e-01
1.58946383e+00 -1.63650823e+00 1.74763232e-01 -2.91077018e-01
2.46190891e-01 5.54937363e-01 -3.60209495e-01 1.11716413e+00
2.72905886e-01 1.76134914e-01 -3.09386045e-01 -6.67771518e-01
1.17539868e-01 4.01209854e-02 -6.35490894e-01 1.59270167e-01
2.73101240e-01 9.92610514e-01 -1.17080331e+00 -5.66584647e-01
-1.89411178e-01 -1.38846353e-01 -2.84355015e-01 2.36865684e-01
-4.98942822e-01 -7.60724097e-02 -5.94849288e-01 4.12151873e-01
3.96233827e-01 -2.31529519e-01 -5.14085889e-02 6.36131223e-03
-2.67551422e-01 4.44264084e-01 -6.49382293e-01 1.67900932e+00
-2.36799553e-01 6.44016206e-01 -3.42364341e-01 -1.18569815e+00
8.64451349e-01 2.91381598e-01 3.69672835e-01 -4.15200144e-01
3.57597888e-01 2.07389325e-01 -4.70385224e-01 -5.77172577e-01
1.20444775e+00 -1.42935395e-01 -4.21807766e-01 9.40044224e-01
3.08647573e-01 -7.19515204e-01 1.08275223e+00 8.27806592e-01
1.27252138e+00 -4.53954250e-01 1.07503915e+00 -8.83633643e-02
2.95991778e-01 4.85286981e-01 3.77029061e-01 9.65191662e-01
2.44598672e-01 9.91923094e-01 7.12253153e-01 -2.53879517e-01
-1.13802648e+00 -5.27190208e-01 3.19627494e-01 1.04756856e+00
-6.85093328e-02 -1.03456903e+00 -7.91841984e-01 -6.37149394e-01
4.89278324e-02 1.15099669e+00 -5.90800405e-01 -2.26311326e-01
-2.93548018e-01 -6.22708559e-01 7.62695372e-02 3.94665956e-01
8.53113309e-02 -1.14022648e+00 -4.98239905e-01 5.02421677e-01
-2.03638583e-01 -9.34052348e-01 -5.32533109e-01 -8.86466075e-03
-7.61641204e-01 -9.18847561e-01 -9.31904435e-01 -5.64949572e-01
5.34379900e-01 3.88535678e-01 1.50785756e+00 -1.61792338e-01
-1.79498345e-02 4.09365773e-01 -8.94093812e-01 -1.03984571e+00
-5.29438436e-01 5.17002046e-01 1.56629443e-01 1.87823991e-03
5.26429534e-01 -5.28451085e-01 -4.37123537e-01 -4.03541446e-01
-9.69696701e-01 -2.22830132e-01 7.67570138e-01 7.57944465e-01
5.49333930e-01 -2.88659483e-01 1.19271541e+00 -1.43331790e+00
1.40057933e+00 -5.82811713e-01 3.43210027e-02 4.57698733e-01
-6.08032286e-01 1.19316325e-01 8.08465302e-01 -3.27757567e-01
-1.05645382e+00 -5.47096014e-01 -6.73065558e-02 1.22171901e-01
6.90283030e-02 1.01975453e+00 3.80727768e-01 7.59547293e-01
8.33941758e-01 2.36581102e-01 -4.09715436e-02 -2.54669130e-01
6.46308064e-01 9.48163509e-01 5.20345569e-01 -1.06455714e-01
4.15429622e-01 2.87687898e-01 -5.77295780e-01 -1.14098418e+00
-1.72259223e+00 -9.48674738e-01 -5.25832832e-01 -1.96320992e-02
6.22626424e-01 -8.65790129e-01 1.19842999e-01 1.42571032e-01
-1.29435515e+00 1.68656901e-01 -1.09799933e+00 2.37140372e-01
-4.08259392e-01 6.37878597e-01 -3.92536014e-01 -6.32694542e-01
-1.41817260e+00 -3.82298440e-01 1.24666739e+00 6.74733698e-01
-7.78704226e-01 -8.39663267e-01 7.49247730e-01 1.39343262e-01
5.36061347e-01 2.12850109e-01 4.25926983e-01 -1.45048678e+00
1.70225114e-01 -7.86875904e-01 1.80124238e-01 4.77483481e-01
2.35053778e-01 3.45879167e-01 -4.07256156e-01 -8.98165554e-02
3.90623249e-02 -6.85374022e-01 1.34496307e+00 6.43516779e-01
3.80926788e-01 -6.19427621e-01 -1.80844277e-01 -3.42690080e-01
1.01075673e+00 -3.64868909e-01 4.22519863e-01 5.48183285e-02
4.62930620e-01 6.28728569e-01 7.70523012e-01 8.62150133e-01
3.52253467e-01 4.46186103e-02 -2.11316243e-01 1.86785832e-01
1.09651931e-01 -7.92290270e-02 3.76051873e-01 1.55771744e+00
1.15658522e-01 -6.10354722e-01 -5.14156818e-01 8.90841901e-01
-2.34302616e+00 -1.44719303e+00 -5.12267016e-02 1.76977181e+00
9.87727225e-01 2.50180036e-01 -2.56877039e-02 -6.88997284e-02
6.75369859e-01 6.05274439e-01 -7.83952296e-01 -8.75925183e-01
-2.51883030e-01 2.74157465e-01 9.80357602e-02 1.71408996e-01
-8.63683343e-01 7.38324523e-01 6.17072153e+00 8.56593192e-01
-7.09901452e-01 -1.08481795e-02 4.87937570e-01 -3.18314284e-01
-4.73713219e-01 -5.68812378e-02 -8.66061747e-01 2.24331081e-01
1.21308661e+00 -9.86494720e-01 -4.16016072e-01 1.03949630e+00
2.23088413e-01 -6.14571214e-01 -1.04563832e+00 7.54065931e-01
9.12500143e-01 -1.71954608e+00 3.10313553e-01 -3.40146929e-01
1.56312501e+00 8.65319595e-02 -3.17439675e-01 4.74927008e-01
4.55601811e-01 -5.91066062e-01 3.17022830e-01 4.88583416e-01
5.23282647e-01 -8.61898005e-01 1.16680157e+00 5.42953610e-01
-9.33879316e-01 1.41774893e-01 -7.44454265e-01 -1.35886252e-01
6.53972745e-01 1.11467779e+00 -5.18948078e-01 8.74660492e-01
2.58608907e-01 1.41108489e+00 -4.73425478e-01 9.47197974e-01
-3.28366041e-01 6.78740680e-01 -1.08476104e-02 -5.15472114e-01
3.18590045e-01 -2.34588012e-01 7.07837999e-01 1.45649838e+00
2.57357031e-01 2.46223122e-01 2.76312768e-01 2.61177957e-01
-3.92738193e-01 4.97475356e-01 -7.52511024e-01 -3.38438302e-01
1.90526500e-01 1.47361863e+00 -6.22434020e-01 -1.19002450e+00
-2.75451869e-01 1.02716005e+00 2.93677151e-01 1.24184266e-01
-2.76015252e-01 -6.99484587e-01 3.52104396e-01 -1.30262762e-01
7.14260578e-01 2.84295022e-01 -2.34123677e-01 -1.57232356e+00
-5.49313873e-02 -1.02788258e+00 2.85009533e-01 -6.49030566e-01
-1.59197724e+00 5.99276721e-01 -4.30135466e-02 -1.23052478e+00
-5.49082041e-01 -8.71954635e-02 -1.47458243e+00 1.96693465e-01
-1.34724319e+00 -7.35403121e-01 -2.40406152e-02 -6.10158704e-02
1.32844007e+00 -2.53953218e-01 7.67694414e-01 -2.57484674e-01
-7.22030222e-01 1.89736232e-01 5.90358377e-01 -3.45970601e-01
1.04277658e+00 -1.47986853e+00 6.72049642e-01 8.29459965e-01
-1.06386445e-03 5.08976936e-01 1.18973112e+00 -8.13495815e-01
-1.23122871e+00 -1.10910976e+00 1.21978259e+00 -4.39251333e-01
7.88485110e-01 -2.20057517e-02 -8.23545396e-01 3.73996258e-01
9.65265691e-01 -4.51229692e-01 1.06264949e+00 2.10703760e-01
-8.86112526e-02 -6.03464283e-02 -7.65638709e-01 5.78233361e-01
5.78660429e-01 -2.54511863e-01 -1.30703390e+00 6.49319470e-01
8.39539409e-01 -8.91474038e-02 -7.09551394e-01 1.01320662e-01
4.25570428e-01 -8.33092213e-01 3.98711592e-01 -7.29180574e-01
1.09158170e+00 7.00020269e-02 2.62674332e-01 -2.03641677e+00
2.71966755e-02 -8.51430535e-01 -4.42692637e-01 1.41045558e+00
5.69149315e-01 -4.23830092e-01 5.66933930e-01 4.07534420e-01
-3.57905090e-01 -7.85064638e-01 -2.59624660e-01 -3.24910372e-01
-8.78516287e-02 1.43631279e-01 2.05128044e-01 5.06541550e-01
5.97407043e-01 1.22161698e+00 -4.03657436e-01 -4.81730044e-01
5.78174472e-01 4.66043770e-01 1.04397893e+00 -1.41165257e+00
2.06176296e-01 -3.93663287e-01 -2.09621266e-01 -9.89082694e-01
2.76434839e-01 -5.00082612e-01 2.51294076e-01 -2.39836192e+00
5.30058742e-01 6.77216709e-01 1.86914518e-01 -8.25582221e-02
-5.51035941e-01 -4.24228259e-04 -9.47427675e-02 1.32826373e-01
-1.41374230e+00 7.46130645e-01 1.01549971e+00 -4.38361764e-01
-3.12010080e-01 6.77553639e-02 -1.23902416e+00 7.61725605e-01
8.47549856e-01 -4.88488555e-01 -3.60359937e-01 -1.15848798e-03
4.37711418e-01 -1.20648809e-01 -6.70221150e-01 -7.45466352e-01
7.36798108e-01 1.16548076e-01 1.11988662e-02 -1.22069120e+00
-7.58792236e-02 5.60727231e-02 -4.21310902e-01 1.05445080e-01
-5.93054712e-01 1.85959131e-01 -9.83404294e-02 8.62449050e-01
-6.03477359e-01 -4.87135440e-01 4.21541214e-01 -4.62378144e-01
-3.50116670e-01 1.05218999e-01 -5.81973612e-01 5.00255644e-01
7.53488123e-01 -7.37484992e-02 -5.66760361e-01 -6.23389542e-01
-2.61686653e-01 4.83751267e-01 2.89274007e-01 1.44814536e-01
6.52255058e-01 -8.60744536e-01 -1.33987343e+00 -3.33570778e-01
3.30199957e-01 1.66543573e-01 2.23062769e-01 8.75551164e-01
-2.26677403e-01 3.37607116e-01 8.14147890e-02 -2.05852181e-01
-1.23529279e+00 4.09016818e-01 -3.69470775e-01 -8.92122388e-01
-7.65544653e-01 5.84156275e-01 -2.23197103e-01 -2.20013335e-01
-5.33455424e-02 -3.83083165e-01 -7.70196915e-01 1.00404489e+00
1.21075737e+00 3.02800626e-01 2.32955560e-01 -3.96318853e-01
3.57424207e-02 5.62599003e-01 -6.82600796e-01 -7.19225453e-03
1.68536115e+00 -2.43557155e-01 -3.31271797e-01 8.20996106e-01
1.08996415e+00 2.36783937e-01 -5.80146074e-01 -3.25912446e-01
1.53394133e-01 1.09663427e-01 -2.59156495e-01 -4.22921091e-01
-4.80537862e-01 5.03244877e-01 -5.55679917e-01 7.26894915e-01
7.71447957e-01 2.15707660e-01 1.04488397e+00 8.81483436e-01
-2.54358888e-01 -1.42825437e+00 3.37037534e-01 7.24626660e-01
9.15746987e-01 -1.44454587e+00 7.12373137e-01 5.63161224e-02
-1.04056227e+00 1.07077241e+00 1.73972294e-01 -5.48684537e-01
3.57852548e-01 -1.41338364e-03 -5.84057793e-02 -4.89471763e-01
-1.25742853e+00 -2.31962606e-01 4.41313356e-01 1.95802212e-01
3.88895005e-01 -2.21591830e-01 -8.07658553e-01 8.59722912e-01
-1.84762597e-01 -2.31188610e-01 9.82342005e-01 1.23214316e+00
-1.00980496e+00 -7.80505657e-01 -6.13241643e-02 1.13058102e+00
-7.17824697e-01 -2.85779417e-01 -7.87551939e-01 2.70008028e-01
-7.16712832e-01 1.21786904e+00 1.94283854e-02 -6.01949878e-02
5.91870666e-01 1.07976750e-01 9.86813605e-02 -1.33738589e+00
-7.74273217e-01 -2.80207433e-02 4.86829787e-01 -1.91437468e-01
-6.89395189e-01 -6.92378879e-01 -1.02385628e+00 -3.99349451e-01
-6.11718357e-01 7.12610483e-01 4.90080774e-01 1.15783763e+00
5.66900671e-01 5.39643228e-01 6.22079253e-01 -1.22827888e+00
-9.23553169e-01 -1.35308719e+00 -6.24246776e-01 5.30481815e-01
4.37217385e-01 1.45293365e-03 -5.37807584e-01 1.23465911e-01] | [12.456968307495117, 9.425171852111816] |
afa5698f-7740-4419-8939-af4dfb0019df | deep-blind-video-decaptioning-by-temporal | 1905.02949 | null | https://arxiv.org/abs/1905.02949v1 | https://arxiv.org/pdf/1905.02949v1.pdf | Deep Blind Video Decaptioning by Temporal Aggregation and Recurrence | Blind video decaptioning is a problem of automatically removing text overlays and inpainting the occluded parts in videos without any input masks. While recent deep learning based inpainting methods deal with a single image and mostly assume that the positions of the corrupted pixels are known, we aim at automatic text removal in video sequences without mask information. In this paper, we propose a simple yet effective framework for fast blind video decaptioning. We construct an encoder-decoder model, where the encoder takes multiple source frames that can provide visible pixels revealed from the scene dynamics. These hints are aggregated and fed into the decoder. We apply a residual connection from the input frame to the decoder output to enforce our network to focus on the corrupted regions only. Our proposed model was ranked in the first place in the ECCV Chalearn 2018 LAP Inpainting Competition Track2: Video decaptioning. In addition, we further improve this strong model by applying a recurrent feedback. The recurrent feedback not only enforces temporal coherence but also provides strong clues on where the corrupted pixels are. Both qualitative and quantitative experiments demonstrate that our full model produces accurate and temporally consistent video results in real time (50+ fps). | ['Joon-Young Lee', 'Dahun Kim', 'Sanghyun Woo', 'In So Kweon'] | 2019-05-08 | deep-blind-video-decaptioning-by-temporal-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Kim_Deep_Blind_Video_Decaptioning_by_Temporal_Aggregation_and_Recurrence_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Kim_Deep_Blind_Video_Decaptioning_by_Temporal_Aggregation_and_Recurrence_CVPR_2019_paper.pdf | cvpr-2019-6 | ['video-denoising', 'video-to-video-synthesis', 'video-inpainting'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.89546424e-01 -1.04977809e-01 -1.28689501e-02 4.65416498e-02
-5.98485172e-01 -5.30044734e-01 4.05953944e-01 -4.38075185e-01
-3.77834082e-01 6.38756931e-01 5.70474148e-01 5.43136373e-02
3.66991580e-01 -1.20498903e-01 -1.22674382e+00 -6.64709747e-01
1.89289629e-01 -5.73529266e-02 2.53276378e-01 -4.27750051e-02
1.70986980e-01 1.68177634e-01 -1.43094254e+00 6.65122390e-01
7.82766640e-01 6.98632538e-01 4.24755841e-01 9.66046154e-01
2.41496772e-01 1.32592928e+00 -6.57668293e-01 -4.37593043e-01
5.66696823e-01 -5.86202562e-01 -4.21092004e-01 4.40098733e-01
8.33473802e-01 -9.14517879e-01 -1.14884305e+00 1.19455767e+00
3.17730457e-01 -3.03057414e-02 2.43146777e-01 -9.03746068e-01
-5.94626307e-01 5.75923264e-01 -7.99377680e-01 3.58272821e-01
4.98149455e-01 5.58039546e-01 6.91411853e-01 -9.22749817e-01
8.52170229e-01 1.07789302e+00 5.00912011e-01 8.15418839e-01
-1.05847573e+00 -4.43028480e-01 4.40351725e-01 4.41148788e-01
-1.17917061e+00 -9.30301964e-01 7.56760836e-01 -4.32277024e-01
6.52992368e-01 3.55506182e-01 6.23412490e-01 1.26969540e+00
4.42333594e-02 1.07626891e+00 7.54683971e-01 -2.79308140e-01
-5.86353876e-02 -2.50579059e-01 -2.51499802e-01 6.45383358e-01
5.29145151e-02 1.34137213e-01 -9.87807751e-01 2.22401008e-01
9.22635674e-01 3.31233382e-01 -8.67258370e-01 -6.47770315e-02
-1.34915388e+00 4.07565027e-01 4.27086890e-01 1.15118206e-01
-3.52385402e-01 3.22221279e-01 2.82249540e-01 4.35059577e-01
4.34371531e-01 2.08409324e-01 -2.02339485e-01 -9.48083401e-02
-1.44555819e+00 8.77560303e-02 3.77290428e-01 7.34181821e-01
6.23392820e-01 5.48442602e-02 -4.89498466e-01 6.79684520e-01
6.39279932e-02 2.03010216e-01 4.69902068e-01 -1.27950537e+00
8.20144236e-01 1.14237010e-01 3.70003253e-01 -9.61509109e-01
1.44676313e-01 -2.59411521e-02 -8.39462757e-01 3.84486556e-01
2.61040419e-01 -3.33524466e-01 -1.22392917e+00 1.52037454e+00
-3.29902582e-02 4.95191038e-01 -2.11241335e-01 1.37307978e+00
4.99341905e-01 7.95213401e-01 -4.52468634e-01 -4.34476584e-01
9.07091260e-01 -1.35527432e+00 -9.37769532e-01 -3.82710248e-01
1.54106081e-01 -7.77939141e-01 7.32304513e-01 6.45128071e-01
-1.43548703e+00 -3.81306559e-01 -9.71739233e-01 -4.02250826e-01
1.21795721e-01 1.71958506e-01 1.03937171e-01 1.01588152e-01
-1.30259287e+00 7.80167639e-01 -9.28608537e-01 -9.14565623e-02
4.45584118e-01 1.41903937e-01 -5.17234027e-01 -2.89485306e-01
-9.62586820e-01 6.27191782e-01 -3.40232477e-02 3.60581309e-01
-1.02296114e+00 -4.99326289e-01 -7.56425500e-01 -8.01079646e-02
4.21168804e-01 -8.05349350e-01 1.15898561e+00 -1.52672219e+00
-1.48697722e+00 6.88967288e-01 -5.55148900e-01 -7.41612136e-01
1.04199612e+00 -5.60677290e-01 -2.15671197e-01 5.90496719e-01
7.46128187e-02 8.63458753e-01 1.53996110e+00 -1.30179918e+00
-5.72977841e-01 -5.91088980e-02 -7.92654511e-03 3.88638645e-01
-2.90307611e-01 8.62014145e-02 -1.06552410e+00 -1.14818168e+00
1.12105340e-01 -5.36495745e-01 -2.53497995e-02 3.91306996e-01
-4.61640596e-01 4.94081825e-01 1.06716073e+00 -1.30008376e+00
1.15170598e+00 -2.19555640e+00 5.23966312e-01 -2.90673137e-01
5.88012040e-01 2.06502169e-01 -1.42339393e-01 2.83077568e-01
-8.14279690e-02 -1.69096187e-01 -4.75700945e-01 -7.47315586e-01
-2.44670361e-01 1.73894510e-01 -7.00793147e-01 6.25144720e-01
1.28808558e-01 9.11949456e-01 -7.40219295e-01 -3.01699549e-01
3.14321280e-01 6.11075997e-01 -7.23869503e-01 4.18058306e-01
-4.31073487e-01 5.50550103e-01 2.43965775e-01 5.17310739e-01
7.06745505e-01 -5.56842498e-02 -1.41036613e-02 -1.98336899e-01
-1.01623185e-01 2.29765907e-01 -9.66994643e-01 1.94980681e+00
-6.21186048e-02 1.10739684e+00 4.57681656e-01 -7.15487599e-01
2.77849346e-01 3.27716768e-01 5.01113057e-01 -8.08489263e-01
2.21769679e-02 1.86950807e-02 -2.13573962e-01 -7.11623847e-01
6.67936206e-01 2.33969614e-01 4.34700727e-01 3.48800659e-01
4.51122746e-02 2.28369832e-01 3.08414161e-01 4.08980250e-01
1.24143291e+00 3.21642637e-01 -3.42598706e-01 3.14689785e-01
2.07010984e-01 -4.10163462e-01 5.36464989e-01 6.61476374e-01
-2.81663686e-01 1.35237360e+00 6.02867007e-01 -3.42919052e-01
-1.04769409e+00 -8.16963315e-01 3.79781008e-01 8.51204276e-01
4.05668229e-01 -5.86977720e-01 -8.73440504e-01 -7.84975648e-01
-1.62639320e-01 3.04027766e-01 -7.93709099e-01 -4.40139621e-02
-7.68430054e-01 -3.83563727e-01 2.76756942e-01 2.16858730e-01
4.90416229e-01 -9.85640526e-01 -3.93762469e-01 3.20162885e-02
-6.47968590e-01 -1.24564171e+00 -1.04360700e+00 4.17007767e-02
-7.75051177e-01 -1.05736148e+00 -1.05133522e+00 -8.07985663e-01
8.33343625e-01 6.33023620e-01 8.10705662e-01 2.57684976e-01
-1.84492990e-01 1.08745590e-01 -3.11937600e-01 1.72450766e-01
-3.62357497e-01 -4.53241974e-01 -1.07577331e-02 4.66955245e-01
-2.61539817e-01 -6.51680589e-01 -8.67169201e-01 1.23651356e-01
-1.04712737e+00 4.50816453e-01 5.90516210e-01 8.26957285e-01
4.65016007e-01 -9.34496000e-02 1.82422088e-03 -6.48028731e-01
2.68684119e-01 -2.36560896e-01 -4.62690592e-01 1.48857728e-01
-6.03837445e-02 1.40219584e-01 7.15525925e-01 -5.07786453e-01
-8.97109687e-01 2.20567465e-01 -1.89927116e-03 -1.20910466e+00
1.08047128e-01 1.85043421e-02 -2.16181785e-01 -1.79336265e-01
4.77153033e-01 4.31009829e-01 -9.29649025e-02 -6.00134790e-01
2.97329217e-01 3.70763153e-01 9.40883696e-01 -1.65539593e-01
8.18954289e-01 6.34506643e-01 -3.78087968e-01 -5.93043864e-01
-5.68800569e-01 -2.30776265e-01 -4.76020128e-01 -2.95579553e-01
6.88690305e-01 -1.08853197e+00 -5.31372309e-01 6.45872176e-01
-1.30909574e+00 -5.88921249e-01 -2.20672265e-01 1.04652561e-01
-5.06892323e-01 8.33871782e-01 -8.01077664e-01 -5.39392889e-01
-2.38799602e-01 -1.13237011e+00 1.20412505e+00 5.66832162e-02
1.13373864e-02 -5.40470660e-01 -1.35997176e-01 3.08495849e-01
1.77654475e-01 -7.58182928e-02 3.12561482e-01 1.12236388e-01
-9.58770454e-01 -3.57446782e-02 -2.49526262e-01 3.22310090e-01
2.05937698e-01 4.19605412e-02 -1.08845508e+00 -3.13052505e-01
1.33591235e-01 1.08092288e-02 1.54565954e+00 4.73397613e-01
1.27902031e+00 -6.59084141e-01 -1.48863658e-01 9.48632061e-01
1.28083789e+00 -1.53559819e-01 8.26772690e-01 1.73985258e-01
9.44020867e-01 4.04849499e-01 2.76499480e-01 3.15959871e-01
9.99919847e-02 7.77464151e-01 7.80896723e-01 -1.27469927e-01
-6.65218830e-01 -5.31395912e-01 9.10948634e-01 7.19359338e-01
4.72387336e-02 -4.87870067e-01 -5.14564991e-01 6.36206865e-01
-2.14993310e+00 -1.08017135e+00 -2.47028451e-02 2.26401401e+00
9.87277269e-01 1.12309419e-01 5.41293398e-02 4.38708579e-03
8.29522610e-01 3.76199961e-01 -6.45791352e-01 5.76623380e-02
-5.84466100e-01 -2.05077529e-02 5.64876139e-01 8.91324699e-01
-1.23087108e+00 9.97892618e-01 5.94180059e+00 4.76097196e-01
-1.16011190e+00 1.28684148e-01 5.79086244e-01 -6.72883689e-01
-7.92374760e-02 -1.15317488e-02 -2.90981233e-01 8.82923841e-01
5.81198335e-01 1.96069852e-01 7.77889073e-01 1.68846563e-01
4.46816325e-01 -2.81291127e-01 -1.05559242e+00 1.30905914e+00
2.58948743e-01 -1.65627897e+00 1.84662521e-01 -1.57652870e-01
9.80137289e-01 5.00272699e-02 1.16195962e-01 -2.64577001e-01
9.24477726e-02 -1.00316143e+00 1.13104594e+00 9.46117222e-01
8.18816662e-01 -5.30630589e-01 3.31087321e-01 2.22759917e-01
-8.80630434e-01 -2.14073136e-01 -2.06337541e-01 -9.54185203e-02
2.72300273e-01 5.13504267e-01 -5.16926467e-01 3.77797902e-01
7.65988350e-01 1.26694214e+00 -6.06414139e-01 1.33008838e+00
-4.61527705e-01 5.65100729e-01 -7.27513758e-03 6.16082788e-01
-9.05056521e-02 -1.04228288e-01 8.10737431e-01 1.17297053e+00
2.85002500e-01 -8.02283734e-03 -1.84795573e-01 7.14741468e-01
-4.49735433e-01 -4.05472994e-01 -4.88413841e-01 4.47931103e-02
-2.65451352e-04 8.31354678e-01 -4.18264687e-01 -3.39743912e-01
-4.88005310e-01 1.81951654e+00 1.53211161e-01 6.94414079e-01
-8.70638013e-01 -1.44225776e-01 8.43089998e-01 1.66734084e-01
6.07003868e-01 -2.06108049e-01 -3.43621761e-01 -1.70004952e+00
3.58770907e-01 -8.97831202e-01 1.20890111e-01 -1.12650061e+00
-9.80258226e-01 3.37638378e-01 -6.49440944e-01 -1.38810813e+00
-1.36495218e-01 -4.24114853e-01 -5.79031944e-01 8.12199831e-01
-1.40925634e+00 -8.71270418e-01 -3.42077076e-01 8.21498454e-01
8.69484067e-01 3.38503532e-02 2.66362786e-01 3.70180815e-01
-7.17432022e-01 6.06988549e-01 1.63268536e-01 2.33078599e-01
1.09111512e+00 -9.65705097e-01 5.20195961e-01 1.61323488e+00
2.27428451e-01 3.46844971e-01 1.01963139e+00 -8.10142577e-01
-1.57304013e+00 -1.13623047e+00 7.00040877e-01 -3.67450804e-01
5.82835555e-01 -4.34594780e-01 -1.05050743e+00 6.80099785e-01
5.03676176e-01 2.13090703e-01 -2.57862389e-01 -7.00083852e-01
-4.05449480e-01 -6.57881126e-02 -8.37612748e-01 6.44090712e-01
1.03248286e+00 -6.13625407e-01 -3.27128261e-01 4.57271397e-01
9.32368577e-01 -6.15915358e-01 -3.22283953e-02 -2.12074816e-02
3.74963433e-01 -1.18576515e+00 7.90640831e-01 -3.67875516e-01
8.19504738e-01 -4.79573280e-01 -3.69087011e-02 -9.65025008e-01
3.90311927e-02 -1.18189609e+00 -6.44737780e-01 9.95028675e-01
1.13187656e-01 -1.38786077e-01 8.52331042e-01 3.79675925e-01
-2.10862756e-01 -3.07684302e-01 -8.92921746e-01 -3.96382660e-01
-2.42036432e-01 -3.90974939e-01 1.03717059e-01 7.71253526e-01
-2.20372379e-01 -8.92349035e-02 -1.04658055e+00 2.92505205e-01
7.14721560e-01 -3.91410515e-02 6.01599336e-01 -6.90232396e-01
-5.07054329e-01 -3.90414983e-01 -2.94887006e-01 -1.47765172e+00
-6.01701848e-02 -4.49892491e-01 2.83517897e-01 -1.46686351e+00
2.90112853e-01 3.39655221e-01 -7.52280280e-02 4.74878371e-01
-3.75322282e-01 4.30171221e-01 3.55463773e-01 5.76863825e-01
-7.27514505e-01 4.72297937e-01 1.17775464e+00 -2.76163280e-01
-4.56848219e-02 -1.41352609e-01 -5.45469999e-01 5.81741750e-01
4.87252712e-01 -3.46303582e-01 -5.74594364e-02 -9.40926790e-01
8.35276246e-02 2.39142403e-01 7.26560473e-01 -9.03242826e-01
5.53508282e-01 4.91509847e-02 5.65694630e-01 -3.23038161e-01
5.18005013e-01 -6.87888086e-01 4.91020903e-02 2.76172876e-01
-2.55155474e-01 1.17964449e-03 1.72847480e-01 7.43841290e-01
-3.39768380e-01 -4.17138226e-02 7.73902595e-01 1.67164905e-03
-7.35027015e-01 4.33466166e-01 -3.09781641e-01 -1.34267092e-01
6.60311043e-01 -2.40645587e-01 -3.87623489e-01 -7.04239964e-01
-8.50838184e-01 1.16953261e-01 8.93281400e-01 4.87122357e-01
8.65234017e-01 -1.02555656e+00 -6.79377079e-01 4.81829077e-01
-3.40990156e-01 -1.96179926e-01 4.77532148e-01 9.88514185e-01
-6.45226240e-01 1.87976018e-01 -1.99303821e-01 -4.89823580e-01
-1.24345720e+00 7.08120704e-01 5.21830857e-01 1.74167052e-01
-1.05345941e+00 1.01039493e+00 4.60747719e-01 4.33428586e-01
6.64329827e-01 -1.89932838e-01 1.70670122e-01 -5.10936752e-02
7.49324262e-01 2.06745729e-01 5.85125526e-03 -5.71482420e-01
-1.38336286e-01 3.39383543e-01 -2.14763373e-01 -4.27086473e-01
1.22978246e+00 -3.91719967e-01 -2.13220030e-01 2.94487536e-01
1.11583579e+00 3.00471604e-01 -2.01527071e+00 -2.69934475e-01
-3.24814230e-01 -6.76083326e-01 1.92207228e-02 -7.68034577e-01
-1.41030335e+00 9.46427464e-01 3.74583155e-01 -5.43368906e-02
1.38543856e+00 -2.03831017e-01 9.14297283e-01 -7.03315716e-03
3.16781225e-03 -9.84118760e-01 2.19485149e-01 3.73793483e-01
1.10584128e+00 -1.11976814e+00 -1.69863224e-01 -1.76666856e-01
-6.34487629e-01 1.20119405e+00 5.60534358e-01 -2.24967644e-01
2.45608523e-01 3.51437718e-01 2.67898683e-02 2.05107525e-01
-9.68632936e-01 -1.23211751e-02 2.40181535e-01 5.13343334e-01
9.07816812e-02 -3.12738866e-01 8.61567110e-02 4.33323324e-01
-1.94038153e-02 -5.05386367e-02 7.74980307e-01 7.17581511e-01
-4.61052090e-01 -9.33171451e-01 -5.64461112e-01 1.70740992e-01
-3.56259346e-01 -3.16042334e-01 -7.38166153e-01 3.41313750e-01
6.43683076e-02 8.84596646e-01 -1.00902170e-01 -3.58384401e-01
9.12296493e-03 -1.25159353e-01 5.38162231e-01 -3.37001413e-01
-4.06847358e-01 3.26508850e-01 -2.41270855e-01 -8.81004393e-01
-3.45877081e-01 -6.38598025e-01 -9.99457777e-01 -3.21052879e-01
1.46076903e-01 -8.18907842e-02 3.75373811e-01 8.19035530e-01
5.55108607e-01 5.89481533e-01 6.18454278e-01 -1.26080632e+00
-6.58539310e-02 -8.39352190e-01 -4.04018164e-01 3.28031451e-01
1.16307569e+00 -2.73284376e-01 -5.36218643e-01 5.26293933e-01] | [10.801613807678223, -1.3425040245056152] |
1130125d-1302-4e15-b536-6b21fd00441c | dr2-diffusion-based-robust-degradation | 2303.06885 | null | https://arxiv.org/abs/2303.06885v3 | https://arxiv.org/pdf/2303.06885v3.pdf | DR2: Diffusion-based Robust Degradation Remover for Blind Face Restoration | Blind face restoration usually synthesizes degraded low-quality data with a pre-defined degradation model for training, while more complex cases could happen in the real world. This gap between the assumed and actual degradation hurts the restoration performance where artifacts are often observed in the output. However, it is expensive and infeasible to include every type of degradation to cover real-world cases in the training data. To tackle this robustness issue, we propose Diffusion-based Robust Degradation Remover (DR2) to first transform the degraded image to a coarse but degradation-invariant prediction, then employ an enhancement module to restore the coarse prediction to a high-quality image. By leveraging a well-performing denoising diffusion probabilistic model, our DR2 diffuses input images to a noisy status where various types of degradation give way to Gaussian noise, and then captures semantic information through iterative denoising steps. As a result, DR2 is robust against common degradation (e.g. blur, resize, noise and compression) and compatible with different designs of enhancement modules. Experiments in various settings show that our framework outperforms state-of-the-art methods on heavily degraded synthetic and real-world datasets. | ['Yanfeng Wang', 'Ya zhang', 'Mingyuan Zhou', 'Huangjie Zheng', 'Ziying Zhang', 'Xiaoyun Zhang', 'Zhixin Wang'] | 2023-03-13 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_DR2_Diffusion-Based_Robust_Degradation_Remover_for_Blind_Face_Restoration_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_DR2_Diffusion-Based_Robust_Degradation_Remover_for_Blind_Face_Restoration_CVPR_2023_paper.pdf | cvpr-2023-1 | ['blind-face-restoration'] | ['computer-vision'] | [ 5.44058979e-01 -4.44357127e-01 2.57054240e-01 -1.50377244e-01
-6.46638453e-01 -3.84525359e-01 5.03912032e-01 -4.82321680e-01
8.89587179e-02 6.81862772e-01 6.08460248e-01 -1.83643121e-03
-1.13945276e-01 -7.22276568e-01 -6.24557257e-01 -1.22349751e+00
2.52703130e-01 -9.20040309e-02 1.31077528e-01 -7.07755834e-02
4.65819538e-02 5.41682303e-01 -1.66734707e+00 5.25606215e-01
1.07210588e+00 9.03625965e-01 5.23986161e-01 7.28328943e-01
1.41190588e-01 8.29324484e-01 -8.67958486e-01 -3.92455459e-01
4.51144069e-01 -4.74966139e-01 -2.89158016e-01 5.95596433e-01
2.52157688e-01 -5.01631737e-01 -5.53907335e-01 1.48778701e+00
6.71570837e-01 -3.15702036e-02 6.74735248e-01 -8.54415059e-01
-1.17744553e+00 1.74340382e-01 -8.14743340e-01 2.45532438e-01
4.06385034e-01 4.92681175e-01 2.95285463e-01 -9.45120752e-01
5.94873488e-01 1.66670191e+00 6.40600920e-01 5.99922419e-01
-1.33138490e+00 -4.18541044e-01 -4.90720682e-02 3.78761411e-01
-1.07170284e+00 -8.48620176e-01 9.14019585e-01 -2.25770652e-01
2.74640173e-01 2.34632373e-01 3.10217202e-01 1.49833965e+00
3.85813147e-01 5.47592103e-01 1.63546753e+00 -2.15527996e-01
4.67530191e-01 -1.02871120e-01 -2.85215974e-01 1.48155227e-01
2.56263584e-01 3.32081944e-01 -4.30131227e-01 -6.33812174e-02
5.81484973e-01 1.27844423e-01 -8.95923018e-01 -3.02283186e-02
-9.21738565e-01 1.90093249e-01 3.51880401e-01 2.63633192e-01
-8.03180754e-01 -5.37238047e-02 7.21869245e-02 4.92712796e-01
5.11029482e-01 -1.35059029e-01 -2.28686988e-01 2.20537737e-01
-1.31325281e+00 9.39340740e-02 3.82539779e-01 6.69530034e-01
5.44431508e-01 1.30905762e-01 -4.10043865e-01 1.10589635e+00
5.55759430e-01 7.14376986e-01 5.16108632e-01 -1.21467555e+00
3.29777181e-01 -7.08125019e-03 3.49236161e-01 -9.52151418e-01
1.52215078e-01 -7.37806916e-01 -1.29322147e+00 6.16307080e-01
4.29538459e-01 2.17554986e-01 -1.27574420e+00 1.47111690e+00
2.48402014e-01 3.99442226e-01 2.44430929e-01 1.25503361e+00
4.07352865e-01 6.95239961e-01 -5.85053898e-02 -6.15283012e-01
1.28041637e+00 -8.36287081e-01 -1.12964916e+00 -3.96886438e-01
-2.32752487e-01 -9.84318614e-01 1.14327884e+00 8.07400584e-01
-1.08564043e+00 -7.31399477e-01 -1.07723212e+00 1.87141690e-02
1.54941604e-01 2.13346869e-01 4.89970706e-02 9.46451008e-01
-1.23593020e+00 8.08006525e-01 -6.68887317e-01 -2.69113600e-01
5.69213688e-01 -1.92737415e-01 -3.46121907e-01 -8.91364932e-01
-1.06275368e+00 8.49195063e-01 -8.43209103e-02 4.83455628e-01
-1.43429768e+00 -6.65265858e-01 -5.04459977e-01 -2.14666516e-01
7.98455551e-02 -8.13151598e-01 9.09094274e-01 -1.11086738e+00
-1.53853130e+00 5.58125615e-01 -1.63003698e-01 -2.48915687e-01
8.24774027e-01 -3.22577536e-01 -7.67171681e-01 2.87311554e-01
-1.50576815e-01 1.19336568e-01 1.66985202e+00 -1.90502584e+00
-1.72099635e-01 -4.60733175e-01 -1.29336953e-01 1.96794301e-01
-2.52918720e-01 1.33421078e-01 -5.16234696e-01 -1.02836239e+00
3.58441144e-01 -4.70668405e-01 -1.54225305e-02 4.15893286e-01
-3.39129329e-01 4.91547555e-01 1.04750204e+00 -1.25747120e+00
1.11876810e+00 -2.31227732e+00 1.28885999e-01 -2.19006017e-02
1.99385077e-01 2.89807081e-01 -2.86973149e-01 3.47700775e-01
-5.90961836e-02 -9.48021784e-02 -7.23349214e-01 -6.88364387e-01
-1.63095891e-01 3.14840555e-01 -4.50856090e-01 9.16580021e-01
-1.84486376e-03 2.64169991e-01 -9.17348623e-01 -2.50045985e-01
1.46057174e-01 9.90082204e-01 -2.97464997e-01 3.94345999e-01
6.98038982e-03 6.08148813e-01 -4.01049890e-02 9.99353111e-01
1.21340704e+00 8.91052857e-02 -7.95922801e-02 -5.37279427e-01
2.74323732e-01 -5.28292246e-02 -1.28182340e+00 1.56058669e+00
-5.03132164e-01 5.41806757e-01 6.56206250e-01 -8.02349210e-01
7.96571553e-01 3.26713771e-01 8.29975009e-02 -4.94909763e-01
-3.47399674e-02 1.76874161e-01 -2.45662138e-01 -7.06241906e-01
2.38590404e-01 -3.81522238e-01 6.30950212e-01 2.61487663e-01
-1.40961424e-01 -7.12078139e-02 -1.96252868e-01 2.31057312e-02
1.41238821e+00 1.45276383e-01 -8.96666422e-02 -8.94175693e-02
6.13645852e-01 -5.43791473e-01 8.13873529e-01 5.90015233e-01
-4.50320691e-01 1.05065215e+00 5.22902459e-02 4.77569290e-02
-1.05784023e+00 -1.16998720e+00 -1.82665605e-03 5.34952819e-01
2.59776831e-01 6.20962232e-02 -1.03346121e+00 -4.29862380e-01
-1.72007173e-01 7.22100616e-01 -3.80639285e-01 -3.42882246e-01
-2.86814392e-01 -1.07427490e+00 5.15267611e-01 4.40181941e-02
9.24708605e-01 -9.28838968e-01 -1.16344035e-01 2.17109680e-01
-2.96204001e-01 -1.13882267e+00 -4.72510755e-01 -2.68764943e-01
-8.29768956e-01 -9.71922517e-01 -9.27262127e-01 -5.65166414e-01
8.12619328e-01 5.67853749e-01 8.94676626e-01 1.46102160e-01
-1.47522524e-01 3.39113563e-01 -3.38750780e-01 2.24297568e-01
-8.34072828e-01 -8.80332053e-01 1.65814802e-01 6.85941517e-01
-1.15643479e-01 -8.38904560e-01 -9.79197204e-01 2.28374839e-01
-1.33902156e+00 -1.13864996e-01 7.20728576e-01 9.48239028e-01
4.19819981e-01 9.09729123e-01 4.94122028e-01 -4.17659521e-01
8.72499764e-01 -5.26773810e-01 -1.89280137e-01 1.87467396e-01
-7.34591842e-01 -2.17261631e-02 7.15789080e-01 -8.11559021e-01
-1.54614544e+00 -1.83502063e-01 2.45203972e-02 -6.71635449e-01
-3.27991277e-01 2.64397740e-01 -7.65744269e-01 3.81370857e-02
7.87739038e-01 5.46377659e-01 -1.82918124e-02 -9.61371481e-01
4.97627348e-01 9.03493166e-01 8.62986147e-01 -3.71560156e-01
1.30332220e+00 6.84640288e-01 -1.43537134e-01 -8.02150309e-01
-5.23804247e-01 -3.42622697e-02 -2.56845027e-01 -3.50498855e-01
4.96842086e-01 -1.08022225e+00 -1.96170360e-01 1.09454584e+00
-1.17356682e+00 -3.29709530e-01 -1.99660748e-01 2.19406843e-01
-3.15702915e-01 7.94848502e-01 -8.98551524e-01 -9.01642382e-01
-2.06313014e-01 -1.29184139e+00 8.50465655e-01 2.71678507e-01
3.93944442e-01 -7.39895701e-01 -3.21611375e-01 5.73085487e-01
7.17615068e-01 -4.15255018e-02 6.44705534e-01 2.08288163e-01
-5.12595713e-01 -7.94549007e-03 -2.99103767e-01 1.02368224e+00
4.90276992e-01 -8.56163725e-02 -1.11655319e+00 -5.68414867e-01
7.00728893e-01 -1.65734515e-02 1.00889075e+00 3.25899869e-01
1.12571013e+00 -4.51291740e-01 -1.63888708e-02 6.09549820e-01
1.43162382e+00 -4.29979004e-02 1.18665636e+00 2.61664927e-01
3.93424332e-01 5.27640402e-01 5.80796480e-01 2.61695296e-01
7.63020590e-02 4.80232894e-01 6.02984548e-01 -1.30895808e-01
-9.10872102e-01 -1.26955032e-01 8.01140070e-01 5.90309381e-01
6.41083047e-02 -3.86414468e-01 -4.58382398e-01 5.80780149e-01
-1.32413661e+00 -1.05558622e+00 -5.96835054e-02 2.17895293e+00
1.20246339e+00 3.30848508e-02 -3.88007373e-01 5.29895663e-01
1.02160954e+00 2.86740363e-01 -5.44756114e-01 1.49374649e-01
-5.51665187e-01 -1.08408734e-01 4.96875674e-01 4.46152180e-01
-8.90818894e-01 6.53859735e-01 5.94219255e+00 1.06478763e+00
-1.02560389e+00 3.21241468e-01 7.69111753e-01 1.32197514e-01
-4.05464560e-01 -4.82302532e-02 -2.29118347e-01 8.32436800e-01
7.88288951e-01 1.28754824e-01 9.01048303e-01 3.51234764e-01
7.94574678e-01 -3.97720039e-02 -6.00608647e-01 1.08248532e+00
9.60262716e-02 -7.37341166e-01 -6.50280565e-02 -7.37857297e-02
8.03971827e-01 -3.92380416e-01 2.69992173e-01 -2.13299319e-01
3.62823278e-01 -9.97484803e-01 7.97006905e-01 9.37656701e-01
7.40985334e-01 -2.99090952e-01 5.54096162e-01 3.51313561e-01
-7.27029264e-01 -3.59388262e-01 -5.40338516e-01 2.15435147e-01
3.33880484e-01 1.30900931e+00 -2.77664691e-01 5.89166462e-01
9.16865110e-01 6.89208210e-01 -4.70186472e-01 1.19980001e+00
-6.38407826e-01 9.05286551e-01 -8.49440694e-02 8.88324797e-01
-3.53226364e-01 -3.44442785e-01 7.93793499e-01 1.07896030e+00
6.19504869e-01 7.64948130e-02 -2.74901867e-01 7.75386870e-01
-2.20843881e-01 -3.57363492e-01 -2.41452143e-01 2.07890466e-01
4.52671140e-01 1.07377958e+00 -4.46966261e-01 -3.25722337e-01
-2.39248767e-01 1.58496940e+00 -2.80736864e-01 8.42124164e-01
-7.66403854e-01 4.81650792e-02 9.39565003e-01 1.13939382e-01
2.34256491e-01 -3.27010453e-02 -2.27595136e-01 -1.15763831e+00
3.46828222e-01 -1.31347334e+00 -9.97796282e-03 -1.14343655e+00
-1.55539179e+00 7.38065183e-01 -4.08904225e-01 -1.38857627e+00
2.95340061e-01 -2.75466949e-01 -5.11008322e-01 1.05984235e+00
-1.85709715e+00 -1.06517124e+00 -5.27253747e-01 6.39418244e-01
7.39897490e-01 -9.38639417e-03 2.92048961e-01 6.10295892e-01
-5.64260900e-01 2.30059311e-01 4.06391621e-01 -4.00401592e-01
1.10690141e+00 -1.08503592e+00 1.76966071e-01 1.35226583e+00
-2.16743037e-01 4.47436959e-01 1.25429678e+00 -7.74757147e-01
-1.50646329e+00 -1.26684237e+00 3.33662629e-01 2.37116944e-02
4.77637023e-01 8.10117833e-03 -1.26312459e+00 1.34084493e-01
2.71238804e-01 3.50460440e-01 1.69468448e-01 -7.53169358e-01
-6.01387024e-01 -3.71698260e-01 -1.61962068e+00 5.47843814e-01
1.04709005e+00 -7.30194747e-01 -5.27715087e-01 2.50970334e-01
6.65748775e-01 -2.80108690e-01 -7.40558147e-01 3.14918578e-01
2.91405827e-01 -1.07620156e+00 1.17617059e+00 -1.48105398e-01
5.95872402e-01 -8.29434156e-01 -3.86502415e-01 -1.54514611e+00
-2.32871264e-01 -7.49925256e-01 -4.10827190e-01 1.43455613e+00
-5.52795921e-03 -4.78171200e-01 2.69472629e-01 3.20571184e-01
-1.97469190e-01 -3.92981946e-01 -8.22732687e-01 -8.97372127e-01
-2.37222895e-01 -3.66034031e-01 6.31105304e-01 8.36390913e-01
-7.30830193e-01 -2.49810442e-01 -7.76152015e-01 7.61588216e-01
1.19031560e+00 -1.23342566e-01 4.48289007e-01 -8.40766788e-01
-3.99656624e-01 -1.16545744e-01 -2.88686067e-01 -1.27497458e+00
-1.63981706e-01 -3.08884770e-01 2.54745036e-01 -1.66600525e+00
-4.80142571e-02 -2.01698795e-01 -2.74961025e-01 1.87646180e-01
-3.87042612e-01 4.87460166e-01 -9.25133377e-02 5.32178342e-01
-5.75567260e-02 7.53768802e-01 1.47350681e+00 -4.40042436e-01
-1.24686072e-02 -1.21030379e-02 -8.29872251e-01 4.74597543e-01
4.94987100e-01 -5.56272626e-01 -4.33280319e-01 -4.98040050e-01
-1.68232232e-01 8.94347206e-02 5.09902418e-01 -1.16732454e+00
1.88781634e-01 -2.62767553e-01 4.91822571e-01 -1.73178166e-01
4.25076008e-01 -1.01809812e+00 4.68140304e-01 2.48063073e-01
-2.89989728e-02 -4.94650304e-01 -8.44523609e-02 8.82156610e-01
-2.66159117e-01 -1.70283824e-01 1.13443303e+00 4.51930165e-02
-3.50551873e-01 1.39754176e-01 -2.23370150e-01 -4.94256735e-01
7.18225241e-01 -4.52897638e-01 -6.22485816e-01 -5.88803172e-01
-7.70891786e-01 -2.30216458e-01 8.72710168e-01 3.93047720e-01
7.73254812e-01 -1.10950339e+00 -9.89739299e-01 2.36491218e-01
-3.40758950e-01 -3.36242944e-01 5.96359670e-01 7.17455089e-01
-4.33886856e-01 -4.83751029e-01 -3.53628024e-02 -3.85313302e-01
-1.20385551e+00 7.86180377e-01 4.73819584e-01 -3.13159227e-02
-7.96813011e-01 5.36113918e-01 3.99249531e-02 4.94740345e-02
2.40184516e-01 -4.79832105e-02 -1.14759155e-01 -1.85140118e-01
9.11533773e-01 4.30325061e-01 2.36746192e-01 -7.07311988e-01
-9.44190621e-02 3.36897820e-01 1.98112756e-01 -1.01311877e-01
1.38282311e+00 -7.33471096e-01 -2.75673956e-01 2.58090696e-03
9.30002153e-01 1.03732325e-01 -1.74733794e+00 -3.43120366e-01
-2.38973275e-01 -8.62243950e-01 4.80113924e-01 -1.00477040e+00
-1.48851335e+00 7.03207910e-01 1.07963312e+00 2.01154709e-01
1.86350417e+00 -3.44132513e-01 7.18693078e-01 -2.05993056e-01
1.23524137e-01 -8.55012774e-01 1.64446965e-01 -1.11792788e-01
1.31056130e+00 -8.86284053e-01 1.25639230e-01 -4.76361722e-01
-4.89275277e-01 9.16479051e-01 2.73046792e-01 8.15106407e-02
5.30494392e-01 2.52788663e-01 3.06504786e-01 1.06504753e-01
-5.95854759e-01 9.00276080e-02 7.72579536e-02 9.71261740e-01
-8.15937072e-02 -2.57064283e-01 -1.81021124e-01 5.72045565e-01
1.27722472e-01 -1.35916034e-02 6.84167922e-01 7.24585712e-01
-3.74896407e-01 -1.10808289e+00 -1.11465132e+00 1.30078852e-01
-5.10594487e-01 -1.03637919e-01 1.28913268e-01 1.65379960e-02
3.15794408e-01 1.53361642e+00 -3.91029477e-01 -2.43790731e-01
2.95534611e-01 -1.42429695e-01 2.85065800e-01 -1.42149389e-01
-2.49857396e-01 2.46950895e-01 -2.15539426e-01 -6.84020996e-01
-4.78871912e-01 -6.75783575e-01 -7.33049572e-01 -3.30868512e-01
-1.26204804e-01 -2.86425054e-01 6.97995782e-01 8.15056860e-01
3.73360336e-01 6.87903762e-01 7.47112811e-01 -9.11962986e-01
-7.48730302e-01 -1.20026648e+00 -7.99262226e-01 6.20816886e-01
7.39490449e-01 -6.58401549e-01 -9.41172838e-01 6.03723884e-01] | [11.469862937927246, -2.186467170715332] |
f181a9ac-b639-481a-bce0-1bb229344b12 | impakt-a-dataset-for-open-schema-knowledge | 2212.10770 | null | https://arxiv.org/abs/2212.10770v1 | https://arxiv.org/pdf/2212.10770v1.pdf | ImPaKT: A Dataset for Open-Schema Knowledge Base Construction | Large language models have ushered in a golden age of semantic parsing. The seq2seq paradigm allows for open-schema and abstractive attribute and relation extraction given only small amounts of finetuning data. Language model pretraining has simultaneously enabled great strides in natural language inference, reasoning about entailment and implication in free text. These advances motivate us to construct ImPaKT, a dataset for open-schema information extraction, consisting of around 2500 text snippets from the C4 corpus, in the shopping domain (product buying guides), professionally annotated with extracted attributes, types, attribute summaries (attribute schema discovery from idiosyncratic text), many-to-one relations between compound and atomic attributes, and implication relations. We release this data in hope that it will be useful in fine tuning semantic parsers for information extraction and knowledge base construction across a variety of domains. We evaluate the power of this approach by fine-tuning the open source UL2 language model on a subset of the dataset, extracting a set of implication relations from a corpus of product buying guides, and conducting human evaluations of the resulting predictions. | ['Sumit Sanghai', 'Patrick Murray', 'Bhargav Kanagal', 'Zach Fisher', 'Luke Vilnis'] | 2022-12-21 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 3.01501483e-01 8.51017892e-01 -5.06488383e-01 -1.03526843e+00
-8.04737628e-01 -9.36427712e-01 5.51603734e-01 5.78264534e-01
-3.85192692e-01 8.95317972e-01 6.60975456e-01 -4.88844067e-01
-3.24451596e-01 -1.03785980e+00 -7.71844864e-01 1.14683159e-01
-6.50872290e-02 1.24554145e+00 -5.27810492e-02 -6.34109855e-01
-1.46514410e-02 3.65033261e-02 -1.64246583e+00 1.01630282e+00
6.27763391e-01 1.15178621e+00 3.94265121e-03 3.71692270e-01
-6.48373067e-01 9.38811600e-01 -1.81098223e-01 -1.25961602e+00
1.90448731e-01 -1.57676600e-02 -1.37883890e+00 -3.37141991e-01
3.40172291e-01 8.06273073e-02 1.63047060e-01 7.52565503e-01
3.15178156e-01 -1.32699742e-03 4.55739558e-01 -1.09960437e+00
-7.36681938e-01 1.55926704e+00 5.67718670e-02 1.72720745e-01
6.27202332e-01 -5.96827082e-02 1.85328352e+00 -4.31657374e-01
1.11558843e+00 1.63369918e+00 7.60297894e-01 5.75871050e-01
-1.54119277e+00 -4.16574776e-01 8.55776072e-02 2.41102532e-01
-1.09088385e+00 -6.03654921e-01 2.57310778e-01 -6.11406900e-02
1.95657051e+00 2.32180998e-01 3.72763783e-01 1.30533516e+00
-4.56693433e-02 1.04569066e+00 8.29206824e-01 -5.77290893e-01
7.31987432e-02 3.97128999e-01 4.54749852e-01 5.67814708e-01
3.49238336e-01 1.26222789e-01 -6.09614193e-01 -1.77021384e-01
4.07942206e-01 -6.54612243e-01 2.91245878e-01 -9.98009294e-02
-1.23978698e+00 8.43716323e-01 1.58023387e-01 -1.04252901e-02
-3.66620481e-01 2.47297939e-02 8.41282427e-01 5.22922516e-01
4.29119378e-01 9.46912110e-01 -1.36298561e+00 -2.58938015e-01
-2.12161675e-01 7.32340813e-01 1.50997102e+00 1.43100679e+00
7.33869553e-01 -5.82005203e-01 4.09900723e-03 1.00788784e+00
8.89808983e-02 3.50982219e-01 5.11340499e-01 -1.05731869e+00
8.32804143e-01 8.19831729e-01 4.25945444e-04 -5.47711551e-01
-6.71009064e-01 -7.27227256e-02 -1.66584119e-01 -4.80424702e-01
4.05260295e-01 -2.60989130e-01 -5.16882360e-01 1.78950608e+00
2.41091967e-01 -3.92112672e-01 5.51941991e-01 5.14595330e-01
1.02254260e+00 2.83682227e-01 7.17757761e-01 1.99302658e-01
1.91516197e+00 -4.33812141e-01 -7.15247631e-01 -6.88415110e-01
1.23470557e+00 -5.80480158e-01 1.20204604e+00 4.08905089e-01
-1.00718343e+00 -4.37884480e-01 -7.66768396e-01 -5.21902561e-01
-8.22781384e-01 -3.71863633e-01 1.44445252e+00 4.08259720e-01
-4.48398530e-01 5.65150619e-01 -5.07921040e-01 -6.27529681e-01
5.54185569e-01 3.57285768e-01 -5.13938665e-01 -2.52703786e-01
-1.74602807e+00 1.06765115e+00 1.06178498e+00 -2.77479500e-01
-1.31345242e-01 -9.93430793e-01 -1.18238473e+00 7.75225982e-02
7.44191647e-01 -1.07693434e+00 1.48882508e+00 -9.47235763e-01
-8.34786534e-01 1.31675220e+00 3.62377241e-02 -8.21925521e-01
6.92882910e-02 -4.87008721e-01 -7.26719022e-01 -4.22659963e-01
4.75919515e-01 8.15603793e-01 1.33789629e-01 -7.33471751e-01
-1.02141702e+00 -7.96684384e-01 3.63096207e-01 2.91490972e-01
9.29722786e-02 2.56415188e-01 -1.36359856e-02 -3.92691433e-01
-5.38146496e-02 -5.80468774e-01 -1.34976953e-01 -6.20681465e-01
-6.90896690e-01 -6.94425404e-01 7.55393878e-02 -4.51230228e-01
9.30428505e-01 -2.00052428e+00 -1.81539848e-01 1.86916422e-02
-1.54973745e-01 -1.61664963e-01 -1.71812788e-01 4.32410717e-01
-1.77287817e-01 1.96961716e-01 6.36339933e-02 -1.69852469e-02
5.47038436e-01 8.00777853e-01 -5.01290739e-01 -3.90487731e-01
4.05911118e-01 1.30103600e+00 -8.60733867e-01 -7.61619151e-01
-1.53302118e-01 -1.67056009e-01 -8.29181314e-01 1.52876005e-01
-9.33689177e-01 -2.72139281e-01 -6.57165706e-01 6.61154628e-01
2.92861074e-01 -4.96819429e-02 3.79437774e-01 -3.39879721e-01
1.13169178e-01 1.05313301e+00 -1.08996868e+00 2.04391742e+00
-7.37841904e-01 1.53537467e-01 -1.62376329e-01 -8.61112654e-01
8.08404803e-01 1.13395490e-01 2.86031812e-01 -8.76427114e-01
6.88750222e-02 1.39935702e-01 -4.82092910e-02 -9.08863723e-01
6.39663517e-01 -5.17853677e-01 -6.97951734e-01 4.22244519e-01
4.93380100e-01 -1.91491872e-01 6.00558281e-01 2.76355565e-01
1.04871321e+00 4.32621181e-01 5.18559098e-01 -5.42398512e-01
4.16435570e-01 4.85887706e-01 4.64078963e-01 6.58771217e-01
2.24956021e-01 -1.65228948e-01 6.30671322e-01 -9.09024239e-01
-1.06070220e+00 -1.01297224e+00 -4.38580930e-01 1.62530065e+00
-2.49033540e-01 -8.45380843e-01 -5.39289773e-01 -9.00332868e-01
5.09039283e-01 1.35907114e+00 -5.86020529e-01 -8.06141943e-02
-7.80807197e-01 -7.68553078e-01 7.84627855e-01 7.40647197e-01
2.72742212e-01 -1.38488877e+00 -2.48922557e-01 4.52751100e-01
-4.28117782e-01 -1.54106617e+00 1.72309369e-01 6.69303000e-01
-7.59169459e-01 -1.09101617e+00 5.86336553e-01 -8.47576678e-01
1.52641073e-01 -7.44926989e-01 2.04311919e+00 -3.74000877e-01
-2.72777587e-01 1.83927380e-02 -2.73165852e-01 -7.28329718e-01
-7.43404686e-01 4.92265522e-01 -1.06329918e-01 -6.11051917e-01
1.40894699e+00 -3.98140967e-01 5.83184930e-03 1.87294558e-01
-7.38949239e-01 1.01712041e-01 5.80666244e-01 7.30855644e-01
6.07336819e-01 2.88141072e-02 7.21818566e-01 -1.79926598e+00
6.47395253e-01 -6.02134407e-01 -4.20901984e-01 2.11510524e-01
-6.96123421e-01 6.24813199e-01 7.81250656e-01 7.52008185e-02
-1.42722595e+00 6.85420781e-02 -5.36343157e-01 5.60482860e-01
-6.63783669e-01 6.35193825e-01 -5.04105270e-01 7.51807332e-01
9.49751198e-01 -1.37894705e-01 -1.71346173e-01 -5.56200147e-01
1.11709762e+00 7.23259509e-01 8.72332573e-01 -1.20376122e+00
1.99473292e-01 1.75722227e-01 -1.73313230e-01 -5.26383340e-01
-1.55651915e+00 -3.70473385e-01 -8.18640292e-01 7.42298543e-01
6.93735242e-01 -9.52488244e-01 -8.70816410e-01 -3.22915167e-01
-8.89847875e-01 -3.88948709e-01 -7.36268342e-01 5.37567697e-02
-8.41647565e-01 -2.20887153e-03 -7.61513412e-01 -3.97705168e-01
-3.77417833e-01 -4.75723356e-01 1.08876300e+00 -2.13706419e-01
-1.01601756e+00 -1.09590709e+00 -7.87753835e-02 5.46843767e-01
1.53464749e-01 -1.19916625e-01 1.53284323e+00 -1.39305389e+00
-3.64808351e-01 -1.92154169e-01 -1.28690854e-01 2.89564550e-01
-2.39172541e-02 -5.26889443e-01 -7.98403800e-01 3.38606328e-01
-3.58316094e-01 -8.36747050e-01 6.89004242e-01 8.67164657e-02
1.16604292e+00 -5.03443122e-01 -4.34133589e-01 7.25062251e-01
1.24459767e+00 -1.61039427e-01 4.81491745e-01 6.83822989e-01
4.78768528e-01 8.66472185e-01 9.09964561e-01 3.50491613e-01
6.77457809e-01 3.48913103e-01 1.13645874e-01 2.60328114e-01
4.71912362e-02 -6.41833603e-01 -8.11381489e-02 -7.03347381e-03
8.91827419e-02 -1.18697800e-01 -7.48539388e-01 5.51777899e-01
-1.56182933e+00 -9.26180005e-01 3.04860603e-02 1.66589999e+00
1.44460809e+00 5.36697686e-01 1.45469010e-02 -1.81630537e-01
1.63852617e-01 -1.90643951e-01 -4.77899075e-01 -6.74689174e-01
-2.59839088e-01 4.11892205e-01 5.21033347e-01 5.49629271e-01
-1.16614711e+00 1.48018813e+00 6.21379852e+00 7.29273140e-01
-3.59564245e-01 -2.52835214e-01 6.06872737e-01 2.90475450e-02
-6.82220161e-01 2.21831784e-01 -1.45779121e+00 2.69691736e-01
1.36684477e+00 -8.93764570e-02 3.99840057e-01 1.00755644e+00
-3.32667500e-01 -6.24601170e-02 -1.79887128e+00 6.47799909e-01
-2.43606314e-01 -1.43771255e+00 2.83488423e-01 -2.59451479e-01
1.33234575e-01 1.10763088e-01 -4.22793448e-01 8.30310822e-01
7.84472644e-01 -1.10039055e+00 4.62944418e-01 1.12496778e-01
7.75827765e-01 -7.89054215e-01 9.59908783e-01 3.63690704e-01
-8.71005118e-01 -3.23916286e-01 -3.52411211e-01 -1.90999389e-01
2.91321486e-01 4.00210798e-01 -1.03471053e+00 5.09968162e-01
5.95000744e-01 8.36364865e-01 -5.56104839e-01 3.81780192e-02
-4.72666562e-01 3.77973706e-01 -3.83194387e-01 -1.20833516e-01
8.90731439e-02 4.32618055e-03 2.73553193e-01 1.51638961e+00
-3.90569270e-01 3.39405686e-01 -5.12238452e-03 1.16210175e+00
-1.34712473e-01 1.50179490e-01 -6.51185930e-01 -2.58717149e-01
6.11431301e-01 1.24826694e+00 -3.73438776e-01 -5.69077849e-01
-6.29741192e-01 7.32295036e-01 4.66862649e-01 1.12158418e-01
-4.35247332e-01 -4.06566769e-01 8.70631576e-01 1.01345897e-01
1.72372371e-01 3.27374309e-01 -5.99742234e-01 -9.79721725e-01
-9.54952911e-02 -1.01477408e+00 1.03962839e+00 -6.59094274e-01
-1.70301688e+00 5.79712451e-01 2.61377305e-01 -3.37897152e-01
-7.94322610e-01 -9.36825454e-01 7.84789696e-02 8.96618009e-01
-1.31614625e+00 -1.34213424e+00 4.58927974e-02 4.25020725e-01
6.46131217e-01 -1.67798445e-01 1.25438499e+00 2.44043097e-01
-2.10433140e-01 6.11571312e-01 -5.97311437e-01 2.83938348e-01
7.40665317e-01 -1.55879271e+00 1.00955117e+00 2.00379252e-01
2.01550215e-01 9.27870333e-01 7.98936307e-01 -7.46079981e-01
-1.37796688e+00 -1.00928628e+00 1.30777144e+00 -1.13156688e+00
9.13346946e-01 -6.87972426e-01 -8.25433075e-01 1.47236252e+00
5.32156229e-02 -1.71807453e-01 1.04025292e+00 9.38609242e-01
-4.68583643e-01 1.15830284e-02 -1.16222382e+00 3.03769588e-01
1.45418060e+00 -4.98966545e-01 -1.18133175e+00 3.17319334e-01
9.96743858e-01 -4.74329561e-01 -1.30699289e+00 3.96562368e-01
5.82816541e-01 -6.01418972e-01 1.14832389e+00 -1.38087964e+00
7.08093762e-01 3.43863815e-01 -4.48874742e-01 -1.36708605e+00
-3.41944426e-01 -3.93598974e-01 -1.19243814e-02 1.46224499e+00
1.09421051e+00 -5.11240542e-01 8.36440742e-01 9.21824813e-01
-3.24018508e-01 -7.68973649e-01 -4.00207937e-01 -6.11796260e-01
1.70369655e-01 -8.19610178e-01 9.89885032e-01 9.22946393e-01
4.69360143e-01 1.06620705e+00 4.35090274e-01 -1.02764711e-01
5.07538855e-01 3.56072813e-01 5.95308959e-01 -1.42773223e+00
-4.87608284e-01 -1.58800840e-01 -2.67721653e-01 -9.08817232e-01
6.26486480e-01 -1.35219419e+00 -1.49452299e-01 -1.48692071e+00
1.26346618e-01 -8.57479930e-01 4.13894840e-02 8.54282081e-01
-9.04906821e-03 -2.30594069e-01 -4.02636826e-01 -1.82664707e-01
-5.91967583e-01 1.47133335e-01 9.04635251e-01 6.72955662e-02
-1.25172269e-02 -9.23992991e-02 -1.30964303e+00 7.98409581e-01
6.06142879e-01 -3.42580646e-01 -5.04577756e-01 -5.19178927e-01
9.42992210e-01 -8.92747864e-02 1.16032854e-01 -3.88499320e-01
-1.58558637e-01 -1.79012582e-01 3.45496982e-01 -4.39959466e-01
1.42651394e-01 -7.52054095e-01 -1.91099390e-01 -3.88162918e-02
-9.84202325e-01 -2.24520769e-02 5.27103543e-01 2.64866263e-01
-1.47305831e-01 -1.49295911e-01 2.95016438e-01 -5.56077361e-01
-1.25486457e+00 -5.62693067e-02 2.18900274e-02 7.35394120e-01
6.18980706e-01 7.37815127e-02 -4.18600023e-01 7.56779686e-02
-8.61970723e-01 3.62107962e-01 1.48263812e-01 6.81682289e-01
1.31969899e-01 -9.78462696e-01 -7.33044267e-01 5.59595525e-01
5.99961400e-01 1.75043419e-01 -9.82521102e-02 3.14326793e-01
-2.48853967e-01 7.88174987e-01 -3.02557468e-01 -2.84841180e-01
-9.83836472e-01 6.81234419e-01 3.96854691e-02 -3.05612296e-01
-6.83359265e-01 1.17843378e+00 3.18637118e-02 -9.02619243e-01
-2.47995257e-02 -6.24036670e-01 -1.40006155e-01 -9.74958166e-02
4.31040823e-01 -1.27417967e-01 2.58286625e-01 -1.95692390e-01
-4.27499592e-01 -9.96166468e-02 -2.96699673e-01 -3.74922692e-03
1.42741859e+00 -1.08719282e-01 -3.49319547e-01 4.15444225e-01
1.00763333e+00 -8.01101550e-02 -7.84706235e-01 -3.82487714e-01
6.77374542e-01 -1.93212911e-01 -2.95402020e-01 -1.34935057e+00
-3.72137070e-01 3.39456260e-01 -1.61843166e-01 2.32712820e-01
7.91566849e-01 6.93329036e-01 9.33558881e-01 8.50740731e-01
4.39871848e-01 -1.18349385e+00 -4.54009801e-01 7.17285156e-01
5.72169960e-01 -1.21997201e+00 -8.80417749e-02 -7.74427533e-01
-7.75559723e-01 1.01455367e+00 6.70593679e-01 7.85981938e-02
3.60419482e-01 6.92302525e-01 3.83676104e-02 -6.35505080e-01
-1.21333408e+00 -3.32342297e-01 4.14328538e-02 7.22592354e-01
9.76512551e-01 2.72670776e-01 -4.60089818e-02 1.14437735e+00
-8.86403203e-01 -9.23284665e-02 7.73421451e-02 6.88714623e-01
-2.56398588e-01 -1.39558971e+00 2.56698698e-01 6.67152822e-01
-5.90859950e-01 -6.30207062e-01 -4.66989070e-01 1.10348034e+00
2.73795426e-01 6.52638555e-01 2.34282777e-01 -1.14085590e-02
7.78683364e-01 5.20752072e-01 6.45376503e-01 -8.78899455e-01
-3.72777373e-01 -5.72299540e-01 1.18385041e+00 -8.51838768e-01
-5.41871376e-02 -7.33280480e-01 -1.54474723e+00 -2.10468605e-01
-5.63359335e-02 2.73805946e-01 5.77771842e-01 1.22796571e+00
4.85383421e-01 3.64522249e-01 -9.11684260e-02 -4.07437086e-02
-7.16935158e-01 -1.06628048e+00 -6.07251167e-01 8.26855004e-01
-2.08308458e-01 -3.78897995e-01 3.45441438e-02 2.06962749e-01] | [10.058904647827148, 8.275453567504883] |
943dd672-f0f9-48f9-a478-557d6f7d4bcb | a-gated-cross-domain-collaborative-network | 2306.14141 | null | https://arxiv.org/abs/2306.14141v1 | https://arxiv.org/pdf/2306.14141v1.pdf | A Gated Cross-domain Collaborative Network for Underwater Object Detection | Underwater object detection (UOD) plays a significant role in aquaculture and marine environmental protection. Considering the challenges posed by low contrast and low-light conditions in underwater environments, several underwater image enhancement (UIE) methods have been proposed to improve the quality of underwater images. However, only using the enhanced images does not improve the performance of UOD, since it may unavoidably remove or alter critical patterns and details of underwater objects. In contrast, we believe that exploring the complementary information from the two domains is beneficial for UOD. The raw image preserves the natural characteristics of the scene and texture information of the objects, while the enhanced image improves the visibility of underwater objects. Based on this perspective, we propose a Gated Cross-domain Collaborative Network (GCC-Net) to address the challenges of poor visibility and low contrast in underwater environments, which comprises three dedicated components. Firstly, a real-time UIE method is employed to generate enhanced images, which can improve the visibility of objects in low-contrast areas. Secondly, a cross-domain feature interaction module is introduced to facilitate the interaction and mine complementary information between raw and enhanced image features. Thirdly, to prevent the contamination of unreliable generated results, a gated feature fusion module is proposed to adaptively control the fusion ratio of cross-domain information. Our method presents a new UOD paradigm from the perspective of cross-domain information interaction and fusion. Experimental results demonstrate that the proposed GCC-Net achieves state-of-the-art performance on four underwater datasets. | ['Mengyuan Liu', 'Pinhao Song', 'Hong Liu', 'Linhui Dai'] | 2023-06-25 | null | null | null | null | ['image-enhancement', 'uie'] | ['computer-vision', 'computer-vision'] | [ 3.54066312e-01 -1.29054010e-01 9.48201895e-01 -1.32929981e-01
-2.25899875e-01 -2.51685917e-01 1.83919668e-01 1.77179247e-01
-7.86292911e-01 4.74684268e-01 -6.22858014e-03 3.74324381e-01
-3.08040917e-01 -1.27321267e+00 -5.36828041e-01 -1.27125156e+00
-3.80646557e-01 -6.37391746e-01 6.23325884e-01 -4.94736284e-01
2.65694916e-01 3.99893343e-01 -2.02191377e+00 8.96156300e-03
1.25314426e+00 1.03397393e+00 8.88888419e-01 4.92841154e-01
-1.32648334e-01 3.24008286e-01 -4.81273443e-01 -5.80096617e-02
4.15213257e-01 -5.10124207e-01 1.28239421e-02 1.88343748e-01
1.90547094e-01 -6.89700842e-01 -2.18892008e-01 1.51788294e+00
8.68929088e-01 3.53518397e-01 3.37314814e-01 -6.62312567e-01
-2.68876880e-01 3.06451380e-01 -5.22601128e-01 2.17024654e-01
8.66068229e-02 1.66086748e-01 5.85416675e-01 -1.02477205e+00
3.32784623e-01 1.10247850e+00 4.81728792e-01 2.53569186e-01
-7.44077265e-01 -6.27298474e-01 5.10357767e-02 1.05201885e-01
-1.02098143e+00 -3.32686365e-01 6.97060943e-01 -2.24893808e-01
2.68869549e-01 1.03586026e-01 8.31042290e-01 9.86361355e-02
4.82671857e-01 4.47569758e-01 1.22728348e+00 -4.01773095e-01
7.85978362e-02 -4.54203710e-02 -7.99915716e-02 6.21285975e-01
4.94540513e-01 3.29176545e-01 -5.91307163e-01 1.88862815e-01
5.52191794e-01 1.32168308e-01 -9.05155420e-01 -1.09909112e-02
-5.96423030e-01 4.25197184e-01 4.61425871e-01 9.60435048e-02
-2.73498505e-01 -4.16784912e-01 2.20005378e-01 3.13891798e-01
3.05453360e-01 2.42577359e-01 -2.68675834e-01 2.55555302e-01
-4.35803860e-01 -7.96097219e-02 5.28750122e-01 6.13911629e-01
1.17596543e+00 1.70448899e-01 1.95431754e-01 8.34964097e-01
6.07743382e-01 8.22619200e-01 3.03009450e-01 -7.64615119e-01
2.90135980e-01 5.74879110e-01 8.60053375e-02 -1.02155876e+00
-5.70267498e-01 -2.23620683e-01 -6.89190626e-01 8.02182853e-01
9.61644351e-02 -3.40040743e-01 -8.16940069e-01 1.28208411e+00
5.01389384e-01 9.30051580e-02 6.40894651e-01 1.18299341e+00
1.31044924e+00 9.58629370e-01 -2.76971191e-01 -4.69289809e-01
1.39363408e+00 -4.17105764e-01 -9.84834850e-01 -8.88878480e-02
2.25072727e-01 -8.31873000e-01 5.08078575e-01 1.88068613e-01
-1.01186061e+00 -6.03910267e-01 -1.47966266e+00 1.89352840e-01
-2.66258359e-01 1.23871649e-02 1.50465831e-01 4.59390938e-01
-7.97793388e-01 4.06012118e-01 -6.56726718e-01 2.32801698e-02
1.06942192e-01 2.88227886e-01 -4.73685801e-01 -3.58143270e-01
-1.20743740e+00 6.45143092e-01 3.81952882e-01 6.45999432e-01
-9.84778583e-01 -5.68206370e-01 -1.21646917e+00 3.60734388e-02
2.43611753e-01 4.56000492e-02 6.68029308e-01 -5.47165394e-01
-1.37178242e+00 1.54247612e-01 3.01751345e-01 -3.13384235e-02
4.37941849e-01 -2.95659572e-01 -2.35157952e-01 5.46356320e-01
-1.39217496e-01 1.89498246e-01 6.84727192e-01 -1.65688407e+00
-1.12903512e+00 -4.08435971e-01 1.10862367e-01 5.52768528e-01
-5.60765505e-01 -1.36025816e-01 -5.08698523e-01 -4.52769756e-01
4.99041378e-01 -4.79638070e-01 -1.75835520e-01 4.84538049e-01
1.95913613e-01 3.09528261e-01 1.17110896e+00 -6.62073314e-01
7.45372593e-01 -2.50820398e+00 -6.57356083e-02 5.39438240e-02
-5.79956658e-02 5.32157183e-01 -2.77710021e-01 4.36900616e-01
4.29049641e-01 -4.49482977e-01 -4.96095270e-01 -2.18820646e-01
-4.14954811e-01 4.79673088e-01 2.77638853e-01 6.47499025e-01
1.65945977e-01 1.18579879e-01 -1.01986444e+00 -7.04372942e-01
3.83248657e-01 6.37846589e-01 -4.18466300e-01 6.22534573e-01
2.49239907e-01 4.80086237e-01 -2.78776050e-01 6.75743401e-01
1.33454752e+00 5.75115621e-01 1.76595792e-01 -4.93652135e-01
-7.89143801e-01 -4.89121079e-01 -1.56090760e+00 1.13694704e+00
-4.25792694e-01 3.54432046e-01 8.34077835e-01 -5.82689524e-01
1.29594040e+00 2.07872182e-01 3.61433446e-01 -8.39296758e-01
1.93580166e-01 3.45203906e-01 -4.30272780e-02 -9.35721695e-01
5.46129704e-01 -2.68975884e-01 3.99562001e-01 -2.76908930e-02
-9.36314017e-02 -1.77278300e-03 2.03307584e-01 -3.79641429e-02
6.34709597e-01 1.97947904e-01 1.71566144e-01 -4.92144406e-01
5.78739882e-01 -4.11094040e-01 9.91903067e-01 6.14901245e-01
-1.81542356e-02 5.71385384e-01 -4.35465761e-02 -2.38496333e-01
-5.93570769e-01 -8.71993780e-01 -4.39472347e-01 7.91238427e-01
1.13529050e+00 2.66920358e-01 -4.42584842e-01 -1.67178094e-01
-9.11271796e-02 3.63007709e-02 -6.30275309e-01 -1.52685732e-01
-4.01278764e-01 -1.04665720e+00 2.21221626e-01 2.25164503e-01
1.03397989e+00 -1.06949365e+00 -8.58943105e-01 4.65747386e-01
-4.94583137e-02 -1.07856011e+00 -1.69035003e-01 2.21791983e-01
-9.22274470e-01 -1.10067356e+00 -7.15104640e-01 -9.66450512e-01
8.79938841e-01 8.59650135e-01 4.85317230e-01 4.13400292e-01
-9.35325176e-02 -7.50507414e-02 -9.00766730e-01 -4.94423658e-01
-3.74571443e-01 -8.67508590e-01 -9.12531018e-02 1.97738275e-01
-7.79290423e-02 -4.34850663e-01 -9.04957771e-01 4.34538275e-01
-1.34348083e+00 -1.78755984e-01 7.58379757e-01 1.00905883e+00
2.29851469e-01 3.69025111e-01 2.78492600e-01 -4.54634398e-01
6.18409403e-02 -3.90306532e-01 -8.80152225e-01 9.65672638e-03
-3.10120195e-01 -3.26074690e-01 3.97913963e-01 -2.09923625e-01
-1.42922091e+00 -6.72220066e-02 -1.65549919e-01 -5.38468063e-02
1.23316646e-01 7.12947786e-01 -4.20893341e-01 -4.39221472e-01
2.03179792e-01 6.07423604e-01 4.04294193e-01 -6.54487252e-01
-3.96816164e-01 7.88873672e-01 5.28651655e-01 -1.25151634e-01
8.91486764e-01 8.42558026e-01 1.18218176e-01 -1.16056132e+00
-4.79085445e-01 -5.07175744e-01 -3.58541042e-01 -6.11407161e-01
8.93716455e-01 -1.26353526e+00 -5.81896305e-01 1.04436958e+00
-9.59724724e-01 -7.91706741e-02 1.42078072e-01 7.62764096e-01
1.83863491e-01 8.52949262e-01 -6.90589547e-01 -9.99447286e-01
-5.38006783e-01 -1.36565351e+00 7.58838236e-01 9.41265821e-01
8.52792084e-01 -6.60097957e-01 -2.90325403e-01 6.94061592e-02
3.56013417e-01 2.53691882e-01 3.69733572e-01 -1.71335429e-01
-6.17399633e-01 6.65814728e-02 -4.28382546e-01 6.54932141e-01
5.08044541e-01 -1.91761719e-04 -8.62461030e-01 -4.69156444e-01
-6.91657811e-02 -4.19808365e-02 7.83281088e-01 2.01398641e-01
4.91396666e-01 1.23227589e-01 -1.10231405e-02 7.39955783e-01
1.69015658e+00 4.70997870e-01 8.55082214e-01 4.79607821e-01
3.22605461e-01 1.02242458e+00 1.14648938e+00 7.95140624e-01
2.92889744e-01 3.81568581e-01 9.68370080e-01 -5.29588342e-01
-6.01037592e-02 2.13993847e-01 4.48208958e-01 5.78962803e-01
-2.95366257e-01 -1.30666092e-01 -1.97008088e-01 7.00704157e-01
-1.47078955e+00 -6.69280529e-01 -3.46706331e-01 2.00624204e+00
6.12474501e-01 -1.31702960e-01 -4.27644134e-01 1.04898877e-01
8.62395108e-01 -3.29847410e-02 -2.73193300e-01 1.58999227e-02
-4.22624648e-01 -7.00000823e-02 5.03719568e-01 3.51391286e-01
-9.35130298e-01 3.79152566e-01 4.73266649e+00 4.70225483e-01
-1.01133370e+00 -1.29868388e-01 -1.90779660e-02 4.42586511e-01
-2.35702470e-01 -1.97345778e-01 -7.42586017e-01 6.39971852e-01
-4.89620790e-02 2.08447501e-01 -1.89215854e-01 6.12840831e-01
5.14387071e-01 -4.46823359e-01 -3.33058625e-01 5.81220746e-01
1.12563938e-01 -9.12793934e-01 -5.78963384e-02 1.93459332e-01
8.06274414e-01 -1.58684403e-01 -3.51052433e-01 -1.63901791e-01
5.55904955e-02 -2.57188052e-01 7.71849990e-01 4.94956195e-01
5.32026231e-01 -7.40605652e-01 1.47154629e+00 1.92161262e-01
-1.32035494e+00 -2.63998717e-01 -5.59056699e-01 -1.60636827e-01
2.35257924e-01 6.98353171e-01 -2.86688745e-01 8.31760883e-01
1.24149954e+00 4.43010539e-01 -1.83109075e-01 1.37163484e+00
-2.34630421e-01 1.64449379e-01 -3.82541984e-01 8.52997303e-02
1.97278604e-01 -6.76062942e-01 8.20677161e-01 1.16970432e+00
5.97620070e-01 6.66052163e-01 8.95145908e-02 3.20845664e-01
1.76560760e-01 1.23249568e-01 -5.57628036e-01 4.20760363e-01
4.05489922e-01 1.28603590e+00 -4.77150053e-01 -1.98438406e-01
-6.19956553e-01 5.92861235e-01 -3.28863263e-01 1.72770932e-01
-4.29542273e-01 -7.91939139e-01 7.13368475e-01 -1.95937902e-01
5.33901155e-01 -2.56245971e-01 4.79961298e-02 -8.66272271e-01
1.03775626e-02 -5.90533078e-01 3.52247685e-01 -6.69687450e-01
-1.00121534e+00 4.93433297e-01 -1.74768046e-01 -1.85650539e+00
6.38277709e-01 -4.89928156e-01 -7.26507485e-01 8.09353769e-01
-2.26960707e+00 -8.79627883e-01 -9.19740558e-01 2.68224418e-01
5.72582662e-01 1.66723698e-01 4.20843035e-01 4.76906657e-01
-3.46945167e-01 1.31041601e-01 4.37279820e-01 1.48717597e-01
4.60928082e-01 -1.02804434e+00 -5.15734255e-01 1.38434982e+00
-5.88905692e-01 3.10996264e-01 8.92104149e-01 -6.76202714e-01
-1.57332838e+00 -9.33135688e-01 1.41159981e-01 6.06957734e-01
3.11400115e-01 1.25296116e-01 -1.20065999e+00 -3.02465335e-02
9.26564187e-02 1.50092125e-01 6.05917573e-01 -7.77103841e-01
3.91011715e-01 -4.31410700e-01 -1.20192540e+00 3.88977736e-01
5.75908720e-01 1.57187387e-01 -4.52787101e-01 -2.07275257e-01
5.23920834e-01 -4.15065438e-01 -1.07907009e+00 9.84064996e-01
6.70153618e-01 -1.14751065e+00 7.54419804e-01 3.09861273e-01
4.79126632e-01 -1.01774001e+00 -3.17442149e-01 -1.33512509e+00
-1.67032145e-02 -2.86644459e-01 4.18115437e-01 1.47302854e+00
1.81813717e-01 -7.07477212e-01 3.58724475e-01 5.13458140e-02
-7.35621512e-01 -2.99250692e-01 -9.85984683e-01 -5.32084823e-01
-4.24673945e-01 3.50389145e-02 4.65002984e-01 5.80855191e-01
-1.59724176e-01 -1.83871001e-01 -5.32188296e-01 1.06570375e+00
1.01324236e+00 3.96262884e-01 8.23530734e-01 -1.24575186e+00
-5.52800931e-02 9.76712182e-02 -5.40412068e-01 -1.04290378e+00
-4.89818513e-01 -1.13580637e-01 7.19118118e-01 -1.79231834e+00
1.39175370e-01 -2.41487488e-01 -2.68118322e-01 2.26167381e-01
-4.79901314e-01 8.10189128e-01 2.00881660e-01 1.89848453e-01
-2.84675777e-01 9.86548722e-01 1.61104178e+00 -2.67011430e-02
-2.70466894e-01 -8.33675042e-02 -4.17085499e-01 5.49028933e-01
3.85545731e-01 -2.28547856e-01 -3.77125777e-02 -6.03806317e-01
-6.70754388e-02 1.02857687e-01 8.00308660e-02 -1.10977304e+00
4.02450979e-01 -1.59866665e-03 3.31011087e-01 -4.72119123e-01
4.20897335e-01 -1.11348939e+00 -1.59608617e-01 8.07663739e-01
4.64704543e-01 -4.39431131e-01 1.48493394e-01 6.53160572e-01
-6.07654333e-01 -4.54617590e-01 1.32690299e+00 -2.45012343e-01
-1.28263235e+00 -3.06742452e-02 -5.49021184e-01 -5.67346871e-01
1.07869470e+00 -5.74494541e-01 -5.36159039e-01 -1.75841227e-01
-3.79146576e-01 5.82702637e-01 4.80297029e-01 1.83837712e-01
1.13434708e+00 -5.44050872e-01 -9.49752986e-01 3.92780870e-01
3.77931923e-01 2.71021128e-01 7.68293679e-01 8.10550928e-01
-1.03491151e+00 -6.52056396e-01 -4.08687532e-01 -5.57575166e-01
-1.46008778e+00 4.81437072e-02 4.15062785e-01 2.09237531e-01
-7.65340745e-01 7.32683837e-01 4.64243680e-01 -2.60597438e-01
-1.27970710e-01 -1.01136137e-03 -8.10631335e-01 3.37187141e-01
1.01233399e+00 4.51132804e-01 -3.01777571e-03 -6.73564076e-01
-6.66603446e-02 8.87692690e-01 4.63378280e-02 2.31590152e-01
1.67766929e+00 -6.97671592e-01 -3.38365257e-01 -9.09157563e-03
8.12631428e-01 2.00015321e-01 -1.79850101e+00 -2.37325758e-01
-6.28998756e-01 -7.08992600e-01 3.42725903e-01 -4.46744412e-01
-1.32517958e+00 7.87807524e-01 7.74523318e-01 2.04055011e-01
1.61324382e+00 -3.46609473e-01 7.93232918e-01 1.35769174e-01
3.69177490e-01 -9.94332790e-01 1.06320009e-01 3.45799774e-01
5.39056599e-01 -1.33174038e+00 3.23864877e-01 -7.00088441e-01
-4.07818317e-01 1.32345676e+00 7.95239806e-01 9.37838182e-02
5.84972858e-01 6.36263788e-01 6.53387070e-01 -2.31966585e-01
-1.86300755e-01 -5.80301583e-01 -2.10261703e-01 6.50558233e-01
-1.32396862e-01 -4.11848664e-01 -5.50763845e-01 3.46026033e-01
4.04968172e-01 -3.58537465e-01 9.17259872e-01 1.17652321e+00
-1.02851915e+00 -6.58266664e-01 -7.13421106e-01 1.24468938e-01
-5.26718497e-01 -7.27871209e-02 3.92472833e-01 6.58058226e-01
4.25565362e-01 1.18253839e+00 -4.06718440e-02 -3.66381913e-01
4.72952873e-01 -7.23083079e-01 -3.60551886e-02 -4.23505664e-01
-3.18397701e-01 3.89993906e-01 -3.12834606e-02 -9.19037238e-02
-8.62244964e-01 -4.50874150e-01 -1.43199480e+00 2.34190330e-01
-7.00578988e-01 6.72358930e-01 8.12642753e-01 7.41830230e-01
-9.13519636e-02 5.43157816e-01 1.05155611e+00 -1.18693805e+00
-2.56543517e-01 -1.11094320e+00 -1.11361086e+00 2.44765833e-01
4.53726739e-01 -8.15150321e-01 -8.53885770e-01 1.27494365e-01] | [10.662798881530762, -3.482184648513794] |
88660fd8-4112-4e24-a6ab-0be91ea65049 | an-evolutionary-approach-to-dynamic | 2205.12755 | null | https://arxiv.org/abs/2205.12755v6 | https://arxiv.org/pdf/2205.12755v6.pdf | An Evolutionary Approach to Dynamic Introduction of Tasks in Large-scale Multitask Learning Systems | Multitask learning assumes that models capable of learning from multiple tasks can achieve better quality and efficiency via knowledge transfer, a key feature of human learning. Though, state of the art ML models rely on high customization for each task and leverage size and data scale rather than scaling the number of tasks. Also, continual learning, that adds the temporal aspect to multitask, is often focused to the study of common pitfalls such as catastrophic forgetting instead of being studied at a large scale as a critical component to build the next generation artificial intelligence.We propose an evolutionary method capable of generating large scale multitask models that support the dynamic addition of new tasks. The generated multitask models are sparsely activated and integrates a task-based routing that guarantees bounded compute cost and fewer added parameters per task as the model expands.The proposed method relies on a knowledge compartmentalization technique to achieve immunity against catastrophic forgetting and other common pitfalls such as gradient interference and negative transfer. We demonstrate empirically that the proposed method can jointly solve and achieve competitive results on 69public image classification tasks, for example improving the state of the art on a competitive benchmark such as cifar10 by achieving a 15% relative error reduction compared to the best model trained on public data. | ['Jeff Dean', 'Andrea Gesmundo'] | 2022-05-25 | null | null | null | null | ['fine-grained-image-classification'] | ['computer-vision'] | [ 1.81520641e-01 -1.47853941e-01 8.27271864e-02 -9.32558402e-02
-8.05373847e-01 -4.62171584e-01 4.40952301e-01 1.30818263e-02
-6.60537481e-01 1.08031094e+00 -2.72938371e-01 -7.93539211e-02
-3.41202497e-01 -5.04914224e-01 -1.08462083e+00 -7.66605675e-01
-5.22468723e-02 7.69346714e-01 5.40667892e-01 -1.04338586e-01
2.59374887e-01 1.74330458e-01 -1.58423042e+00 4.06780660e-01
1.42887139e+00 8.82872820e-01 5.80459237e-01 4.06714141e-01
-5.30257002e-02 6.17143691e-01 -7.59571552e-01 -4.89514679e-01
2.50962019e-01 1.06132038e-01 -7.03686655e-01 -1.49569243e-01
4.10352230e-01 1.32685289e-01 1.75990760e-01 7.91924715e-01
6.83200419e-01 6.42783269e-02 5.92976213e-01 -1.34554887e+00
-7.49256372e-01 5.08790910e-01 -6.41757429e-01 2.70371884e-01
-3.51923376e-01 2.60588020e-01 5.78529835e-01 -1.07327247e+00
3.58624160e-01 1.06215024e+00 7.87215471e-01 4.89881903e-01
-1.31810415e+00 -1.06996763e+00 3.18760663e-01 2.78453827e-01
-1.32446504e+00 -4.34263796e-01 5.82904518e-01 -3.93130332e-01
1.22935677e+00 -4.54372503e-02 5.56706667e-01 1.21054518e+00
6.85632050e-01 7.72068679e-01 1.24466383e+00 -3.37667644e-01
4.36087668e-01 6.44630671e-01 -3.46485637e-02 7.80557036e-01
4.40650672e-01 -1.91118568e-01 -1.00136244e+00 -1.49692565e-01
3.11121792e-01 1.11593552e-01 -2.01209430e-02 -4.04589832e-01
-1.06113029e+00 7.13665247e-01 2.43998900e-01 3.28928769e-01
-5.07140458e-01 1.92998365e-01 3.04440081e-01 6.66468680e-01
6.83193088e-01 6.12087965e-01 -8.26401114e-01 8.41598585e-02
-9.84255731e-01 7.63154328e-02 5.43585062e-01 8.20611775e-01
1.02112079e+00 1.11187249e-01 -3.42618674e-01 7.40554869e-01
-1.18263774e-01 5.77121556e-01 9.77058470e-01 -6.18950844e-01
5.01202583e-01 8.42880189e-01 -8.96247253e-02 -5.80683708e-01
-4.04076368e-01 -1.07906234e+00 -8.23762298e-01 1.85980931e-01
1.22693367e-02 -3.17740321e-01 -9.97250617e-01 2.01636171e+00
2.49419108e-01 3.46649021e-01 1.92288514e-02 3.63243520e-01
2.18183681e-01 4.09849435e-01 1.94931418e-01 -2.95846403e-01
1.16332591e+00 -1.09858775e+00 -4.49199200e-01 -6.53449237e-01
6.97465241e-01 -5.61196029e-01 1.27174020e+00 6.55890465e-01
-1.01062572e+00 -5.60927927e-01 -1.16455269e+00 2.46398330e-01
-4.85636204e-01 -1.00028962e-01 7.65365958e-01 8.46621811e-01
-1.11152530e+00 2.88786203e-01 -6.30843163e-01 -1.38702452e-01
5.70514977e-01 4.53736871e-01 -1.34491280e-01 -1.25755072e-01
-1.22701848e+00 1.24079847e+00 6.60789430e-01 -1.92329854e-01
-1.20891476e+00 -1.18768477e+00 -2.90591687e-01 2.57588267e-01
6.01668358e-01 -1.00806987e+00 8.57030988e-01 -1.17626703e+00
-1.29569185e+00 5.34046113e-01 -8.75899717e-02 -6.71755135e-01
7.22447693e-01 -5.12325406e-01 -2.65759915e-01 -2.54999518e-01
-3.57978307e-02 6.97757542e-01 1.40855122e+00 -1.25379252e+00
-7.88377821e-01 -3.54563385e-01 -2.69584954e-01 4.09250826e-01
-1.06326950e+00 -3.85728806e-01 -1.83087051e-01 -5.86362064e-01
-1.91808775e-01 -1.08189654e+00 7.60857090e-02 -3.59240264e-01
9.78100747e-02 -2.74531752e-01 9.77984488e-01 -3.95170897e-01
1.16786313e+00 -1.97636664e+00 3.97906929e-01 -1.29696861e-01
3.28401208e-01 3.83980691e-01 -3.59601736e-01 2.00324401e-01
4.21758667e-02 2.40408435e-01 -7.62047619e-02 -6.13369167e-01
-2.77631491e-01 7.36694708e-02 -2.34723017e-01 1.11079253e-01
2.55962908e-01 1.06514978e+00 -6.60351694e-01 -3.21815848e-01
-2.96957105e-01 2.89584309e-01 -6.11260414e-01 -1.16062604e-01
-2.18607336e-01 1.84313014e-01 -2.48543277e-01 4.59602177e-01
5.35300791e-01 -6.18132114e-01 8.31373781e-02 1.43986093e-02
1.51214242e-01 -1.34405926e-01 -1.10997987e+00 2.01013017e+00
-8.22910130e-01 1.73912898e-01 -9.89819169e-02 -9.91521537e-01
7.06828177e-01 3.35001856e-01 1.91136762e-01 -6.13305807e-01
-2.41696924e-01 3.72459531e-01 1.59322500e-01 -3.03188190e-02
3.38968664e-01 -1.51318356e-01 -4.32110578e-02 6.79205120e-01
3.36307824e-01 -1.37675971e-01 3.87857482e-02 3.00993502e-01
1.16324985e+00 -1.26793712e-01 7.61630312e-02 -5.18834710e-01
3.26747507e-01 2.54860185e-02 6.03185177e-01 1.02554166e+00
-8.79541337e-02 -2.44633742e-02 2.70023989e-03 -5.75843334e-01
-9.72943008e-01 -9.95207965e-01 6.38657361e-02 1.75141013e+00
-2.16540024e-01 8.53339285e-02 -4.43564206e-01 -8.05783629e-01
2.79979080e-01 7.78888166e-01 -8.53402853e-01 -6.80251539e-01
-4.15495306e-01 -1.20347571e+00 3.92821282e-01 1.76977977e-01
6.04991674e-01 -1.04409015e+00 -8.60725284e-01 3.06016892e-01
-1.72833711e-01 -7.26884604e-01 -4.66131479e-01 7.41295934e-01
-1.01603055e+00 -7.89415121e-01 -7.67359972e-01 -7.77535081e-01
4.75151479e-01 3.24569523e-01 1.15873551e+00 9.18463990e-02
-4.09488976e-01 2.19361052e-01 -1.29983529e-01 -7.39763975e-01
-2.02143505e-01 6.86130583e-01 3.72419447e-01 7.87312686e-02
-6.89325556e-02 -8.11688721e-01 -4.37712818e-01 2.69172966e-01
-9.01525080e-01 2.30343670e-01 1.00931871e+00 1.13786602e+00
4.05530840e-01 2.38307238e-01 1.23671103e+00 -9.76354957e-01
8.10542285e-01 -4.91962045e-01 -3.68826300e-01 7.96813846e-01
-1.08649528e+00 2.99057901e-01 6.01478457e-01 -8.78462791e-01
-1.42141211e+00 -2.96635949e-03 5.73371768e-01 -5.81699848e-01
2.58475691e-01 3.42478335e-01 1.91320106e-01 -2.96419591e-01
8.54526818e-01 6.75707638e-01 -1.92071557e-01 -2.87998080e-01
3.97489041e-01 3.00964206e-01 7.54275844e-02 -7.16782212e-01
7.76613414e-01 3.57278734e-01 -1.11319564e-01 -4.81708199e-01
-1.02843225e+00 -1.07994929e-01 -5.82850277e-01 -1.91259772e-01
4.41724330e-01 -1.11341333e+00 -6.95595026e-01 6.64523304e-01
-9.90746915e-01 -4.99595731e-01 -1.83611572e-01 1.92345366e-01
-3.52114797e-01 -1.09847829e-01 -3.28629792e-01 -6.53775871e-01
-6.63658500e-01 -1.04981244e+00 6.67131603e-01 2.02543542e-01
1.74635872e-01 -1.06149781e+00 1.43396715e-02 4.78459746e-01
8.96417677e-01 -3.59542400e-01 1.28612554e+00 -7.31854737e-01
-6.02283239e-01 1.55360758e-01 -9.09551084e-02 3.38701785e-01
4.44066748e-02 -6.16253436e-01 -1.04179311e+00 -7.91383922e-01
1.76433846e-01 -9.04428005e-01 1.14470553e+00 1.88913465e-01
8.97609293e-01 -2.60321259e-01 -5.20313323e-01 5.20460665e-01
1.50547349e+00 2.34774545e-01 3.67241949e-01 4.37127024e-01
4.30605561e-01 4.75332975e-01 4.89657640e-01 4.19690728e-01
3.19927871e-01 3.77654046e-01 3.62186253e-01 1.26167893e-01
-2.30653256e-01 -2.55085975e-02 2.99162418e-01 1.03663456e+00
1.56353086e-01 -1.01434693e-01 -1.15086412e+00 5.73470950e-01
-1.94413126e+00 -6.17405951e-01 3.92938107e-01 2.27640605e+00
1.12417865e+00 4.32534158e-01 -2.46485755e-01 -8.61410946e-02
5.94652653e-01 -2.34345272e-01 -1.10109711e+00 -9.81490240e-02
-2.23851889e-01 1.71992347e-01 5.23239374e-01 2.69897521e-01
-8.98884833e-01 1.20823097e+00 6.15165567e+00 1.00766122e+00
-1.21230114e+00 6.50363326e-01 7.95614123e-01 -5.24664760e-01
-9.83787403e-02 -3.41924399e-01 -1.01481891e+00 3.66064698e-01
8.63436282e-01 -5.44072986e-01 5.45690477e-01 8.75313878e-01
-2.90420771e-01 -2.65095353e-01 -9.59040523e-01 8.19053948e-01
1.89756602e-01 -1.26326275e+00 3.10706764e-01 -1.14687853e-01
1.08390522e+00 1.85448185e-01 5.38800061e-01 7.10781515e-01
5.33250868e-01 -9.64109838e-01 8.24794531e-01 5.13835907e-01
6.36593044e-01 -7.97109663e-01 5.25848448e-01 8.39228332e-01
-9.39995885e-01 -5.61768770e-01 -3.58328134e-01 5.69273420e-02
-9.99225453e-02 6.58014596e-01 -9.68143046e-01 4.54573095e-01
8.79619896e-01 1.21845394e-01 -9.69084084e-01 9.56108928e-01
7.71176592e-02 5.59461117e-01 -2.51437902e-01 9.01595801e-02
2.61434745e-02 2.72819251e-01 3.41743171e-01 9.10350740e-01
4.50583994e-01 -4.61167365e-01 2.50331789e-01 5.01609325e-01
-2.24089235e-01 1.55952141e-01 -5.82169712e-01 1.61449671e-01
7.04988420e-01 1.22891903e+00 -5.81133604e-01 -3.57732743e-01
-1.48043856e-01 1.06256807e+00 8.36859405e-01 2.56638467e-01
-9.23614383e-01 -1.74660608e-01 2.21528545e-01 -2.37485990e-02
2.96143413e-01 -1.78871855e-01 -6.26295686e-01 -9.99293029e-01
1.23441564e-02 -1.00672019e+00 3.82298917e-01 -4.52777982e-01
-1.34607959e+00 5.45846343e-01 -2.78549314e-01 -6.75661504e-01
3.79972607e-02 -1.60417646e-01 -4.58329350e-01 8.64693820e-01
-1.76179802e+00 -1.28210735e+00 -2.62425303e-01 8.50796819e-01
7.34966934e-01 -5.68224728e-01 7.40702450e-01 3.36361915e-01
-5.73272347e-01 8.15435112e-01 1.90294027e-01 -6.92720234e-01
1.19638419e+00 -1.08339036e+00 5.13902679e-02 6.78502977e-01
-3.50346453e-02 5.87454438e-01 3.98087144e-01 -8.38954508e-01
-1.52758849e+00 -1.23377466e+00 7.82296956e-01 -5.56119084e-01
6.26383662e-01 -6.77618980e-01 -1.00258303e+00 6.68543041e-01
8.79814699e-02 -2.13634193e-01 5.25024354e-01 4.12271231e-01
-3.95917952e-01 -4.54931289e-01 -1.03967309e+00 3.56853783e-01
1.14242053e+00 -2.11911455e-01 -6.04169190e-01 4.91345108e-01
8.28670382e-01 8.43794867e-02 -5.86386323e-01 3.77765328e-01
3.91662359e-01 -7.97635138e-01 8.63086760e-01 -6.15706801e-01
1.58826724e-01 1.01777809e-02 1.67830333e-01 -1.65447080e+00
-5.47577739e-01 -6.04103804e-01 -1.66090280e-01 1.03496850e+00
7.46096909e-01 -8.61707628e-01 7.93273449e-01 3.21086884e-01
-1.65736362e-01 -6.11589730e-01 -1.10424078e+00 -1.12739050e+00
2.41254568e-01 -1.19644858e-01 2.70436317e-01 9.60379124e-01
-4.71695691e-01 6.58835471e-01 -7.03659832e-01 7.74456933e-02
7.59016454e-01 -1.32459089e-01 5.93281388e-01 -1.55787897e+00
-3.55424762e-01 -2.91982323e-01 3.12207103e-01 -5.36822617e-01
1.39004529e-01 -9.95660126e-01 -5.17984256e-02 -1.16024137e+00
4.81634349e-01 -7.99594581e-01 -7.01442778e-01 7.53994703e-01
-3.52236122e-01 1.34634197e-01 2.59361923e-01 4.32578593e-01
-9.28087711e-01 7.01625049e-01 1.23730981e+00 -1.25869662e-01
-3.88924330e-01 -2.37493664e-01 -7.94222891e-01 6.17452204e-01
7.57669210e-01 -8.76731277e-01 -6.89533770e-01 -7.10987449e-01
4.93029058e-01 -1.41661137e-01 3.97588052e-02 -1.44493651e+00
6.54870868e-01 -6.88536838e-02 5.02172291e-01 -2.35264644e-01
4.07403380e-01 -8.39390814e-01 2.82748133e-01 9.42151427e-01
-2.84449995e-01 2.50762433e-01 3.85051012e-01 8.97188246e-01
1.40111566e-01 -3.91651206e-02 8.27809036e-01 -3.51935059e-01
-5.35001874e-01 2.23687723e-01 -1.69147000e-01 1.20723769e-01
1.48336887e+00 6.71772361e-02 -5.86007357e-01 1.98360998e-02
-6.20489955e-01 4.55326140e-01 1.59638971e-01 6.78932250e-01
5.37363410e-01 -1.01874804e+00 -8.23775649e-01 1.93827882e-01
-2.36162581e-02 -3.34228933e-01 5.32154143e-01 8.68416429e-01
7.56853372e-02 4.85796064e-01 -3.51473004e-01 -5.10533512e-01
-1.18318594e+00 6.50734186e-01 2.11069629e-01 -8.90255392e-01
-2.45202050e-01 1.06108844e+00 4.57857579e-01 -3.78604949e-01
3.38818103e-01 1.60928458e-01 6.68425784e-02 3.83250535e-01
2.95363218e-01 2.90940732e-01 3.76779377e-01 1.51566297e-01
-4.34351176e-01 2.08386302e-01 -6.86731994e-01 5.44824898e-02
1.43623638e+00 -9.62686837e-02 5.74232489e-02 5.69121778e-01
6.23283684e-01 -4.19867873e-01 -1.36505520e+00 -5.48456073e-01
1.78860202e-01 -1.78625196e-01 1.30302981e-01 -1.41602349e+00
-8.61963689e-01 8.54492962e-01 8.83523464e-01 -1.02950729e-01
1.24291778e+00 -2.40333825e-01 6.11911952e-01 7.85205364e-01
7.36374438e-01 -1.37695539e+00 7.68946886e-01 6.82314277e-01
8.70671988e-01 -1.24473310e+00 -6.05023094e-02 -2.84401607e-02
-8.64368260e-01 5.56534767e-01 8.74679089e-01 9.40184370e-02
6.35453641e-01 3.30768943e-01 -1.82983637e-01 3.13850231e-02
-1.41684139e+00 2.11227506e-01 1.88687742e-01 4.82637256e-01
2.62257904e-02 -1.58625916e-01 -1.61811441e-01 6.44708574e-01
2.11013883e-01 1.65491134e-01 2.11041898e-01 1.00850821e+00
-7.02749848e-01 -1.04754758e+00 -1.19397968e-01 5.65307736e-01
-4.72128987e-01 -4.88690645e-01 7.61314854e-02 5.33069134e-01
3.55278999e-01 6.78760469e-01 -1.87889323e-01 -2.48101160e-01
9.50979218e-02 6.70601845e-01 4.65639800e-01 -6.82205617e-01
-1.04024279e+00 -2.01218024e-01 -1.95831269e-01 -2.15337321e-01
-1.15176551e-01 -3.64620596e-01 -7.67306030e-01 -1.59114093e-01
-3.92927706e-01 6.37657493e-02 7.60712922e-01 9.01445448e-01
8.14872503e-01 7.16349065e-01 4.77486432e-01 -5.77516437e-01
-8.21419179e-01 -1.01368666e+00 -4.62138891e-01 3.73342901e-01
6.92029223e-02 -9.23447311e-01 -1.85003102e-01 1.17381345e-02] | [9.642034530639648, 3.419278383255005] |
72d82952-cd62-409a-9e42-deae5e4d0970 | transformers-in-healthcare-a-survey | 2307.00067 | null | https://arxiv.org/abs/2307.00067v1 | https://arxiv.org/pdf/2307.00067v1.pdf | Transformers in Healthcare: A Survey | With Artificial Intelligence (AI) increasingly permeating various aspects of society, including healthcare, the adoption of the Transformers neural network architecture is rapidly changing many applications. Transformer is a type of deep learning architecture initially developed to solve general-purpose Natural Language Processing (NLP) tasks and has subsequently been adapted in many fields, including healthcare. In this survey paper, we provide an overview of how this architecture has been adopted to analyze various forms of data, including medical imaging, structured and unstructured Electronic Health Records (EHR), social media, physiological signals, and biomolecular sequences. Those models could help in clinical diagnosis, report generation, data reconstruction, and drug/protein synthesis. We identified relevant studies using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We also discuss the benefits and limitations of using transformers in healthcare and examine issues such as computational cost, model interpretability, fairness, alignment with human values, ethical implications, and environmental impact. | ['Parisa Rashidi', 'Kia Khezeli', 'Azra Bihorac', 'Benjamin Shickel', 'Jessica Sena', 'Brandon Silva', 'Aysegul Bumin', 'Scott Siegel', 'Miguel Contreras', 'Jiaqing Zhang', 'Sabyasachi Bandyopadhyay', 'Subhash Nerella'] | 2023-06-30 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [ 3.82992595e-01 4.49879616e-01 -2.53251731e-01 -3.98134202e-01
-4.12600607e-01 -3.05751324e-01 1.11020513e-01 6.56623840e-01
-6.50861442e-01 1.04920566e+00 4.87331122e-01 -6.22524798e-01
-3.12824547e-01 -6.00679576e-01 -4.14524555e-01 -5.98430157e-01
-1.90651789e-01 5.74608505e-01 -6.26039922e-01 1.95643112e-01
1.77874312e-01 4.50933486e-01 -9.86265659e-01 5.41914165e-01
7.58360624e-01 1.11031961e+00 -2.83460498e-01 2.38086745e-01
-2.87052751e-01 1.18047118e+00 -7.12729454e-01 -6.08995855e-01
-2.05647364e-01 -2.21290886e-01 -9.36181962e-01 -3.65802258e-01
-3.53820056e-01 -2.61341184e-01 1.61070414e-02 9.13956463e-01
9.90251601e-01 -2.34597251e-01 5.50946593e-01 -1.04343522e+00
-1.07879722e+00 5.01819253e-01 -1.46513611e-01 1.44510632e-02
4.28577304e-01 3.34115446e-01 5.07820725e-01 -7.79225171e-01
7.85833538e-01 1.10914028e+00 7.91119814e-01 6.53402805e-01
-6.67666376e-01 -7.29535758e-01 -2.05331028e-01 2.51742303e-01
-1.21612513e+00 -4.06328857e-01 3.96896511e-01 -5.68589866e-01
1.38454521e+00 1.95479751e-01 7.79873967e-01 1.37671196e+00
8.85217249e-01 5.66975713e-01 8.98956358e-01 -1.61595508e-01
3.32284421e-01 1.27463698e-01 -2.20383424e-03 5.11251867e-01
4.08239663e-01 -3.63264978e-02 -3.94599915e-01 -6.73614264e-01
4.63954389e-01 6.49993345e-02 -1.03437908e-01 2.92381942e-02
-1.45504606e+00 9.99228716e-01 4.23987418e-01 4.34197336e-01
-9.16233540e-01 -1.28750816e-01 8.33951354e-01 1.67502657e-01
3.95506710e-01 6.77080333e-01 -7.08378911e-01 -1.39935344e-01
-5.68273723e-01 1.04295976e-01 7.37301290e-01 4.66820776e-01
-8.53648260e-02 1.56316519e-01 -1.15195312e-01 7.09199250e-01
4.58692819e-01 3.02115768e-01 6.95479572e-01 -1.01913846e+00
1.28875077e-01 6.40012264e-01 3.75770554e-02 -1.15321028e+00
-9.07775104e-01 -3.22323769e-01 -1.49567854e+00 -4.18012202e-01
-4.91439477e-02 -3.43997091e-01 -7.33438432e-01 1.44187987e+00
3.00284594e-01 -1.38517842e-01 2.36316368e-01 8.14717770e-01
1.32405031e+00 3.82092506e-01 6.07721150e-01 -4.29989159e-01
1.47939825e+00 -4.46869671e-01 -1.15323901e+00 2.58093216e-02
6.84534967e-01 -3.21315974e-01 6.99480474e-01 6.58893764e-01
-1.21225846e+00 -6.10130019e-02 -6.86507106e-01 -5.65486193e-01
-7.75662959e-01 -2.81083643e-01 5.66064239e-01 6.14931583e-01
-9.64706957e-01 4.37003314e-01 -7.29003608e-01 -4.13046539e-01
8.96898627e-01 5.27431607e-01 -3.09416920e-01 1.94727778e-01
-1.53323483e+00 1.04738247e+00 1.43833995e-01 1.80220813e-01
-5.04306078e-01 -7.10605562e-01 -7.09847927e-01 9.09720436e-02
1.79221407e-02 -1.22168791e+00 9.84272301e-01 -8.62723231e-01
-1.26372898e+00 9.82260585e-01 -1.29812956e-01 -7.20141947e-01
3.04666013e-01 6.55373186e-02 -4.00146753e-01 -4.55581322e-02
-6.35959283e-02 6.15327716e-01 8.36639777e-02 -4.66793269e-01
-2.32658833e-01 -5.77865183e-01 -4.79866415e-01 9.03773233e-02
-1.95323125e-01 4.51044381e-01 1.45698860e-01 -5.17078578e-01
-5.46032667e-01 -5.94257355e-01 -6.16798699e-01 2.71205187e-01
-3.37988019e-01 -2.24836424e-01 3.59602481e-01 -8.21743429e-01
1.10512757e+00 -1.90485370e+00 8.28542039e-02 2.56652296e-01
5.65341413e-01 3.50507528e-01 5.19061908e-02 3.59392434e-01
-1.18190996e-01 5.06375790e-01 -3.05718303e-01 1.49770654e-04
-1.36652291e-01 2.42035836e-02 7.72265866e-02 3.47554505e-01
2.53970951e-01 1.31720638e+00 -8.31875861e-01 -3.42996866e-01
2.98302412e-01 7.74039626e-01 -3.73074114e-01 -4.57474738e-02
-4.65734787e-02 5.98891139e-01 -6.67682469e-01 6.25079095e-01
1.53582066e-01 -6.02256894e-01 4.70275372e-01 -1.28583580e-01
1.19848303e-01 4.31854248e-01 -5.45088351e-01 1.57562447e+00
-1.89557403e-01 5.40679991e-01 2.11207032e-01 -1.09408963e+00
7.25384176e-01 7.21618235e-01 9.58416939e-01 -9.64166999e-01
5.33976376e-01 -6.78573921e-02 2.63280958e-01 -8.78236830e-01
-1.59633961e-02 -3.14892083e-01 9.45747122e-02 5.60372531e-01
-1.74256310e-01 1.62562773e-01 -3.23205799e-01 -8.14838633e-02
1.06603611e+00 -3.05064142e-01 9.42294836e-01 -1.04575150e-01
3.12386721e-01 3.28725949e-03 6.23980165e-01 4.69109833e-01
-6.34946942e-01 3.54269236e-01 4.21672076e-01 -1.05689740e+00
-1.01599240e+00 -6.33665800e-01 -1.10053390e-01 5.68317413e-01
-5.73239446e-01 -4.00558710e-01 -8.17780137e-01 -9.48237181e-02
-2.38072887e-01 6.33474052e-01 -8.20505261e-01 -3.13935041e-01
-1.49940401e-01 -1.14545405e+00 6.65085435e-01 3.02862376e-01
2.49085218e-01 -1.56609130e+00 -1.11747062e+00 4.82688606e-01
-2.65529841e-01 -1.06097102e+00 -1.63967395e-03 9.35265943e-02
-6.93413854e-01 -1.20427847e+00 -6.15927935e-01 -5.09442091e-01
1.88754320e-01 -5.57448328e-01 1.08219492e+00 -1.04724243e-01
-5.19719839e-01 2.65220016e-01 -5.80077842e-02 -1.14167821e+00
-5.70731163e-01 -1.68323368e-01 1.61513403e-01 -3.50908548e-01
6.99400306e-01 -6.30634427e-02 -7.75039494e-01 -2.25918204e-01
-1.05254281e+00 1.26403987e-01 4.95744973e-01 8.70879710e-01
4.81866032e-01 -3.84142362e-02 8.24465871e-01 -1.03305972e+00
1.14018965e+00 -7.32766628e-01 8.44881684e-02 2.01621160e-01
-7.86701918e-01 -3.51719052e-01 5.42541564e-01 -2.20284402e-01
-6.92563117e-01 -2.58939475e-01 -4.07911837e-01 -1.08740494e-01
-2.46508852e-01 1.11920798e+00 -4.85990532e-02 1.63041130e-01
7.68630743e-01 1.73264381e-03 2.25509152e-01 -1.43277049e-01
-8.53753742e-03 1.01552725e+00 5.89554831e-02 -9.20897201e-02
-2.04792649e-01 1.99553668e-01 -2.53399491e-01 -7.23972738e-01
-4.51411039e-01 1.00971356e-01 -9.04504508e-02 -7.36068040e-02
9.94354010e-01 -6.89122856e-01 -1.18411124e+00 3.33020955e-01
-1.21339786e+00 -1.78596750e-01 -2.23033100e-01 4.34875637e-01
-2.65852064e-01 1.23024665e-01 -6.70518517e-01 -6.60442352e-01
-1.18899226e+00 -1.15204179e+00 7.41096318e-01 5.40904403e-02
-7.67841935e-01 -8.53346348e-01 1.06836958e-02 4.94814336e-01
6.92670107e-01 6.38211250e-01 1.33933365e+00 -7.20145643e-01
8.99325386e-02 -1.00568399e-01 -1.56405061e-01 1.55139297e-01
3.68706375e-01 -7.06381053e-02 -8.84240270e-01 8.11474547e-02
2.26522520e-01 -3.99584115e-01 1.21323295e-01 8.79789233e-01
1.55474818e+00 -7.97296047e-01 -3.64806116e-01 4.34481025e-01
9.00895417e-01 8.84093165e-01 7.78836310e-01 1.92473173e-01
6.13938630e-01 8.39007735e-01 2.52486914e-01 6.29117668e-01
4.99626309e-01 1.24663934e-01 1.49438486e-01 -2.66204476e-01
4.84257758e-01 2.51578361e-01 2.14805417e-02 7.53689647e-01
5.81236498e-04 -3.20675641e-01 -1.40409172e+00 4.31090146e-01
-1.60329735e+00 -6.63359821e-01 8.92343223e-02 1.83111203e+00
9.01351094e-01 -9.62932557e-02 3.77712250e-02 -1.23565622e-01
3.79005402e-01 -3.71051490e-01 -8.79167974e-01 -9.35292363e-01
-5.88728264e-02 4.22277331e-01 2.66002804e-01 -5.75200375e-03
-6.49619758e-01 5.03258765e-01 7.50818348e+00 2.58347332e-01
-1.33016121e+00 1.46202296e-01 1.19592583e+00 -1.60037493e-03
-2.53076017e-01 -8.22742999e-01 3.01384896e-01 3.63736004e-01
1.52877963e+00 -5.67638934e-01 4.51944053e-01 5.44615209e-01
7.35787272e-01 2.29722396e-01 -1.05589592e+00 1.01249528e+00
-2.96459585e-01 -1.66109347e+00 5.14728539e-02 9.18096974e-02
2.97896922e-01 4.51304287e-01 2.70180881e-01 -1.52396755e-02
1.65725783e-01 -1.41637945e+00 2.91272193e-01 5.29865563e-01
9.81087327e-01 -5.60991943e-01 1.05446911e+00 3.99027914e-02
-3.90012085e-01 -2.08229750e-01 -9.76307839e-02 -1.65468007e-01
8.96443352e-02 6.62255287e-01 -9.83479977e-01 5.35953999e-01
1.04345334e+00 7.73271203e-01 -7.56080076e-02 8.32800031e-01
4.22767490e-01 4.46293414e-01 1.91914346e-02 -2.23056599e-01
1.89945981e-01 -1.72770292e-01 1.09211184e-01 1.15803742e+00
1.24258548e-01 3.21122766e-01 -2.89560050e-01 9.15805101e-01
-1.23756386e-01 4.75585192e-01 -7.82389522e-01 -6.86960638e-01
1.73390448e-01 8.21350098e-01 -5.50901115e-01 -5.75944424e-01
-6.11542284e-01 4.26688254e-01 -1.19944170e-01 3.25524747e-01
-7.03013361e-01 -1.96709037e-02 5.46657562e-01 3.66910622e-02
-3.51146787e-01 3.54547232e-01 -6.54442310e-01 -9.36167359e-01
-2.51571387e-01 -1.44636023e+00 5.42863131e-01 -9.97417390e-01
-1.28744245e+00 6.05702460e-01 -5.24232574e-02 -9.68338609e-01
-2.80643046e-01 -6.85229003e-01 1.70857366e-02 7.39959538e-01
-1.03808737e+00 -7.37797260e-01 5.83832152e-02 4.20978665e-01
2.12296039e-01 -3.54547143e-01 1.36141860e+00 4.07493055e-01
-6.15064502e-01 4.06591117e-01 2.58616477e-01 1.39718354e-01
6.12037599e-01 -9.04124200e-01 2.47124404e-01 1.23802729e-01
-5.98998487e-01 9.57607806e-01 5.16240239e-01 -5.23947179e-01
-1.21403539e+00 -1.16442895e+00 1.15240753e+00 -1.36911333e-01
4.77050900e-01 -1.57538116e-01 -7.38791704e-01 7.08033979e-01
3.97753000e-01 -2.77888209e-01 1.33301091e+00 -2.49187008e-01
-4.09780219e-02 -1.54593661e-02 -1.71728504e+00 6.84774816e-01
5.88377655e-01 -4.30632383e-01 -3.57883871e-01 4.46867645e-01
6.93788052e-01 -3.64742994e-01 -1.34571433e+00 3.09358299e-01
7.52083898e-01 -6.12891316e-01 9.25336719e-01 -9.42204237e-01
6.26482427e-01 1.89940915e-01 2.20205128e-01 -1.07798409e+00
-5.09855866e-01 -6.22189581e-01 -1.58056080e-01 5.32703996e-01
3.09692800e-01 -9.56630647e-01 6.11801684e-01 1.23885798e+00
-1.68157890e-01 -9.77701306e-01 -1.05914915e+00 1.88991785e-01
3.06484461e-01 -3.24916840e-01 8.21045876e-01 1.41477609e+00
2.71325827e-01 4.29492712e-01 -3.12999755e-01 -3.67540061e-01
3.88149261e-01 -3.52488846e-01 2.59240866e-01 -1.27941275e+00
1.66563585e-01 -4.45072412e-01 -3.36384237e-01 9.88000184e-02
-1.66543588e-01 -7.31710792e-01 -4.15119261e-01 -1.86534035e+00
3.51979226e-01 -1.39595894e-02 -5.79124928e-01 7.35807776e-01
1.53220728e-01 -5.69190383e-02 -1.39679134e-01 -1.70673311e-01
-3.35584790e-01 3.75654250e-01 1.13186288e+00 -4.52381045e-01
-6.30098060e-02 -4.51817960e-01 -1.28227341e+00 4.24460620e-01
9.25122499e-01 -5.47605097e-01 -1.97518528e-01 -4.01610315e-01
5.52370369e-01 1.94413036e-01 8.31256956e-02 -5.54682910e-01
4.72915322e-02 -3.98102969e-01 6.31185830e-01 -2.02906668e-01
-7.31141195e-02 -1.08038592e+00 7.52595425e-01 8.89109373e-01
-7.03546166e-01 1.67730600e-01 4.36639249e-01 1.25890344e-01
-1.82578743e-01 2.73574471e-01 6.40966594e-01 -4.86384094e-01
-1.59304634e-01 2.96806127e-01 -1.01065505e+00 -2.38672078e-01
9.95042205e-01 -1.47062317e-01 -3.07726145e-01 -4.61941570e-01
-6.96869373e-01 4.36556220e-01 1.20247588e-01 5.30143976e-01
8.35688233e-01 -1.05962157e+00 -6.14841640e-01 2.31148414e-02
2.01265022e-01 1.07830346e-01 4.44722384e-01 1.04660678e+00
-6.35766566e-01 7.41877258e-01 -3.98339063e-01 -1.86422080e-01
-9.91120338e-01 7.74220705e-01 4.61048663e-01 -3.17487925e-01
-5.03623545e-01 3.93194735e-01 3.08352172e-01 -6.31830335e-01
3.69483650e-01 -6.39657199e-01 -3.47034752e-01 -1.48888603e-01
6.89310193e-01 1.83342636e-01 1.52708143e-01 -4.14988101e-01
-6.75851226e-01 1.09243236e-01 -1.16681024e-01 1.01222411e-01
1.62998652e+00 -1.04562946e-01 -4.83103156e-01 4.64709848e-01
1.17173624e+00 -7.03176558e-01 -2.87563682e-01 8.99751410e-02
-2.01137513e-02 3.57805610e-01 1.49148360e-01 -1.22298622e+00
-1.13658941e+00 8.86076152e-01 7.75859594e-01 9.53595266e-02
1.20394599e+00 -4.05517817e-01 7.97583759e-01 4.61124003e-01
2.19623446e-01 -1.02013421e+00 -3.28654319e-01 2.84589916e-01
7.55580425e-01 -1.10856724e+00 4.15893756e-02 1.05706207e-01
-8.73108208e-01 1.01878500e+00 2.69500971e-01 4.54420567e-01
8.67333233e-01 2.84837008e-01 2.87138700e-01 -5.83662152e-01
-8.84414375e-01 5.38582563e-01 -1.74485948e-02 7.72224426e-01
9.34554517e-01 1.09879941e-01 -6.44174993e-01 8.95101905e-01
3.61863859e-02 7.83063948e-01 4.86793488e-01 9.86660063e-01
2.05735922e-01 -8.71034026e-01 -3.63455862e-01 9.38566267e-01
-9.76415694e-01 -2.96466172e-01 -8.40695441e-01 2.78894037e-01
1.16864428e-01 8.51562321e-01 -3.39830443e-02 -1.78183600e-01
8.55458155e-02 1.77619666e-01 8.86006877e-02 -3.69496495e-01
-1.08553815e+00 -4.71439660e-02 2.62543112e-01 -5.86902797e-01
-7.00588405e-01 -4.96325344e-01 -1.25371170e+00 -3.27493459e-01
1.69455588e-01 1.92755744e-01 6.53959453e-01 9.10518587e-01
9.80343282e-01 7.58817255e-01 4.05089604e-03 -2.46298328e-01
9.01325122e-02 -9.31619287e-01 -2.40836158e-01 1.77123785e-01
3.60656977e-01 -3.26979965e-01 1.88949138e-01 -2.37845257e-02] | [7.9151458740234375, 6.972306251525879] |
7a157816-dbbb-4ac2-ab75-d3b3c50c2668 | snore-scalable-unsupervised-learning-of | 2009.04535 | null | https://arxiv.org/abs/2009.04535v2 | https://arxiv.org/pdf/2009.04535v2.pdf | SNoRe: Scalable Unsupervised Learning of Symbolic Node Representations | Learning from complex real-life networks is a lively research area, with recent advances in learning information-rich, low-dimensional network node representations. However, state-of-the-art methods are not necessarily interpretable and are therefore not fully applicable to sensitive settings in biomedical or user profiling tasks, where explicit bias detection is highly relevant. The proposed SNoRe (Symbolic Node Representations) algorithm is capable of learning symbolic, human-understandable representations of individual network nodes, based on the similarity of neighborhood hashes which serve as features. SNoRe's interpretable features are suitable for direct explanation of individual predictions, which we demonstrate by coupling it with the widely used instance explanation tool SHAP to obtain nomograms representing the relevance of individual features for a given classification. To our knowledge, this is one of the first such attempts in a structural node embedding setting. In the experimental evaluation on eleven real-life datasets, SNoRe proved to be competitive to strong baselines, such as variational graph autoencoders, node2vec and LINE. The vectorized implementation of SNoRe scales to large networks, making it suitable for contemporary network learning and analysis tasks. | ['Blaž Škrlj', 'Nada Lavrač', 'Sebastian Mežnar'] | 2020-09-08 | null | null | null | null | ['structural-node-embedding'] | ['graphs'] | [ 5.03482580e-01 9.22922492e-01 -3.89980137e-01 -4.16414201e-01
6.41437545e-02 -2.00268030e-01 7.95003057e-01 8.45986009e-01
-3.18023890e-01 7.21853077e-01 3.41916949e-01 -4.62076783e-01
-6.00230098e-01 -8.93144429e-01 -5.81999242e-01 -7.39437401e-01
-4.05056059e-01 1.03895283e+00 5.15709817e-02 -4.82368499e-01
-2.04275586e-02 6.42576039e-01 -1.30783641e+00 1.61234960e-01
1.33768290e-01 7.06526577e-01 -4.90212947e-01 6.18279457e-01
2.53659859e-02 4.77726877e-01 -4.14802253e-01 -6.74812675e-01
-4.03303862e-01 -2.63111055e-01 -1.05866313e+00 -3.70674819e-01
1.27242982e-01 3.94182019e-02 -5.50323188e-01 8.13187182e-01
3.50641549e-01 8.98782760e-02 1.03542137e+00 -1.58789146e+00
-8.16891253e-01 1.16848040e+00 -9.39823091e-02 3.87181640e-01
1.92174062e-01 -3.07689458e-01 1.65712082e+00 -8.02290618e-01
8.38108599e-01 1.55038965e+00 7.38809109e-01 7.31546700e-01
-1.57254910e+00 -3.59528363e-01 4.16801944e-02 3.14944178e-01
-1.04368603e+00 -1.03300557e-01 7.61837780e-01 -3.87282103e-01
1.02529168e+00 4.88023758e-01 6.03710949e-01 1.68332303e+00
2.29673728e-01 4.57207948e-01 6.16784215e-01 -3.78761649e-01
1.95935503e-01 5.14484823e-01 5.71405649e-01 8.08946967e-01
5.46070933e-01 9.47454572e-02 -2.36481234e-01 -5.34477055e-01
7.31290817e-01 4.46712375e-01 -1.53580233e-01 -6.17632926e-01
-1.30420756e+00 1.37391293e+00 9.96996582e-01 6.92147493e-01
-4.80924428e-01 4.89874989e-01 6.84429288e-01 3.67908418e-01
6.00164711e-01 6.14286900e-01 -4.80503470e-01 3.03525418e-01
-4.10033733e-01 2.99351662e-02 8.15712988e-01 4.21205461e-01
5.82792401e-01 1.02831602e-01 -9.99311879e-02 4.19608414e-01
3.09188068e-01 -8.21126178e-02 5.87687492e-01 -5.48745751e-01
-3.68206613e-02 7.54728019e-01 -6.14648402e-01 -1.30362236e+00
-7.87019253e-01 -6.00991368e-01 -1.22692394e+00 -1.89528279e-02
1.29244477e-01 1.18440820e-03 -5.29651403e-01 1.63047767e+00
2.92977095e-01 3.92974883e-01 1.16151474e-01 6.42006278e-01
1.03039122e+00 2.92461753e-01 2.28926048e-01 -5.17464019e-02
1.43172634e+00 -5.21607816e-01 -6.46020174e-01 -4.40163761e-02
7.82924354e-01 1.11172805e-02 5.48750579e-01 1.43298820e-01
-4.96913671e-01 -3.39221805e-01 -8.55600238e-01 -2.65176315e-02
-8.16703498e-01 -3.53695512e-01 1.20859838e+00 5.96083999e-01
-1.03050601e+00 1.08866096e+00 -7.12270260e-01 -6.99607909e-01
6.32162511e-01 7.41049290e-01 -8.21231604e-01 9.78001282e-02
-1.51257586e+00 9.06500816e-01 4.53755260e-01 1.72442779e-01
-5.92186272e-01 -5.68376541e-01 -1.09459949e+00 5.09919763e-01
3.47672552e-01 -7.17612207e-01 5.93399882e-01 -8.11827540e-01
-1.04740810e+00 7.65659392e-01 6.02766275e-02 -7.10659564e-01
5.31685986e-02 3.41150641e-01 -5.57343662e-01 -5.42056039e-02
-2.75092304e-01 6.25388086e-01 9.12477553e-01 -1.08123755e+00
1.00677475e-01 -5.42559922e-01 -5.63983172e-02 -2.42755368e-01
-7.32051432e-01 -4.68973249e-01 3.09387207e-01 -3.57446402e-01
1.70054864e-02 -9.00495648e-01 -6.01294518e-01 2.04026312e-01
-7.32278049e-01 -5.07319868e-01 6.08415723e-01 -3.50577235e-01
9.16915357e-01 -1.92833710e+00 5.80092430e-01 7.79494286e-01
9.00099456e-01 2.70204365e-01 -1.07329339e-01 6.75107300e-01
-4.96483654e-01 3.78240705e-01 -1.58640832e-01 -1.48213819e-01
1.68492869e-01 5.06649077e-01 -8.68880376e-02 5.36269367e-01
3.76310140e-01 1.07972455e+00 -1.00844240e+00 -3.86205465e-01
3.84447485e-01 1.02948368e+00 -5.63015938e-01 -1.16408393e-01
-1.70627069e-02 5.25567383e-02 -5.73391557e-01 2.34226435e-01
-1.81414559e-01 -5.22026479e-01 4.73485410e-01 -1.69745401e-01
7.15905666e-01 -1.08474851e-01 -7.67210841e-01 1.16165721e+00
-2.02355236e-01 9.98839259e-01 -4.53941494e-01 -1.51838005e+00
9.98830438e-01 4.45913762e-01 3.25850129e-01 -7.61984289e-02
2.88661122e-01 3.07717510e-02 2.06765786e-01 -3.45291913e-01
1.74723729e-01 -1.18120186e-01 6.91700578e-02 2.97741950e-01
3.03503811e-01 3.69991899e-01 -1.67407438e-01 4.27928150e-01
1.24664605e+00 -5.97495019e-01 7.36167789e-01 -1.92221120e-01
5.73920131e-01 -4.45175886e-01 6.75488338e-02 6.37729943e-01
-1.25504553e-01 3.61127049e-01 1.00955379e+00 -6.86901450e-01
-1.04575515e+00 -8.58850837e-01 -2.56462157e-01 1.00792313e+00
-2.72362173e-01 -4.56642598e-01 -4.63907540e-01 -8.10931027e-01
3.39832217e-01 7.48775601e-01 -1.24834120e+00 -5.41738749e-01
-2.74045855e-01 -6.76114202e-01 5.24193168e-01 5.40731609e-01
-5.33468246e-01 -1.19278729e+00 -4.86766919e-02 2.82160640e-01
4.15867388e-01 -1.01986253e+00 1.49729192e-01 3.03893834e-01
-1.02443051e+00 -1.31417727e+00 -3.71238410e-01 -5.63557565e-01
7.71181047e-01 -1.61783829e-01 1.24792469e+00 5.95881701e-01
-5.52644193e-01 4.35867667e-01 -2.54261732e-01 -4.06858593e-01
-5.59039831e-01 2.90905505e-01 2.00143233e-01 1.78302363e-01
4.65387970e-01 -8.55693161e-01 -5.96685231e-01 -6.56941682e-02
-9.99593318e-01 -4.07836586e-01 6.52014554e-01 1.23168862e+00
4.91355568e-01 -2.07703784e-01 6.12239480e-01 -1.49918020e+00
8.24179590e-01 -8.50780427e-01 -2.41748663e-03 8.80238637e-02
-1.01303875e+00 4.01605308e-01 8.26180518e-01 -3.34903419e-01
-2.58355975e-01 -2.01682970e-01 -2.28297502e-01 -3.08683336e-01
-1.95006639e-01 5.99054813e-01 2.21643344e-01 -1.20046414e-01
8.55731428e-01 7.48193786e-02 2.84628272e-01 -2.68245786e-01
4.80346680e-01 4.34167296e-01 2.56937265e-01 5.75220808e-02
7.64020801e-01 3.70444298e-01 5.81193149e-01 -1.06195760e+00
-6.52683377e-01 -3.65744740e-01 -7.38642991e-01 6.88919425e-02
7.78592110e-01 -4.54595268e-01 -1.00198078e+00 -5.31272471e-01
-9.99060392e-01 2.04031780e-01 -2.92646736e-01 3.67169172e-01
-2.90322691e-01 2.19901264e-01 -5.18617988e-01 -5.91731727e-01
-4.75955039e-01 -8.59450459e-01 1.02038383e+00 -1.12688094e-01
-6.13488674e-01 -1.73924398e+00 9.29240063e-02 1.64362535e-01
3.57339948e-01 5.75249016e-01 1.25147426e+00 -1.36554301e+00
-6.86132461e-02 -5.52015007e-01 -1.16699129e-01 2.60015558e-02
-1.91281304e-01 5.89320548e-02 -9.33248580e-01 -2.45867401e-01
-6.99215651e-01 -1.62070379e-01 1.00615406e+00 2.22302124e-01
1.31472349e+00 -3.90537769e-01 -7.14893401e-01 2.86944836e-01
1.27072859e+00 -5.19340277e-01 4.86529559e-01 -7.58285895e-02
8.58791232e-01 6.21330023e-01 -4.29335460e-02 3.00448447e-01
1.73560992e-01 8.14363539e-01 8.53298545e-01 -1.36143252e-01
3.94785441e-02 -3.10440838e-01 1.37138098e-01 5.37034333e-01
-8.78344327e-02 -3.94937992e-01 -8.61231327e-01 5.53636670e-01
-1.96748018e+00 -9.82697308e-01 -4.52537835e-01 1.68039787e+00
3.40802550e-01 2.91774571e-01 3.51664908e-02 3.79443437e-01
6.28802180e-01 2.59325951e-01 -5.39048016e-01 -7.05232561e-01
1.00217044e-01 2.83426911e-01 3.16325903e-01 5.07087886e-01
-8.93984318e-01 7.00617194e-01 5.92960453e+00 4.94457722e-01
-7.82280922e-01 1.02955744e-01 6.18667126e-01 2.86975890e-01
-6.67520702e-01 -1.29055679e-01 -4.24941301e-01 5.78133240e-02
1.39432192e+00 -2.53645927e-02 3.37957680e-01 9.55952406e-01
-9.14947465e-02 6.68565333e-01 -1.49382925e+00 8.34197104e-01
1.68614000e-01 -1.64867270e+00 1.84248596e-01 2.20207736e-01
2.89057732e-01 -1.79033689e-02 1.23911500e-02 3.13005626e-01
1.43517971e-01 -1.58516264e+00 -4.09829058e-02 4.65601265e-01
4.90783811e-01 -5.58566034e-01 1.11172473e+00 7.14762509e-02
-7.32149839e-01 -1.08542427e-01 -6.56120837e-01 -1.20281152e-01
-1.16223924e-01 7.03039050e-01 -1.33212817e+00 4.16782856e-01
2.64977515e-01 1.05031013e+00 -7.17222095e-01 3.86260062e-01
-1.99823901e-01 7.03664780e-01 -2.30485156e-01 -5.55395305e-01
2.48191506e-01 2.58883953e-01 4.66724515e-01 1.05758679e+00
5.39019797e-03 -1.44286066e-01 -2.97905266e-01 8.71394217e-01
-1.92581996e-01 1.53771594e-01 -9.89264071e-01 -2.89292663e-01
2.79026389e-01 1.50154471e+00 -7.56982625e-01 -2.10634574e-01
-8.22446123e-02 8.42589557e-01 5.86877584e-01 1.45533308e-01
-5.30118704e-01 -2.91281193e-01 7.31898010e-01 1.07746862e-01
3.42577040e-01 1.73421249e-01 -7.00923353e-02 -9.60396409e-01
-4.48819011e-01 -6.21215582e-01 6.24042273e-01 -4.63430852e-01
-1.50845456e+00 8.14609289e-01 -1.14423804e-01 -7.33620584e-01
-4.63445336e-01 -1.01273406e+00 -6.72920287e-01 4.20477122e-01
-1.25771129e+00 -1.10301673e+00 -1.15076825e-01 6.20109200e-01
9.00731087e-02 -3.51642370e-01 1.46013439e+00 5.26519939e-02
-6.95664704e-01 6.23918116e-01 2.25644514e-01 2.40066290e-01
9.74069089e-02 -1.42905331e+00 3.89848113e-01 1.10795997e-01
8.03523660e-01 6.56841934e-01 9.61906791e-01 -2.38907754e-01
-1.28335631e+00 -9.55636144e-01 1.05816913e+00 -5.89552641e-01
8.98504972e-01 -6.09124899e-01 -1.01483870e+00 8.84828806e-01
-1.40692487e-01 4.05454785e-01 1.01186466e+00 5.65835953e-01
-4.93937612e-01 -5.49934059e-02 -1.00795090e+00 4.85690027e-01
1.05528295e+00 -5.84498584e-01 -4.89376515e-01 7.70648301e-01
7.02344537e-01 1.54849216e-01 -1.19881511e+00 1.30087733e-01
4.25059766e-01 -8.17099094e-01 1.28397441e+00 -1.33414185e+00
6.36519015e-01 3.32117051e-01 1.88738555e-01 -1.37779820e+00
-5.10464191e-01 -3.24737936e-01 -4.53403920e-01 1.02103829e+00
6.95292056e-01 -7.85493553e-01 8.14756095e-01 4.02491242e-01
3.24285358e-01 -9.18081582e-01 -9.17059004e-01 -2.96548694e-01
-1.16303764e-01 -1.36875078e-01 6.82156503e-01 1.11796093e+00
2.62856752e-01 6.88684285e-01 -3.56261134e-01 7.50801489e-02
7.50901878e-01 -1.11495823e-01 4.43865597e-01 -1.92431486e+00
-3.83718997e-01 -5.16021848e-01 -1.20940769e+00 -9.17944759e-02
6.41146183e-01 -1.40835762e+00 -5.50783932e-01 -1.45112765e+00
1.24534689e-01 -1.31145567e-01 -5.42804837e-01 4.46052790e-01
-1.46769419e-01 2.96008259e-01 -1.43391222e-01 -9.48450938e-02
-3.76768917e-01 6.17912352e-01 9.08184290e-01 -2.69319922e-01
2.37540871e-01 7.70794749e-02 -8.27336907e-01 6.15593374e-01
7.74703205e-01 -6.75679624e-01 -5.27465165e-01 3.49161997e-02
3.43929052e-01 3.39621082e-02 6.93222702e-01 -5.49211144e-01
1.22049088e-02 2.95260668e-01 4.80105609e-01 4.22364101e-02
3.31803679e-01 -1.05382490e+00 1.89091817e-01 7.20800638e-01
-8.68947804e-01 -8.16061646e-02 -2.18103036e-01 1.09800959e+00
-8.98304116e-03 -1.36696920e-01 2.88859427e-01 1.23206325e-01
-6.68463051e-01 5.78176737e-01 -1.27886340e-01 -1.18725739e-01
1.00209260e+00 -1.38482764e-01 -2.04214469e-01 -5.91973066e-01
-9.78167474e-01 3.96397859e-02 -1.22805968e-01 5.44343412e-01
8.37064147e-01 -1.27961302e+00 -8.02666843e-01 1.76816314e-01
2.31157929e-01 -4.21753347e-01 1.42727047e-01 6.79053068e-01
-3.47308248e-01 5.82956970e-01 -2.13680327e-01 -5.58892548e-01
-1.36443067e+00 7.76264369e-01 1.79500848e-01 -4.18760329e-01
-7.68119872e-01 8.01867843e-01 -1.41810969e-01 -4.97546554e-01
1.61329091e-01 -9.67101231e-02 -6.97803795e-01 9.68101844e-02
3.14962029e-01 2.57480800e-01 -7.63809010e-02 -5.74565887e-01
-5.01348495e-01 2.08814397e-01 -6.22286648e-02 3.45692933e-01
1.76970565e+00 3.22659999e-01 -2.20481649e-01 4.81769204e-01
1.43282795e+00 -6.07640505e-01 -5.27034163e-01 -2.93581873e-01
1.99584946e-01 -3.98560353e-02 1.08809680e-01 -3.72555315e-01
-1.17405450e+00 1.10135949e+00 3.73258084e-01 5.97483337e-01
4.75252926e-01 2.91840583e-01 3.96921098e-01 6.68458939e-01
6.43854812e-02 -4.99362111e-01 -1.82926983e-01 1.04521982e-01
9.35005903e-01 -1.37289917e+00 1.56417623e-01 -5.20966947e-01
-3.89277846e-01 1.40063250e+00 2.41371706e-01 -1.98752925e-01
8.38780761e-01 -3.77518564e-01 -2.78414994e-01 -6.63569510e-01
-9.27882254e-01 6.16763122e-02 5.16338110e-01 7.55593240e-01
4.71675247e-01 2.27101430e-01 -1.12212248e-01 5.67765772e-01
-1.45233423e-01 -4.92039472e-01 5.68058848e-01 1.93522289e-01
-2.04051271e-01 -1.03234255e+00 4.44697104e-02 8.94362628e-01
-2.84323424e-01 -8.65973607e-02 -6.20580375e-01 9.61038888e-01
-3.02825600e-01 5.49613059e-01 -9.13063879e-04 -4.19638574e-01
2.90253907e-01 2.32058182e-01 1.34592921e-01 -5.97006083e-01
-7.97901511e-01 -6.14736319e-01 1.36191204e-01 -6.58584714e-01
-3.36289823e-01 -5.50135374e-01 -1.22713363e+00 -5.21812737e-01
-3.13603818e-01 1.32746369e-01 5.77162445e-01 9.17824984e-01
3.44347268e-01 7.51905978e-01 4.06271845e-01 -8.43128264e-01
-6.61202729e-01 -7.85404325e-01 -6.41274869e-01 7.73764968e-01
3.71258467e-01 -6.26497447e-01 -5.12028813e-01 -4.67752814e-01] | [7.244147777557373, 6.466592311859131] |
fcca6d3a-d99b-4607-8a83-a9e934fab7f2 | shapelet-based-counterfactual-explanations | 2208.10462 | null | https://arxiv.org/abs/2208.10462v1 | https://arxiv.org/pdf/2208.10462v1.pdf | Shapelet-Based Counterfactual Explanations for Multivariate Time Series | As machine learning and deep learning models have become highly prevalent in a multitude of domains, the main reservation in their adoption for decision-making processes is their black-box nature. The Explainable Artificial Intelligence (XAI) paradigm has gained a lot of momentum lately due to its ability to reduce models opacity. XAI methods have not only increased stakeholders' trust in the decision process but also helped developers ensure its fairness. Recent efforts have been invested in creating transparent models and post-hoc explanations. However, fewer methods have been developed for time series data, and even less when it comes to multivariate datasets. In this work, we take advantage of the inherent interpretability of shapelets to develop a model agnostic multivariate time series (MTS) counterfactual explanation algorithm. Counterfactuals can have a tremendous impact on making black-box models explainable by indicating what changes have to be performed on the input to change the final decision. We test our approach on a real-life solar flare prediction dataset and prove that our approach produces high-quality counterfactuals. Moreover, a comparison to the only MTS counterfactual generation algorithm shows that, in addition to being visually interpretable, our explanations are superior in terms of proximity, sparsity, and plausibility. | ['Shah Muhammad Hamdi', 'Soukaina Filali Boubrahimi', 'Omar Bahri'] | 2022-08-22 | null | null | null | null | ['counterfactual-explanation', 'solar-flare-prediction'] | ['miscellaneous', 'time-series'] | [ 3.29621851e-01 5.47508657e-01 -3.03693175e-01 -3.94199580e-01
-1.42527074e-01 -6.36443019e-01 1.00211108e+00 1.63019553e-01
2.79499143e-01 9.51278090e-01 5.36036253e-01 -8.85150135e-01
-4.61946070e-01 -7.95307040e-01 -6.86518312e-01 -1.86651587e-01
-8.68347511e-02 2.99117386e-01 -4.58491772e-01 -1.04889832e-01
2.55760491e-01 4.37579453e-01 -1.61619687e+00 4.04865026e-01
1.07114267e+00 9.13856268e-01 -3.72226000e-01 1.79648608e-01
-1.90767124e-01 9.46728766e-01 -5.39302945e-01 -6.86670840e-01
3.28529805e-01 -4.62871045e-01 -5.54254115e-01 -1.53373003e-01
1.07081562e-01 -1.55306337e-02 1.24624826e-01 6.85865462e-01
2.13801503e-01 -2.51493275e-01 7.18227446e-01 -1.68473411e+00
-8.20609272e-01 7.75012076e-01 -5.25933743e-01 -1.77469745e-01
1.36725575e-01 2.97745526e-01 1.16307771e+00 -4.74041402e-01
4.15836215e-01 1.27820921e+00 7.78492272e-01 4.23909038e-01
-1.43668115e+00 -6.65709317e-01 2.90227056e-01 3.59636128e-01
-6.27539873e-01 -1.37576297e-01 1.00418091e+00 -5.43173075e-01
1.05666149e+00 7.50873625e-01 9.92802620e-01 9.94594038e-01
5.46919882e-01 6.60261929e-01 1.48015082e+00 -6.35628104e-01
5.56250036e-01 1.43630607e-02 -1.18162774e-01 3.53876621e-01
6.62357092e-01 5.25343299e-01 -3.31767887e-01 -1.22399315e-01
4.31276262e-01 1.72799736e-01 -2.60999382e-01 -4.83744234e-01
-1.18647575e+00 1.13956964e+00 5.73788941e-01 2.36002788e-01
-4.34449524e-01 2.81635731e-01 2.69558221e-01 3.93825382e-01
5.61333537e-01 8.97781670e-01 -5.98280430e-01 8.85348581e-03
-7.61889637e-01 4.67377305e-01 6.64343655e-01 2.09713295e-01
3.96855831e-01 4.13655192e-01 -5.00682518e-02 3.67602646e-01
2.94513822e-01 4.85395461e-01 5.50287306e-01 -9.43414509e-01
3.26428354e-01 1.02087998e+00 2.45023683e-01 -1.07891512e+00
-4.10074592e-01 -5.13055801e-01 -8.77474248e-01 9.80418622e-01
2.97136843e-01 -2.30363920e-01 -8.03382337e-01 1.57125413e+00
1.15212552e-01 1.22534810e-02 1.07530884e-01 9.08257365e-01
2.74983138e-01 3.83494169e-01 -7.41618732e-03 -9.33595374e-02
9.07591462e-01 -4.09717500e-01 -6.04599535e-01 -3.07781786e-01
6.20193779e-01 -5.01026928e-01 1.01267493e+00 4.67301846e-01
-8.58592987e-01 -3.85373831e-01 -1.19968009e+00 2.82134861e-01
-4.95618582e-01 -3.38661700e-01 1.29362559e+00 7.52494335e-01
-6.54467762e-01 8.44615996e-01 -5.61290681e-01 -2.27103755e-02
6.54998541e-01 3.30094904e-01 -2.65195847e-01 1.73438549e-01
-1.00635576e+00 1.04777932e+00 1.62951007e-01 -2.56273374e-02
-3.66464019e-01 -8.68815899e-01 -7.14244366e-01 3.68658662e-01
3.72111708e-01 -9.74309087e-01 1.11932945e+00 -1.54795337e+00
-1.13959336e+00 4.51198369e-01 8.97759274e-02 -1.08794844e+00
8.20035040e-01 1.24345675e-01 -5.97633421e-01 -5.29544771e-01
1.51319534e-01 5.75592041e-01 7.34108269e-01 -1.14788496e+00
-7.19928086e-01 -4.07637894e-01 3.24972868e-01 -2.13315398e-01
9.10081267e-02 -3.01552176e-01 4.42727000e-01 -8.66757452e-01
-7.36868978e-02 -9.52772260e-01 -3.38247269e-01 6.74702227e-02
-4.34424669e-01 1.20013123e-02 8.45310688e-01 -4.92230415e-01
1.26474893e+00 -1.70302141e+00 -2.16370776e-01 1.95112318e-01
2.26688847e-01 1.13666601e-01 1.95869923e-01 2.97280639e-01
-6.01570427e-01 5.80931485e-01 -4.63105321e-01 -7.97994286e-02
3.30105782e-01 2.82990158e-01 -8.10459256e-01 1.47490859e-01
3.89913797e-01 9.67669725e-01 -6.50704682e-01 -7.14672580e-02
4.06433135e-01 2.34844431e-01 -5.17904758e-01 -1.52581215e-01
-5.49989641e-01 2.48517454e-01 -2.00050518e-01 3.42791796e-01
4.93553728e-01 -1.74863517e-01 3.61015528e-01 2.19185397e-01
-3.72745097e-01 2.70919830e-01 -1.03665698e+00 1.18940377e+00
-6.29878402e-01 9.54388142e-01 -5.63797653e-01 -9.08590376e-01
7.25347579e-01 3.85450602e-01 3.30905318e-01 -9.85627890e-01
-1.30138353e-01 2.58560061e-01 3.44381988e-01 -3.04551899e-01
2.11098447e-01 -5.01773596e-01 5.02924807e-02 7.91933417e-01
-6.98968291e-01 -2.30151325e-01 3.19886543e-02 -1.07840516e-01
8.16627920e-01 2.99925804e-01 7.65876174e-01 -1.45637766e-01
2.08981797e-01 3.57547402e-01 6.64032102e-01 4.55091357e-01
1.65479124e-01 7.28310287e-01 5.60509920e-01 -1.12317681e+00
-9.74764585e-01 -7.50885308e-01 -1.08436625e-02 3.28107297e-01
-3.48550588e-01 -7.38230674e-03 -5.50321221e-01 -8.58116090e-01
3.59579325e-01 1.43276525e+00 -9.59762633e-01 -1.06229715e-01
-2.59211093e-01 -6.32181585e-01 1.79547757e-01 6.60218596e-01
3.61878395e-01 -1.21235490e+00 -1.17517161e+00 2.52401203e-01
5.35563678e-02 -4.04321522e-01 4.76377755e-02 2.52777133e-02
-9.95653927e-01 -1.20564759e+00 -3.48942369e-01 1.53961718e-01
4.82629627e-01 2.25940272e-01 1.26284134e+00 9.89784598e-02
-2.69573122e-01 1.20396778e-01 -7.38290548e-02 -1.13763642e+00
-5.39046884e-01 -4.40149754e-01 -6.79375529e-02 -3.28283161e-02
3.59138995e-01 -6.29661262e-01 -5.09513319e-01 7.44761899e-02
-1.06830263e+00 5.08151531e-01 5.69788337e-01 7.75153279e-01
3.27956349e-01 8.45530331e-02 7.40441382e-01 -1.03795838e+00
6.76983833e-01 -4.47282314e-01 -6.99395180e-01 3.15425962e-01
-1.27993476e+00 3.58735025e-01 7.26689875e-01 -2.75539637e-01
-1.30844390e+00 -2.78690100e-01 3.31970930e-01 -2.33140588e-01
-1.79903582e-01 8.07030201e-01 -1.92528330e-02 3.47898543e-01
7.49867141e-01 -3.29148293e-01 4.82860208e-02 -3.75225157e-01
5.86297631e-01 3.08137417e-01 3.28629553e-01 -2.09117606e-01
9.03192639e-01 7.14532852e-01 1.39048994e-01 -2.29640067e-01
-6.17922485e-01 3.44222635e-01 -4.17785823e-01 -1.86260507e-01
5.86291850e-01 -4.93898928e-01 -6.50126874e-01 -1.27845649e-02
-1.09569561e+00 -2.31571972e-01 -6.05080068e-01 2.83199668e-01
-5.42459548e-01 1.00757845e-01 2.99552858e-01 -9.74693418e-01
-2.63265520e-01 -7.28538454e-01 4.14170206e-01 1.15337804e-01
-8.18858564e-01 -1.24142814e+00 1.37192458e-01 3.50695908e-01
5.60981929e-01 8.11293304e-01 1.32507491e+00 -5.53007483e-01
-5.22660673e-01 -1.95629299e-01 -2.84622580e-01 2.51949243e-02
2.15928614e-01 6.84019923e-02 -1.12128568e+00 4.19005901e-02
2.43570292e-04 1.32390663e-01 6.33683085e-01 6.75350487e-01
9.45366502e-01 -7.90573835e-01 -2.49196067e-01 3.44518989e-01
1.29949975e+00 3.36009204e-01 6.16135955e-01 6.61746681e-01
4.07609642e-01 8.57225716e-01 7.36641109e-01 5.08293986e-01
3.16966534e-01 6.70216978e-01 8.43005300e-01 -3.92333329e-01
-1.35417417e-01 -3.16930860e-01 2.64298171e-01 -6.66366071e-02
-2.59898275e-01 -1.13676436e-01 -1.07893717e+00 6.00091577e-01
-2.08140850e+00 -1.34965193e+00 -4.01143789e-01 2.26156592e+00
3.17680717e-01 3.64014447e-01 1.39430957e-02 4.85203594e-01
2.44499385e-01 4.09253277e-02 -7.63701439e-01 -7.42908001e-01
-2.87074089e-01 1.88182443e-02 2.17498541e-01 2.87281066e-01
-8.07705462e-01 3.31712633e-01 6.04321241e+00 2.94435650e-01
-1.35675454e+00 -1.62809744e-01 8.41928661e-01 -2.77769148e-01
-1.02619624e+00 2.19077989e-01 1.28750220e-01 4.97521579e-01
7.97196925e-01 -5.73870122e-01 3.16661775e-01 8.88115287e-01
7.50815749e-01 5.39476350e-02 -1.31584311e+00 7.48153985e-01
-2.97458440e-01 -1.99346304e+00 2.80685782e-01 2.36262679e-01
8.65434885e-01 -3.11860800e-01 2.40203172e-01 1.27863422e-01
4.05756682e-01 -1.40472138e+00 1.12210226e+00 6.08267128e-01
4.81060863e-01 -9.15244222e-01 7.96164572e-01 1.95582435e-01
-7.91881680e-01 -4.84295726e-01 -1.97685927e-01 -8.24868143e-01
-5.34564927e-02 6.63319767e-01 -1.16943395e+00 6.72662377e-01
5.67237437e-01 5.43959558e-01 -4.45160747e-01 8.98843646e-01
-4.23714668e-01 8.62119317e-01 -6.12267815e-02 9.17916521e-02
5.06075211e-02 -1.70881242e-01 4.17050421e-01 8.85795176e-01
3.98202688e-01 -4.70579453e-02 -4.37458307e-01 1.35518610e+00
1.52974293e-01 -1.97940037e-01 -9.32568967e-01 -1.31323367e-01
2.73618639e-01 8.18806946e-01 -4.87931728e-01 -2.32515916e-01
-5.17825246e-01 6.24569774e-01 -4.54243347e-02 3.45204562e-01
-8.32331717e-01 -6.72888290e-03 8.55271816e-01 3.48128796e-01
3.92032489e-02 1.88111752e-01 -1.09942508e+00 -9.49352443e-01
7.28554651e-02 -1.14345610e+00 4.60315287e-01 -1.10138893e+00
-1.13712823e+00 3.69431198e-01 -1.44290492e-01 -1.40089583e+00
-7.01161623e-01 -4.59712595e-01 -1.04731131e+00 9.22629476e-01
-1.48589134e+00 -1.32604957e+00 -1.00897521e-01 4.44535799e-02
4.66326565e-01 -1.72849998e-01 8.67473662e-01 -3.12267095e-01
-3.19715947e-01 1.01779319e-01 4.38006371e-02 -2.99274087e-01
2.08229974e-01 -1.41857123e+00 6.28465712e-01 9.39212620e-01
4.56651300e-01 5.79763353e-01 1.06787562e+00 -5.28746367e-01
-1.13598299e+00 -9.71427143e-01 9.63530064e-01 -5.95495164e-01
6.21491432e-01 4.71050059e-03 -7.42827594e-01 7.42834210e-01
3.01515073e-01 -5.06575286e-01 6.21834874e-01 2.89200515e-01
-3.13764215e-01 -1.34936765e-01 -1.11834598e+00 8.39324474e-01
6.78340316e-01 -2.06874207e-01 -9.17494476e-01 2.03683116e-02
6.14481390e-01 1.13370784e-01 -4.99886632e-01 3.95110846e-01
7.16320038e-01 -1.47778988e+00 8.23788285e-01 -8.13037813e-01
7.01485574e-01 -3.89397532e-01 1.47903457e-01 -1.39276433e+00
-3.06320697e-01 -7.00633109e-01 1.85695570e-02 1.15767801e+00
7.04590559e-01 -8.63148689e-01 7.52065122e-01 1.15511262e+00
-6.59189075e-02 -7.74998665e-01 -8.28489900e-01 -6.53875053e-01
3.67030315e-02 -8.84614408e-01 1.24114776e+00 1.10618770e+00
7.68073723e-02 1.98481753e-01 -3.25081199e-01 7.30477422e-02
4.32426304e-01 8.21894288e-01 9.46474195e-01 -1.57446814e+00
-3.66363555e-01 -6.42399490e-01 -2.79834181e-01 -2.24167258e-01
-3.23686786e-02 -7.80572534e-01 -5.60335279e-01 -1.64672458e+00
1.51941732e-01 -3.12509269e-01 -2.31312022e-01 8.82441044e-01
-8.99144635e-02 8.52491185e-02 3.06825995e-01 2.16369301e-01
-1.63229089e-02 4.05405551e-01 8.10783565e-01 -2.73957670e-01
-2.20057741e-01 1.74527392e-01 -9.30294633e-01 1.00192678e+00
8.90583456e-01 -4.96521473e-01 -4.68168199e-01 -3.47154915e-01
5.33047080e-01 4.53055128e-02 6.95466816e-01 -8.93365979e-01
-2.56687641e-01 -4.81163591e-01 4.48472232e-01 -3.38450164e-01
2.12535635e-02 -9.02060330e-01 8.97903442e-01 7.62281299e-01
-3.00047547e-01 1.06796190e-01 4.12464082e-01 5.55783153e-01
-1.91056907e-01 2.41482202e-02 4.07916456e-01 4.41990867e-02
-6.72789335e-01 -6.42460808e-02 -1.57446906e-01 -3.71593684e-01
1.01043737e+00 -5.51321149e-01 -3.02536249e-01 -5.64616561e-01
-4.66054589e-01 6.45094812e-02 7.76227772e-01 5.71599960e-01
3.61772597e-01 -1.16420305e+00 -7.66080976e-01 1.08836219e-01
1.28338829e-01 -4.17594969e-01 1.13832347e-01 5.85963726e-01
-2.58981615e-01 7.18700349e-01 -3.70659471e-01 -2.04697937e-01
-1.13787639e+00 6.38192415e-01 2.30964229e-01 -3.65657717e-01
-6.26982212e-01 2.51591355e-01 2.82922357e-01 -2.56156772e-01
-1.04137942e-01 -6.47702277e-01 -1.58249527e-01 -2.73376871e-02
4.06988710e-01 4.41933483e-01 -5.96893243e-02 -1.13452390e-01
-1.63341358e-01 1.36852667e-01 3.21895152e-01 -1.14307448e-01
1.78085959e+00 1.62448943e-01 4.06765155e-02 4.26329732e-01
4.62045670e-01 7.59817585e-02 -1.40451419e+00 3.69208783e-01
3.03624898e-01 -6.15799129e-01 -5.57421222e-02 -1.36985743e+00
-9.09532011e-01 9.52460825e-01 4.65710431e-01 8.29783916e-01
1.08671963e+00 -2.77111530e-01 2.45008156e-01 1.32898763e-01
3.00165296e-01 -8.05136979e-01 -3.55914742e-01 -1.06886424e-01
1.29833639e+00 -1.33665299e+00 2.02742219e-01 -4.00681086e-02
-8.45661104e-01 1.04464531e+00 9.46532041e-02 1.89665288e-01
1.02578878e-01 7.90160373e-02 2.54667789e-01 -2.35432968e-01
-9.72529888e-01 1.04705639e-01 4.23463404e-01 7.82673597e-01
6.44635797e-01 3.23832512e-01 -2.68180579e-01 7.65550077e-01
-4.66273129e-01 2.17868328e-01 5.91073632e-01 4.26235169e-01
-1.71310142e-01 -1.00743747e+00 -5.77439666e-01 6.18535995e-01
-4.03892189e-01 2.30759569e-02 -6.02834046e-01 1.16104376e+00
1.39144048e-01 1.09364593e+00 -9.14871469e-02 -3.55965495e-02
3.45138162e-01 9.10428837e-02 7.93143958e-02 -3.43380302e-01
-5.84866166e-01 -3.95319313e-01 2.58568257e-01 -5.91295660e-01
-2.73644537e-01 -7.39048600e-01 -1.27763546e+00 -4.98791575e-01
-1.43604279e-01 1.73868164e-01 8.32328379e-01 1.09640837e+00
6.42441928e-01 6.45811915e-01 4.99769002e-01 -6.91127896e-01
-6.48719728e-01 -6.46380007e-01 -2.55084306e-01 5.55079699e-01
1.99074090e-01 -6.19075060e-01 -3.97220194e-01 4.47336026e-02] | [8.730120658874512, 5.642559051513672] |
9d5f2d13-321c-4022-9fb7-700f2cd37059 | statistical-user-simulation-for-spoken | null | null | https://aclanthology.org/W12-1805 | https://aclanthology.org/W12-1805.pdf | Statistical User Simulation for Spoken Dialogue Systems: What for, Which Data, Which Future? | null | ['Olivier Pietquin'] | 2012-06-01 | null | null | null | ws-2012-6 | ['user-simulation'] | ['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] | [-7.238363265991211, 3.7262535095214844] |
1b383c88-4a43-4fcc-9bf5-6cbba3700ea5 | kleister-key-information-extraction-datasets | 2105.05796 | null | https://arxiv.org/abs/2105.05796v1 | https://arxiv.org/pdf/2105.05796v1.pdf | Kleister: Key Information Extraction Datasets Involving Long Documents with Complex Layouts | The relevance of the Key Information Extraction (KIE) task is increasingly important in natural language processing problems. But there are still only a few well-defined problems that serve as benchmarks for solutions in this area. To bridge this gap, we introduce two new datasets (Kleister NDA and Kleister Charity). They involve a mix of scanned and born-digital long formal English-language documents. In these datasets, an NLP system is expected to find or infer various types of entities by employing both textual and structural layout features. The Kleister Charity dataset consists of 2,788 annual financial reports of charity organizations, with 61,643 unique pages and 21,612 entities to extract. The Kleister NDA dataset has 540 Non-disclosure Agreements, with 3,229 unique pages and 2,160 entities to extract. We provide several state-of-the-art baseline systems from the KIE domain (Flair, BERT, RoBERTa, LayoutLM, LAMBERT), which show that our datasets pose a strong challenge to existing models. The best model achieved an 81.77% and an 83.57% F1-score on respectively the Kleister NDA and the Kleister Charity datasets. We share the datasets to encourage progress on more in-depth and complex information extraction tasks. | ['Przemysław Biecek', 'Bartosz Topolski', 'Paulina Rosalska', 'Agnieszka Kaliska', 'Dawid Lipiński', 'Anna Wróblewska', 'Filip Graliński', 'Tomasz Stanisławek'] | 2021-05-12 | null | null | null | null | ['key-information-extraction'] | ['natural-language-processing'] | [-2.58140057e-01 4.96387929e-01 -4.15895879e-01 -2.48396188e-01
-1.19394505e+00 -1.10666525e+00 8.87229502e-01 4.37144816e-01
-6.22608781e-01 1.10065711e+00 3.50293070e-01 -5.67908645e-01
-4.65215534e-01 -8.81722689e-01 -7.61192799e-01 -7.32174665e-02
-3.19082588e-01 8.50059032e-01 -7.02202171e-02 8.15702975e-02
4.52446431e-01 6.73060060e-01 -1.01585007e+00 6.41453862e-01
6.32176638e-01 1.09483182e+00 3.83495614e-02 5.99539518e-01
-4.38336849e-01 8.24237108e-01 -6.75682247e-01 -9.49068308e-01
3.66074473e-01 2.90224224e-01 -1.30604053e+00 -4.19494629e-01
4.83889163e-01 -5.71317319e-03 -1.31675065e-01 6.82177722e-01
3.21248263e-01 -3.83515120e-01 9.50082481e-01 -1.25183535e+00
-6.89786434e-01 9.98248577e-01 -7.41188049e-01 4.26149279e-01
5.70387781e-01 -2.66754806e-01 1.73943353e+00 -1.29781842e+00
1.31445372e+00 1.00525224e+00 5.07405221e-01 5.01751453e-02
-6.84290290e-01 -7.79917240e-01 -7.93213118e-03 2.16722161e-01
-1.38806581e+00 -5.51174998e-01 2.47446448e-01 -4.27663654e-01
1.36811388e+00 2.86214083e-01 1.54775679e-01 7.94198275e-01
2.38444954e-01 9.24579084e-01 1.33061206e+00 -5.83638966e-01
-2.22296417e-01 2.77573794e-01 3.95617902e-01 6.77185178e-01
9.75486755e-01 -3.48967195e-01 -5.69774151e-01 -4.13016289e-01
5.03078878e-01 -4.46958214e-01 -9.07324776e-02 -1.31138265e-01
-1.40449929e+00 6.58733070e-01 6.82850406e-02 5.60563326e-01
-4.15619642e-01 -3.27381700e-01 2.63250649e-01 3.01385671e-01
3.06796283e-01 1.13836169e+00 -1.03920209e+00 -2.10488230e-01
-7.45047390e-01 6.81503773e-01 1.28887808e+00 1.39560270e+00
7.59554625e-01 -6.78605378e-01 -7.12383240e-02 5.47189355e-01
2.60014117e-01 3.57468456e-01 4.43233503e-03 -7.78075039e-01
1.41722405e+00 9.02133882e-01 4.04911339e-01 -1.32729125e+00
-7.16177583e-01 -1.61280036e-01 -7.14909732e-01 -4.38551337e-01
3.76997650e-01 -1.92541152e-01 -7.97940373e-01 1.03872848e+00
-7.76798204e-02 -5.22735476e-01 2.64025778e-01 4.86182064e-01
1.20219719e+00 6.92505538e-01 -1.18303210e-01 -2.38000691e-01
1.75216568e+00 -9.78054106e-01 -8.78020704e-01 -4.95418161e-01
7.45345652e-01 -9.88914728e-01 5.28413594e-01 4.50668544e-01
-1.32311237e+00 -1.57500416e-01 -9.94436622e-01 -4.32662398e-01
-1.07606936e+00 5.55277288e-01 9.08681333e-01 3.42581064e-01
-6.40143156e-01 3.27484518e-01 -2.09802732e-01 -3.21972579e-01
4.38269645e-01 1.65273577e-01 -8.70702446e-01 -1.62318602e-01
-1.26614892e+00 9.68273878e-01 6.89688921e-01 -1.16803152e-02
-1.69984758e-01 -8.61226916e-01 -8.61048281e-01 1.81618914e-01
8.62200499e-01 -5.26833832e-01 1.12326312e+00 5.73598109e-02
-7.36731708e-01 1.16946375e+00 4.92025204e-02 -4.87297893e-01
5.23000300e-01 -4.44721699e-01 -7.46802866e-01 8.79139379e-02
4.43821222e-01 4.45672065e-01 9.98094901e-02 -1.17401588e+00
-8.85623932e-01 -2.90192217e-01 -1.98830497e-02 -1.14093177e-01
-3.32785398e-01 4.73778754e-01 -6.29424691e-01 -6.52166545e-01
-2.60092199e-01 -7.91242957e-01 1.78536698e-01 -6.36885405e-01
-8.48377883e-01 -4.96865809e-01 4.30569023e-01 -1.06080055e+00
1.90469897e+00 -1.53708625e+00 -1.53224766e-01 3.50114614e-01
3.73484612e-01 1.52781904e-01 -8.60578120e-02 8.99633944e-01
-2.41480723e-01 7.10666716e-01 -1.13594323e-01 -1.42446324e-01
3.67493331e-01 9.77776945e-02 -4.59474146e-01 5.56220002e-02
3.73599261e-01 1.15249944e+00 -6.79712594e-01 -5.78094363e-01
-4.29627895e-01 -1.23203374e-01 -1.76621005e-01 4.05931734e-02
-1.06057376e-01 -3.32456052e-01 -6.27322495e-01 1.02718902e+00
7.39531815e-01 -3.67280811e-01 4.51343507e-01 -2.82794207e-01
-3.92528206e-01 7.55635619e-01 -1.29284036e+00 1.35577404e+00
-1.89276531e-01 7.26368368e-01 1.74474522e-01 -6.23511136e-01
8.75406682e-01 1.59786269e-01 6.24847829e-01 -6.89012289e-01
-1.31656617e-01 4.61871326e-01 -1.98691204e-01 -6.29943728e-01
9.90148723e-01 4.68439162e-01 -8.26478839e-01 5.30929327e-01
-1.38061106e-01 3.57086072e-03 8.28163266e-01 5.60261607e-01
1.59494436e+00 -2.22971573e-01 6.92093492e-01 -2.77341813e-01
2.88680017e-01 3.10473979e-01 7.50671029e-01 7.34625101e-01
1.45022899e-01 4.35801297e-01 7.70450771e-01 -6.13603711e-01
-1.11045778e+00 -7.80517220e-01 -2.39293694e-01 6.70109391e-01
-4.23479378e-01 -9.36819136e-01 -4.44284201e-01 -1.10429776e+00
3.39413613e-01 6.03508592e-01 -6.24395728e-01 5.53222299e-01
-7.84846008e-01 -5.75771272e-01 7.81453311e-01 5.68783343e-01
6.51857316e-01 -9.89361167e-01 -2.03451544e-01 3.12136889e-01
-3.83855879e-01 -1.58227003e+00 -1.98731199e-01 1.56887844e-01
-2.55654842e-01 -1.36903846e+00 -5.05838871e-01 -5.52404881e-01
4.85822946e-01 -2.93591648e-01 1.69680667e+00 -1.21025622e-01
-4.10262495e-01 9.27198902e-02 -2.57722080e-01 -5.51302075e-01
-1.69141039e-01 6.90636396e-01 -2.01952145e-01 -6.67023003e-01
5.85023880e-01 -5.13789011e-03 -6.43091947e-02 1.99055538e-01
-6.92307830e-01 2.88650412e-02 1.14750326e+00 6.99459553e-01
4.40988302e-01 2.64748514e-01 5.90122044e-01 -1.43759871e+00
7.51197338e-01 -5.86393356e-01 -6.19581282e-01 6.60989404e-01
-1.07446170e+00 5.22121973e-03 4.65322226e-01 2.36012846e-01
-1.14762247e+00 -2.35217869e-01 6.75067157e-02 1.97770983e-01
-1.16029263e-01 1.21107340e+00 -2.60398090e-01 4.33466345e-01
3.91767025e-01 -4.00177717e-01 -8.14701378e-01 -7.45434344e-01
2.99235523e-01 8.56576681e-01 7.80671895e-01 -8.15029919e-01
1.09647739e+00 1.26030773e-01 -6.58343658e-02 -5.09313762e-01
-9.11879420e-01 -8.00498128e-01 -7.94311166e-01 2.91564375e-01
5.47280550e-01 -8.42367709e-01 -5.25084198e-01 3.14226270e-01
-1.32851052e+00 -6.54262584e-03 -1.55895548e-02 1.27952427e-01
-1.66980609e-01 2.51959652e-01 -6.69930160e-01 -6.42543614e-01
-5.02587080e-01 -6.10100985e-01 1.02273202e+00 7.01852813e-02
-3.98326278e-01 -8.46370339e-01 -2.78359018e-02 7.12559044e-01
-1.92111935e-02 3.55670333e-01 1.24477005e+00 -9.25568521e-01
-7.63713360e-01 -3.26693088e-01 -6.55518234e-01 -4.68885675e-02
4.91240956e-02 2.06398875e-01 -5.60755968e-01 -1.24466002e-01
-7.60214627e-01 -2.94900447e-01 8.84222150e-01 -2.25355402e-01
9.30447400e-01 -6.31331563e-01 -6.45946920e-01 2.07539797e-01
1.33953536e+00 1.79759920e-01 8.13776672e-01 8.55634570e-01
5.00036240e-01 8.83536518e-01 1.04190981e+00 5.30114591e-01
7.77671933e-01 4.41139847e-01 -1.98704712e-02 2.25565918e-02
9.14278254e-02 -4.05117393e-01 9.92565230e-02 8.40469837e-01
-3.14306051e-01 -3.30980062e-01 -1.50892055e+00 8.12089682e-01
-1.78362036e+00 -9.41321790e-01 -3.86686027e-01 1.63947117e+00
1.13642108e+00 2.35729069e-01 -1.33112207e-01 -2.04187632e-02
2.77783871e-01 5.21262921e-02 -6.32678196e-02 -1.78041339e-01
-4.07581955e-01 2.45687783e-01 8.41073275e-01 1.38290852e-01
-1.29453099e+00 9.49801326e-01 6.03178406e+00 8.05372477e-01
-3.60203594e-01 -3.41437459e-01 6.36553586e-01 1.60190657e-01
-4.19919521e-01 1.56615585e-01 -1.51413560e+00 2.64087141e-01
1.17985988e+00 -5.43870628e-01 -2.51089055e-02 6.40255034e-01
-1.77039310e-01 -1.78326815e-01 -1.14553022e+00 7.53602982e-01
-9.10626799e-02 -1.81637931e+00 -1.68222442e-01 4.92462039e-01
9.08832967e-01 -1.12087816e-01 -9.41611305e-02 4.71573621e-01
6.46421134e-01 -1.22659492e+00 5.90547800e-01 7.15215027e-01
7.40158498e-01 -8.39537263e-01 9.57085490e-01 4.26565647e-01
-1.28224885e+00 -9.77944508e-02 -3.11983198e-01 2.04668179e-01
2.87615191e-02 8.47472131e-01 -1.05345595e+00 1.14028156e+00
1.07297790e+00 5.73401332e-01 -8.30239534e-01 6.51640415e-01
-3.87637049e-01 2.22698823e-01 -2.95909286e-01 -1.23031475e-01
3.24152678e-01 -5.04808389e-02 1.67979404e-01 1.48811412e+00
3.22023630e-01 1.54774517e-01 -4.28991951e-02 7.35163152e-01
-8.55972111e-01 2.62278676e-01 -6.29896283e-01 -6.06730819e-01
7.28666902e-01 1.69086230e+00 -5.50397038e-01 -4.15556461e-01
-5.51896274e-01 3.91212046e-01 5.83855331e-01 2.23094359e-01
-4.44462568e-01 -7.42053628e-01 2.92511940e-01 2.85201550e-01
2.98511147e-01 -3.92925262e-01 -3.47995698e-01 -1.16514683e+00
2.82372236e-01 -9.42552924e-01 8.13105881e-01 -7.48732805e-01
-1.40859914e+00 4.24515337e-01 2.90488273e-01 -6.61207616e-01
-6.16256297e-01 -8.94893289e-01 -1.32087067e-01 8.96696746e-01
-1.98468268e+00 -1.35870564e+00 -7.35516250e-02 1.84319511e-01
2.44410485e-01 -3.60287756e-01 7.51444638e-01 4.54291999e-01
-5.32943547e-01 4.04282957e-01 8.57934207e-02 8.09789777e-01
1.05214036e+00 -1.57762194e+00 7.37855434e-01 5.97350955e-01
3.24480295e-01 1.05527139e+00 2.14078605e-01 -7.63736606e-01
-1.51940727e+00 -9.20318782e-01 1.82010615e+00 -8.96724701e-01
9.95639622e-01 -5.31855106e-01 -7.74978578e-01 9.66407776e-01
5.33790588e-01 -2.34621301e-01 5.61088681e-01 3.81007373e-01
-4.01324034e-01 1.21623948e-01 -1.05283666e+00 3.11402023e-01
1.07245469e+00 -5.71470737e-01 -7.10970938e-01 4.20544446e-01
3.72664511e-01 -4.58775520e-01 -1.49148250e+00 4.33842361e-01
4.58020747e-01 -6.43091142e-01 8.39840829e-01 -5.65941453e-01
6.16334796e-01 -9.29244235e-02 2.06223801e-01 -9.00412381e-01
-3.84746730e-01 -6.14986420e-01 -3.43737870e-01 1.73505807e+00
1.04204786e+00 -6.44397974e-01 6.00642681e-01 9.60874677e-01
6.87347874e-02 -8.72098505e-01 -7.33800590e-01 -8.00237715e-01
2.59028792e-01 -2.64777422e-01 6.79972172e-01 1.32396078e+00
7.43593201e-02 5.88697672e-01 -5.66918626e-02 -6.52648602e-03
2.89167732e-01 4.55067098e-01 9.56618726e-01 -1.44674850e+00
5.85448407e-02 -3.15613180e-01 3.97063605e-02 -8.22957695e-01
4.42562014e-01 -9.14182603e-01 -4.14497733e-01 -1.93957067e+00
4.07951444e-01 -6.85077786e-01 -8.49956796e-02 9.15665090e-01
1.54446051e-01 -3.59599710e-01 7.28775933e-02 4.18319464e-01
-5.31129122e-01 -1.06593519e-01 9.63928580e-01 -2.34006301e-01
2.15715975e-01 -2.47382328e-01 -1.01802397e+00 7.01846957e-01
5.24893820e-01 -3.64395946e-01 6.51879162e-02 -2.98477978e-01
8.15784395e-01 -1.19114958e-01 2.90206708e-02 -4.17401880e-01
6.60195410e-01 -2.75819898e-01 5.06883562e-01 -1.04173684e+00
5.08848839e-02 -6.83188498e-01 -4.74051423e-02 -8.31462443e-02
-3.35320652e-01 5.51613510e-01 1.81161880e-01 2.91138023e-01
-4.07392651e-01 -5.24954498e-01 6.74060360e-02 -2.29274884e-01
-8.63209069e-01 2.19116375e-01 -6.78814948e-02 4.11838800e-01
9.06545103e-01 1.14439808e-01 -9.46192443e-01 -7.45171160e-02
-2.61942595e-01 3.23738992e-01 2.92275190e-01 5.18441200e-01
4.72232223e-01 -1.12390792e+00 -8.68027151e-01 3.52868848e-02
2.78060645e-01 1.71930537e-01 -2.34090462e-01 6.22615516e-01
-4.25609887e-01 1.05734444e+00 -4.45424244e-02 1.74325347e-01
-1.17850792e+00 4.10154492e-01 -3.86974961e-01 -1.16078520e+00
-2.27843523e-01 6.28147483e-01 -3.18657190e-01 -4.40719694e-01
-1.63642764e-01 -4.29416239e-01 -3.32225293e-01 5.19381046e-01
4.35606033e-01 4.30233777e-01 4.55843627e-01 -4.21913475e-01
-4.84260082e-01 8.32957178e-02 -5.56188285e-01 -1.26604974e-01
1.69087231e+00 4.23417352e-02 -4.20736611e-01 7.53021613e-02
1.03516126e+00 5.71991563e-01 -4.37700331e-01 -2.59144783e-01
8.73291612e-01 -3.60513091e-01 -5.65651692e-02 -1.27490520e+00
-8.69657159e-01 4.75866824e-01 -1.46642968e-01 5.37022412e-01
7.38930166e-01 2.34440908e-01 6.04856670e-01 7.57912815e-01
5.91321886e-01 -1.48622525e+00 -1.93533257e-01 7.33301759e-01
1.15045667e+00 -1.10533082e+00 5.11465669e-01 -7.37534165e-01
-5.30909717e-01 1.24857998e+00 4.17476952e-01 3.25514853e-01
4.70111102e-01 7.64036894e-01 -1.48729563e-01 -4.98942107e-01
-1.05655253e+00 -2.11660564e-02 4.86342698e-01 3.07778984e-01
5.41274011e-01 2.46024132e-03 -4.05214369e-01 1.14432704e+00
-4.56929207e-01 -1.32747665e-01 5.00233054e-01 1.21616518e+00
-1.91369355e-01 -1.39576161e+00 -3.48497182e-01 7.07645774e-01
-7.50581622e-01 -3.66058320e-01 -8.68549347e-01 1.38911676e+00
-2.32233927e-01 7.91949987e-01 -2.53743321e-01 3.56956273e-02
6.51858330e-01 1.78750813e-01 2.00103506e-01 -6.87280655e-01
-7.16159284e-01 -7.61267990e-02 1.01310670e+00 -3.07338804e-01
-3.15965205e-01 -8.56919348e-01 -1.30572057e+00 -5.26403725e-01
-2.09184349e-01 1.83862939e-01 6.09614968e-01 7.05027580e-01
6.01298153e-01 2.67834127e-01 2.09562019e-01 -7.08529055e-02
-4.12239760e-01 -1.06063628e+00 -7.11271703e-01 2.95760542e-01
-1.00715354e-01 -5.68144739e-01 -1.61427617e-01 1.02797143e-01] | [9.415894508361816, 8.620926856994629] |
23a0bd31-6772-49d8-88ab-093f9dbe5613 | towards-automatic-pulmonary-nodule-management | 1610.09157 | null | http://arxiv.org/abs/1610.09157v2 | http://arxiv.org/pdf/1610.09157v2.pdf | Towards automatic pulmonary nodule management in lung cancer screening with deep learning | The introduction of lung cancer screening programs will produce an
unprecedented amount of chest CT scans in the near future, which radiologists
will have to read in order to decide on a patient follow-up strategy. According
to the current guidelines, the workup of screen-detected nodules strongly
relies on nodule size and nodule type. In this paper, we present a deep
learning system based on multi-stream multi-scale convolutional networks, which
automatically classifies all nodule types relevant for nodule workup. The
system processes raw CT data containing a nodule without the need for any
additional information such as nodule segmentation or nodule size and learns a
representation of 3D data by analyzing an arbitrary number of 2D views of a
given nodule. The deep learning system was trained with data from the Italian
MILD screening trial and validated on an independent set of data from the
Danish DLCST screening trial. We analyze the advantage of processing nodules at
multiple scales with a multi-stream convolutional network architecture, and we
show that the proposed deep learning system achieves performance at classifying
nodule type that surpasses the one of classical machine learning approaches and
is within the inter-observer variability among four experienced human
observers. | ['Alfonso Marchiano', 'Mathilde M. W. Wille', 'Cornelia Schaefer-Prokop', 'Colin Jacobs', 'Paul K. Gerke', 'Arnaud Arindra Adiyoso Setio', 'Mathias Prokop', 'Bram van Ginneken', 'Sarah J. van Riel', 'Kaman Chung', 'Ugo Pastorino', 'Francesco Ciompi', 'Ernst Th. Scholten'] | 2016-10-28 | null | null | null | null | ['pulmonary-nodules-classification', 'lung-nodule-detection', 'lung-nodule-classification'] | ['medical', 'medical', 'medical'] | [ 3.77164215e-01 3.61716777e-01 -1.73163712e-01 -4.77794915e-01
-8.36340904e-01 -5.55701911e-01 3.43652397e-01 5.40533841e-01
-6.25721633e-01 -5.31537645e-02 7.47334212e-02 -8.10521066e-01
-1.65356010e-01 -1.04635847e+00 -5.30835450e-01 -4.67991650e-01
-5.82306609e-02 9.70409751e-01 7.82541454e-01 1.26121391e-03
-5.23585737e-01 5.82309425e-01 -1.14737713e+00 8.56192172e-01
3.74548733e-01 1.27193999e+00 4.63584960e-01 1.23438096e+00
1.04391947e-01 7.68591881e-01 -1.22860931e-01 -2.77549624e-01
5.61838269e-01 -5.75943530e-01 -9.59182739e-01 3.34393680e-01
4.79727805e-01 -6.84641540e-01 -2.87206709e-01 7.75617898e-01
4.82034475e-01 -6.49411082e-01 6.36649489e-01 -3.02510828e-01
-3.21736634e-01 4.89473850e-01 -4.78400923e-02 5.42225838e-01
-3.65786165e-01 1.67807788e-01 8.31011832e-01 -5.96081376e-01
4.26728874e-01 6.41849101e-01 9.46014285e-01 3.32189411e-01
-7.75096774e-01 4.10740674e-02 -4.60835040e-01 -1.08116299e-01
-1.01040518e+00 1.89751580e-01 -1.15168080e-01 -5.67920268e-01
4.00025487e-01 5.60618877e-01 1.03786337e+00 7.67655909e-01
4.95963424e-01 3.99605274e-01 9.04170752e-01 -3.26835930e-01
1.28617600e-01 1.17697679e-01 -2.42246419e-01 1.14372349e+00
6.88563049e-01 2.65312437e-02 3.64753067e-01 -2.04447031e-01
1.20146501e+00 2.59179741e-01 1.35098577e-01 -6.22708976e-01
-1.45208776e+00 1.00942564e+00 1.03145957e+00 7.47354031e-01
-5.41256428e-01 2.06043512e-01 5.37729383e-01 4.96193692e-02
2.15867713e-01 3.99063110e-01 -2.51339048e-01 4.03369099e-01
-9.48244035e-01 -2.13648632e-01 6.54220164e-01 2.95447946e-01
1.44879237e-01 -4.50519919e-01 -5.63060522e-01 6.09286427e-01
2.35582709e-01 5.45965195e-01 1.03779042e+00 -6.63954020e-01
-8.18772912e-02 9.81010675e-01 -1.03590518e-01 -3.54127079e-01
-9.62858021e-01 -5.40619791e-01 -1.12571347e+00 1.13778934e-01
5.90790570e-01 -2.24836111e-01 -1.30418217e+00 9.01061237e-01
1.67492822e-01 -2.94642091e-01 -2.60683179e-01 9.31717873e-01
1.12856197e+00 -3.23418714e-02 -6.46829382e-02 2.41356775e-01
1.79581392e+00 -8.21207285e-01 -1.51319489e-01 -1.88767135e-01
9.15782571e-01 -5.88001192e-01 6.73199952e-01 8.49099904e-02
-1.02302051e+00 -6.43722296e-01 -8.87445688e-01 7.66941383e-02
-1.68514073e-01 9.57873285e-01 5.96778274e-01 9.50567842e-01
-1.15676987e+00 4.54950899e-01 -1.37848032e+00 -7.89427221e-01
5.98684728e-01 6.82182848e-01 -3.16472560e-01 6.31979406e-02
-9.53172088e-01 9.94413257e-01 4.78369415e-01 3.31040144e-01
-1.04973507e+00 -7.03380108e-01 -4.86206323e-01 2.84191787e-01
5.73392689e-01 -8.71429682e-01 1.87046754e+00 -1.05520046e+00
-1.06427193e+00 9.96558011e-01 2.15363607e-01 -8.04557085e-01
8.59591484e-01 2.67868191e-01 -2.51576483e-01 4.45317984e-01
9.46636370e-04 4.44712281e-01 7.03428090e-01 -4.99289185e-01
-9.08467770e-01 -2.94733435e-01 4.17764531e-03 1.24937497e-01
-2.09591821e-01 -1.45784348e-01 -5.18956423e-01 -2.04555020e-01
1.56472266e-01 -1.23552072e+00 -8.57964635e-01 2.29443088e-01
-3.14473927e-01 9.02803689e-02 5.99051833e-01 -4.53738928e-01
7.79489219e-01 -1.76909649e+00 -1.51224598e-01 2.79780149e-01
7.08186269e-01 3.91168982e-01 1.46076471e-01 -1.34988591e-01
-2.15367883e-01 3.25534225e-01 -4.44590971e-02 4.47054744e-01
-4.60340828e-01 3.65685746e-02 5.69239020e-01 4.43560570e-01
4.40024287e-01 1.22482896e+00 -7.84675837e-01 -6.65332496e-01
5.17000139e-01 1.97521403e-01 -3.26142728e-01 1.37675837e-01
5.27032092e-02 2.34304890e-01 -6.71343148e-01 3.47836882e-01
3.27532262e-01 -1.04437113e+00 3.34607065e-01 -2.64777035e-01
2.66235955e-02 -8.41929298e-03 -8.59913886e-01 1.36236286e+00
-5.09914219e-01 4.12170053e-01 -2.78255362e-02 -6.59334898e-01
4.65509683e-01 7.23728001e-01 7.57947505e-01 -3.54121864e-01
2.23144397e-01 6.65289938e-01 8.45662951e-01 -6.04597151e-01
-4.11891453e-02 -1.82902768e-01 1.12800337e-01 5.03351152e-01
1.23951055e-01 -2.62529254e-01 3.38551909e-01 -4.74546663e-02
1.71453691e+00 -6.07684135e-01 7.32348323e-01 -3.51494730e-01
5.97945392e-01 2.74943680e-01 -1.24536447e-01 9.21636283e-01
-2.33613402e-01 8.26043248e-01 7.41460145e-01 -8.65956783e-01
-1.21893311e+00 -1.07149363e+00 -4.37747508e-01 8.57670724e-01
-3.10507357e-01 3.49158272e-02 -5.51447988e-01 -1.04326773e+00
-8.17445740e-02 -3.84116694e-02 -1.31430447e+00 -2.51135882e-02
-6.50365591e-01 -9.01905000e-01 4.54700530e-01 7.05074847e-01
4.73601043e-01 -1.01063013e+00 -1.15783823e+00 1.58573747e-01
1.51405297e-02 -8.33645523e-01 -8.11453387e-02 6.02830946e-01
-9.99339283e-01 -1.52735341e+00 -1.25519907e+00 -7.70557702e-01
7.90087581e-01 1.09494239e-01 1.22054756e+00 3.21957648e-01
-6.87552094e-01 4.60510328e-02 -2.63350457e-01 -4.81672764e-01
-1.02778411e+00 6.74203575e-01 -5.44179499e-01 -1.69223621e-01
1.84424832e-01 1.95087507e-01 -7.78815091e-01 2.27635682e-01
-9.39489245e-01 1.21948048e-01 1.50719631e+00 7.42093861e-01
6.90048158e-01 -7.62574226e-02 1.67497948e-01 -1.34605837e+00
2.78908938e-01 -5.19257069e-01 -4.24982190e-01 3.79278123e-01
-9.74373147e-02 1.56416669e-01 4.21705633e-01 -7.39607885e-02
-9.06928062e-01 4.49795842e-01 -2.61719316e-01 -1.14660010e-01
-2.51844376e-01 4.98387188e-01 6.48545861e-01 -1.77333698e-01
1.02693355e+00 -2.62438450e-02 1.69371329e-02 -1.68760464e-01
3.73337455e-02 5.48861504e-01 3.42410684e-01 3.47562611e-01
7.69226789e-01 5.91110468e-01 4.64039922e-01 -6.51279032e-01
-1.00885248e+00 -8.32643926e-01 -1.17953873e+00 -1.40226752e-01
1.10312212e+00 -8.07164073e-01 -2.10558131e-01 7.96701983e-02
-7.61756837e-01 -2.80210376e-01 -6.50648892e-01 7.04625726e-01
-3.67892116e-01 -7.34354332e-02 -7.14775980e-01 -1.06653981e-01
-3.87477040e-01 -1.22351289e+00 1.21242976e+00 7.87251294e-02
-1.00277811e-01 -8.67623031e-01 4.70103957e-02 -3.58099043e-02
7.48930991e-01 1.84831962e-01 8.34840059e-01 -1.22645354e+00
-5.67877591e-01 -7.55482197e-01 -4.28645134e-01 1.34313717e-01
4.32352841e-01 -8.55778754e-02 -6.66943669e-01 -4.58100326e-02
1.06363587e-01 -3.75316501e-01 9.59101200e-01 7.95955598e-01
1.37070453e+00 2.31595457e-01 -4.09066558e-01 4.57397550e-01
1.39891684e+00 -9.72424299e-02 1.32734939e-01 2.76484102e-01
5.98891020e-01 1.00693502e-01 5.89017719e-02 1.28926069e-01
-9.71747562e-02 1.17807277e-01 7.25825012e-01 -5.38786054e-01
-2.80828357e-01 2.03400388e-01 -4.59212661e-01 2.97678709e-01
-4.31065023e-01 -7.99792558e-02 -1.19882441e+00 5.10411561e-01
-1.41415322e+00 -6.51329219e-01 -4.19519514e-01 2.02603769e+00
4.47614282e-01 2.53989607e-01 1.93811730e-01 1.20612187e-02
6.25623345e-01 -1.59412414e-01 -4.87210900e-01 -2.26307228e-01
3.23914945e-01 3.82834375e-01 9.66551960e-01 -1.45262526e-03
-1.44715250e+00 1.89144924e-01 6.74150848e+00 3.88627529e-01
-1.34632051e+00 3.07665706e-01 8.77152264e-01 5.85492700e-02
1.64013028e-01 -6.76428616e-01 -3.32273096e-01 -9.92780998e-02
1.20115101e+00 -1.37391075e-01 -2.54485577e-01 9.13770497e-01
1.03250034e-01 -2.82309502e-01 -1.26775444e+00 3.09598446e-01
-1.62007779e-01 -1.50925183e+00 -1.17423929e-01 1.90212876e-01
5.75346351e-01 6.63956165e-01 9.79965702e-02 1.40713751e-01
2.20963538e-01 -1.21855021e+00 1.87813029e-01 1.74917266e-01
9.06890988e-01 -1.46358475e-01 1.35106826e+00 2.81409562e-01
-1.28488731e+00 -2.11456105e-01 -2.01237991e-01 5.97476900e-01
-2.60125011e-01 3.71966511e-01 -1.71296132e+00 5.50379932e-01
6.68352127e-01 4.27363008e-01 -1.38511145e+00 1.25931633e+00
4.11732011e-02 6.99532926e-01 -4.31536019e-01 -1.35029078e-01
4.82113034e-01 4.14234966e-01 -4.24864478e-02 1.22213995e+00
7.70409346e-01 2.36427542e-02 -2.87610237e-02 5.42123020e-01
-7.21578971e-02 1.40900418e-01 -3.51260215e-01 -1.61652923e-01
-6.14065230e-02 1.63217556e+00 -1.28709042e+00 -5.63419819e-01
-6.71751797e-01 7.21491337e-01 -1.45896047e-01 -2.78705865e-01
-7.34860301e-01 3.93283963e-02 -3.35383356e-01 6.03172123e-01
6.22594953e-01 2.79695690e-01 -2.25519031e-01 -6.91933393e-01
-1.27239943e-01 -5.78232348e-01 5.74407578e-01 -4.51352626e-01
-1.04438818e+00 9.02528286e-01 -2.19083026e-01 -1.37961113e+00
-4.76896137e-01 -1.07468867e+00 -7.34230101e-01 7.78168976e-01
-1.11678898e+00 -1.08582759e+00 -6.72949255e-01 2.88967520e-01
4.26930517e-01 -8.69899839e-02 8.63181472e-01 1.14946544e-01
-1.17865600e-01 8.59691501e-02 7.12157860e-02 5.78404963e-01
5.73727310e-01 -1.52691901e+00 3.24097335e-01 4.47829992e-01
-3.49945426e-02 -4.54157218e-02 1.43945828e-01 -5.34049153e-01
-1.07317007e+00 -1.38944972e+00 7.44605064e-01 -5.10289788e-01
6.40326202e-01 1.74237609e-01 -7.14874864e-01 7.48109937e-01
-1.13343216e-01 7.80252695e-01 7.27390647e-01 -1.96910396e-01
1.73891008e-01 4.48861942e-02 -1.05394375e+00 1.23154581e-01
4.49577063e-01 -1.62625954e-01 -4.51521456e-01 6.25827670e-01
2.68621594e-01 -7.24667132e-01 -9.88219142e-01 6.60219789e-01
4.20277357e-01 -1.20789623e+00 9.24383163e-01 -4.63077694e-01
5.44330180e-01 1.13318726e-01 2.39073485e-01 -1.06278157e+00
-6.12869382e-01 2.74430335e-01 2.71059811e-01 -2.45090015e-02
6.54714167e-01 -1.54854760e-01 1.17048740e+00 2.60164499e-01
-1.48084447e-01 -8.42054844e-01 -9.27603841e-01 -3.66543800e-01
1.26623750e-01 -3.98747846e-02 2.76244700e-01 2.67857581e-01
-5.50133705e-01 -9.16690528e-02 3.47215325e-01 1.50879323e-01
2.54434496e-01 1.15829676e-01 2.63579041e-01 -1.41804695e+00
-4.98269111e-01 -6.16112053e-01 -4.71739620e-01 -3.97044897e-01
-8.15049052e-01 -1.17166853e+00 -1.59697197e-02 -1.93084037e+00
5.51040769e-01 -1.09458856e-01 -4.08748180e-01 3.56453151e-01
-1.79494888e-01 4.43012208e-01 3.76190692e-02 2.74840385e-01
-5.72713554e-01 -2.21212104e-01 1.62486279e+00 -2.83065625e-02
4.40066420e-02 6.46934807e-01 -4.42340821e-01 9.55684543e-01
6.14090860e-01 -5.01274288e-01 -1.42421335e-01 -4.22878951e-01
7.10475594e-02 2.81487614e-01 5.72531044e-01 -1.21366823e+00
8.10985491e-02 1.74811617e-01 1.01905930e+00 -5.99217534e-01
9.49686617e-02 -9.73215044e-01 9.42179114e-02 1.37710440e+00
-4.74440038e-01 -2.86942452e-01 8.80803689e-02 5.77247798e-01
-1.57503799e-01 -4.60170835e-01 8.96462440e-01 -6.42765284e-01
-3.48906934e-01 5.49409926e-01 -7.43513167e-01 -3.03102642e-01
1.05783367e+00 -1.21097282e-01 9.92418751e-02 2.43135057e-02
-1.07691729e+00 -1.17009297e-01 1.04746610e-01 1.02203563e-01
2.27159619e-01 -1.43789673e+00 -9.97198582e-01 2.65647382e-01
1.41503289e-01 3.18119735e-01 1.88439831e-01 1.14069116e+00
-1.05787289e+00 7.97707081e-01 -3.08756828e-01 -1.01943135e+00
-1.17990637e+00 2.16765270e-01 9.30593789e-01 -8.61433268e-01
-5.28139293e-01 8.20146978e-01 2.19635695e-01 -4.28243190e-01
-2.04324603e-01 -9.97519970e-01 -1.67488948e-01 -8.73551797e-03
3.58599722e-01 6.62053153e-02 7.35415876e-01 -2.45401949e-01
-2.46554062e-01 1.82673946e-01 -1.77398562e-01 1.82099625e-01
1.14153659e+00 3.26235294e-01 1.76581219e-01 4.41182792e-01
9.42799747e-01 -4.04517919e-01 -8.91683578e-01 -2.63813853e-01
1.32492542e-01 -8.33922476e-02 1.32910937e-01 -8.57459068e-01
-1.11698592e+00 8.32449675e-01 1.09022892e+00 6.66763306e-01
9.80877995e-01 3.17736268e-01 4.49244380e-01 7.54570723e-01
-1.67560771e-01 -5.57570338e-01 2.27629542e-01 3.26750994e-01
7.60808647e-01 -1.81602204e+00 5.28549850e-02 -3.32505405e-01
-6.13331974e-01 1.59050488e+00 5.38618147e-01 -2.94051170e-01
7.73920536e-01 2.71648347e-01 1.65915772e-01 -4.23753798e-01
-6.73635781e-01 -3.92177731e-01 6.45985007e-01 2.87878335e-01
8.14569712e-01 3.26487094e-01 -6.78668544e-02 5.72216570e-01
4.57994640e-02 2.72298306e-01 5.87376654e-01 8.32566023e-01
-8.68028402e-01 -9.30936098e-01 -4.47219968e-01 1.27409065e+00
-6.95426881e-01 1.72819197e-01 -4.35616463e-01 1.14069283e+00
2.70882308e-01 3.35128814e-01 2.74825662e-01 1.10439464e-01
4.31082249e-01 -9.38655362e-02 4.42972064e-01 -1.15402174e+00
-9.54836309e-01 2.27424234e-01 7.18245283e-02 -3.73044908e-01
-3.49065006e-01 -6.23377085e-01 -1.16392577e+00 2.69034863e-01
-2.29844257e-01 -7.84621462e-02 4.92815524e-01 8.46372664e-01
-1.26030594e-01 1.06653917e+00 4.74816024e-01 -8.06198955e-01
-7.02129424e-01 -1.15732515e+00 -4.59601164e-01 1.46019667e-01
3.59116852e-01 -1.27616793e-01 -3.46807018e-02 1.26508772e-01] | [15.37608814239502, -2.154320478439331] |
55c4784d-5aab-429a-9bf4-c3f439b0e3fe | filtered-noise-shaping-for-time-domain-room | 2107.07503 | null | https://arxiv.org/abs/2107.07503v1 | https://arxiv.org/pdf/2107.07503v1.pdf | Filtered Noise Shaping for Time Domain Room Impulse Response Estimation From Reverberant Speech | Deep learning approaches have emerged that aim to transform an audio signal so that it sounds as if it was recorded in the same room as a reference recording, with applications both in audio post-production and augmented reality. In this work, we propose FiNS, a Filtered Noise Shaping network that directly estimates the time domain room impulse response (RIR) from reverberant speech. Our domain-inspired architecture features a time domain encoder and a filtered noise shaping decoder that models the RIR as a summation of decaying filtered noise signals, along with direct sound and early reflection components. Previous methods for acoustic matching utilize either large models to transform audio to match the target room or predict parameters for algorithmic reverberators. Instead, blind estimation of the RIR enables efficient and realistic transformation with a single convolution. An evaluation demonstrates our model not only synthesizes RIRs that match parameters of the target room, such as the $T_{60}$ and DRR, but also more accurately reproduces perceptual characteristics of the target room, as shown in a listening test when compared to deep learning baselines. | ['Paul Calamia', 'Vamsi Krishna Ithapu', 'Christian J. Steinmetz'] | 2021-07-15 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 1.87082574e-01 -3.74817342e-01 9.95278537e-01 -1.88376874e-01
-1.45878518e+00 -6.40022993e-01 2.47969091e-01 -2.81569332e-01
-1.38326466e-01 3.61063212e-01 9.78727877e-01 -2.90296286e-01
7.57739693e-02 -4.65749234e-01 -7.95915127e-01 -5.90576172e-01
-6.53082654e-02 -9.59089622e-02 -1.37707040e-01 -3.39603364e-01
-4.08103794e-01 2.14991391e-01 -1.54081059e+00 5.46343923e-01
5.07049859e-01 1.37470758e+00 3.55624706e-01 1.23145819e+00
3.51420105e-01 7.81124651e-01 -1.08154225e+00 -6.56490549e-02
5.04081488e-01 -5.13812125e-01 -2.47378305e-01 -4.50293720e-01
7.05296457e-01 -6.35291100e-01 -8.41967583e-01 8.33216429e-01
1.16851830e+00 6.41940594e-01 6.35433137e-01 -5.77338278e-01
-5.99712193e-01 7.87239552e-01 1.54041931e-01 -6.46831393e-02
7.92827606e-01 1.59010038e-01 8.24667633e-01 -8.75434697e-01
-6.00010194e-02 1.22768819e+00 1.33159602e+00 4.35712427e-01
-1.33870709e+00 -9.50058281e-01 -2.95012861e-01 5.10034487e-02
-1.32902217e+00 -9.02462482e-01 8.80952179e-01 -1.89328447e-01
9.39383984e-01 4.14571494e-01 4.87309903e-01 1.54763222e+00
1.10289976e-01 2.39158779e-01 8.23695362e-01 -6.18302643e-01
2.01341853e-01 -2.88092315e-01 -3.29663932e-01 1.32324100e-01
-5.87109566e-01 6.42045438e-01 -6.29291415e-01 -1.50639758e-01
7.21893787e-01 -4.60929483e-01 -8.61523688e-01 2.54659474e-01
-1.22451222e+00 2.29172662e-01 5.73550761e-01 3.69851261e-01
-3.73968273e-01 5.77605128e-01 2.17208475e-01 3.07727337e-01
2.38314226e-01 6.23056710e-01 -4.20831949e-01 -1.11394756e-01
-9.86034751e-01 1.48648620e-01 9.62260902e-01 8.35125864e-01
2.61264950e-01 7.20904410e-01 -3.98607016e-01 1.08906245e+00
3.90154183e-01 8.37051690e-01 4.21177059e-01 -1.21446621e+00
2.70359725e-01 -5.36521494e-01 5.19371390e-01 -7.22076356e-01
-4.83633846e-01 -7.97173440e-01 -8.62952232e-01 1.76554561e-01
4.74091560e-01 -3.44958365e-01 -9.99517381e-01 1.94284654e+00
2.14922312e-03 6.84671164e-01 1.76971614e-01 9.53371644e-01
9.20400262e-01 9.76545155e-01 -4.43542838e-01 1.22065507e-01
1.03658307e+00 -8.85675073e-01 -9.89509761e-01 -3.53077412e-01
-4.71275225e-02 -1.34717977e+00 1.25038898e+00 6.95413649e-01
-1.24557304e+00 -1.16751826e+00 -1.06802952e+00 -5.05287349e-02
1.27468571e-01 2.15899162e-02 -1.16569111e-02 8.68183076e-01
-1.42267716e+00 5.81115425e-01 -6.28173351e-01 1.69306636e-01
-2.33398452e-01 3.71966287e-02 3.23993736e-03 1.61710922e-02
-1.27116168e+00 5.86487949e-01 -3.38842481e-01 3.71066660e-01
-1.41482997e+00 -1.24388194e+00 -8.45191360e-01 2.27943033e-01
-1.67924911e-01 -7.36581564e-01 1.92065310e+00 -9.35790718e-01
-1.82219040e+00 1.76950842e-01 -1.08025745e-01 -5.73528230e-01
3.26270223e-01 -5.18487573e-01 -1.14142263e+00 -1.16369210e-01
-2.21283272e-01 8.12517777e-02 1.26770270e+00 -1.45056987e+00
-2.70817697e-01 2.45331466e-01 -2.44977742e-01 2.02156872e-01
-8.73831287e-02 -5.03100082e-02 -5.08403704e-02 -1.01001179e+00
2.89555788e-01 -5.21617651e-01 -3.18607390e-01 -1.55018374e-01
-4.49136585e-01 6.03900433e-01 1.98225066e-01 -1.22312927e+00
1.16860020e+00 -2.48610377e+00 -4.53827113e-01 3.51429194e-01
1.11615621e-01 3.21044028e-01 -4.41617936e-01 5.10069072e-01
-4.00515229e-01 -4.59245861e-01 8.85257050e-02 -4.51663613e-01
2.69667983e-01 -5.16022682e-01 -8.33861768e-01 3.13054174e-01
-2.19238177e-01 2.69165784e-01 -8.46574545e-01 1.85832113e-01
1.68102607e-01 1.01130819e+00 -7.85795331e-01 6.55500948e-01
8.70129317e-02 7.41627753e-01 2.57456541e-01 9.81272087e-02
6.48980975e-01 4.11293030e-01 -1.07011363e-01 -7.14807153e-01
-8.02219752e-03 6.79997504e-01 -1.23659754e+00 1.75365412e+00
-1.12519944e+00 7.63283193e-01 7.59190619e-01 -2.74186641e-01
1.33178544e+00 6.78069711e-01 1.89677551e-01 -9.73887146e-01
2.14520916e-01 6.08532965e-01 1.02700561e-01 -2.59273052e-01
4.20680404e-01 -3.19617450e-01 2.25290015e-01 4.68524516e-01
8.56049731e-02 -6.27846301e-01 -6.49635792e-01 -3.48930418e-01
1.42457914e+00 2.91610900e-02 -1.19536780e-01 1.58546552e-01
4.96671498e-01 -8.17364156e-01 1.16712637e-01 9.24477398e-01
-1.98880471e-02 1.16118348e+00 -4.05883431e-01 -1.95794195e-01
-9.75736260e-01 -1.66276371e+00 1.62331790e-01 1.15682244e+00
-2.38925174e-01 -1.45365164e-01 -1.00637329e+00 2.36853957e-01
-3.59949589e-01 1.15151501e+00 -2.43730187e-01 -6.11109972e-01
-8.19273889e-01 1.38678953e-01 8.53752196e-01 5.49023688e-01
2.09065348e-01 -9.27178800e-01 -3.87882963e-02 4.65571880e-01
-6.64538503e-01 -8.67532551e-01 -1.04311991e+00 5.30983686e-01
-2.58005768e-01 -5.24213076e-01 -7.43622005e-01 -8.88191640e-01
2.05074325e-01 3.22386622e-01 1.18901503e+00 -3.95676881e-01
-2.15258151e-01 6.30953848e-01 -1.80729941e-01 -6.05973125e-01
-7.95645535e-01 -5.66465616e-01 4.34383035e-01 2.49073759e-01
-2.21747652e-01 -9.93516803e-01 -8.10393274e-01 4.84705448e-01
-6.51216567e-01 -2.42897242e-01 2.06108958e-01 5.74615777e-01
3.62617016e-01 1.52781919e-01 5.39665222e-01 3.81568633e-02
9.37538505e-01 1.09324642e-02 -3.32470357e-01 -4.29915846e-04
7.71203497e-03 -2.17693821e-01 1.07641935e+00 -7.42728591e-01
-1.45685327e+00 -1.55123621e-01 -7.36714542e-01 -4.93803620e-01
-1.02671713e-01 1.26392007e-01 -3.00940841e-01 1.67074516e-01
1.27986884e+00 2.01154813e-01 -2.96609432e-01 -6.36719167e-01
2.75465429e-01 9.69005942e-01 1.11157179e+00 -4.99003500e-01
1.14739454e+00 2.58617491e-01 -3.45350176e-01 -7.41684854e-01
-7.68569827e-01 -3.85471314e-01 -1.40094131e-01 -1.57170996e-01
4.71719593e-01 -1.14928496e+00 -8.18138242e-01 4.10954177e-01
-1.38156044e+00 -8.21169794e-01 -5.92510045e-01 1.07860363e+00
-8.49467754e-01 -1.66838601e-01 -7.43246675e-01 -9.59444761e-01
-3.88072073e-01 -8.64257514e-01 7.84134448e-01 -1.66890115e-01
-5.10862708e-01 -6.46789253e-01 3.74681562e-01 1.92077354e-01
1.06576097e+00 -2.05809921e-01 7.50836611e-01 -2.86731571e-01
-2.61251330e-01 -3.38575274e-01 2.48855487e-01 8.36595178e-01
3.21941137e-01 -3.36686522e-01 -1.69417167e+00 -3.06789130e-01
4.76972401e-01 -3.93989272e-02 5.27496397e-01 8.52868080e-01
9.14142668e-01 -3.97705138e-01 2.66122818e-01 9.07928765e-01
1.04688025e+00 5.50745308e-01 9.53791201e-01 -5.89776896e-02
4.06448334e-01 2.63729513e-01 2.44600579e-01 4.66731191e-01
-1.34088099e-01 5.58173358e-01 3.73670876e-01 -3.51285219e-01
-8.84306729e-01 -5.21550477e-01 5.07236600e-01 1.02756441e+00
1.85447693e-01 -2.62561977e-01 -7.18074322e-01 3.17186594e-01
-1.01237261e+00 -9.61867154e-01 7.72146955e-02 2.44566822e+00
1.00820160e+00 -1.57819852e-01 -2.49932244e-01 3.19173157e-01
5.52763164e-01 1.19921185e-01 -4.64675128e-01 -4.97821242e-01
3.86634097e-02 8.47150207e-01 9.35205445e-02 9.78304625e-01
-6.45016491e-01 3.38790119e-01 6.80458117e+00 5.73724926e-01
-1.24596405e+00 8.40964466e-02 4.18650955e-01 -2.04011366e-01
-4.41453159e-01 -5.91734111e-01 -2.62297928e-01 -5.30229472e-02
1.33982611e+00 -1.05371788e-01 1.15867448e+00 4.65993285e-01
5.40292144e-01 4.71463323e-01 -1.40660453e+00 1.07616782e+00
7.36800730e-02 -9.15544152e-01 -3.41222167e-01 -3.80372167e-01
5.61543465e-01 -2.02339798e-01 6.83661878e-01 3.74828964e-01
3.30686390e-01 -1.38538802e+00 1.36087060e+00 7.74354994e-01
9.25489902e-01 -4.86616403e-01 3.45578104e-01 2.82495935e-02
-1.36325216e+00 -3.16278756e-01 -1.64167687e-01 -1.03801943e-01
4.39669713e-02 4.68815327e-01 -1.37612641e+00 1.50236934e-01
8.19369078e-01 -2.65221912e-02 -1.42962784e-02 1.26360202e+00
-9.84494314e-02 9.42027926e-01 -1.81014046e-01 4.51229542e-01
-1.39433056e-01 2.04630598e-01 5.91595054e-01 1.19484735e+00
7.90921271e-01 -6.63086399e-02 -1.66196853e-01 7.93465197e-01
-3.06179255e-01 -2.65843552e-02 -3.26475114e-01 2.47643560e-01
5.95767021e-01 1.00721788e+00 -4.51468825e-02 1.97405770e-01
-6.88345954e-02 7.50459969e-01 -4.01280493e-01 9.27360654e-01
-1.02488172e+00 -5.83385110e-01 8.01353335e-01 1.73660383e-01
3.62696737e-01 -2.60070831e-01 5.84913359e-04 -4.80172664e-01
-1.16996147e-01 -1.17471087e+00 -3.67360681e-01 -1.39739001e+00
-1.09768355e+00 1.11309922e+00 -5.73075891e-01 -1.67062044e+00
-3.20262223e-01 -4.54610169e-01 -6.07440889e-01 1.29775858e+00
-1.07237172e+00 -8.98165226e-01 -3.23121130e-01 8.27212334e-01
2.66800433e-01 1.42012229e-02 1.24146068e+00 6.98590338e-01
4.59605083e-03 7.94937015e-01 4.24402833e-01 1.58758581e-01
1.08530581e+00 -1.06290925e+00 6.92224979e-01 7.72380054e-01
1.65857241e-01 6.31832540e-01 1.24322474e+00 -1.40542969e-01
-1.09732246e+00 -1.21706545e+00 4.98333454e-01 -3.32007170e-01
5.66058099e-01 -6.35532320e-01 -9.45703208e-01 4.36153620e-01
9.96632501e-02 1.58405691e-01 7.19927013e-01 -1.22817211e-01
-7.91204393e-01 -6.82518423e-01 -1.02302015e+00 7.30493605e-01
9.14899647e-01 -9.18527782e-01 -3.90359461e-01 1.63999453e-01
1.05232930e+00 -7.39954591e-01 -5.66372931e-01 -1.20678298e-01
6.84155524e-01 -1.10885918e+00 1.47964251e+00 -1.64484859e-01
1.98419034e-01 -4.06599283e-01 -7.37113118e-01 -1.81232417e+00
-3.56488466e-01 -1.09311867e+00 -1.30952686e-01 1.24781072e+00
3.38715285e-01 -3.81919801e-01 2.25158513e-01 3.54985118e-01
-7.77237773e-01 4.98126373e-02 -8.29700828e-01 -8.40563595e-01
-1.90223619e-01 -1.02857435e+00 8.20875525e-01 3.41494173e-01
-5.85260272e-01 1.94664866e-01 -3.82559270e-01 5.95959663e-01
5.68975747e-01 -2.38350168e-01 6.98313653e-01 -9.26860631e-01
-6.89794242e-01 -1.49134934e-01 1.78452671e-01 -1.18721867e+00
-4.53611575e-02 -4.34095293e-01 8.59175742e-01 -1.55694938e+00
-7.95772195e-01 -5.01964271e-01 -4.18536574e-01 -4.78529632e-02
1.28158882e-01 2.21985146e-01 1.32632568e-01 -3.93577933e-01
-3.74686420e-02 5.95143557e-01 1.20355558e+00 -1.87960371e-01
-4.29438859e-01 4.70329285e-01 -6.55981600e-01 8.87615800e-01
5.19543827e-01 -2.75741935e-01 -3.30937654e-01 -7.06919193e-01
1.55358329e-01 4.01425302e-01 4.09791231e-01 -1.65995121e+00
3.23927432e-01 3.09024423e-01 5.27884424e-01 -2.98251629e-01
9.02370393e-01 -1.14186943e+00 6.26801074e-01 2.61854768e-01
-7.11000264e-01 -3.80684495e-01 4.39964950e-01 6.19145811e-01
-2.46679023e-01 1.34497315e-01 7.79448450e-01 5.15079051e-02
-2.46812385e-02 -7.12909102e-02 -4.65730399e-01 1.21988393e-02
9.39791426e-02 -9.90674123e-02 -1.31831080e-01 -1.15169013e+00
-6.83552265e-01 -6.60485983e-01 -1.57650933e-01 1.79937124e-01
7.59884417e-01 -1.44770980e+00 -7.94620693e-01 3.92939419e-01
-2.71287948e-01 -1.50829837e-01 5.08696139e-01 3.47516328e-01
-2.97647506e-01 3.03151608e-01 8.70198160e-02 -3.05484176e-01
-9.51251507e-01 3.90837997e-01 9.97121453e-01 1.07813314e-01
-5.27403831e-01 1.26037669e+00 4.79379922e-01 -5.25238574e-01
8.28891158e-01 -6.90233946e-01 1.83413297e-01 -3.77302408e-01
8.93372059e-01 3.99034530e-01 3.44531178e-01 -5.10915518e-01
-1.61706284e-01 3.98561954e-01 5.47514677e-01 -5.87436199e-01
1.12308025e+00 -2.68697917e-01 2.08623439e-01 5.36668956e-01
1.51686656e+00 5.86142600e-01 -1.36582863e+00 -6.31416917e-01
-6.52638376e-01 -4.21165198e-01 2.62099594e-01 -1.24210870e+00
-7.75129557e-01 9.17448997e-01 9.62400377e-01 1.73474818e-01
1.59725809e+00 -4.09436941e-01 1.00120521e+00 4.43756104e-01
3.01931947e-01 -7.93429077e-01 6.60609365e-01 6.90163374e-01
1.24795926e+00 -5.72617769e-01 -6.61739349e-01 3.39467078e-02
-2.03815490e-01 1.12552583e+00 2.24094927e-01 -1.10714780e-02
7.46895552e-01 6.92624569e-01 6.27968550e-01 4.43484813e-01
-4.51254189e-01 1.11426197e-01 4.12666172e-01 1.02840614e+00
5.87208986e-01 -7.40042925e-02 9.10056949e-01 9.21720624e-01
-1.00881827e+00 -3.00634086e-01 3.07921648e-01 2.14610562e-01
-4.34473842e-01 -6.14922106e-01 -1.13408852e+00 1.94588304e-01
-4.54758197e-01 -5.74394166e-01 1.91795733e-02 -3.93884117e-03
-6.23866282e-02 1.41267240e+00 1.71533555e-01 -6.29933000e-01
8.37138236e-01 2.04156786e-02 5.03097892e-01 -3.66164774e-01
-9.56994951e-01 6.20600104e-01 7.68795982e-02 -4.32942063e-01
2.10040644e-01 -3.07332814e-01 -1.15423143e+00 -2.49578372e-01
-2.05564678e-01 1.09449983e-01 8.45068276e-01 3.95501107e-01
-5.01329079e-02 1.32704186e+00 9.40932393e-01 -1.23127425e+00
-8.55049431e-01 -8.45724344e-01 -7.97097862e-01 1.65010870e-01
1.07203567e+00 1.62880253e-02 -5.95369399e-01 1.73717245e-01] | [15.21505355834961, 5.9222636222839355] |
bca46701-17dc-477d-8f21-ff3fc6edcbbc | fedftn-personalized-federated-learning-with | 2304.00570 | null | https://arxiv.org/abs/2304.00570v1 | https://arxiv.org/pdf/2304.00570v1.pdf | FedFTN: Personalized Federated Learning with Deep Feature Transformation Network for Multi-institutional Low-count PET Denoising | Low-count PET is an efficient way to reduce radiation exposure and acquisition time, but the reconstructed images often suffer from low signal-to-noise ratio (SNR), thus affecting diagnosis and other downstream tasks. Recent advances in deep learning have shown great potential in improving low-count PET image quality, but acquiring a large, centralized, and diverse dataset from multiple institutions for training a robust model is difficult due to privacy and security concerns of patient data. Moreover, low-count PET data at different institutions may have different data distribution, thus requiring personalized models. While previous federated learning (FL) algorithms enable multi-institution collaborative training without the need of aggregating local data, addressing the large domain shift in the application of multi-institutional low-count PET denoising remains a challenge and is still highly under-explored. In this work, we propose FedFTN, a personalized federated learning strategy that addresses these challenges. FedFTN uses a local deep feature transformation network (FTN) to modulate the feature outputs of a globally shared denoising network, enabling personalized low-count PET denoising for each institution. During the federated learning process, only the denoising network's weights are communicated and aggregated, while the FTN remains at the local institutions for feature transformation. We evaluated our method using a large-scale dataset of multi-institutional low-count PET imaging data from three medical centers located across three continents, and showed that FedFTN provides high-quality low-count PET images, outperforming previous baseline FL reconstruction methods across all low-count levels at all three institutions. | ['Chi Liu', 'James S. Duncan', 'Kuangyu Shi', 'Axel Rominger', 'Biao Li', 'S. Kevin Zhou', 'Zhicheng Feng', 'Xueqi Guo', 'Xiongchao Chen', 'Qiong Liu', 'Huidong Xie', 'Bo Zhou'] | 2023-04-02 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [ 2.88405232e-02 -2.95083642e-01 -1.76547602e-01 -5.72764695e-01
-1.40401256e+00 -5.29312372e-01 1.33320615e-01 2.41934985e-01
-7.62346923e-01 8.27542126e-01 3.17534506e-01 -1.94489226e-01
-3.40240419e-01 -8.64943445e-01 -6.95706367e-01 -1.11118686e+00
2.30529979e-01 4.30766433e-01 1.05189309e-01 1.96430087e-01
-5.76695502e-01 6.38788998e-01 -7.15455115e-01 5.89407146e-01
6.64971352e-01 9.63697791e-01 5.65944493e-01 4.51723069e-01
4.56646420e-02 8.86074543e-01 -3.55645835e-01 -2.68693686e-01
3.56190085e-01 -4.85047251e-01 -6.23002708e-01 -3.73565018e-01
1.14106767e-01 -5.82045257e-01 -8.46188068e-01 9.36910570e-01
8.75086308e-01 1.62820220e-02 2.81243294e-01 -1.01713538e+00
-4.26249385e-01 6.59498751e-01 -1.95913315e-01 3.63151282e-01
-4.94733751e-01 4.89733219e-01 5.53718388e-01 -4.88664001e-01
5.31328380e-01 4.94175792e-01 9.60625529e-01 5.05258322e-01
-1.33005738e+00 -7.57616103e-01 -2.18171895e-01 6.89738244e-02
-9.53355074e-01 -1.41801283e-01 2.66285181e-01 -2.94606835e-01
7.48023331e-01 2.74694473e-01 8.55912268e-01 1.10649550e+00
5.67641377e-01 4.02496099e-01 1.07729292e+00 1.00537635e-01
1.54410496e-01 -2.18539625e-01 -1.92965329e-01 4.88368928e-01
2.40790009e-01 -6.07118756e-02 -4.93028611e-01 -3.05738807e-01
8.07642639e-01 4.25986558e-01 -4.76607889e-01 -3.23306680e-01
-1.33675516e+00 8.74712825e-01 8.12225461e-01 6.88468039e-01
-7.08714664e-01 3.57290089e-01 5.51788330e-01 2.93017119e-01
3.65675241e-01 -1.56949554e-02 -5.36749780e-01 2.14846045e-01
-9.09339190e-01 -1.15500696e-01 3.85505080e-01 6.12731099e-01
5.66947579e-01 -2.35025600e-01 -5.15794039e-01 8.39668572e-01
-6.28383085e-02 4.33214813e-01 8.21715951e-01 -9.87744927e-01
3.77648681e-01 3.80452186e-01 -2.74687767e-01 -5.74074030e-01
-7.32437372e-01 -6.60665452e-01 -1.15139985e+00 -1.64392230e-04
4.14009601e-01 -2.62208700e-01 -8.87833357e-01 1.85304320e+00
4.01407838e-01 5.13255447e-02 -2.55298913e-01 1.05855334e+00
8.52593780e-01 2.66126066e-01 3.72535437e-01 -1.67317227e-01
1.32094216e+00 -8.64057302e-01 -6.26691937e-01 -6.04174435e-02
7.34935701e-01 -4.64399993e-01 6.89794898e-01 3.46332461e-01
-9.79697108e-01 -2.73261629e-02 -8.71325672e-01 -6.24377802e-02
-2.03509942e-01 -7.23007023e-02 7.96150684e-01 6.92615449e-01
-1.02082968e+00 8.51612628e-01 -1.17431581e+00 -7.35523775e-02
1.09339130e+00 6.84244692e-01 -6.71545267e-01 -5.63586950e-01
-1.01041746e+00 9.37433600e-01 4.39665578e-02 -5.18521480e-03
-1.14179778e+00 -1.29195464e+00 -5.18883407e-01 3.21209393e-02
1.97693542e-01 -1.19199347e+00 1.41991127e+00 -6.55681908e-01
-1.09064078e+00 4.74107713e-01 3.25905889e-01 -5.91035724e-01
6.96049809e-01 4.54084575e-01 -4.05319750e-01 2.74734318e-01
3.93779516e-01 4.86021310e-01 6.40772641e-01 -6.34225488e-01
-5.98083615e-01 -6.30623817e-01 -6.33629799e-01 1.15447804e-01
-2.53584564e-01 -4.07636352e-02 -2.08366990e-01 -5.38323820e-01
2.34383136e-01 -5.22949278e-01 -5.49345732e-01 3.15725207e-01
1.68936461e-01 2.27026820e-01 8.70409727e-01 -5.91929972e-01
7.22398400e-01 -2.20626831e+00 -2.08198190e-01 4.18576822e-02
6.67468488e-01 -3.31673585e-02 -9.41223130e-02 -2.59353131e-01
-3.80176425e-01 3.38019207e-02 -3.34664851e-01 -1.87385410e-01
-2.08141863e-01 4.13750619e-01 4.72066939e-01 8.47283721e-01
-2.82810271e-01 1.15653074e+00 -1.11529326e+00 -4.43578273e-01
2.93196768e-01 5.59056580e-01 -4.56038415e-01 -4.84782532e-02
8.87853056e-02 8.49798322e-01 -5.85950136e-01 4.96099293e-01
6.99298739e-01 -4.07028258e-01 2.06454054e-01 -6.07134163e-01
-5.32076955e-02 -1.52666166e-01 -5.70133865e-01 2.11281967e+00
-5.54759562e-01 1.61162093e-01 4.10875857e-01 -9.46595550e-01
5.20566344e-01 4.38906670e-01 1.29449046e+00 -1.16752100e+00
4.27317023e-01 4.91975605e-01 -3.47820576e-03 -3.45408559e-01
6.41635731e-02 -7.84008265e-01 -6.48844093e-02 5.90353787e-01
3.45117211e-01 -1.02771148e-01 -3.08789432e-01 1.45966366e-01
1.64793110e+00 -4.57214832e-01 -1.55436993e-01 -5.41265942e-02
9.77639258e-02 2.66594272e-02 7.27468252e-01 9.03777003e-01
-6.46915257e-01 6.95180118e-01 3.31345648e-01 -4.69195426e-01
-1.03956473e+00 -1.10804749e+00 -5.00784755e-01 8.93325686e-01
-3.19294274e-01 -2.19604317e-02 -5.47087967e-01 -9.07413304e-01
-1.22930249e-02 4.24283028e-01 -4.71401632e-01 -2.23834231e-01
-5.94403863e-01 -1.18611586e+00 6.37416661e-01 4.45952207e-01
4.31664199e-01 -7.98247159e-01 -5.55605292e-01 6.28556073e-01
-5.35999358e-01 -1.07412207e+00 -5.82198560e-01 8.90289009e-01
-9.13613915e-01 -1.10232711e+00 -9.04847443e-01 -4.78899240e-01
5.00753164e-01 2.68755645e-01 1.05630231e+00 -2.36565158e-01
-4.66834813e-01 3.18364501e-01 -1.15314707e-01 -4.59617436e-01
-4.19693619e-01 3.43407355e-02 -1.64155230e-01 -9.97122675e-02
8.21589753e-02 -4.91026312e-01 -9.08224344e-01 4.43440706e-01
-1.08930516e+00 -1.85559019e-01 5.58592796e-01 1.16471279e+00
1.07320261e+00 8.83427039e-02 7.85948455e-01 -9.95257437e-01
2.84544617e-01 -7.17421591e-01 -4.77986485e-01 1.65191010e-01
-2.55103052e-01 -2.79433966e-01 7.51839697e-01 -1.29403099e-01
-9.55446064e-01 2.48393044e-01 -2.93076456e-01 -3.80287111e-01
1.29273668e-01 2.70183355e-01 -1.23993143e-01 -4.08072650e-01
7.89330840e-01 -1.14417020e-02 2.41327643e-01 -2.54802823e-01
2.04800650e-01 4.94126171e-01 6.41587853e-01 -2.89055526e-01
4.01220828e-01 6.51701331e-01 9.18743014e-02 -3.99166346e-01
-8.09412956e-01 -2.51925379e-01 -4.41417217e-01 -1.17263600e-01
1.08538055e+00 -1.10225070e+00 -5.35417378e-01 5.51559389e-01
-7.61806071e-01 -4.66816247e-01 -7.86936104e-01 7.55133569e-01
-5.55172145e-01 8.73022154e-02 -6.51991904e-01 5.88872321e-02
-7.30928004e-01 -1.49816918e+00 7.51221299e-01 6.06601909e-02
1.98401529e-02 -8.92656684e-01 6.39035925e-02 5.42294443e-01
8.21243882e-01 4.68938023e-01 1.08413017e+00 -5.35166323e-01
-7.06547081e-01 -1.01255998e-01 -2.87900418e-01 3.86017680e-01
3.57431918e-01 -8.31142962e-01 -1.01432431e+00 -3.28853786e-01
5.33291817e-01 -4.75624770e-01 6.71484172e-01 8.66258562e-01
1.65823150e+00 1.73954606e-01 -1.40951380e-01 1.13515723e+00
1.41736782e+00 -7.03878999e-02 4.50932145e-01 3.27801973e-01
6.56082213e-01 -6.62929565e-02 2.10737824e-01 5.90155602e-01
2.49356925e-01 3.27147126e-01 6.08499587e-01 -3.27498734e-01
-4.18061554e-01 -9.09517705e-02 -8.53828266e-02 5.04845142e-01
3.53530258e-01 -9.98011231e-02 -7.06675589e-01 5.07268131e-01
-1.62096465e+00 -7.75080025e-01 -2.07647264e-01 2.00190449e+00
6.35077059e-01 -3.94235790e-01 4.15752046e-02 -3.12014014e-01
6.19860828e-01 1.79431159e-02 -1.01648343e+00 -1.24395201e-02
-3.41337472e-01 5.50616145e-01 9.35314775e-01 -1.18546031e-01
-9.10271823e-01 2.72755980e-01 6.23764086e+00 8.32441092e-01
-1.32218444e+00 9.78893340e-01 9.59049225e-01 -6.57133520e-01
-2.86601901e-01 -3.59947890e-01 -1.35922238e-01 4.61237788e-01
9.98463035e-01 -2.72795975e-01 3.69524539e-01 5.99049866e-01
3.01885545e-01 -1.21001370e-01 -1.06578481e+00 1.20060873e+00
-3.58001322e-01 -1.56622481e+00 -2.89011925e-01 1.65016323e-01
7.49190748e-01 8.51618767e-01 9.06293392e-02 2.41072252e-01
3.47367436e-01 -1.17836010e+00 3.71505708e-01 4.95380282e-01
1.12214375e+00 -8.85527492e-01 7.29863107e-01 4.11623001e-01
-7.64123499e-01 -2.48793215e-01 -3.65872771e-01 8.12153280e-01
9.87693146e-02 8.40684295e-01 -6.13507330e-01 7.45212495e-01
9.34958518e-01 5.09758174e-01 -3.99729073e-01 9.46497262e-01
2.41075873e-01 3.51776689e-01 -4.58252281e-01 3.77424121e-01
3.11304748e-01 5.12643531e-02 1.04022883e-01 9.06577945e-01
5.22483110e-01 2.76057035e-01 -1.25911072e-01 7.12133229e-01
-6.28954649e-01 -4.42051962e-02 -3.85721177e-01 1.91431656e-01
5.91344871e-02 1.57425046e+00 -6.24883890e-01 -1.35597438e-02
-5.64798295e-01 8.73549521e-01 1.59650341e-01 -1.56484321e-01
-1.00553036e+00 1.68563887e-01 4.54605281e-01 2.68293411e-01
2.19788119e-01 1.25697538e-01 -3.68835956e-01 -1.08280599e+00
-2.59065628e-01 -9.65610802e-01 9.40588713e-01 -5.60979128e-01
-1.62148201e+00 4.99405533e-01 -3.20684314e-01 -1.06347716e+00
2.35510662e-01 -3.50128561e-01 -3.89512777e-01 9.95523989e-01
-1.49776506e+00 -1.12606728e+00 -5.03508925e-01 1.19685066e+00
7.22580999e-02 -1.22010760e-01 7.42915213e-01 8.48683298e-01
-4.80812281e-01 6.12601042e-01 3.82117629e-01 -4.15780582e-02
8.19014966e-01 -8.91788185e-01 -1.97030544e-01 5.24810672e-01
-2.04839483e-01 1.34120092e-01 -2.48384178e-02 -5.47841191e-01
-1.59679914e+00 -1.43689406e+00 4.29454267e-01 -3.94954532e-01
7.07109392e-01 -9.00437385e-02 -7.80770302e-01 5.58232069e-01
3.82933132e-02 8.88121128e-01 9.26453948e-01 -4.08971518e-01
5.39608784e-02 -4.89070773e-01 -1.88420546e+00 1.64884061e-01
8.34940732e-01 -3.88478875e-01 4.71765734e-03 6.66192830e-01
6.11769795e-01 -6.04567111e-01 -1.28817356e+00 1.69383109e-01
2.85374045e-01 -9.50050294e-01 8.64529908e-01 -2.18826875e-01
6.10857047e-02 -3.26015241e-02 -2.80606717e-01 -1.47632432e+00
-6.51552379e-01 -2.86607176e-01 2.37319231e-01 6.94946945e-01
1.71478808e-01 -8.99958432e-01 1.06860161e+00 8.05716395e-01
-4.89690572e-01 -6.66033745e-01 -1.49677265e+00 -4.64125007e-01
3.37626934e-01 -6.05303347e-01 7.67378211e-01 8.96028399e-01
-5.16659141e-01 -8.26634243e-02 -9.48129818e-02 1.18958317e-02
8.12936902e-01 -2.28513896e-01 3.67369980e-01 -7.52302766e-01
-4.71352786e-01 -2.07335249e-01 -3.47046673e-01 -3.52712244e-01
-4.79090102e-02 -1.35194337e+00 -3.06552887e-01 -1.57034039e+00
4.85060006e-01 -6.75537944e-01 -6.50826097e-01 7.85272598e-01
3.02906662e-01 4.21196759e-01 1.50070637e-01 1.69089660e-02
-4.46269453e-01 5.17158031e-01 1.65881813e+00 -2.82965750e-01
2.35369876e-01 -1.14122098e-02 -8.92433643e-01 5.73109329e-01
6.52530670e-01 -9.35455203e-01 -2.99857557e-01 -7.02273548e-01
-1.44838795e-01 4.31306928e-01 6.51260614e-01 -1.02759862e+00
4.82369512e-01 4.67224978e-02 9.45629358e-01 -5.92555463e-01
8.35851803e-02 -1.18734539e+00 6.02676392e-01 5.14941812e-01
1.81021422e-01 -1.95231661e-01 1.84546009e-01 4.08014327e-01
-1.66175738e-01 -9.05091763e-02 1.22933161e+00 -4.88163978e-01
-2.55844295e-01 8.55112195e-01 -2.81869620e-01 -2.31266078e-02
9.89853799e-01 1.77587599e-01 -2.58716792e-01 -1.26121283e-01
-9.67720985e-01 2.10517883e-01 2.22920761e-01 -6.90188110e-02
4.87629294e-01 -1.32269883e+00 -6.03690386e-01 2.39386097e-01
-1.94561020e-01 3.65717471e-01 9.50367510e-01 1.25666034e+00
-3.56572211e-01 1.20973594e-01 -1.32334128e-01 -7.86139965e-01
-7.48735189e-01 1.09378830e-01 8.73916030e-01 -5.05026460e-01
-9.34172511e-01 6.29235804e-01 1.79264039e-01 -7.65656650e-01
-1.78203762e-01 -3.14170778e-01 4.40487653e-01 2.33064760e-02
4.12846088e-01 2.85219848e-01 8.89696836e-01 -4.13316578e-01
-3.54201347e-01 1.21624529e-01 -2.39047706e-01 2.19497830e-01
1.97588253e+00 1.00996427e-01 -1.50601879e-01 -7.25545287e-02
1.52259314e+00 -5.12512565e-01 -1.25796294e+00 -2.45433703e-01
-5.26745260e-01 -3.92458916e-01 6.24618948e-01 -9.99492943e-01
-1.63949347e+00 3.39451939e-01 9.28086042e-01 -4.85064089e-01
1.44883049e+00 5.85245788e-02 1.14039218e+00 2.08520874e-01
5.87383151e-01 -8.68670881e-01 2.26529036e-03 2.95183271e-01
6.30816877e-01 -1.19006777e+00 9.95594040e-02 1.04909003e-01
-5.82594335e-01 8.75953972e-01 4.17726725e-01 2.08256587e-01
6.01816416e-01 5.68934143e-01 1.44459769e-01 -3.08499545e-01
-4.87052798e-01 4.42925572e-01 -2.99267501e-01 6.43377841e-01
1.67405993e-01 8.17959681e-02 -2.97355894e-02 9.88404095e-01
-1.15978137e-01 1.90299109e-01 3.60557050e-01 9.65513587e-01
-3.80050540e-02 -1.25525475e+00 -4.18702453e-01 9.24687028e-01
-6.47807896e-01 1.08762078e-01 2.24916801e-01 5.90735912e-01
4.75323707e-01 6.37770534e-01 -1.88410431e-01 9.58775580e-02
5.88278770e-01 -1.35156244e-01 6.16308153e-01 -4.77324396e-01
-1.13569438e+00 1.76815078e-01 -4.24910635e-01 -7.76197970e-01
-1.71648815e-01 -6.76808178e-01 -1.24863613e+00 -3.65655869e-01
-9.21177343e-02 -1.52727440e-01 9.15209532e-01 6.94890201e-01
2.70351142e-01 9.13259685e-01 8.15189779e-01 -6.59272492e-01
-7.31603026e-01 -4.77716386e-01 -8.30539823e-01 2.92781353e-01
4.23122674e-01 -2.21241653e-01 -1.21627517e-01 -4.62480277e-01] | [6.0938262939453125, 6.441249370574951] |
41f5961e-2d92-4353-9aaf-4dfe1f1699e5 | 3d-siammask-vision-based-multi-rotor-aerial | null | null | https://www.mdpi.com/1945298 | https://www.mdpi.com/1945298 | 3D-SiamMask: Vision-Based Multi-Rotor Aerial-Vehicle Tracking for a Moving Object | This paper aims to develop a multi-rotor-based visual tracker for a specified moving object. Visual object-tracking algorithms for multi-rotors are challenging due to multiple issues such as occlusion, quick camera motion, and out-of-view scenarios. Hence, algorithmic changes are required for dealing with images or video sequences obtained by multi-rotors. Therefore, we propose two approaches: a generic object tracker and a class-specific tracker. Both tracking settings require the object bounding box to be selected in the first frame. As part of the later steps, the object tracker uses the updated template set and the calibrated RGBD sensor data as inputs to track the target object using a Siamese network and a machine-learning model for depth estimation. The class-specific tracker is quite similar to the generic object tracker but has an additional auxiliary object classifier. The experimental study and validation were carried out in a robot simulation environment. The simulation environment was designed to serve multiple case scenarios using Gazebo. According to the experiment results, the class-specific object tracker performed better than the generic object tracker in terms of stability and accuracy. Experiments show that the proposed generic tracker achieves promising results on three challenging datasets. Our tracker runs at approximately 36 fps on GPU. | ['Alexandr Klimchik', 'Geesara Kulathunga', 'Mohamad Al Mdfaa'] | 2022-11-14 | null | null | null | remote-sensing-2022-11 | ['3d-single-object-tracking', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-3.46312761e-01 -4.48370099e-01 2.92854309e-02 3.64193507e-02
-2.66844600e-01 -3.43893439e-01 3.60241622e-01 -8.61740485e-02
-5.92446566e-01 3.86457950e-01 -5.91485143e-01 1.10315777e-01
1.84491307e-01 -2.98345506e-01 -6.75565839e-01 -8.78674448e-01
2.76019037e-01 4.76338357e-01 1.00883424e+00 1.10016219e-01
3.96605283e-01 7.20531225e-01 -1.85745442e+00 -3.15449208e-01
6.00427747e-01 1.03218734e+00 7.96074271e-01 7.28558779e-01
-6.22998029e-02 4.55426216e-01 -7.26531804e-01 1.42041206e-01
5.12891531e-01 -1.25985190e-01 -9.84705314e-02 3.32513094e-01
4.79131937e-01 -2.72537947e-01 -1.33218346e-02 1.06574202e+00
5.52528858e-01 1.96896493e-01 3.79121095e-01 -1.68797565e+00
5.08977175e-02 -7.77290314e-02 -9.19764757e-01 4.13350701e-01
3.33652645e-01 2.62961775e-01 2.03728974e-01 -8.92924249e-01
4.87177163e-01 1.35011005e+00 6.53582275e-01 5.69844246e-01
-5.77301860e-01 -7.64829934e-01 2.69948959e-01 3.35746467e-01
-1.42032194e+00 -1.90541714e-01 7.11994469e-01 -5.90927303e-01
5.23598969e-01 -9.41923857e-02 8.60205233e-01 7.23363638e-01
5.02342343e-01 5.98362923e-01 9.58036423e-01 -3.52741569e-01
1.94758564e-01 4.28852469e-01 1.71899334e-01 6.84599936e-01
5.98560810e-01 2.80128777e-01 -3.28680903e-01 -3.43981199e-02
9.82629299e-01 -4.15025428e-02 -3.36232394e-01 -1.00188613e+00
-1.31226194e+00 5.96248209e-01 3.25694531e-01 1.48198947e-01
-8.82813781e-02 2.70006746e-01 3.80299777e-01 -1.72802538e-01
8.20322260e-02 -1.26837417e-01 -4.23190594e-01 -1.40921801e-01
-6.62106395e-01 1.56143978e-01 6.59605861e-01 1.20886219e+00
3.67633492e-01 1.85428724e-01 1.59706846e-01 3.49782854e-01
8.33397269e-01 8.43516290e-01 5.76786578e-01 -5.74520111e-01
3.47781599e-01 3.48367363e-01 5.08027017e-01 -1.05582094e+00
-6.01911306e-01 -3.78808737e-01 -4.31250691e-01 7.62632072e-01
7.69134760e-01 -5.07995933e-02 -9.02940214e-01 1.27334869e+00
9.59864497e-01 1.73522651e-01 -1.39462113e-01 1.39881110e+00
8.69592547e-01 5.71288228e-01 -3.68627086e-02 -3.36766183e-01
1.52837479e+00 -1.11954737e+00 -9.24509346e-01 -9.01340321e-02
3.47262144e-01 -9.40147281e-01 7.33314335e-01 4.11363572e-01
-8.68431926e-01 -1.01457560e+00 -1.17976820e+00 3.23460370e-01
-2.87652612e-01 6.36398554e-01 2.24318430e-01 8.30299437e-01
-7.19385684e-01 1.29110783e-01 -1.02742064e+00 -7.13771760e-01
6.81429803e-02 3.07305455e-01 -1.90743566e-01 3.00330073e-01
-5.78667402e-01 1.09012389e+00 6.19868159e-01 2.01495528e-01
-9.02766943e-01 -3.11786801e-01 -6.88829362e-01 -2.17429414e-01
3.99191409e-01 -7.34346449e-01 1.16922224e+00 -9.80412722e-01
-1.69488895e+00 6.10147536e-01 -9.84127223e-02 -2.70678937e-01
6.76759779e-01 -2.76338249e-01 -2.43583158e-01 1.02448590e-01
1.25091240e-01 3.62118065e-01 1.13394904e+00 -1.29261029e+00
-9.71025467e-01 -3.35748166e-01 -4.43866476e-02 4.53087479e-01
4.94599566e-02 1.40734583e-01 -8.09528828e-01 -4.90061849e-01
2.43450642e-01 -1.03736973e+00 -4.83402275e-02 3.80980968e-01
8.49518999e-02 -2.07494944e-01 1.33510387e+00 -2.06668913e-01
7.41701603e-01 -2.09737277e+00 -1.74756087e-02 -1.27182066e-01
-2.37993896e-02 3.80145788e-01 2.38969684e-01 -3.10521275e-02
2.65248753e-02 -6.20518744e-01 5.18566310e-01 -3.13701868e-01
-4.54809964e-01 -4.85088825e-02 1.75218694e-02 8.85652959e-01
-1.82109386e-01 2.89035857e-01 -8.29288065e-01 -8.56471062e-01
6.65657520e-01 3.59260768e-01 -8.18304494e-02 2.96535671e-01
-1.55190647e-01 5.79626977e-01 -5.42500436e-01 8.85675430e-01
8.56331587e-01 -1.84002995e-01 -2.18731493e-01 -5.10137737e-01
-3.90720725e-01 -4.57370102e-01 -1.72936428e+00 1.51682770e+00
-1.86804473e-01 5.96872509e-01 1.77509382e-01 -5.49213409e-01
1.03218138e+00 1.95653424e-01 5.29717803e-01 -3.42259407e-01
5.12006164e-01 3.26845020e-01 3.21264757e-04 -7.69755125e-01
5.79889536e-01 2.14231640e-01 3.58494639e-01 1.14190824e-01
-2.15578422e-01 2.26840470e-02 2.73513496e-01 -2.69127451e-02
6.33217275e-01 7.27187812e-01 3.63290906e-01 -1.05955042e-01
7.11492896e-01 2.14220151e-01 7.29076147e-01 5.88392913e-01
-5.30414999e-01 5.08655608e-01 -2.64228463e-01 -4.68305320e-01
-8.12973559e-01 -7.83734679e-01 -1.85026988e-01 6.76800728e-01
9.90187764e-01 -2.97585074e-02 -4.30440545e-01 -5.48524976e-01
2.98002493e-02 1.05889559e-01 -2.33815596e-01 1.23608291e-01
-6.84174478e-01 -4.96153891e-01 3.33910175e-02 5.58762252e-01
5.59684396e-01 -8.49141479e-01 -1.58807802e+00 9.06363875e-02
1.02954730e-01 -1.28524375e+00 -5.93957067e-01 1.79025263e-01
-8.63170505e-01 -1.53906953e+00 -7.24370122e-01 -7.31668591e-01
6.78516328e-01 6.95817173e-01 4.25729394e-01 3.73173580e-02
-2.75800288e-01 5.07800400e-01 -2.89854079e-01 -5.06756186e-01
-1.34527832e-01 -2.21241832e-01 3.57111245e-01 6.74750432e-02
2.43458062e-01 1.07126720e-02 -5.67660630e-01 7.03775227e-01
-5.74603140e-01 -8.45779385e-03 4.86352265e-01 5.49604714e-01
4.03299034e-01 -1.59054428e-01 1.62389189e-01 -2.10741498e-02
9.55924019e-02 -1.38168901e-01 -1.38328207e+00 2.13773534e-01
-3.06733757e-01 -1.90691367e-01 4.16692823e-01 -8.67014229e-01
-8.82119358e-01 5.90937555e-01 3.38383406e-01 -9.32330728e-01
-2.89286803e-02 -3.54450196e-02 -3.57609332e-01 -4.49457228e-01
4.93313760e-01 8.57043639e-02 2.06497371e-01 -3.38718504e-01
8.41121525e-02 7.30427504e-01 4.71663892e-01 -2.26264179e-01
9.60063815e-01 4.96403188e-01 1.07271641e-01 -8.09666514e-01
-5.83749115e-01 -8.51403713e-01 -6.54383063e-01 -9.29524481e-01
9.25393760e-01 -9.79561746e-01 -1.24965930e+00 7.52038002e-01
-1.33558202e+00 1.15874253e-01 8.56532529e-02 9.32593703e-01
-5.20702839e-01 4.21241015e-01 -1.17524423e-01 -1.00443137e+00
-3.64747733e-01 -1.34183550e+00 1.13266587e+00 8.07275057e-01
3.16977441e-01 -8.55766654e-01 2.38989610e-02 1.43573731e-01
1.75311998e-01 3.23075950e-01 -5.56065179e-02 -3.02544087e-01
-1.00340509e+00 -1.53804883e-01 -1.77092582e-01 1.48791187e-02
1.62568003e-01 2.22213924e-01 -8.05413783e-01 -5.38466156e-01
2.22390860e-01 -2.77876332e-02 4.25801814e-01 5.20392537e-01
7.60457397e-01 2.83117235e-01 -8.59151006e-01 6.89204335e-01
1.52839088e+00 7.04326391e-01 1.64800689e-01 6.88603997e-01
7.10430861e-01 1.34939119e-01 1.46502149e+00 4.49123859e-01
3.87494951e-01 1.18917131e+00 7.46576905e-01 2.38768548e-01
-1.38017446e-01 1.07903883e-01 5.32551289e-01 4.85516459e-01
1.01521067e-01 -4.09446694e-02 -8.82688999e-01 5.52425921e-01
-2.03791213e+00 -8.12268972e-01 -5.56014717e-01 2.27434301e+00
3.67981762e-01 1.30387023e-01 3.65628034e-01 -4.56157848e-02
9.91495192e-01 -1.69753700e-01 -7.14371145e-01 4.62312177e-02
1.22346520e-01 -4.35712665e-01 7.48282015e-01 1.73130333e-01
-1.19333649e+00 7.85800755e-01 5.30024147e+00 5.18112183e-01
-1.28924656e+00 2.23647714e-01 -2.74244189e-01 -1.62403971e-01
7.30205715e-01 -7.66284242e-02 -1.40491486e+00 6.61665201e-01
6.30630136e-01 -6.11759350e-02 -3.79641280e-02 1.05945218e+00
2.57722169e-01 -6.29879594e-01 -8.06986928e-01 1.37980056e+00
2.84152567e-01 -9.63127017e-01 -4.77887422e-01 -1.04249537e-01
3.06363523e-01 4.22982611e-02 -2.16560632e-01 -5.76816872e-02
-7.84660429e-02 -3.77321213e-01 1.05453253e+00 5.37342727e-01
2.72056133e-01 -4.94650453e-01 7.22163856e-01 4.60703164e-01
-1.60623276e+00 -8.01166818e-02 -5.42301476e-01 1.57429278e-01
1.72085613e-01 2.08668381e-01 -7.50568032e-01 5.98086417e-01
9.94165063e-01 6.99153066e-01 -5.66276193e-01 1.75480497e+00
6.00137934e-03 -7.07495064e-02 -6.08255267e-01 -3.26740354e-01
4.54719663e-02 -2.13379532e-01 7.53094077e-01 8.19694400e-01
3.34700167e-01 -9.58820060e-02 3.82261962e-01 3.69692594e-01
6.09690666e-01 1.53894052e-01 -4.51278746e-01 4.90748912e-01
4.00421679e-01 1.53441310e+00 -1.10086632e+00 -3.24296772e-01
-3.33174080e-01 7.41565228e-01 1.92699432e-02 6.40196279e-02
-1.44633472e+00 -4.09390032e-01 4.78481978e-01 6.79430142e-02
7.34895766e-01 -3.74336958e-01 -5.27785113e-03 -1.10117853e+00
1.01716258e-01 -5.86128414e-01 2.18570501e-01 -1.19351351e+00
-6.66402102e-01 5.91538548e-01 1.33970648e-01 -1.77380240e+00
1.69804115e-02 -8.02102804e-01 -3.05199504e-01 6.37993395e-01
-1.24820435e+00 -1.17849338e+00 -9.12392676e-01 6.40487432e-01
6.27971768e-01 -2.31528953e-01 2.22128168e-01 2.98676699e-01
-6.39270902e-01 3.17717493e-01 1.44010976e-01 -9.33784992e-02
8.31844985e-01 -9.20185626e-01 -9.17416438e-02 9.62745786e-01
-1.16559699e-01 3.88650179e-01 8.54346395e-01 -7.92204976e-01
-1.60330939e+00 -9.07084703e-01 2.22357333e-01 -6.27615988e-01
5.31285048e-01 -2.66438276e-01 -6.27510607e-01 5.59198141e-01
9.98575613e-02 4.64156061e-01 5.70887327e-02 -4.72145766e-01
2.52906740e-01 -3.00185740e-01 -1.14699042e+00 2.31679529e-01
7.88141131e-01 1.56921923e-01 -5.82666159e-01 3.49840432e-01
3.71910185e-01 -1.00324357e+00 -7.91634500e-01 2.00263247e-01
6.85404599e-01 -7.77282000e-01 7.73022830e-01 1.33562312e-02
-4.61006790e-01 -1.14768898e+00 1.26392096e-01 -9.86843467e-01
1.28530443e-01 -3.69088203e-01 -3.38228017e-01 1.02490008e+00
-4.02220279e-01 -6.33474588e-01 7.26835549e-01 1.40986785e-01
-5.38310222e-03 -4.13031131e-01 -1.02530456e+00 -1.10055113e+00
-6.85211360e-01 -1.16894029e-01 2.33480111e-01 4.27057922e-01
-4.36288178e-01 1.82636708e-01 -3.36358249e-01 5.90964019e-01
9.45401490e-01 2.90936351e-01 1.35660625e+00 -1.22373855e+00
-2.08799001e-02 -9.18003637e-03 -7.40532935e-01 -1.20027697e+00
-9.94196907e-02 -2.58260816e-01 1.36459783e-01 -1.34463835e+00
2.97180787e-02 -4.64006335e-01 7.05445930e-02 2.08481520e-01
-1.46778464e-01 2.02664822e-01 4.08596098e-01 3.02736819e-01
-6.90031767e-01 6.17491245e-01 1.13033926e+00 6.64111748e-02
-2.29923263e-01 3.93175393e-01 1.06161565e-01 7.23831475e-01
6.08733296e-01 -5.92650533e-01 -3.11014205e-01 -2.53785998e-01
-2.27675602e-01 9.86031443e-02 5.26250005e-01 -1.45739079e+00
7.07051814e-01 -1.12335950e-01 4.23195958e-01 -1.21004045e+00
6.60364568e-01 -1.34701312e+00 4.40660805e-01 8.99212182e-01
4.91843909e-01 2.58241385e-01 3.99739861e-01 5.56204200e-01
2.77886167e-02 -3.75124454e-01 1.05960417e+00 8.34222957e-02
-9.10961151e-01 1.62387162e-01 -2.03812078e-01 -2.86822706e-01
1.76611543e+00 -8.55126441e-01 -3.90706390e-01 -5.51243126e-02
-5.33905029e-01 3.27686101e-01 6.77791238e-01 8.56671274e-01
6.58353508e-01 -1.28598630e+00 -2.51281321e-01 1.80262625e-01
2.21034110e-01 -4.02085930e-02 -6.73685148e-02 1.17123008e+00
-6.32753074e-01 5.74499309e-01 -4.07686949e-01 -1.14698517e+00
-1.75049436e+00 8.43714833e-01 6.97543204e-01 2.23407716e-01
-6.69603229e-01 4.62765157e-01 1.37221172e-01 -1.03787772e-01
4.16707039e-01 -6.87717497e-01 -3.22605222e-01 1.31136896e-02
4.46129799e-01 4.90184397e-01 -1.60994306e-01 -8.62063825e-01
-6.59207404e-01 1.13341618e+00 9.91263539e-02 9.01806876e-02
9.22088385e-01 -2.52806693e-01 2.17179060e-01 5.77706635e-01
7.17171609e-01 -9.73349214e-02 -1.40142930e+00 -2.23236196e-02
-1.27661079e-01 -7.74965465e-01 -7.25116357e-02 -3.60965103e-01
-1.20037329e+00 4.57293421e-01 1.14596987e+00 -9.89217684e-02
9.58552063e-01 -2.56950527e-01 2.19453856e-01 1.45175859e-01
8.17623794e-01 -1.02616191e+00 2.01006919e-01 3.56194019e-01
7.06798971e-01 -1.34871042e+00 1.72010317e-01 -3.15103769e-01
-3.42195064e-01 1.23010695e+00 1.25826895e+00 -1.19171344e-01
5.63228548e-01 4.27326143e-01 3.12152207e-01 -3.61216739e-02
-4.60207134e-01 -2.56033778e-01 1.70825884e-01 5.93024790e-01
-8.20284858e-02 -4.76492822e-01 -3.90485615e-01 -1.32904619e-01
1.30761504e-01 1.02157474e-01 5.62438726e-01 1.01023030e+00
-6.40940011e-01 -7.30630934e-01 -1.05634618e+00 -3.53096388e-02
-2.19756588e-01 5.63504040e-01 2.57319719e-01 1.22999465e+00
1.90415680e-01 9.17602420e-01 1.38827860e-01 -1.07061662e-01
3.33683819e-01 -3.51716965e-01 6.94296360e-01 -3.28974009e-01
-7.26266384e-01 1.29551724e-01 -3.45698565e-01 -7.29768038e-01
-8.50558996e-01 -8.33516479e-01 -1.39949000e+00 1.49430156e-01
-8.09725642e-01 9.64663830e-03 1.20489526e+00 8.35610569e-01
1.59368411e-01 4.53947246e-01 4.64548945e-01 -1.09619248e+00
-4.05259311e-01 -9.26678538e-01 -3.98769677e-01 1.15433559e-01
3.19393128e-01 -1.20169663e+00 -2.60797203e-01 2.41520209e-03] | [6.850841522216797, -1.9749561548233032] |
9679e364-6b1d-4323-aea4-6e60dfbf7ba2 | an-updated-duet-model-for-passage-re-ranking | 1903.07666 | null | http://arxiv.org/abs/1903.07666v1 | http://arxiv.org/pdf/1903.07666v1.pdf | An Updated Duet Model for Passage Re-ranking | We propose several small modifications to Duet---a deep neural ranking
model---and evaluate the updated model on the MS MARCO passage ranking task. We
report significant improvements from the proposed changes based on an ablation
study. | ['Bhaskar Mitra', 'Nick Craswell'] | 2019-03-18 | null | null | null | null | ['passage-ranking', 'passage-re-ranking'] | ['natural-language-processing', 'natural-language-processing'] | [-3.69197428e-02 -1.66206792e-01 -1.28663465e-01 -5.52958071e-01
-1.51409960e+00 -5.21188796e-01 5.42914510e-01 4.92115915e-01
-1.15797567e+00 1.16079247e+00 7.94559062e-01 -4.91197228e-01
-3.80820841e-01 -1.95443049e-01 -7.08514392e-01 1.72382474e-01
-7.37085640e-01 7.68815637e-01 3.14365983e-01 -9.95285332e-01
7.74012089e-01 -1.29556701e-01 -9.77460206e-01 7.37422884e-01
1.17622817e+00 3.57106894e-01 1.23951375e-01 1.25976336e+00
1.41593695e-01 1.37320852e+00 -1.04397881e+00 -5.72811127e-01
5.58704436e-02 -2.50123858e-01 -1.38333082e+00 -1.23909259e+00
9.02300954e-01 -7.59575367e-01 -8.07099104e-01 5.51461041e-01
8.66901934e-01 1.01914036e+00 6.94110453e-01 -5.37920535e-01
-1.05098820e+00 1.21766067e+00 1.26136959e-01 8.30442131e-01
8.40928376e-01 -7.32392609e-01 1.52988827e+00 -8.12895536e-01
6.51211023e-01 9.50521946e-01 1.02224445e+00 7.07389057e-01
-5.48845589e-01 -1.04762241e-01 -1.94517404e-01 5.77713966e-01
-9.54318821e-01 -6.00669920e-01 2.88124382e-01 2.83780038e-01
1.85613108e+00 5.49955308e-01 1.27468690e-01 1.24627793e+00
1.68890387e-01 1.09862888e+00 8.25245440e-01 -5.66414058e-01
-2.62171030e-01 -3.51507455e-01 5.82992375e-01 6.33923769e-01
2.44228169e-01 7.62301981e-02 -4.64649826e-01 -2.18407810e-01
1.62761688e-01 -6.17480099e-01 -4.30627704e-01 2.81274378e-01
-1.16200280e+00 5.03178120e-01 7.13241339e-01 4.14092392e-01
8.94775614e-02 3.91454279e-01 8.08602154e-01 9.12622631e-01
8.40113103e-01 1.19115591e+00 -6.59344494e-01 -7.63965428e-01
-1.17429459e+00 3.31005871e-01 9.91705120e-01 6.82228386e-01
-5.93520142e-02 -2.06489429e-01 -9.01463091e-01 1.11128950e+00
1.37408525e-02 1.60284102e-01 7.15711534e-01 -7.14752436e-01
8.05361629e-01 -7.98407793e-02 4.19728428e-01 -6.60509825e-01
-8.40847015e-01 -8.56882215e-01 -3.41710657e-01 -4.56092864e-01
-1.46933645e-01 3.81775163e-02 -7.18563855e-01 1.59148502e+00
-5.97714126e-01 7.24751055e-02 2.62053907e-01 8.78719151e-01
1.43943727e+00 6.80497229e-01 1.74303904e-01 9.02237296e-02
6.85416102e-01 -1.65065598e+00 -8.21733832e-01 6.06343411e-02
1.22865236e+00 -6.65372431e-01 1.36031079e+00 -5.92008643e-02
-1.34904790e+00 -4.05113071e-01 -1.51531124e+00 -6.54812396e-01
-6.81794941e-01 4.14678156e-01 6.04507804e-01 4.58782673e-01
-1.74535012e+00 1.23222113e+00 -3.77403110e-01 -3.91074896e-01
-4.53612655e-02 5.02860725e-01 -1.81310505e-01 -8.92010797e-03
-1.90177822e+00 1.35353804e+00 4.08499569e-01 2.22076967e-01
-8.06812763e-01 -5.52422583e-01 -6.74822509e-01 1.74363717e-01
-1.64108068e-01 -1.09312785e+00 2.02843308e+00 -7.29827285e-02
-1.79793096e+00 6.15280509e-01 -2.23385915e-01 -5.57240844e-01
2.73502171e-01 -7.39907563e-01 -7.06344604e-01 1.29040614e-01
1.03028752e-01 3.91692638e-01 1.68729806e-03 -8.18364501e-01
-6.01553857e-01 4.82877523e-01 7.17014849e-01 8.03120673e-01
-4.38166052e-01 2.60287434e-01 -2.67434210e-01 -8.00448298e-01
-3.70732844e-01 -7.58683980e-01 -2.67677277e-01 -9.55303788e-01
-3.74480188e-01 -4.34379041e-01 4.64255624e-02 -9.32016850e-01
1.47599745e+00 -1.48192203e+00 2.77935602e-02 -1.62549898e-01
3.28805149e-01 2.85483956e-01 -7.22155690e-01 4.60033715e-01
-2.02701557e-02 4.14363474e-01 1.13676272e-01 -5.42650938e-01
2.74470542e-02 -2.62152195e-01 -2.46229827e-01 4.79608402e-02
-2.20805421e-01 1.28254652e+00 -1.43231797e+00 -2.57907480e-01
-2.08715901e-01 8.42149332e-02 -4.78840888e-01 1.98532298e-01
-2.66763657e-01 -1.22686280e-02 -5.45554876e-01 1.79141343e-01
2.54343837e-01 -2.97510684e-01 -2.80557722e-01 3.08271021e-01
1.22586787e-01 1.37832677e+00 -1.86465591e-01 2.32166004e+00
-4.97544259e-01 9.97575879e-01 -6.98931098e-01 -5.70098042e-01
2.71396399e-01 5.73136628e-01 3.46722566e-02 -8.11419427e-01
5.04935272e-02 4.28761065e-01 8.99824053e-02 -2.84911215e-01
1.53865826e+00 3.49123806e-01 -2.14711696e-01 4.79312092e-01
2.17766583e-01 2.50610024e-01 6.14668131e-01 9.18917418e-01
1.39187706e+00 3.64843011e-01 5.11899181e-02 -4.57352936e-01
3.67099226e-01 1.93253502e-01 -1.73387334e-01 1.46240962e+00
-7.93678686e-02 6.85759902e-01 2.99010485e-01 -1.02231175e-01
-8.25512111e-01 -8.54967535e-01 5.00472412e-02 1.53519738e+00
2.95272935e-02 -6.69954717e-01 -5.39055705e-01 -9.94347453e-01
-4.66767520e-01 8.88431311e-01 -8.65675807e-01 -1.30383670e-01
-6.69577241e-01 -8.24283361e-01 8.06159258e-01 5.75343609e-01
4.29141521e-01 -1.29393232e+00 -5.43841645e-02 2.84121573e-01
-5.69787681e-01 -9.65230525e-01 -6.97058082e-01 7.24781603e-02
-9.27733541e-01 -6.19550347e-01 -8.74122441e-01 -1.17255950e+00
3.45450968e-01 3.56074125e-01 1.75544536e+00 3.31016302e-01
1.81570709e-01 2.54377395e-01 -8.00984740e-01 -3.67314816e-02
-4.59556580e-01 7.93037891e-01 1.03658304e-01 -1.05099249e+00
1.02806628e-01 -2.64930189e-01 -5.63386798e-01 -2.15541631e-01
-4.72454548e-01 -2.42442816e-01 4.85603869e-01 8.64792049e-01
1.99007429e-03 -3.80889475e-01 8.50482702e-01 -1.15943921e+00
1.53405547e+00 -2.91826218e-01 -2.97591202e-02 4.69952255e-01
-5.50739586e-01 1.95139080e-01 5.73624849e-01 -6.88077286e-02
-9.29011285e-01 -9.21130061e-01 -5.53944588e-01 1.42324820e-01
4.78846580e-01 1.03918171e+00 3.58211100e-01 -7.01516196e-02
9.14921224e-01 -4.94471677e-02 -8.54308367e-01 -5.89087844e-01
6.27951026e-01 6.95667684e-01 6.33490741e-01 -7.71823347e-01
4.51456994e-01 -2.93608844e-01 -4.67289627e-01 -2.00594500e-01
-1.19764698e+00 -4.31452930e-01 -4.27377492e-01 -1.36396199e-01
2.29991838e-01 -1.00767171e+00 -3.59663397e-01 1.73963159e-01
-1.29244578e+00 -4.80489135e-01 -6.79712817e-02 6.84931695e-01
-2.68183291e-01 4.46638256e-01 -1.41954195e+00 -1.28287077e-01
-7.62870371e-01 -8.13356757e-01 5.72136700e-01 4.76666510e-01
-1.13479264e-01 -1.12512755e+00 8.94836128e-01 2.25945264e-01
8.84339094e-01 -3.94120067e-01 8.37730706e-01 -1.08691180e+00
4.90935240e-03 -5.29199004e-01 -2.46344373e-01 1.42047301e-01
-3.28525901e-01 -2.33450487e-01 -8.03971052e-01 -5.52788496e-01
-6.75948322e-01 -6.72781050e-01 1.35276508e+00 3.53252262e-01
1.08789408e+00 1.37344375e-02 -3.72295827e-02 3.26675326e-01
1.00028276e+00 -1.04555093e-01 5.18725872e-01 8.09820235e-01
4.69157457e-01 -2.19374336e-02 4.79779065e-01 5.80429286e-02
6.17334127e-01 6.62511706e-01 7.18482062e-02 1.60708904e-01
-3.73190045e-01 -4.41940665e-01 5.57879388e-01 1.89458358e+00
-4.72793937e-01 -5.99301636e-01 -3.47543597e-01 5.59925556e-01
-1.78467500e+00 -9.65054870e-01 1.03031561e-01 2.05230117e+00
8.79060447e-01 2.94021815e-01 -8.12938958e-02 -4.06984091e-01
4.59013969e-01 4.08700645e-01 -2.46578917e-01 -7.35935807e-01
-3.57341468e-01 6.04619026e-01 2.39729360e-01 8.73777032e-01
-1.15604007e+00 1.13440931e+00 8.63535023e+00 7.29609728e-01
-6.61938787e-01 3.23271006e-01 1.98439389e-01 -2.65050918e-01
-3.60863596e-01 -1.39074400e-01 -5.20932734e-01 -5.84385451e-03
1.47190666e+00 -4.90708172e-01 5.09320199e-01 5.94184875e-01
-1.78890660e-01 3.29324961e-01 -1.11798096e+00 4.04469043e-01
4.85524923e-01 -1.34805942e+00 3.16137254e-01 -3.56442988e-01
8.25421572e-01 7.53148556e-01 2.13062540e-01 1.21221101e+00
6.41681850e-01 -1.00177491e+00 4.31414038e-01 3.66644681e-01
8.29627156e-01 -5.87181330e-01 1.02283478e+00 -2.38262489e-01
-8.17188501e-01 1.93393514e-01 -7.12186396e-01 -3.24172884e-01
5.23166172e-02 1.56694397e-01 -6.97570801e-01 8.36636186e-01
5.42098701e-01 9.51119483e-01 -9.06392157e-01 1.42965865e+00
-5.75097382e-01 7.73211598e-01 8.98205787e-02 -5.39865553e-01
2.50824749e-01 3.95608693e-01 8.26427519e-01 1.62429190e+00
2.94737518e-01 -6.45978302e-02 -6.57749772e-02 2.04214051e-01
-7.86826611e-01 3.19750875e-01 -3.09920788e-01 6.31509861e-03
4.27667886e-01 1.17476821e+00 -4.94231749e-03 -5.71247339e-01
-2.18974873e-01 1.22470200e+00 1.11578536e+00 6.16120696e-01
-9.05125499e-01 -1.01291871e+00 1.94908008e-01 -6.11171663e-01
1.62515521e-01 -3.31447065e-01 1.58931002e-01 -1.50730646e+00
1.07794181e-01 -6.47504508e-01 5.27965784e-01 -9.07346010e-01
-1.32271433e+00 1.11484027e+00 -5.29236980e-02 -1.28707790e+00
-3.50313723e-01 -4.72413152e-01 -5.61201572e-01 8.48064959e-01
-1.96242726e+00 -6.27012551e-01 3.28048825e-01 2.36691862e-01
5.40685475e-01 -3.58104974e-01 1.01956201e+00 5.06482422e-01
-5.18637657e-01 1.18953502e+00 8.05445910e-01 2.65254855e-01
9.07502830e-01 -1.73469698e+00 7.67618299e-01 6.95667863e-01
4.84612167e-01 9.31153655e-01 5.17884552e-01 -3.08941364e-01
-6.33363605e-01 -8.36646855e-01 1.59365690e+00 -8.18718374e-01
5.87165892e-01 -1.07228935e-01 -6.26039386e-01 7.81910360e-01
7.61527181e-01 -4.01139170e-01 7.29182005e-01 8.37341845e-01
-3.06640208e-01 2.50212669e-01 -7.98561811e-01 8.06919813e-01
1.16305935e+00 -7.39150226e-01 -1.20523798e+00 4.83432233e-01
1.10538280e+00 -7.05700099e-01 -9.35787916e-01 4.67048496e-01
5.16302586e-01 -3.75456750e-01 1.16482985e+00 -1.14129353e+00
6.44369841e-01 -2.37356164e-02 -9.51056741e-03 -2.05668783e+00
-6.06539190e-01 -8.04004133e-01 -4.92891192e-01 5.24986327e-01
8.98443937e-01 -4.02166158e-01 4.25311655e-01 2.01875761e-01
-8.49087238e-01 -5.78997374e-01 -1.28669274e+00 -7.87844360e-01
4.88258213e-01 -1.42558903e-01 3.94552439e-01 8.67256641e-01
7.23869860e-01 8.28869343e-01 -4.24983889e-01 -1.93848565e-01
1.35178924e-01 -2.65140146e-01 1.40688851e-01 -9.18668926e-01
-3.77118170e-01 -7.18470454e-01 1.58981502e-01 -1.50777423e+00
2.51714826e-01 -1.15465963e+00 1.87013730e-01 -1.95104480e+00
5.91221154e-01 -1.17809549e-01 -1.30367029e+00 4.18133587e-01
-6.07844830e-01 4.71144527e-01 -7.77910575e-02 2.44809255e-01
-1.66002440e+00 8.34150910e-01 1.12124670e+00 -3.15049022e-01
1.87596530e-02 -2.37357393e-01 -9.41377461e-01 7.74035081e-02
7.71050572e-01 -5.72813749e-01 -4.30713743e-01 -1.15619421e+00
8.39747608e-01 2.14308500e-01 -2.26272538e-01 -1.12312984e+00
1.85445607e-01 5.61582267e-01 2.93593556e-01 -5.10385096e-01
1.28054157e-01 1.51602586e-03 -9.89972949e-01 1.69066668e-01
-1.29212534e+00 5.95275819e-01 5.07139683e-01 3.60525608e-01
-1.36703953e-01 -5.10052800e-01 5.40875234e-02 -6.55223355e-02
-7.50099182e-01 9.21599716e-02 -4.90782946e-01 3.85115027e-01
-3.55593264e-02 2.79326588e-01 -1.27779019e+00 -8.09377611e-01
-5.95388174e-01 4.57176059e-01 5.15119433e-02 8.12551975e-01
6.49994850e-01 -1.32728291e+00 -1.35360038e+00 -6.16378725e-01
2.81249911e-01 -6.71048641e-01 1.84506819e-01 6.05252326e-01
-8.32334816e-01 1.12926149e+00 -2.42795935e-03 2.50210851e-01
-1.01404786e+00 1.34197161e-01 5.67126751e-01 -1.11344182e+00
-5.66212177e-01 9.97120440e-01 -4.14266825e-01 -9.05473650e-01
1.56695694e-01 -4.47771102e-01 -9.60174859e-01 -4.34290677e-01
7.10284770e-01 3.46642673e-01 5.88629007e-01 -2.23423049e-01
-3.10487151e-01 3.89903709e-02 -7.88224280e-01 -4.05913651e-01
1.12034726e+00 -2.55523145e-01 -2.28298262e-01 3.61020923e-01
1.29440188e+00 4.58011255e-02 -2.95548320e-01 -2.13969022e-01
2.05615088e-01 -1.93446457e-01 3.23843002e-01 -1.64026999e+00
-2.38016114e-01 5.95051348e-01 3.22381973e-01 1.90342206e-03
6.79395556e-01 -3.74228895e-01 9.31618512e-01 1.29194224e+00
2.27313623e-01 -1.00829959e+00 -1.34020224e-01 1.20841861e+00
9.32688117e-01 -9.18002605e-01 2.31464893e-01 6.19666874e-01
-2.73091078e-01 9.52165425e-01 7.09753156e-01 -1.50308609e-01
4.25531894e-01 -6.15053058e-01 1.30861998e-01 -2.00497121e-01
-9.83232617e-01 -2.69514732e-02 7.82030046e-01 1.71550751e-01
1.04907417e+00 8.62075365e-04 -7.85677373e-01 5.60682595e-01
-4.45196748e-01 1.69655934e-01 8.35522473e-01 8.72400939e-01
-2.51546264e-01 -1.17095780e+00 3.45339477e-01 7.67414570e-01
-6.04515553e-01 -8.87219489e-01 -3.72077674e-01 5.54130852e-01
-7.43089199e-01 1.19197392e+00 -1.60114601e-01 -7.33418703e-01
3.54238719e-01 -1.47091627e-01 6.76199317e-01 -6.14316225e-01
-8.95366848e-01 -4.81326312e-01 8.37503731e-01 -4.66762781e-01
-4.92288738e-01 -4.57499683e-01 -9.35426652e-01 -2.24713638e-01
-4.24634516e-01 9.00247991e-01 5.70048094e-01 8.98556232e-01
4.34053123e-01 9.64874864e-01 5.06025791e-01 -8.39285493e-01
-6.88014686e-01 -1.69058228e+00 -4.32086468e-01 3.15116078e-01
5.37460268e-01 -3.06400001e-01 -3.93498629e-01 -7.35620022e-01] | [11.494245529174805, 7.683866024017334] |
2aad236f-764d-4de0-89c5-f1292ffe7dea | on-boosting-single-frame-3d-human-pose | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Li_On_Boosting_Single-Frame_3D_Human_Pose_Estimation_via_Monocular_Videos_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Li_On_Boosting_Single-Frame_3D_Human_Pose_Estimation_via_Monocular_Videos_ICCV_2019_paper.pdf | On Boosting Single-Frame 3D Human Pose Estimation via Monocular Videos | The premise of training an accurate 3D human pose estimation network is the possession of huge amount of richly annotated training data. Nonetheless, manually obtaining rich and accurate annotations is, even not impossible, tedious and slow. In this paper, we propose to exploit monocular videos to complement the training dataset for the single-image 3D human pose estimation tasks. At the beginning, a baseline model is trained with a small set of annotations. By fixing some reliable estimations produced by the resulting model, our method automatically collects the annotations across the entire video as solving the 3D trajectory completion problem. Then, the baseline model is further trained with the collected annotations to learn the new poses. We evaluate our method on the broadly-adopted Human3.6M and MPI-INF-3DHP datasets. As illustrated in experiments, given only a small set of annotations, our method successfully makes the model to learn new poses from unlabelled monocular videos, promoting the accuracies of the baseline model by about 10%. By contrast with previous approaches, our method does not rely on either multi-view imagery or any explicit 2D keypoint annotations.
| [' Peilin Jiang', ' Fei Wang', ' Xuan Wang', 'Zhi Li'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['weakly-supervised-3d-human-pose-estimation'] | ['computer-vision'] | [-1.83897801e-02 1.66282237e-01 -1.21894889e-01 -2.93839216e-01
-9.45673525e-01 -6.25558734e-01 4.10470575e-01 -4.24989045e-01
-7.46690571e-01 7.48318613e-01 8.88722390e-02 3.36076885e-01
4.88414049e-01 -1.34573013e-01 -9.24450099e-01 -4.22611237e-01
1.53684229e-01 7.38550425e-01 2.90577173e-01 -5.60597889e-02
-6.50244430e-02 3.82652223e-01 -1.46067810e+00 -6.72634393e-02
5.05280077e-01 1.11091638e+00 3.48810405e-01 6.84127450e-01
5.27320862e-01 7.41563201e-01 -3.84795040e-01 -3.55278313e-01
4.94804859e-01 -1.45603910e-01 -8.73286784e-01 5.49794555e-01
5.51572502e-01 -8.62560630e-01 -5.22505283e-01 7.85696924e-01
5.49330473e-01 9.98285562e-02 3.47931296e-01 -1.16870868e+00
-1.88559722e-02 -3.09636444e-02 -6.22505665e-01 -7.37581179e-02
8.14823568e-01 2.38226488e-01 6.43733561e-01 -1.25370586e+00
9.43797350e-01 9.74867821e-01 7.47789919e-01 5.81297755e-01
-8.63354862e-01 -4.44829911e-01 2.14531183e-01 7.02030957e-02
-1.62210357e+00 -5.16569018e-01 6.56621397e-01 -4.97866213e-01
8.20743084e-01 5.01117297e-02 9.84598994e-01 1.24498475e+00
-1.85416043e-01 9.98412490e-01 7.35642314e-01 -2.78854966e-01
-3.57468636e-03 1.99729595e-02 -2.38346636e-01 8.01538646e-01
1.32380456e-01 4.74513993e-02 -5.56088030e-01 5.33522330e-02
8.39344442e-01 8.05324316e-02 -4.57746089e-01 -8.60382617e-01
-1.40304279e+00 3.56111676e-01 5.09742796e-01 -1.17624015e-01
-4.38436240e-01 1.29227683e-01 2.88415760e-01 -7.33447224e-02
5.58108270e-01 2.41728351e-01 -7.00332224e-01 -2.00838730e-01
-8.47848237e-01 4.14786339e-01 4.75642085e-01 1.14236903e+00
8.59215319e-01 -3.83602351e-01 1.22093439e-01 5.12247145e-01
1.68222129e-01 5.56195557e-01 1.06437564e-01 -1.05959570e+00
6.68224454e-01 6.37968898e-01 5.74053943e-01 -1.01252294e+00
-5.61561465e-01 -4.94287342e-01 -5.56725383e-01 -1.46186665e-01
6.26550317e-01 -2.03948110e-01 -9.31577623e-01 1.53363574e+00
7.75883019e-01 3.40772122e-02 -2.59069204e-01 1.21498096e+00
4.28130180e-01 3.92981380e-01 -2.97083527e-01 -1.59467291e-02
1.04593229e+00 -1.21240687e+00 -4.42953497e-01 -3.79879475e-01
7.75479496e-01 -5.91918528e-01 7.99459159e-01 4.19592768e-01
-1.00961435e+00 -7.05467403e-01 -8.51991057e-01 -1.75077528e-01
4.71877605e-02 5.46083748e-01 3.49853456e-01 2.01105520e-01
-8.73904943e-01 3.61602396e-01 -1.03763592e+00 -5.19665480e-01
2.45521411e-01 3.52318138e-01 -8.72338474e-01 -4.33180094e-01
-9.55107212e-01 1.00135684e+00 5.36773622e-01 5.05824983e-01
-1.12036347e+00 -3.54874760e-01 -1.03752851e+00 -3.30296904e-01
9.03633952e-01 -8.33404541e-01 1.37114799e+00 -7.50243068e-01
-1.36422646e+00 9.20906007e-01 -5.50215840e-02 -2.14826465e-01
9.34587121e-01 -8.12455654e-01 3.50839198e-01 3.37468237e-01
9.79416743e-02 1.08285081e+00 6.18814111e-01 -1.45693636e+00
-7.10477889e-01 -5.61492920e-01 3.19571465e-01 5.07605493e-01
-2.15568691e-02 -3.33467543e-01 -1.04972398e+00 -3.09895456e-01
2.14366958e-01 -1.31917155e+00 -3.24258834e-01 1.12929702e-01
-5.10068655e-01 9.55172032e-02 6.31725132e-01 -9.09445584e-01
7.42830873e-01 -1.81693649e+00 6.68381929e-01 4.73119989e-02
3.24717075e-01 1.13079719e-01 7.49659464e-02 2.87247628e-01
2.03870058e-01 -4.45974320e-01 -1.17903352e-01 -6.24907970e-01
-8.72182325e-02 2.18387142e-01 1.40670866e-01 7.39993334e-01
2.95082177e-03 9.92594361e-01 -9.68570888e-01 -5.16595304e-01
3.48411173e-01 5.17708540e-01 -7.80975223e-01 5.16594827e-01
-1.84339941e-01 9.99773502e-01 -2.54619479e-01 6.55672073e-01
4.56590831e-01 -4.98473465e-01 2.57752359e-01 -3.16191435e-01
-2.24517891e-03 -3.45527492e-02 -1.27065003e+00 2.31457305e+00
-2.11928457e-01 3.23377579e-01 -1.33708999e-01 -9.03529584e-01
4.76720452e-01 3.71293902e-01 6.63860261e-01 -2.80661672e-01
3.04601789e-01 2.17298269e-01 -3.36805463e-01 -6.93899870e-01
4.56476629e-01 1.81927413e-01 -2.36365050e-01 1.88927457e-01
3.90089542e-01 -1.92372426e-02 2.76821805e-03 8.17204416e-02
9.89495993e-01 7.37033367e-01 4.26339120e-01 2.93865293e-01
5.67120135e-01 7.81418458e-02 6.15273297e-01 3.75305444e-01
-3.28999251e-01 7.78023124e-01 2.10670784e-01 -7.59496450e-01
-1.32401729e+00 -8.92754436e-01 2.22496957e-01 8.38025153e-01
1.98816791e-01 -5.39757252e-01 -9.25359726e-01 -9.85022366e-01
-1.11916833e-01 2.54265927e-02 -5.38512588e-01 1.78472728e-01
-6.03608429e-01 -4.60726857e-01 4.51712221e-01 6.22362018e-01
7.30806947e-01 -6.04767263e-01 -8.02596271e-01 -1.44083366e-01
-6.49779320e-01 -1.55708742e+00 -6.21341765e-01 -2.44836688e-01
-6.21687889e-01 -1.21294105e+00 -1.10832262e+00 -6.45036221e-01
8.91354620e-01 3.59537214e-01 8.35638404e-01 1.09517857e-01
-2.46410027e-01 4.15112615e-01 -4.16620642e-01 -8.12911242e-02
5.53104095e-02 1.97242722e-01 2.96526432e-01 2.61791628e-02
2.13274926e-01 -3.27948898e-01 -7.19438612e-01 4.44681793e-01
-4.69530195e-01 3.51775587e-01 6.54733002e-01 8.28984916e-01
6.04928076e-01 -3.46778065e-01 1.88670814e-01 -6.72823966e-01
-2.94111788e-01 -1.12849005e-01 -5.97009122e-01 5.21978289e-02
-6.69098198e-02 -1.75128996e-01 3.11513245e-01 -3.93347114e-01
-1.00992346e+00 8.78259718e-01 -1.59204096e-01 -7.59490728e-01
-3.03107411e-01 4.17366624e-01 -2.56714106e-01 -1.39607862e-01
5.63073874e-01 6.78892061e-02 -4.21656892e-02 -5.46198726e-01
3.63437414e-01 5.36940336e-01 7.91126490e-01 -3.91565233e-01
1.06546187e+00 6.04660571e-01 -2.08873078e-01 -6.18033290e-01
-1.29405570e+00 -6.36562467e-01 -1.44980419e+00 -4.85624224e-01
1.02135468e+00 -1.26501572e+00 -5.66497386e-01 6.33825779e-01
-1.24405384e+00 -2.82160610e-01 8.04215893e-02 6.97918594e-01
-7.45260358e-01 5.46674371e-01 -5.06588399e-01 -7.92741179e-01
6.33308478e-03 -1.15599847e+00 1.57153153e+00 -2.74826586e-01
-2.53757775e-01 -6.66105449e-01 1.51615530e-01 7.68574178e-01
-1.55395776e-01 5.42247355e-01 2.22086281e-01 -3.59669685e-01
-7.34426081e-01 -5.98257065e-01 -4.26814295e-02 2.58425802e-01
1.22044664e-02 -4.23575729e-01 -9.88374114e-01 -4.36565459e-01
1.68277342e-02 -7.18187571e-01 6.30527556e-01 1.15503885e-01
9.87796485e-01 -1.45876974e-01 -3.99658591e-01 6.19763196e-01
1.05932724e+00 -1.65356576e-01 3.89362514e-01 2.65793890e-01
1.15799403e+00 5.82016110e-01 9.81047511e-01 4.53527302e-01
6.92293942e-01 9.74857628e-01 5.10963976e-01 -2.85271481e-02
2.63756514e-01 -5.59695899e-01 3.02372485e-01 8.20751131e-01
-4.58357662e-01 3.26791294e-02 -9.89935696e-01 3.76763165e-01
-1.90384972e+00 -8.15233052e-01 3.29990685e-01 2.24863148e+00
6.61354125e-01 1.31298434e-02 2.78966397e-01 8.34495425e-02
5.15023828e-01 2.53559407e-02 -6.16161823e-01 6.71479583e-01
3.01049352e-01 -2.96078414e-01 5.04102767e-01 5.15196204e-01
-1.29851449e+00 1.07145810e+00 5.86507082e+00 3.24440241e-01
-9.19993818e-01 1.12495467e-01 3.24449956e-01 -4.67666090e-01
2.74055272e-01 -3.88961323e-02 -7.49106109e-01 1.89221159e-01
5.09706855e-01 3.23646545e-01 3.34841192e-01 8.80736232e-01
1.47266224e-01 -2.99262404e-01 -1.26618397e+00 1.28590035e+00
3.60749394e-01 -9.59478617e-01 -9.33359042e-02 1.70036316e-01
8.05659533e-01 -1.50947943e-02 -2.45463341e-01 2.16868028e-01
1.76289584e-02 -8.65767002e-01 8.99173379e-01 4.62853253e-01
7.80419111e-01 -8.22470784e-01 8.79862368e-01 8.62715721e-01
-1.09152448e+00 -2.36509331e-02 -3.03453922e-01 -2.84633160e-01
3.99441630e-01 3.88536602e-01 -7.31221199e-01 7.13904381e-01
7.68668711e-01 9.03194129e-01 -6.28086746e-01 9.13883626e-01
-3.51622403e-01 6.84668124e-02 -3.04266632e-01 4.23234820e-01
1.09474964e-01 6.50002137e-02 3.20416063e-01 6.82821810e-01
2.61941403e-01 1.54005140e-01 7.01152325e-01 4.01800931e-01
-2.10284442e-01 -8.07425752e-02 -6.55670345e-01 2.12717906e-01
2.87053198e-01 1.28752232e+00 -5.71445644e-01 -3.68398815e-01
-4.79743689e-01 1.29437411e+00 6.33023620e-01 2.02353761e-01
-9.35292244e-01 1.07041188e-01 2.84220368e-01 8.97682011e-02
2.36036867e-01 -4.88442183e-01 2.35765725e-01 -1.56885374e+00
3.15072507e-01 -9.06498015e-01 2.12446526e-01 -9.61044312e-01
-8.91982377e-01 6.10747337e-01 2.51476318e-01 -1.25153482e+00
-5.75201929e-01 -6.31877482e-01 3.82666439e-02 5.23616552e-01
-1.29625094e+00 -1.34888089e+00 -7.64878750e-01 4.45048630e-01
5.78732491e-01 1.01137392e-01 5.10245502e-01 4.06330764e-01
-5.04434943e-01 3.88317406e-01 -3.60212535e-01 3.66959721e-01
9.20365214e-01 -1.04855287e+00 4.84340698e-01 7.38395751e-01
9.72428396e-02 3.85184944e-01 6.13140941e-01 -6.61150753e-01
-1.43242872e+00 -1.06480408e+00 7.75067151e-01 -1.02908051e+00
2.26681873e-01 -5.34642696e-01 -6.76555216e-01 1.04826379e+00
-1.84734777e-01 2.03552112e-01 3.75900328e-01 -3.55005711e-02
-1.50020331e-01 1.52012929e-01 -9.08581793e-01 4.51847047e-01
1.46111822e+00 -3.58636767e-01 -5.54158449e-01 4.80590761e-01
7.38559783e-01 -1.05355299e+00 -8.25267315e-01 3.79418999e-01
6.83720052e-01 -7.92619526e-01 1.08082438e+00 -4.92364079e-01
3.60581487e-01 -4.14183438e-01 -3.33067328e-01 -1.16723764e+00
2.39283927e-02 -4.11613137e-01 -1.31236598e-01 6.61011100e-01
1.61369324e-01 -1.97962582e-01 1.00707483e+00 4.79728162e-01
5.40552568e-03 -7.31975853e-01 -9.05645967e-01 -6.80799127e-01
-3.19856197e-01 -4.04234886e-01 3.63493532e-01 6.70206249e-01
-6.28022701e-02 3.22026074e-01 -8.69855642e-01 3.09307963e-01
8.45258117e-01 -1.51238039e-01 1.52260196e+00 -1.20063436e+00
-3.27022374e-01 3.85328442e-01 -4.80923474e-01 -1.49881017e+00
2.52496123e-01 -6.32873535e-01 3.17604691e-01 -1.22667277e+00
4.27881032e-01 -2.08215415e-01 1.19823672e-01 3.71574610e-01
-3.03720295e-01 5.82737744e-01 2.59804636e-01 3.50473315e-01
-9.99746323e-01 6.56518042e-01 1.38067091e+00 2.00971931e-01
-5.84407002e-02 -3.74323428e-02 -2.39152506e-01 1.06569970e+00
4.62891042e-01 -2.56146133e-01 -4.40982580e-01 -4.92624819e-01
2.69628078e-01 1.92618623e-01 8.40443015e-01 -1.09056687e+00
2.75321096e-01 -7.48511851e-02 6.85930610e-01 -8.88065636e-01
7.32119739e-01 -8.59538853e-01 2.32658148e-01 3.87132913e-01
-1.62007764e-01 1.32416382e-01 -8.91537890e-02 6.55042350e-01
-5.37053235e-02 3.67261581e-02 4.83803660e-01 -5.00734687e-01
-8.44287515e-01 5.45575559e-01 1.04373395e-01 1.97316676e-01
1.12755501e+00 -1.68345794e-01 9.97853875e-02 -4.59175169e-01
-8.14868271e-01 3.90963942e-01 8.84204924e-01 4.06868249e-01
6.10705972e-01 -1.40691483e+00 -3.67229462e-01 1.10340498e-01
2.53681958e-01 5.81313014e-01 3.52676362e-01 9.04840827e-01
-5.59991598e-01 5.84408820e-01 -2.02192396e-01 -9.32499886e-01
-1.19572520e+00 7.25493431e-01 4.11347210e-01 -2.38095030e-01
-7.13751912e-01 5.48815370e-01 3.07733715e-01 -7.71646142e-01
3.06829393e-01 -1.74592242e-01 4.25893404e-02 -1.27307281e-01
4.80610430e-01 3.87608051e-01 -5.32048419e-02 -1.00804472e+00
-3.17917675e-01 7.54932284e-01 1.86626706e-03 -2.66019762e-01
1.36611283e+00 -3.05599868e-01 2.15795472e-01 5.05055904e-01
1.28081667e+00 -2.05640405e-01 -1.75997603e+00 -2.17191055e-01
-2.47668162e-01 -7.64593363e-01 -3.41031760e-01 -5.49472451e-01
-9.71730769e-01 9.04862642e-01 2.71584392e-01 -5.61401308e-01
8.81832480e-01 4.65361178e-02 8.97123396e-01 6.44127905e-01
9.73864198e-01 -1.12778020e+00 3.77437562e-01 4.68253374e-01
8.21471810e-01 -1.45412922e+00 6.00501485e-02 -5.30202150e-01
-7.02127934e-01 8.46905231e-01 8.85099053e-01 -9.86436680e-02
2.99456060e-01 -9.31902155e-02 1.54370487e-01 -8.42891112e-02
-5.35340309e-01 -2.18598202e-01 3.56384009e-01 4.87002045e-01
1.63945973e-01 -2.84832239e-01 1.62109807e-01 2.62183785e-01
-1.72053680e-01 2.75781274e-01 3.96272510e-01 9.68330204e-01
-4.21157569e-01 -8.89029741e-01 -4.12694842e-01 5.16453050e-02
-2.79591680e-01 3.41567129e-01 -4.15716261e-01 1.04507327e+00
4.37779799e-02 6.82296574e-01 -1.33182511e-01 -6.06770337e-01
5.23677409e-01 1.04596570e-01 7.12747216e-01 -6.51793480e-01
-1.20527603e-01 7.41156563e-02 2.53046658e-02 -8.24036121e-01
-6.93148136e-01 -7.54939079e-01 -1.06199718e+00 -1.97547197e-01
-3.23763043e-01 -1.16586737e-01 3.32311422e-01 1.31455743e+00
2.58560836e-01 2.49015316e-01 3.51957083e-01 -1.58003426e+00
-5.97933292e-01 -9.13817108e-01 -2.70159036e-01 5.61637163e-01
3.55419576e-01 -1.14071918e+00 -1.55236736e-01 2.28173867e-01] | [7.080169677734375, -0.9424672722816467] |
fa8aafb1-429e-402f-a981-d77d99c52768 | esrgan-enhanced-super-resolution-generative | 1809.00219 | null | http://arxiv.org/abs/1809.00219v2 | http://arxiv.org/pdf/1809.00219v2.pdf | ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks | The Super-Resolution Generative Adversarial Network (SRGAN) is a seminal work
that is capable of generating realistic textures during single image
super-resolution. However, the hallucinated details are often accompanied with
unpleasant artifacts. To further enhance the visual quality, we thoroughly
study three key components of SRGAN - network architecture, adversarial loss
and perceptual loss, and improve each of them to derive an Enhanced SRGAN
(ESRGAN). In particular, we introduce the Residual-in-Residual Dense Block
(RRDB) without batch normalization as the basic network building unit.
Moreover, we borrow the idea from relativistic GAN to let the discriminator
predict relative realness instead of the absolute value. Finally, we improve
the perceptual loss by using the features before activation, which could
provide stronger supervision for brightness consistency and texture recovery.
Benefiting from these improvements, the proposed ESRGAN achieves consistently
better visual quality with more realistic and natural textures than SRGAN and
won the first place in the PIRM2018-SR Challenge. The code is available at
https://github.com/xinntao/ESRGAN . | ['Xiaoou Tang', 'Yu Qiao', 'Xintao Wang', 'Shixiang Wu', 'Jinjin Gu', 'Chen Change Loy', 'Chao Dong', 'Yihao Liu', 'Ke Yu'] | 2018-09-01 | null | null | null | null | ['face-hallucination'] | ['computer-vision'] | [ 5.42773843e-01 1.54262125e-01 1.71308815e-01 -2.76891470e-01
-7.69023299e-01 -2.92305648e-01 4.63113040e-01 -7.84654737e-01
8.28170255e-02 9.10218418e-01 3.91533822e-01 4.86056320e-03
2.26708546e-01 -9.43512857e-01 -8.54323924e-01 -1.05323911e+00
2.43937761e-01 -4.04992074e-01 -8.74841437e-02 -5.41426957e-01
-2.09186748e-01 3.43431443e-01 -1.35199594e+00 4.67820078e-01
1.08006585e+00 1.01237237e+00 8.51411521e-02 5.13244569e-01
3.14202160e-01 1.12978208e+00 -6.28484011e-01 -4.68213588e-01
5.96381962e-01 -8.71353030e-01 -5.04061937e-01 7.19256420e-03
4.75893110e-01 -6.53564334e-01 -7.03200698e-01 1.29418302e+00
7.42556274e-01 2.18645081e-01 2.14913011e-01 -1.11654866e+00
-1.15327585e+00 7.31586397e-01 -7.71113634e-01 1.02905771e-02
1.12429425e-01 5.03914058e-01 7.39538431e-01 -8.30575347e-01
6.25510573e-01 1.38345563e+00 4.23303157e-01 8.06291699e-01
-1.24987280e+00 -8.64605129e-01 -4.25779112e-02 2.02177376e-01
-1.32305288e+00 -6.96383893e-01 9.24492121e-01 6.30803555e-02
2.71314263e-01 4.23868150e-01 4.09067035e-01 1.47118533e+00
-1.00832330e-02 7.44243860e-01 1.47471249e+00 -1.69802725e-01
-8.04797858e-02 -1.68536097e-01 -6.40728056e-01 3.89342606e-01
-9.85835306e-03 4.87078875e-01 -2.71586537e-01 3.25005770e-01
1.33110404e+00 2.50679888e-02 -6.55996084e-01 -1.86563790e-01
-1.16657615e+00 6.37956202e-01 8.69065106e-01 1.82718381e-01
-2.27265209e-01 2.64358282e-01 1.90429121e-01 3.75054538e-01
5.03002822e-01 3.00369799e-01 7.37636611e-02 2.28317425e-01
-7.69481778e-01 1.02644145e-01 1.05087139e-01 7.75758505e-01
6.20335340e-01 6.12716436e-01 -5.11172652e-01 1.01039529e+00
-6.22155070e-02 5.53270757e-01 2.71026611e-01 -1.13716102e+00
3.62957478e-01 1.25683591e-01 2.85335593e-02 -1.04402685e+00
9.56110805e-02 -6.99378610e-01 -1.53908682e+00 5.81647217e-01
1.39101341e-01 -1.23374268e-01 -1.08126163e+00 1.95156384e+00
1.23441055e-01 3.19556624e-01 1.74268484e-01 1.17001331e+00
9.61647153e-01 7.43731022e-01 -5.41428402e-02 3.99597026e-02
1.01679242e+00 -1.13719356e+00 -7.96684027e-01 -1.72243312e-01
-5.97861111e-02 -6.80537164e-01 1.11818039e+00 3.12274814e-01
-1.37215984e+00 -8.75430942e-01 -1.12785661e+00 -2.92128235e-01
5.44123491e-03 8.31875578e-02 5.12479961e-01 3.37801516e-01
-1.22216594e+00 6.92722440e-01 -5.43846011e-01 2.13557929e-01
6.62102699e-01 1.38806757e-02 -3.02083343e-01 -3.53973359e-01
-1.52435946e+00 6.44041657e-01 2.95948476e-01 3.84546250e-01
-9.51303184e-01 -6.50518715e-01 -9.42793846e-01 -5.65669462e-02
3.03706586e-01 -7.87773788e-01 7.95991302e-01 -1.36287951e+00
-1.72740436e+00 6.91294730e-01 8.35776627e-02 -3.76372963e-01
7.97178328e-01 -7.55268708e-02 -6.78744197e-01 1.13586791e-01
-1.75483152e-01 7.84082234e-01 1.16053629e+00 -1.48559165e+00
-2.42439166e-01 -4.99063507e-02 7.49805123e-02 2.05766112e-01
1.96027439e-02 -1.48532495e-01 -3.24381351e-01 -1.10871160e+00
-2.34872818e-01 -6.04524314e-01 -2.30516538e-01 -2.62146350e-02
-5.21855950e-01 3.12264234e-01 6.35335445e-01 -1.02605700e+00
8.57991397e-01 -2.54181075e+00 3.47853631e-01 -3.18720192e-02
4.52517122e-01 2.32101947e-01 -4.07252401e-01 1.85462292e-02
-3.44834447e-01 1.49805903e-01 -4.17250574e-01 -2.84440756e-01
-1.40792534e-01 1.03918105e-01 -4.64015573e-01 3.34094614e-01
3.29918891e-01 1.15075839e+00 -8.21786165e-01 -1.55499876e-01
3.29328030e-01 9.93850052e-01 -3.20963174e-01 2.89632142e-01
-1.11860642e-02 8.86534333e-01 -2.09033549e-01 4.36142236e-01
1.08863473e+00 -2.01165617e-01 -9.66708213e-02 -4.25305068e-01
1.00479789e-01 4.49629612e-02 -1.05341482e+00 1.67679930e+00
-5.14515817e-01 5.97382903e-01 1.62127092e-01 -7.22979367e-01
1.07256889e+00 9.14202854e-02 1.05235934e-01 -1.07775950e+00
8.84872451e-02 1.05476364e-01 -1.12911351e-01 -8.10711831e-02
5.72115183e-01 -2.27966547e-01 7.78680295e-02 6.99984506e-02
-1.48456842e-01 -3.33568491e-02 -1.42064288e-01 1.55279264e-01
7.37362325e-01 2.92033702e-01 1.34835958e-01 1.21300565e-02
6.23448908e-01 -6.16371691e-01 7.67242372e-01 5.48095822e-01
4.59311157e-02 1.07813442e+00 4.53300416e-01 -3.40128690e-01
-1.42131412e+00 -1.24457943e+00 7.12773427e-02 8.30741107e-01
3.76032054e-01 1.21986761e-03 -6.68516755e-01 -4.30955648e-01
-3.44926268e-01 5.80908597e-01 -7.47245371e-01 -3.18719029e-01
-6.24636173e-01 -6.29241288e-01 4.86522138e-01 4.86752987e-01
1.01827121e+00 -1.24771750e+00 -9.12717357e-02 -4.81062271e-02
-3.93908918e-01 -1.04568422e+00 -5.84105551e-01 -1.66055381e-01
-4.39868182e-01 -6.33473575e-01 -9.82176781e-01 -5.69438398e-01
6.03016794e-01 2.31872961e-01 1.08368480e+00 1.52651623e-01
-1.85514748e-01 -3.01259875e-01 -4.77294981e-01 -9.54973325e-02
-5.21483064e-01 -1.62850365e-01 -2.57060707e-01 1.65314093e-01
-3.73276085e-01 -8.13290238e-01 -9.82417703e-01 2.26026341e-01
-1.08535409e+00 5.81998885e-01 8.05396259e-01 1.01362693e+00
7.28707850e-01 2.36197144e-01 5.26293933e-01 -9.01275754e-01
3.15617472e-01 -3.36847782e-01 -4.02146131e-01 -2.37668818e-03
-4.14795518e-01 -6.29551383e-03 9.24417138e-01 -3.61649960e-01
-1.31416976e+00 -2.89564908e-01 -4.17578369e-01 -5.30116498e-01
-6.42765835e-02 -1.51354834e-01 -5.21452904e-01 -2.03248113e-01
4.35031831e-01 5.26146173e-01 -6.97521716e-02 -4.85775739e-01
5.68767071e-01 4.36783433e-01 8.57566416e-01 -2.97715366e-01
1.11480391e+00 5.81878543e-01 -1.07016467e-01 -3.21879804e-01
-9.87228513e-01 2.20311046e-01 -1.72127485e-01 -4.75839227e-02
7.46059775e-01 -1.18339586e+00 -5.16273379e-01 7.22968102e-01
-8.95964801e-01 -7.57123053e-01 -5.24506211e-01 4.76438645e-03
-5.30499756e-01 3.35168868e-01 -8.28229964e-01 -5.48265874e-01
-5.92496216e-01 -9.81111467e-01 9.75043654e-01 4.30797607e-01
4.69639421e-01 -6.99747562e-01 -2.29501620e-01 3.11946869e-01
9.17072117e-01 8.46139014e-01 6.09501898e-01 5.84528781e-02
-5.70646465e-01 3.01103413e-01 -5.56232512e-01 8.04202139e-01
2.00978041e-01 -2.26635247e-01 -9.57424343e-01 -5.38778603e-01
1.20016024e-01 -2.50138491e-01 1.07967508e+00 2.07602948e-01
1.43624210e+00 -4.93164599e-01 3.14295113e-01 1.22481787e+00
1.66220284e+00 4.06598412e-02 1.28100622e+00 3.16084474e-01
9.30887640e-01 3.52952123e-01 2.62851626e-01 2.44260326e-01
1.27946958e-01 6.85049653e-01 5.59975922e-01 -7.32178509e-01
-6.76894486e-01 -4.24536377e-01 5.59738278e-01 5.46003580e-01
-2.83720404e-01 -2.76563883e-01 -2.84191906e-01 2.82378882e-01
-1.50517595e+00 -1.15196013e+00 1.33976102e-01 1.94341719e+00
9.47869956e-01 4.97800708e-02 -3.69300023e-02 8.02401155e-02
8.58428419e-01 6.14946961e-01 -7.59929240e-01 -2.09310129e-01
-7.34709501e-01 4.18101966e-01 5.61923921e-01 4.98638660e-01
-8.42226982e-01 9.89584625e-01 5.32693195e+00 1.07830226e+00
-1.22464848e+00 1.86342016e-01 1.05628455e+00 -3.77219077e-03
-4.28968728e-01 -2.63611615e-01 -4.08790380e-01 6.83769703e-01
6.10641122e-01 4.21204371e-03 9.33200121e-01 5.93173027e-01
2.04221159e-01 2.50104815e-01 -5.21937072e-01 9.58890140e-01
7.59976506e-02 -1.27147663e+00 3.32410514e-01 -4.13647071e-02
1.03318560e+00 -1.32928059e-01 5.38147032e-01 1.46386638e-01
4.98617351e-01 -1.43253076e+00 6.55614674e-01 5.69828928e-01
1.37120640e+00 -9.25056040e-01 7.64722705e-01 -1.53222397e-01
-1.05945146e+00 3.95242348e-02 -5.24140179e-01 2.50866741e-01
1.60424620e-01 6.25376403e-01 -1.05577081e-01 9.37437892e-01
6.89051807e-01 8.92602503e-01 -5.50478578e-01 5.83757102e-01
-6.59607112e-01 4.65262085e-01 7.27939829e-02 7.25553036e-01
-3.27265188e-02 -2.01495290e-01 5.62870085e-01 9.82761860e-01
3.27522814e-01 1.90259069e-01 -2.79578656e-01 1.18512416e+00
-3.07553351e-01 -1.93439081e-01 -3.41592997e-01 3.04598033e-01
3.25601786e-01 1.36152935e+00 -3.23655427e-01 -2.71602690e-01
-1.92970946e-01 1.33480549e+00 8.36461410e-02 5.73019981e-01
-9.55911160e-01 -4.49387133e-01 8.16497028e-01 1.46384155e-02
4.24639910e-01 9.30857956e-02 -1.99458897e-01 -1.36028409e+00
6.22212254e-02 -1.29048836e+00 4.05836530e-04 -1.12413216e+00
-1.27217424e+00 9.55678642e-01 -3.73187959e-01 -1.30411494e+00
-3.35963294e-02 -2.25908726e-01 -6.27368033e-01 1.10091221e+00
-1.79325604e+00 -1.27729225e+00 -7.63364732e-01 7.92362392e-01
3.38194728e-01 1.77927818e-02 4.88802463e-01 4.40875351e-01
-6.19362831e-01 8.76169264e-01 1.95267364e-01 2.78531045e-01
8.60608816e-01 -1.16370857e+00 6.73197746e-01 1.12881756e+00
-2.05367133e-01 2.97009259e-01 7.25025535e-01 -3.87043804e-01
-1.10876369e+00 -1.39289880e+00 1.67966172e-01 -1.31682903e-01
4.05759364e-01 -4.45825487e-01 -1.07672322e+00 6.14050806e-01
3.31748128e-01 2.16126561e-01 1.07930072e-01 -5.72993994e-01
-5.26809216e-01 -2.81242698e-01 -1.31320274e+00 7.27241397e-01
1.18596637e+00 -4.73838061e-01 -1.00634180e-01 1.79516766e-02
1.06607342e+00 -6.77744687e-01 -9.20241654e-01 5.56282163e-01
3.39576393e-01 -1.28712785e+00 1.20241666e+00 -1.55915782e-01
1.02269757e+00 -5.47305405e-01 -8.97253156e-02 -1.38982403e+00
-4.92953449e-01 -7.96992302e-01 2.40454841e-02 1.35212564e+00
1.45221680e-01 -6.58475161e-01 5.41317046e-01 -2.71533839e-02
-5.81946410e-02 -6.52408540e-01 -5.26209831e-01 -7.64705658e-01
1.68487594e-01 4.68108617e-03 1.00292885e+00 9.51251805e-01
-6.23235047e-01 1.05200000e-01 -1.08079112e+00 8.37885067e-02
9.17611361e-01 3.92156988e-02 7.86107898e-01 -5.72252035e-01
-5.49888730e-01 -3.06990653e-01 -2.75542706e-01 -1.07822788e+00
-1.13810703e-01 -7.14724243e-01 7.12075830e-02 -1.27289140e+00
2.78039634e-01 -3.69611919e-01 -4.93271679e-01 4.36220199e-01
-3.76146019e-01 8.02013814e-01 3.28576326e-01 1.58188939e-01
-4.62960869e-01 7.79155552e-01 1.83204877e+00 -2.04307511e-02
1.27661489e-02 -2.43621260e-01 -1.12341177e+00 3.46191883e-01
9.42005694e-01 -1.21875666e-01 -2.77071625e-01 -5.19602060e-01
1.48547679e-01 1.20468669e-01 6.98787808e-01 -8.38862121e-01
-1.77383721e-01 -1.22253388e-01 7.22197175e-01 -9.71821845e-02
2.88187206e-01 -5.32108605e-01 6.15986645e-01 2.88245142e-01
-4.30041611e-01 -1.96834937e-01 7.71874189e-02 4.31581706e-01
-2.98303068e-01 3.80562872e-01 1.26266634e+00 -2.68473420e-02
-6.98719025e-01 4.34416711e-01 2.51399368e-01 9.99822095e-03
7.77288854e-01 -9.63072926e-02 -6.49038017e-01 -5.98639846e-01
-6.08166516e-01 -1.43438624e-03 8.14895153e-01 4.87396300e-01
7.13390768e-01 -1.55700910e+00 -1.15542936e+00 3.06788296e-01
-1.40630081e-01 1.66852087e-01 7.68726528e-01 6.66682303e-01
-4.85036016e-01 -8.08367431e-02 -6.27450228e-01 -2.78320253e-01
-8.73086274e-01 7.01945126e-01 3.86304647e-01 -2.71075994e-01
-9.52940941e-01 8.30754042e-01 5.55046260e-01 -7.01900274e-02
-5.17321900e-02 8.21505934e-02 -4.44425903e-02 -5.47054112e-01
7.79534400e-01 2.47422934e-01 -3.55950534e-01 -5.69936812e-01
2.49680467e-02 4.15092677e-01 -1.75510577e-04 1.75593141e-02
1.43680668e+00 -3.74404997e-01 -1.38335615e-01 1.46556422e-02
1.08254242e+00 1.94260374e-01 -1.70192206e+00 -3.31404895e-01
-7.11522162e-01 -6.79697156e-01 1.43437505e-01 -8.33814263e-01
-1.68340528e+00 6.00725532e-01 6.98648870e-01 -6.13495708e-02
1.68379986e+00 -2.08819941e-01 1.10945594e+00 -3.69558305e-01
1.94807097e-01 -5.44423878e-01 1.63571874e-03 1.02280416e-01
1.15003920e+00 -1.25160944e+00 -1.23567052e-01 -3.69423121e-01
-7.60073543e-01 8.15612197e-01 5.68118155e-01 -4.53346878e-01
1.59911335e-01 3.53375584e-01 5.08890860e-03 1.65655866e-01
-5.31308174e-01 -4.13189679e-02 3.56024235e-01 6.82460427e-01
2.28566617e-01 5.56391925e-02 -7.78432116e-02 3.89122128e-01
-3.15939397e-01 -1.30049214e-01 5.59888780e-01 2.94990480e-01
-5.34024797e-02 -9.66520727e-01 -2.58181781e-01 1.62423417e-01
-6.16349041e-01 -4.41201687e-01 -1.46112189e-01 6.16477251e-01
2.01570481e-01 7.82455325e-01 -8.62823874e-02 -5.52882850e-01
2.19286606e-01 -5.09053171e-01 4.53574687e-01 -2.17441931e-01
-3.61348331e-01 -1.07130257e-03 -1.93927929e-01 -8.37719977e-01
-3.96684051e-01 -3.08357686e-01 -8.33517373e-01 -6.77844524e-01
5.73086133e-03 -6.06222861e-02 3.49968404e-01 5.23903549e-01
4.48628247e-01 7.95484483e-01 9.34349716e-01 -8.09359729e-01
-3.12056839e-01 -9.46611345e-01 -6.51565075e-01 5.63485384e-01
6.05951130e-01 -3.09742987e-01 -5.78808546e-01 9.24699828e-02] | [11.277533531188965, -1.7270007133483887] |
985addc5-1b4d-430d-8e21-86c382d196b6 | alexsis-a-dataset-for-lexical-simplification | null | null | https://aclanthology.org/2022.lrec-1.383 | https://aclanthology.org/2022.lrec-1.383.pdf | ALEXSIS: A Dataset for Lexical Simplification in Spanish | Lexical Simplification is the process of reducing the lexical complexity of a text by replacing difficult words with easier to read (or understand) expressions while preserving the original information and meaning. In this paper we introduce ALEXSIS, a new dataset for this task, and we use ALEXSIS to benchmark Lexical Simplification systems in Spanish. The paper describes the evaluation of three kind of approaches to Lexical Simplification, a thesaurus-based approach, a single transformers-based approach, and a combination of transformers. We also report state of the art results on a previous Lexical Simplification dataset for Spanish. | ['Horacio Saggion', 'Daniel Ferrés'] | null | null | null | null | lrec-2022-6 | ['lexical-simplification'] | ['natural-language-processing'] | [ 5.56216575e-02 4.74735141e-01 -2.76516825e-02 -3.88063967e-01
-5.37914157e-01 -3.06860983e-01 5.62473714e-01 4.60250318e-01
-7.59349823e-01 1.11745405e+00 6.06319249e-01 -1.23543210e-01
-4.92839925e-02 -8.66865396e-01 -2.19128907e-01 -1.93311512e-01
7.28297949e-01 9.21602428e-01 2.14749485e-01 -1.02812433e+00
5.88191301e-02 6.85568869e-01 -1.56022799e+00 3.74755740e-01
1.00747991e+00 3.32918197e-01 -9.02581960e-02 2.64063776e-01
-6.06368542e-01 7.12854862e-01 -1.03579485e+00 -1.05056858e+00
3.16854149e-01 -3.67855728e-01 -9.62680817e-01 -4.06693012e-01
6.05802059e-01 2.29030773e-01 -3.07857305e-01 1.30963004e+00
6.26607060e-01 3.76416832e-01 6.54293597e-01 -8.33742082e-01
-6.06323123e-01 1.17110813e+00 -8.71198252e-02 6.39500916e-02
9.04413939e-01 -4.77528244e-01 9.06070471e-01 -8.03813875e-01
1.18411303e+00 1.65934706e+00 1.20944476e+00 4.80555654e-01
-1.20931590e+00 -5.81142664e-01 1.07505217e-01 3.15720201e-01
-1.83468628e+00 -4.84562129e-01 1.87473193e-01 -7.26015400e-03
1.75545871e+00 5.60446501e-01 8.86675477e-01 8.52289379e-01
4.17250663e-01 4.17737305e-01 8.58900011e-01 -9.11607265e-01
-2.12371796e-01 1.12822056e-01 4.19457406e-01 6.48781598e-01
7.22389340e-01 -3.37305516e-01 -3.41248542e-01 -1.95290148e-01
-8.08006618e-03 -2.54864484e-01 -4.07028347e-02 -2.61994526e-02
-9.25788045e-01 9.35688198e-01 -3.04142963e-02 3.51689160e-01
-3.54333341e-01 -3.62485498e-01 8.40311408e-01 7.18010306e-01
4.31073219e-01 9.24717724e-01 -6.05003476e-01 2.20975906e-01
-1.06664324e+00 5.64674258e-01 1.36150885e+00 1.38615847e+00
5.52774429e-01 8.22643638e-02 -2.38875747e-01 9.70157087e-01
-1.82630345e-01 7.53398359e-01 6.93545759e-01 -7.58225679e-01
3.83932561e-01 6.31397784e-01 -2.23475367e-01 -7.07507908e-01
-7.30643988e-01 2.73049116e-01 -7.06015766e-01 1.44357784e-02
1.21693257e-02 4.44393158e-02 -8.63560796e-01 1.67577779e+00
1.83991298e-01 -6.62050605e-01 4.25727218e-02 1.79079339e-01
1.35624194e+00 6.02688670e-01 5.01361430e-01 -6.53039455e-01
1.43310308e+00 -1.05184913e+00 -1.35784328e+00 2.07351163e-01
1.02426147e+00 -1.32152343e+00 1.31035376e+00 6.45165682e-01
-1.41544855e+00 -2.35835597e-01 -1.09396315e+00 -7.72506595e-01
-1.23366535e+00 -5.74672632e-02 5.12521148e-01 8.60327363e-01
-8.36914837e-01 8.35072100e-01 -2.38784999e-01 -9.07218516e-01
3.15777492e-03 6.12806380e-01 -7.13932037e-01 2.26769745e-01
-1.63100088e+00 1.70767057e+00 6.75318122e-01 -6.95337236e-01
-1.91418320e-01 -8.81420791e-01 -1.02993286e+00 1.30396441e-01
3.89766067e-01 -1.06344640e+00 1.39579129e+00 -8.91032577e-01
-1.69792521e+00 1.26641428e+00 -2.28675574e-01 -6.14035726e-01
2.81589299e-01 -4.59106565e-01 -5.56485713e-01 -3.56156260e-01
4.01981860e-01 5.51659822e-01 7.20942557e-01 -6.33879304e-01
-7.53719985e-01 -1.28678232e-01 1.12971589e-01 3.10299516e-01
-2.68710136e-01 3.58702898e-01 -3.47324908e-01 -1.23812521e+00
-3.21786821e-01 -8.66054296e-01 -1.16812568e-02 -5.03111959e-01
-4.84887838e-01 -4.92528319e-01 3.99011850e-01 -1.03179896e+00
1.96333420e+00 -1.84087288e+00 7.68802106e-01 1.24441735e-01
2.36206546e-01 7.15917706e-01 -6.48664460e-02 5.70510209e-01
-4.50216353e-01 5.54546595e-01 -1.95225433e-01 -7.23510027e-01
2.38232911e-01 4.53849167e-01 -1.78987563e-01 3.16942446e-02
-2.32307047e-01 7.94105768e-01 -6.46660626e-01 -8.05737734e-01
3.77562016e-01 2.73652554e-01 -6.24487400e-01 -2.95605361e-01
-9.78222769e-03 -4.41509128e-01 1.03919670e-01 6.03417695e-01
5.73173523e-01 8.19876194e-01 2.20573992e-01 -3.94645751e-01
-1.27413660e-01 6.21275961e-01 -1.17494023e+00 1.76320887e+00
-7.34081209e-01 5.23815811e-01 -3.21217626e-01 -4.12477195e-01
5.72742462e-01 3.08495015e-01 7.86714703e-02 -8.83129478e-01
3.32313538e-01 3.34535629e-01 -1.86169729e-01 -2.29728863e-01
9.36919510e-01 -4.05529410e-01 -5.62024057e-01 4.13876295e-01
4.02443349e-01 -8.04792464e-01 9.36043024e-01 2.30102181e-01
9.01013315e-01 3.48929055e-02 1.51186395e+00 -6.65596664e-01
9.22073305e-01 1.25727713e-01 4.57946807e-01 6.02821231e-01
7.12932646e-02 4.70547855e-01 3.70323539e-01 -7.70609796e-01
-1.27739537e+00 -1.05704796e+00 -1.51773281e-02 1.33511508e+00
-2.12364510e-01 -1.28527808e+00 -1.14106023e+00 -7.19582796e-01
1.46565244e-01 1.40175843e+00 -7.63695300e-01 -4.26732749e-01
-9.01001155e-01 -4.77205276e-01 1.10471976e+00 1.19798295e-01
1.34868279e-01 -1.34183609e+00 -2.72680730e-01 1.80564865e-01
-4.05232787e-01 -9.04478967e-01 -3.54781568e-01 4.02652860e-01
-5.06761551e-01 -9.48991537e-01 -4.19743545e-02 -8.23694170e-01
2.77092725e-01 1.14689365e-01 1.73986447e+00 2.44185627e-01
-3.36268127e-01 -1.98982105e-01 -5.26209593e-01 -1.03507912e+00
-7.00125098e-01 6.57975495e-01 1.38417706e-01 -8.58880222e-01
8.04152369e-01 -2.62088388e-01 3.85140717e-01 -1.81719393e-01
-1.07196093e+00 -4.44673896e-01 3.29075426e-01 7.14281082e-01
8.04874182e-01 7.95208961e-02 5.71927354e-02 -1.50043333e+00
1.19235015e+00 -1.48197934e-01 -2.08889738e-01 4.37203497e-01
-4.74024206e-01 2.27679282e-01 9.80681479e-01 -2.92807013e-01
-1.18978703e+00 -8.78016930e-03 -5.72994113e-01 6.03508800e-02
4.49543893e-02 3.49728763e-01 -4.24836934e-01 -2.27619395e-01
8.35265458e-01 -2.59231567e-01 -2.48209000e-01 -8.01707208e-01
8.47262800e-01 3.97559464e-01 3.44012469e-01 -4.09385413e-01
5.17666936e-01 4.20559853e-01 -1.90468431e-02 -1.19771421e+00
-9.94745016e-01 -3.84029955e-01 -9.82083559e-01 4.30798262e-01
6.54051304e-01 -6.15773916e-01 1.70448571e-02 1.70034111e-01
-1.31929886e+00 2.67079026e-02 -1.22283483e+00 1.00863263e-01
-4.91547883e-01 5.78434765e-01 -5.56364298e-01 1.52418897e-01
-7.63763428e-01 -9.12600338e-01 1.07514119e+00 2.04299372e-02
-1.19345045e+00 -1.09215927e+00 3.84885699e-01 -5.91246709e-02
1.66355029e-01 -3.35578248e-02 1.25743318e+00 -9.49200690e-01
4.27819490e-01 -3.30142468e-01 9.57915839e-03 3.82131308e-01
9.82893929e-02 2.43520319e-01 -5.03430784e-01 6.92693293e-02
-5.10052370e-04 -3.62481587e-02 8.12363803e-01 2.98726410e-01
5.65552175e-01 -1.40743420e-01 -2.97744691e-01 9.02015150e-01
1.15691388e+00 3.09752598e-02 1.00394177e+00 8.15674007e-01
6.86667383e-01 2.65121967e-01 7.48649359e-01 2.56314307e-01
4.06302035e-01 8.83519530e-01 -1.76469907e-01 -2.43506446e-01
-6.82365477e-01 -1.52103409e-01 1.58935115e-01 1.31086040e+00
-1.01084337e-02 -3.77621293e-01 -6.32841706e-01 3.14628184e-01
-1.54915118e+00 -7.25508749e-01 -2.28378922e-01 1.92406857e+00
1.18694091e+00 3.47838737e-02 4.12873030e-02 4.21650797e-01
4.96500731e-01 -1.49310753e-01 8.15565512e-02 -1.30834424e+00
-7.71038115e-01 9.35101986e-01 4.97448355e-01 9.20896113e-01
-1.23278356e+00 1.78930736e+00 7.43260241e+00 1.35978246e+00
-7.33014703e-01 3.00778031e-01 -3.38413835e-01 -1.05793133e-01
-1.99740455e-01 3.08769364e-02 -9.91437793e-01 -5.88448159e-02
7.77245998e-01 -7.42074788e-01 4.54604656e-01 5.86795449e-01
-5.89210652e-02 -9.95039940e-02 -9.77896094e-01 8.10592055e-01
6.47574902e-01 -9.57425952e-01 9.79812980e-01 -4.31068420e-01
6.25752211e-01 -3.68809670e-01 -4.31614459e-01 6.49117291e-01
3.59336317e-01 -1.03170335e+00 9.71280336e-01 7.02415347e-01
6.81438029e-01 -8.60057890e-01 9.73305464e-01 -3.98045890e-02
-1.25075555e+00 2.74042636e-01 -6.50055647e-01 -3.24962020e-01
1.83768734e-01 2.63856947e-01 -2.50186801e-01 6.71992004e-01
7.68431306e-01 5.84960759e-01 -1.01965499e+00 8.44398499e-01
-5.65543532e-01 -6.70272484e-02 -4.91818398e-01 -8.70227963e-02
-1.40750855e-01 -2.31400460e-01 8.40006471e-01 1.83652627e+00
6.41127899e-02 1.38297230e-01 -3.97642888e-02 2.00816423e-01
-2.22721294e-01 9.89853263e-01 -9.31568980e-01 2.83843607e-01
5.92292070e-01 1.11108088e+00 -6.75538421e-01 -8.85055482e-01
-2.84133703e-01 9.79263484e-01 3.54485810e-01 -1.47497341e-01
-7.30915844e-01 -1.12832260e+00 6.24715745e-01 -1.35685012e-01
1.55196130e-01 2.14810058e-01 -8.23392034e-01 -1.29313004e+00
-1.29601970e-01 -1.44338644e+00 7.23287523e-01 -7.67967224e-01
-1.08356524e+00 7.14754343e-01 3.50476235e-01 -7.82739401e-01
-4.38663624e-02 -4.27199185e-01 -2.68405467e-01 9.29532468e-01
-1.19015658e+00 -1.06511486e+00 -1.73462644e-01 5.59305251e-01
8.11735570e-01 -3.89201075e-01 1.34544647e+00 5.32881796e-01
-4.69935298e-01 7.89697051e-01 -1.49308704e-02 -4.47426707e-01
1.01557779e+00 -1.35948706e+00 6.41376495e-01 7.14140177e-01
-4.43546809e-02 7.85846293e-01 9.47642088e-01 -8.05933595e-01
-7.31265008e-01 -1.07164776e+00 1.89921689e+00 -6.59409106e-01
7.09815502e-01 -1.59775674e-01 -8.88683558e-01 9.49731112e-01
5.92767119e-01 -9.59158123e-01 7.36599445e-01 1.02686577e-01
-3.79185915e-01 -7.53554627e-02 -1.30592167e+00 1.06798244e+00
1.25295782e+00 -3.83394808e-01 -1.45254922e+00 4.50066626e-01
9.86153245e-01 -3.73010427e-01 -1.05863798e+00 3.27978581e-01
5.13645351e-01 -7.15765834e-01 1.07557762e+00 -8.04506898e-01
1.03186511e-01 -3.22035216e-02 -2.31230464e-02 -1.70505476e+00
-5.19944072e-01 -8.73637378e-01 3.09997231e-01 1.07902217e+00
5.28373003e-01 -5.06320357e-01 2.68336773e-01 4.60598059e-02
-3.16087931e-01 -2.73385584e-01 -9.24887061e-01 -6.46016479e-01
5.79429746e-01 -3.03733628e-02 8.73408377e-01 1.01901805e+00
1.98389068e-01 8.30334067e-01 -1.27912894e-01 -4.69437689e-01
2.22239047e-01 -4.02446985e-01 8.07775974e-01 -1.43924975e+00
4.81221437e-01 -8.74349475e-01 -5.35706997e-01 -3.03183317e-01
5.78201711e-01 -1.07970726e+00 -4.91347015e-01 -1.36517000e+00
2.79259868e-02 1.22892208e-01 3.67582738e-01 5.34422278e-01
-2.27700233e-01 3.23046088e-01 4.96780783e-01 6.02369942e-02
-6.99969053e-01 5.18640697e-01 7.79063761e-01 -1.45254001e-01
-1.58266678e-01 -4.68138874e-01 -7.27298021e-01 1.03730309e+00
6.86810434e-01 -9.41029489e-01 -2.75827825e-01 -5.14423788e-01
5.44267237e-01 -6.92060530e-01 -3.90869737e-01 -7.71771312e-01
5.20783663e-02 1.43500552e-01 -4.46354412e-02 -5.53190410e-01
1.81013465e-01 -6.56132817e-01 2.68013626e-01 5.89343369e-01
-9.45543498e-02 6.20279372e-01 5.71846366e-01 -2.44566426e-01
-2.93589473e-01 -6.53063178e-01 1.02688551e+00 -3.26557159e-01
-6.30520701e-01 -2.44863167e-01 -6.82337403e-01 3.91525507e-01
1.05172706e+00 -1.46966681e-01 -1.70402095e-01 -1.04778543e-01
-1.08935881e+00 -1.36658773e-01 6.33550525e-01 4.89614099e-01
2.61979550e-01 -1.21550608e+00 -7.01038480e-01 1.88423321e-01
1.34427175e-02 -5.50332844e-01 -3.18120658e-01 5.86144269e-01
-1.31186092e+00 7.56425083e-01 -5.09993672e-01 2.02633560e-01
-1.61258972e+00 7.62240827e-01 3.47952306e-01 -6.16770387e-01
-5.55696011e-01 7.33809352e-01 -6.78797811e-02 -8.63984048e-01
2.46752158e-01 -6.53309405e-01 -6.91392958e-01 4.15873885e-01
4.66130555e-01 7.97409534e-01 7.18272150e-01 -8.50692987e-01
-2.84721404e-01 7.83952534e-01 -1.59557730e-01 1.62118122e-01
8.93524408e-01 -3.11072707e-01 -7.10712075e-01 3.84154409e-01
8.70833158e-01 3.11440855e-01 2.80546308e-01 -3.32053602e-01
2.58013159e-01 -7.90705904e-03 -1.46180928e-01 -7.85653412e-01
-6.80780590e-01 6.06745899e-01 8.39791968e-02 3.95874888e-01
1.25297999e+00 -3.80551994e-01 9.05011535e-01 1.01221001e+00
2.09319368e-01 -1.54774439e+00 -7.33408809e-01 1.01723146e+00
1.04166245e+00 -7.45974600e-01 6.19428456e-01 -9.23693955e-01
-6.67473674e-01 1.18548250e+00 1.53134376e-01 -2.59960890e-01
8.09072971e-01 3.82094622e-01 1.43921345e-01 -1.91701666e-01
-7.02419937e-01 -5.20091951e-01 1.72112972e-01 7.49508023e-01
5.55699766e-01 2.66092181e-01 -1.16044581e+00 8.40257049e-01
-9.51175451e-01 -1.61329895e-01 5.50077617e-01 9.22541380e-01
-2.62661874e-01 -1.45934606e+00 -3.42938185e-01 3.97306561e-01
-5.81720352e-01 -7.34843552e-01 -1.23432457e+00 1.51207280e+00
4.63649482e-01 6.25516355e-01 -2.12022141e-01 -1.61842972e-01
1.37384713e+00 3.35649550e-01 6.70612514e-01 -1.02919376e+00
-1.12666738e+00 -1.37629539e-01 6.38867617e-01 -5.31928420e-01
-1.46558926e-01 -8.37618351e-01 -1.06632376e+00 -1.00013328e+00
-2.40565106e-01 1.17555782e-01 4.79113996e-01 9.32024837e-01
1.24512322e-01 6.05890214e-01 -1.01204447e-01 -8.18929732e-01
-4.37412471e-01 -1.02815354e+00 -5.67052603e-01 7.37033963e-01
-2.95756161e-01 -7.62328506e-01 -2.26683840e-01 2.11409256e-01] | [10.917851448059082, 10.431655883789062] |
1de07644-633f-4dc1-9ae3-d9046d237803 | text-to-face-generation-with-stylegan2 | 2205.12512 | null | https://arxiv.org/abs/2205.12512v1 | https://arxiv.org/pdf/2205.12512v1.pdf | Text-to-Face Generation with StyleGAN2 | Synthesizing images from text descriptions has become an active research area with the advent of Generative Adversarial Networks. The main goal here is to generate photo-realistic images that are aligned with the input descriptions. Text-to-Face generation (T2F) is a sub-domain of Text-to-Image generation (T2I) that is more challenging due to the complexity and variation of facial attributes. It has a number of applications mainly in the domain of public safety. Even though several models are available for T2F, there is still the need to improve the image quality and the semantic alignment. In this research, we propose a novel framework, to generate facial images that are well-aligned with the input descriptions. Our framework utilizes the high-resolution face generator, StyleGAN2, and explores the possibility of using it in T2F. Here, we embed text in the input latent space of StyleGAN2 using BERT embeddings and oversee the generation of facial images using text descriptions. We trained our framework on attribute-based descriptions to generate images of 1024x1024 in resolution. The images generated exhibit a 57% similarity to the ground truth images, with a face semantic distance of 0.92, outperforming state-of-the-artwork. The generated images have a FID score of 118.097 and the experimental results show that our model generates promising images. | ['Sarasi Munasinghe', 'D. M. A. Ayanthi'] | 2022-05-25 | null | null | null | null | ['text-to-face-generation'] | ['computer-vision'] | [ 4.56115931e-01 4.69194174e-01 2.81505674e-01 -3.49378496e-01
-6.42299056e-01 -4.37619686e-01 9.66966748e-01 -1.02167487e+00
9.24693048e-02 8.28630805e-01 2.00329587e-01 3.32765281e-01
2.11505219e-01 -1.08696258e+00 -8.16456378e-01 -8.09929430e-01
4.31281924e-01 4.01796967e-01 -3.03031445e-01 -2.92157233e-01
-1.15302674e-01 5.71699202e-01 -1.66472173e+00 4.32385653e-01
6.69525445e-01 1.11967361e+00 -5.56320325e-02 3.69792432e-01
-1.54467747e-01 3.83088738e-01 -8.41705441e-01 -9.09691870e-01
5.91050088e-01 -6.98849499e-01 -4.79442090e-01 3.32733601e-01
6.99900210e-01 -4.89260197e-01 -4.58247691e-01 1.16663933e+00
6.19874597e-01 -2.67747313e-01 8.70510340e-01 -1.77635598e+00
-9.89651918e-01 1.21246457e-01 -4.67118710e-01 -5.33634901e-01
3.72407705e-01 2.22815737e-01 3.88320595e-01 -9.65482652e-01
9.27634835e-01 1.72746289e+00 3.06602478e-01 1.09594703e+00
-1.17542911e+00 -1.17583406e+00 -4.16770339e-01 -7.28423055e-03
-1.62817538e+00 -7.71646321e-01 6.23522103e-01 -3.89210045e-01
2.79074013e-01 1.50558293e-01 3.06144714e-01 1.51743686e+00
2.17858523e-01 3.43572795e-01 1.25153100e+00 -4.52958047e-01
4.77036946e-02 3.61439914e-01 -1.07546806e+00 6.70760870e-01
6.29533976e-02 4.02164340e-01 -5.10528505e-01 1.27945051e-01
9.88490641e-01 -3.03190928e-02 -1.50391087e-01 1.09885320e-01
-1.02258825e+00 1.03857994e+00 3.32844824e-01 8.42907950e-02
-3.83812517e-01 2.04012737e-01 -5.06806076e-02 1.99537069e-01
4.40493584e-01 3.17943126e-01 4.15089875e-01 3.73364657e-01
-9.18216050e-01 4.56056625e-01 4.28759962e-01 1.39609027e+00
7.66971767e-01 5.66340208e-01 -4.71712381e-01 6.93811178e-01
3.47135067e-01 1.06242847e+00 4.42355096e-01 -9.50823605e-01
3.01578790e-01 4.55004156e-01 6.32114783e-02 -1.10557973e+00
3.65012378e-01 1.77805573e-02 -9.97571886e-01 5.78279018e-01
6.85743093e-02 -2.11755574e-01 -1.11067164e+00 1.69708228e+00
2.73068041e-01 1.66069731e-01 2.78272599e-01 7.38661230e-01
9.98611569e-01 7.87242413e-01 -1.05217293e-01 -6.09224215e-02
1.15210915e+00 -9.70486462e-01 -1.03773248e+00 -1.24630891e-01
-9.72063616e-02 -1.11812854e+00 7.75855005e-01 2.77199186e-02
-1.03266418e+00 -7.61844695e-01 -8.49254370e-01 2.09876746e-01
-2.60345668e-01 2.19919592e-01 1.64165869e-01 9.04555202e-01
-1.27074468e+00 2.36854821e-01 -1.57222688e-01 -4.20192599e-01
6.19550467e-01 2.10006624e-01 -7.81557918e-01 -2.45833918e-01
-1.24720573e+00 6.41625762e-01 3.99339527e-01 -3.69520634e-02
-1.33290386e+00 -4.92674381e-01 -8.27917755e-01 -1.09460875e-01
2.29748040e-01 -6.39973998e-01 7.56642282e-01 -1.23643482e+00
-1.55651009e+00 9.74729717e-01 -1.04865864e-01 -1.94217950e-01
8.16516757e-01 1.79182172e-01 -5.70668757e-01 2.82320529e-01
3.43981177e-01 1.28662539e+00 1.37178004e+00 -1.40215254e+00
-4.07982975e-01 -2.42584169e-01 -1.09499343e-01 -1.86675621e-04
-3.47512513e-01 1.88989285e-02 -4.14044976e-01 -8.15630794e-01
-3.07392180e-01 -1.23041153e+00 5.81542738e-02 1.94180727e-01
-6.10075772e-01 5.48119992e-02 1.10145926e+00 -5.81702292e-01
6.00864589e-01 -2.16532779e+00 -1.62747160e-01 -9.35663097e-03
-8.87021981e-03 3.92752737e-01 -4.37921405e-01 5.69767773e-01
-1.45470813e-01 4.23911035e-01 -1.74483716e-01 -4.18346465e-01
-5.14452197e-02 7.49586448e-02 -5.25598645e-01 2.46532142e-01
5.37285030e-01 1.05184805e+00 -5.39219201e-01 -5.98667741e-01
1.99663669e-01 7.74479151e-01 -4.94258106e-01 5.74233174e-01
-2.18470693e-01 6.24018133e-01 -3.58902723e-01 6.55563116e-01
9.08421278e-01 1.97694853e-01 -1.34289443e-01 -2.45731547e-01
2.27368459e-01 -5.51997542e-01 -9.58322644e-01 1.52603590e+00
-3.59694451e-01 6.24009788e-01 -2.27870047e-01 -5.52218854e-01
1.29410613e+00 5.17950177e-01 2.97245711e-01 -5.96963465e-01
5.58395721e-02 1.65706068e-01 -2.15436980e-01 -6.03488505e-01
2.24505916e-01 -4.19896781e-01 1.52185280e-02 2.61545122e-01
1.19735368e-01 -4.95252520e-01 5.21235280e-02 1.72614714e-03
5.89242101e-01 3.07161331e-01 -9.82985646e-02 4.89389151e-02
5.44979095e-01 -2.76909620e-01 3.13849360e-01 3.33970487e-01
1.51564628e-01 1.07805347e+00 3.16758662e-01 -2.82533854e-01
-1.47983909e+00 -1.11775696e+00 -7.68430084e-02 2.47389778e-01
-5.82564548e-02 -2.24346533e-01 -1.21941423e+00 -6.11981571e-01
-2.36993149e-01 6.38802767e-01 -7.55099714e-01 -2.24614531e-01
-3.02789569e-01 -3.25291187e-01 8.94647360e-01 2.37982452e-01
1.08637321e+00 -1.33630729e+00 -1.22580454e-01 3.79156284e-02
-3.83320957e-01 -1.50727272e+00 -5.34913540e-01 -8.88578594e-01
-3.71026099e-01 -8.78177822e-01 -8.67463827e-01 -8.93950224e-01
1.05240178e+00 4.54795128e-03 8.37412596e-01 -1.44753486e-01
-5.21671295e-01 2.91807801e-01 -3.40819031e-01 -5.98859012e-01
-8.80588531e-01 -3.03259224e-01 8.33559781e-02 5.10889769e-01
1.76568399e-03 -3.15213382e-01 -6.18766665e-01 5.29426217e-01
-1.28259540e+00 3.42782497e-01 6.66618824e-01 8.54786396e-01
5.38096368e-01 2.14375168e-01 6.90860689e-01 -7.91205347e-01
3.43160331e-01 -2.93190867e-01 -5.60999572e-01 1.86341718e-01
-4.50391442e-01 -8.17082971e-02 8.65342975e-01 -3.80840361e-01
-1.25925708e+00 3.84273715e-02 -1.47568896e-01 -7.82082558e-01
-3.24641049e-01 -1.80938214e-01 -6.81817472e-01 -8.46073478e-02
5.14663160e-01 3.63977700e-01 3.70241761e-01 5.83605394e-02
4.73091602e-01 7.12611198e-01 6.02380574e-01 -4.85684872e-01
1.36253047e+00 6.57754779e-01 1.83004156e-01 -8.30726326e-01
-4.93879169e-01 3.79203230e-01 -1.91355914e-01 -3.57230127e-01
1.12942171e+00 -9.72967505e-01 -5.45709908e-01 6.56950295e-01
-1.20082808e+00 1.20979734e-02 -1.95033357e-01 1.93920001e-01
-7.42965400e-01 1.21439192e-02 -1.26555145e-01 -6.25979245e-01
-4.17178959e-01 -1.39155686e+00 1.29069161e+00 3.00325602e-01
1.51864916e-01 -6.77562594e-01 -3.32194179e-01 6.18986189e-01
5.15945554e-01 7.80510664e-01 7.43566155e-01 -2.81363368e-01
-7.92476892e-01 -1.44610658e-01 -2.07956970e-01 5.84695458e-01
4.45257574e-01 -2.99711563e-02 -1.16604424e+00 -3.23596120e-01
-5.10358550e-02 -3.32365274e-01 3.52031678e-01 4.42271829e-02
1.14098239e+00 -6.62461221e-01 -1.57794505e-01 7.86215365e-01
1.49374914e+00 3.48901898e-01 1.22819853e+00 6.70531392e-02
6.86396718e-01 9.03266609e-01 6.99565828e-01 3.26484770e-01
9.61462408e-02 9.00265813e-01 5.90739071e-01 -2.38991782e-01
-5.06789088e-01 -6.69906616e-01 3.65496159e-01 1.22634456e-01
-9.72321033e-02 -3.72939229e-01 -5.55621386e-01 4.18377519e-01
-1.43671715e+00 -1.32808971e+00 2.14624070e-02 1.97400677e+00
7.66198754e-01 -2.39010483e-01 -1.45762011e-01 4.58239391e-03
1.07316661e+00 8.70575663e-03 -3.67325097e-01 -3.36321682e-01
-1.68089330e-01 4.49397087e-01 2.83903569e-01 2.89280593e-01
-8.47012222e-01 1.20775306e+00 5.80213261e+00 1.05764270e+00
-1.23657405e+00 3.45903374e-02 1.03951013e+00 4.31304201e-02
-3.28358740e-01 -2.48944432e-01 -1.00705278e+00 7.60511100e-01
8.42339218e-01 -3.18133950e-01 4.04112965e-01 7.56161034e-01
2.74214476e-01 3.98705870e-01 -9.33550000e-01 1.11596024e+00
5.27897894e-01 -1.38874400e+00 6.57109916e-01 2.38879561e-01
1.18069887e+00 -8.12438667e-01 4.39965695e-01 8.62931833e-02
6.43547401e-02 -1.68060219e+00 7.44706750e-01 4.58124846e-01
1.62491143e+00 -8.71418297e-01 7.53096581e-01 6.77752271e-02
-9.44904923e-01 9.27002281e-02 -4.56605315e-01 5.11759758e-01
1.24969617e-01 3.42790306e-01 -1.19026244e+00 5.04025698e-01
6.86111748e-01 2.98891693e-01 -4.66670364e-01 4.76127625e-01
-3.31139237e-01 2.80750483e-01 5.42573743e-02 2.95507640e-01
6.08274303e-02 -3.81260276e-01 2.86095500e-01 6.91296399e-01
8.18803847e-01 4.07364219e-02 -1.37042403e-01 1.24075902e+00
-3.74483019e-01 2.15688255e-02 -1.10836983e+00 -1.94233730e-01
4.82785553e-01 1.21350503e+00 -2.61934876e-01 -2.55698949e-01
-1.34321719e-01 1.17061532e+00 -3.23887050e-01 2.57452935e-01
-1.04074466e+00 -4.23871815e-01 5.62181473e-01 4.31793153e-01
1.85199708e-01 1.20871983e-01 9.46952775e-02 -7.86487937e-01
-3.73363058e-04 -1.16298139e+00 -2.45461762e-01 -1.01816952e+00
-1.10757399e+00 1.08450258e+00 4.87049446e-02 -1.29244697e+00
-5.30190110e-01 -3.91112149e-01 -5.64225554e-01 1.06999278e+00
-1.48166931e+00 -1.70284295e+00 -7.96285987e-01 7.88789570e-01
6.19383395e-01 -7.44223416e-01 9.41965520e-01 2.27520257e-01
-2.07187951e-01 9.35619295e-01 1.42697310e-02 2.33177990e-01
8.54728460e-01 -7.95130730e-01 5.58067858e-01 8.26729298e-01
-4.61011678e-02 3.17533702e-01 5.78724921e-01 -6.59661829e-01
-1.25007403e+00 -1.66067076e+00 7.08963811e-01 -2.81786114e-01
8.02991912e-02 -4.76756364e-01 -4.64436620e-01 5.11011541e-01
4.08963293e-01 5.88851757e-02 5.31217873e-01 -9.86233532e-01
-1.97003976e-01 -3.18244725e-01 -1.66922092e+00 6.62865579e-01
1.16778529e+00 -3.69713515e-01 -1.54522732e-01 2.80832291e-01
5.65078437e-01 -2.57532388e-01 -8.31978738e-01 3.74254078e-01
5.27525365e-01 -1.04669476e+00 9.83889341e-01 -1.80057094e-01
8.21054816e-01 -3.91701549e-01 -1.56003118e-01 -1.32756388e+00
-1.23719499e-02 -7.87924945e-01 5.49085677e-01 1.67453432e+00
1.43821225e-01 -6.41210139e-01 8.96127343e-01 4.30140138e-01
8.68444219e-02 -4.36045855e-01 -7.43938267e-01 -8.49369109e-01
-8.71528988e-04 2.09568679e-01 1.05187535e+00 7.52756834e-01
-9.04320657e-01 1.79553255e-01 -6.47879183e-01 6.13656975e-02
7.63267756e-01 -1.90287419e-02 1.04202330e+00 -9.79006946e-01
1.52015328e-01 -8.10675919e-02 -6.79221332e-01 -4.49486583e-01
4.85249192e-01 -8.74206662e-01 3.54082733e-02 -1.34551668e+00
3.15362662e-02 -3.31955761e-01 4.18677539e-01 4.66756612e-01
6.55280575e-02 8.61028492e-01 3.66180629e-01 1.28239334e-01
2.16452017e-01 8.95681918e-01 1.60842848e+00 -3.02952558e-01
3.75935018e-01 -2.85403788e-01 -6.69713020e-01 4.50555503e-01
8.38173568e-01 -4.39343721e-01 -6.51649892e-01 -2.73654312e-01
-1.79716080e-01 9.49116945e-02 3.46122682e-01 -1.14740551e+00
-2.10439697e-01 -3.35137755e-01 6.55730486e-01 -3.26034516e-01
5.83135605e-01 -9.01987195e-01 7.39665627e-01 2.05890432e-01
-2.34930485e-01 5.65768871e-03 1.87154463e-03 2.81178206e-01
-3.46554309e-01 -2.29033262e-01 1.05759072e+00 -2.18799397e-01
-5.06332397e-01 5.93019426e-01 4.11371626e-02 2.61262506e-02
1.39073706e+00 -2.76284605e-01 -3.56471539e-01 -6.38695478e-01
-3.04724753e-01 -2.76384979e-01 7.36792386e-01 7.32037902e-01
1.04763150e+00 -1.89023757e+00 -1.11866248e+00 5.39825201e-01
1.99363738e-01 -7.17925504e-02 1.20241381e-01 3.08626890e-02
-5.70609212e-01 3.71577203e-01 -7.05156028e-01 -4.92584109e-01
-1.21958375e+00 4.17969555e-01 2.06804141e-01 2.25416899e-01
-4.18141365e-01 5.67855000e-01 6.67810857e-01 -2.37774625e-01
-1.51880339e-01 4.84676778e-01 -1.43170998e-01 -2.09392488e-01
5.97217500e-01 1.21269092e-01 -1.89725295e-01 -1.01422703e+00
-4.42597419e-02 6.59076929e-01 1.90944001e-01 -2.13217229e-01
1.16923058e+00 1.54175609e-01 -1.07641056e-01 -3.19711238e-01
1.24373400e+00 8.45459998e-02 -1.25773740e+00 1.47147968e-01
-5.76297641e-01 -9.26236093e-01 -3.39257151e-01 -6.15213931e-01
-1.43856299e+00 7.98668087e-01 8.73792946e-01 -1.36143774e-01
1.10050201e+00 -1.84638977e-01 8.75845969e-01 -4.61593829e-02
3.92316550e-01 -7.30040729e-01 5.50311983e-01 4.90094908e-03
1.32727277e+00 -1.23166382e+00 -3.32861274e-01 -6.06870294e-01
-7.42587924e-01 9.67033684e-01 8.43271971e-01 -9.15179178e-02
2.05487087e-01 1.79191098e-01 1.92815170e-01 1.36495959e-02
-4.39202368e-01 3.76239396e-03 1.71647370e-01 9.68929887e-01
1.86166704e-01 -6.83926865e-02 -1.71024706e-02 1.50316492e-01
-5.38119614e-01 -7.78012071e-03 6.48377001e-01 4.29587990e-01
-2.18617693e-02 -1.38584912e+00 -7.35767961e-01 1.95077211e-01
-4.63228762e-01 -1.96805447e-02 -2.89810836e-01 7.25047648e-01
3.65286946e-01 1.24446309e+00 -8.29267353e-02 -4.54052210e-01
1.47894815e-01 6.53348938e-02 4.93784249e-01 -4.92827177e-01
-1.34424344e-01 -9.87041518e-02 -2.35759523e-02 -3.75156760e-01
-4.12166744e-01 -2.58121520e-01 -8.58792543e-01 -5.14576674e-01
-7.70177618e-02 6.44074306e-02 9.08253968e-01 5.56919873e-01
4.13224131e-01 3.42833549e-01 1.15949082e+00 -7.62482584e-01
-3.60258698e-01 -8.63675833e-01 -4.27367747e-01 7.69445956e-01
-5.68492487e-02 -6.85098529e-01 -2.56988078e-01 4.69318837e-01] | [12.317617416381836, -0.16529656946659088] |
81bbfd0a-59c1-461e-98c7-d2fa844dad05 | explainable-multi-agent-reinforcement | 2305.10378 | null | https://arxiv.org/abs/2305.10378v1 | https://arxiv.org/pdf/2305.10378v1.pdf | Explainable Multi-Agent Reinforcement Learning for Temporal Queries | As multi-agent reinforcement learning (MARL) systems are increasingly deployed throughout society, it is imperative yet challenging for users to understand the emergent behaviors of MARL agents in complex environments. This work presents an approach for generating policy-level contrastive explanations for MARL to answer a temporal user query, which specifies a sequence of tasks completed by agents with possible cooperation. The proposed approach encodes the temporal query as a PCTL logic formula and checks if the query is feasible under a given MARL policy via probabilistic model checking. Such explanations can help reconcile discrepancies between the actual and anticipated multi-agent behaviors. The proposed approach also generates correct and complete explanations to pinpoint reasons that make a user query infeasible. We have successfully applied the proposed approach to four benchmark MARL domains (up to 9 agents in one domain). Moreover, the results of a user study show that the generated explanations significantly improve user performance and satisfaction. | ['Lu Feng', 'Sarit Kraus', 'Kayla Boggess'] | 2023-05-17 | null | null | null | null | ['multi-agent-reinforcement-learning'] | ['methodology'] | [-3.21650878e-02 7.50032961e-01 -6.91379458e-02 -2.56394982e-01
-7.13054478e-01 -5.99073172e-01 9.91497755e-01 5.18027127e-01
-1.32498488e-01 1.28832352e+00 -1.53691217e-01 -3.37583959e-01
-3.72985959e-01 -6.88082933e-01 -6.22873724e-01 -3.83211195e-01
-4.49912697e-01 9.70309019e-01 4.41838592e-01 -3.56355488e-01
1.33452982e-01 1.76359385e-01 -1.82430685e+00 2.07895264e-01
1.07666779e+00 5.34105420e-01 1.93290755e-01 8.88423383e-01
2.81247824e-01 1.22757304e+00 -7.40891755e-01 -2.83258455e-03
6.34574145e-02 -5.81784844e-01 -9.62132096e-01 3.25867027e-01
-3.16934347e-01 -4.34569001e-01 1.14394523e-01 1.15735304e+00
-9.53750387e-02 1.83313087e-01 5.09279132e-01 -2.14305782e+00
-4.79888916e-01 8.95684659e-01 -2.49058679e-01 -2.31003463e-01
1.00971448e+00 5.87728918e-01 1.01993072e+00 3.40473689e-02
4.52254176e-01 1.52278173e+00 -5.61576262e-02 7.63701677e-01
-9.56579089e-01 -1.90020606e-01 3.58267188e-01 4.84326065e-01
-9.60345209e-01 -2.07229406e-01 4.93035913e-01 -3.68120611e-01
1.36670530e+00 4.19228494e-01 7.21509516e-01 9.72453713e-01
7.68886268e-01 6.69562697e-01 1.41948950e+00 -4.07483816e-01
8.01902592e-01 2.88154781e-01 -8.05211067e-02 8.05491269e-01
5.20276487e-01 4.02339190e-01 -4.59281713e-01 -6.42241359e-01
4.73854601e-01 -3.23052615e-01 3.89075279e-01 -2.18852013e-01
-1.13953233e+00 7.31871247e-01 -2.36873016e-01 7.76654854e-02
-8.54086280e-01 3.62805039e-01 2.79126108e-01 2.42140800e-01
-2.12018743e-01 5.43725312e-01 -4.90803450e-01 -4.35508907e-01
-1.28521398e-01 6.85598254e-01 1.02871525e+00 1.01703894e+00
5.66823900e-01 2.36562982e-01 -5.28412312e-02 -6.42018318e-02
7.35765994e-01 7.54044890e-01 2.24567533e-01 -1.30212176e+00
6.29620673e-03 7.82040775e-01 8.53940666e-01 -7.14950085e-01
-4.57370549e-01 -1.00867622e-01 -2.03444004e-01 6.24073803e-01
2.34291658e-01 -4.58300442e-01 -2.08808452e-01 1.64689422e+00
4.30740595e-01 2.02884927e-01 8.28271210e-01 7.71302998e-01
3.91898930e-01 6.65844560e-01 1.73084632e-01 -5.11915207e-01
1.41221929e+00 -6.59329891e-01 -7.78563201e-01 -3.68292689e-01
4.81601506e-01 -2.58189470e-01 8.49607587e-01 5.86504221e-01
-1.12676406e+00 -8.08252618e-02 -9.57496464e-01 7.66051769e-01
9.22380090e-02 -4.96643297e-02 6.78243220e-01 3.90183002e-01
-9.66812313e-01 1.23782847e-02 -1.00755131e+00 -4.31497216e-01
-1.81626767e-01 6.16880834e-01 1.58946868e-02 3.03478450e-01
-1.15893042e+00 8.81249547e-01 7.88759947e-01 -1.35162190e-01
-1.47121310e+00 1.14370346e-01 -1.05851722e+00 1.04020476e-01
7.26330519e-01 -2.13127151e-01 1.76590300e+00 -9.55140471e-01
-1.70748866e+00 2.50068158e-01 5.89574799e-02 -8.60258639e-01
5.87013066e-01 1.43727198e-01 -5.79739332e-01 -2.79447436e-02
1.28982261e-01 5.86325765e-01 4.77844924e-01 -1.45716739e+00
-1.15214026e+00 -3.36249508e-02 7.82207847e-01 3.11373889e-01
3.33783507e-01 -1.73354074e-01 2.93145537e-01 2.17223674e-01
-5.25898099e-01 -1.04452395e+00 -3.83684576e-01 -6.72167480e-01
-2.65525192e-01 -6.38686121e-01 6.01953804e-01 -1.30869523e-01
6.57062471e-01 -1.62511158e+00 -1.32301778e-01 3.47060084e-01
-5.69194905e-04 -1.98254228e-01 -4.66209501e-01 7.73206174e-01
2.83375293e-01 -6.26527295e-02 1.80113828e-03 1.38065115e-01
5.38029015e-01 5.16562343e-01 -3.62604171e-01 3.27518255e-01
1.41731799e-01 4.80375677e-01 -1.08742666e+00 -4.19917762e-01
3.32370847e-01 -1.17727578e-01 -5.99893987e-01 5.89398801e-01
-1.14379919e+00 4.61392075e-01 -7.62247562e-01 4.57718313e-01
3.17144960e-01 -5.65740280e-02 6.08132064e-01 5.46073735e-01
-6.84552267e-02 3.81425321e-02 -1.24740684e+00 1.10384381e+00
-3.06871355e-01 3.08228493e-01 -3.72923255e-01 -4.36617404e-01
7.75236070e-01 6.37467086e-01 4.79903996e-01 -6.54613137e-01
1.30073830e-01 8.74115899e-02 2.72280216e-01 -6.84136748e-01
4.16743875e-01 8.44961554e-02 -4.76852775e-01 7.87969530e-01
-3.18857670e-01 -1.98268980e-01 4.20788348e-01 8.30495805e-02
1.25996125e+00 2.85228312e-01 8.09925556e-01 -1.37098536e-01
8.63806307e-01 4.50178325e-01 7.91533589e-01 1.18285728e+00
-4.60556269e-01 -5.32308400e-01 9.16272163e-01 -7.03625739e-01
-6.36636078e-01 -8.18897545e-01 7.88563251e-01 7.78449535e-01
4.76575464e-01 -2.48804748e-01 -7.59691954e-01 -8.87593091e-01
-1.96136922e-01 1.29744232e+00 -4.45852965e-01 -1.15996286e-01
-1.71242744e-01 -2.56747693e-01 3.43857139e-01 -1.81634002e-03
5.42125821e-01 -1.60402918e+00 -1.64198077e+00 2.37128928e-01
-2.00127617e-01 -1.16127789e+00 -8.96692276e-02 -1.44065276e-01
-2.95753896e-01 -1.52262235e+00 3.30361784e-01 -4.33221430e-01
7.03153193e-01 -1.49026230e-01 9.49698269e-01 3.84437799e-01
1.12797439e-01 6.72912002e-01 -5.09275198e-01 -5.00106335e-01
-1.04533887e+00 -2.73965269e-01 5.59246778e-01 -2.35861763e-01
3.94096136e-01 -1.20508052e-01 -1.70097128e-01 3.02811921e-01
-8.44574153e-01 2.04795197e-01 3.12101722e-01 5.16356707e-01
2.59362876e-01 5.76610506e-01 8.08236539e-01 -6.87937498e-01
1.33324313e+00 -4.43911225e-01 -1.23212028e+00 7.09221959e-01
-7.04905868e-01 4.75342125e-01 7.53225744e-01 -3.58454704e-01
-1.16050708e+00 -3.46085578e-02 2.42336169e-01 3.70890170e-01
-6.25799596e-01 6.38414383e-01 8.33478421e-02 3.55485171e-01
6.62649989e-01 2.54507124e-01 -4.45650555e-02 3.65789682e-01
2.87752580e-02 5.25068641e-01 2.30719656e-01 -1.00033438e+00
7.37864614e-01 1.08595856e-01 -1.71856396e-02 -3.85117799e-01
-2.94546574e-01 2.14707285e-01 1.65628582e-01 -5.53080380e-01
7.11457610e-01 -6.09543920e-01 -1.85042715e+00 -2.80700810e-02
-1.21014369e+00 -6.74566388e-01 -1.77741230e-01 4.18408126e-01
-1.12151361e+00 1.48377880e-01 -2.23220080e-01 -1.46984637e+00
9.57934186e-03 -1.59226334e+00 6.62843347e-01 4.78983223e-01
-4.46017861e-01 -7.17446029e-01 2.89178222e-01 6.79768920e-02
1.48641139e-01 2.03459620e-01 1.13512003e+00 -1.01493478e+00
-5.93924224e-01 2.20014043e-02 2.88350642e-01 -4.40898597e-01
2.60485917e-01 1.48485497e-01 -7.62706161e-01 -3.98090065e-01
-3.44130635e-01 -4.87434506e-01 -5.64491034e-01 3.00501019e-01
5.30887067e-01 -7.69119263e-01 -2.22605616e-02 -6.47714674e-01
1.28822243e+00 8.41983259e-01 4.52212185e-01 4.37836468e-01
-3.17377478e-01 5.24507582e-01 1.18493569e+00 1.03143454e+00
7.68613756e-01 4.89017516e-01 9.09277320e-01 4.01888132e-01
5.75996161e-01 -1.69939920e-01 8.89005899e-01 3.26589756e-02
-7.05220550e-02 -4.12767053e-01 -8.93756926e-01 3.46024543e-01
-2.51602769e+00 -1.14783204e+00 1.34717152e-01 2.00274301e+00
5.47261536e-01 3.18882227e-01 2.02020377e-01 -1.81491911e-01
3.60843778e-01 -2.17773885e-01 -9.02685702e-01 -7.65348375e-01
2.57521570e-01 -3.82100344e-01 1.43247947e-01 1.09447253e+00
-6.07581198e-01 9.59276617e-01 6.03191566e+00 2.95573592e-01
-5.72200596e-01 -3.74115482e-02 2.31235936e-01 3.37152481e-01
-4.74205106e-01 2.38416895e-01 -4.90625858e-01 -5.75867407e-02
1.03089809e+00 -5.77281117e-01 7.98505008e-01 8.34451318e-01
7.56339610e-01 -6.81552827e-01 -1.32078922e+00 5.36152959e-01
-2.49494165e-01 -1.11890101e+00 -1.26231670e-01 -4.85556200e-02
6.41386271e-01 -2.63873011e-01 -1.42999738e-01 6.04682684e-01
1.01396799e+00 -7.83807635e-01 1.01033294e+00 5.00893056e-01
6.88188225e-02 -1.06151676e+00 8.87854874e-01 9.34484601e-01
-9.01659071e-01 -3.76461267e-01 1.83395911e-02 -4.76444006e-01
-1.54841274e-01 -2.00802222e-01 -1.43782163e+00 5.04151702e-01
4.59546417e-01 2.04925865e-01 -3.23511720e-01 7.51606286e-01
-5.74884713e-01 1.89877823e-01 -1.29701672e-02 -4.84709948e-01
3.16497356e-01 -2.14920625e-01 6.53100729e-01 5.78440785e-01
3.00251037e-01 9.93786100e-03 4.25171345e-01 9.15519774e-01
5.27515829e-01 -4.19584215e-01 -7.55771756e-01 -1.71520770e-01
5.26525199e-01 9.29931998e-01 -6.18794024e-01 -4.28958148e-01
-1.71706989e-01 6.87131345e-01 1.03902921e-01 5.23093283e-01
-1.17754221e+00 1.89556077e-01 7.81737030e-01 -2.62172043e-01
-3.82831067e-01 -3.06686312e-01 1.65270850e-01 -7.90679276e-01
-1.07278064e-01 -1.39929020e+00 5.07221818e-01 -1.18685603e+00
-8.14761579e-01 9.14017200e-01 1.98880985e-01 -1.12148750e+00
-8.78881693e-01 -1.75814107e-01 -4.29164320e-01 3.87742072e-01
-1.42939878e+00 -7.96436608e-01 -1.78513184e-01 4.00514871e-01
8.04908156e-01 -3.96901995e-01 1.13342094e+00 -3.46828669e-01
-4.41658854e-01 -2.59947516e-02 -7.28741050e-01 -5.65231621e-01
3.58297825e-01 -1.18500614e+00 -6.04091883e-02 9.66324449e-01
-2.56428778e-01 4.22101021e-01 1.32175207e+00 -9.75172400e-01
-1.60463297e+00 -8.78977478e-01 6.73447967e-01 -1.91407412e-01
5.41896045e-01 1.48450419e-01 -4.58332032e-01 9.95137453e-01
9.48437929e-01 -4.51911926e-01 4.09032017e-01 -3.71672362e-01
1.75402611e-01 1.36061028e-01 -1.50588822e+00 9.64709282e-01
3.82047087e-01 -2.14694649e-01 -6.97054803e-01 4.13415015e-01
7.43234456e-01 -4.33759719e-01 -4.98837948e-01 -8.54843203e-03
2.42347613e-01 -9.99362230e-01 2.28014052e-01 -9.29951072e-01
1.87286422e-01 -8.40068698e-01 -1.84906051e-01 -1.46383870e+00
7.13583380e-02 -9.56652582e-01 -1.06812552e-01 8.26576293e-01
4.39240247e-01 -6.92440867e-01 6.14938796e-01 1.27539372e+00
1.44697368e-01 -2.73731858e-01 -8.84653151e-01 -6.97577596e-01
-6.85993910e-01 -5.42170703e-01 9.42644358e-01 5.63554525e-01
5.99562824e-01 -1.44630924e-01 -4.15255994e-01 8.76733601e-01
7.43961632e-01 7.93559775e-02 5.69392860e-01 -1.07113159e+00
-4.50548232e-01 -1.11898333e-01 -4.22521792e-02 -3.52149308e-01
6.27223194e-01 -4.89070266e-01 5.25453389e-01 -1.73578417e+00
9.60963294e-02 -3.94800097e-01 1.47249978e-02 8.93753052e-01
3.32649320e-01 -5.17544627e-01 1.02247678e-01 -1.20491847e-01
-1.06324995e+00 5.87496519e-01 9.39557254e-01 -1.61171436e-01
-4.70097572e-01 1.26658902e-01 -4.22681987e-01 8.55791807e-01
1.05288434e+00 -5.01210630e-01 -7.81263590e-01 -1.75137505e-01
6.72520518e-01 8.43315542e-01 3.22640270e-01 -1.07239032e+00
3.36713493e-01 -1.19531631e+00 -4.09559488e-01 -2.33892947e-01
1.13401555e-01 -1.31035161e+00 4.72190022e-01 1.01570225e+00
-5.91182828e-01 5.49598396e-01 2.96883106e-01 5.26165605e-01
-1.33942189e-02 -2.69323409e-01 3.57897103e-01 1.27853438e-01
-8.88236105e-01 -1.26391992e-01 -1.27984011e+00 -2.89951533e-01
1.74449205e+00 1.07571363e-01 -3.47889721e-01 -8.00104022e-01
-5.71594238e-01 7.27219641e-01 2.58298457e-01 4.46015716e-01
9.83869076e-01 -1.10308194e+00 -6.07675314e-01 1.89828292e-01
2.25102887e-01 -4.81888026e-01 1.66497310e-03 4.82853621e-01
-4.42032695e-01 4.80875105e-01 -6.49262130e-01 -2.38712430e-01
-1.31563032e+00 3.15632284e-01 5.84587038e-01 -3.01656008e-01
-3.04930687e-01 -1.48836256e-03 -3.03064406e-01 -5.04409671e-01
2.64913112e-01 -4.87012804e-01 -4.00835305e-01 -5.90687692e-01
5.21993756e-01 1.66338652e-01 -3.66616935e-01 -2.49301657e-01
-6.46182120e-01 -3.08029726e-02 2.41338223e-01 -6.15466475e-01
1.22052252e+00 -7.03218356e-02 -5.78882806e-02 2.30254516e-01
-1.57171980e-01 -1.58495590e-01 -1.31076050e+00 -1.94337349e-02
2.30956092e-01 -1.19895115e-01 -4.67798054e-01 -9.94722843e-01
-4.37446833e-01 1.72760338e-01 4.27703023e-01 6.51951432e-01
8.65335584e-01 -2.53102239e-02 1.15051657e-01 7.88054585e-01
9.68865752e-01 -1.05090714e+00 5.73410951e-02 5.98438203e-01
8.48247409e-01 -1.26024246e+00 -1.35008633e-01 -6.28203750e-02
-1.07232678e+00 1.06759417e+00 1.05924368e+00 4.96786833e-02
-1.55610805e-02 3.28089505e-01 -9.54907089e-02 -1.92122757e-01
-1.41219389e+00 -1.69470936e-01 -2.89748847e-01 7.50176311e-01
-8.72139353e-03 4.33559954e-01 -3.07566792e-01 4.44261789e-01
7.43474588e-02 1.53374553e-01 1.32668638e+00 1.05368924e+00
-6.58504307e-01 -1.17760694e+00 -5.00194430e-01 5.93923330e-02
1.10641539e-01 5.22484303e-01 -4.54319626e-01 8.14565539e-01
-1.47208840e-01 1.58581662e+00 -1.63234726e-01 -3.40324104e-01
2.62288958e-01 -1.82220548e-01 4.28676605e-01 -5.99771082e-01
-5.21958828e-01 -1.33616090e-01 3.35297257e-01 -6.77906752e-01
-5.38602591e-01 -4.99376982e-01 -2.03696156e+00 -1.59854367e-02
2.03153834e-01 6.67675018e-01 3.18632483e-01 1.25361967e+00
2.13029519e-01 6.58484161e-01 4.55173492e-01 -2.74347365e-01
-7.18191803e-01 -5.92398524e-01 -3.23360443e-01 5.94833493e-02
4.17360276e-01 -5.80405772e-01 -9.34670568e-02 -6.05814792e-02] | [4.50866174697876, 1.871673822402954] |
dd54bfac-3aa6-497f-ac53-a42a3edbc0c5 | rola-a-real-time-online-lightweight-anomaly | 2305.16509 | null | https://arxiv.org/abs/2305.16509v1 | https://arxiv.org/pdf/2305.16509v1.pdf | RoLA: A Real-Time Online Lightweight Anomaly Detection System for Multivariate Time Series | A multivariate time series refers to observations of two or more variables taken from a device or a system simultaneously over time. There is an increasing need to monitor multivariate time series and detect anomalies in real time to ensure proper system operation and good service quality. It is also highly desirable to have a lightweight anomaly detection system that considers correlations between different variables, adapts to changes in the pattern of the multivariate time series, offers immediate responses, and provides supportive information regarding detection results based on unsupervised learning and online model training. In the past decade, many multivariate time series anomaly detection approaches have been introduced. However, they are unable to offer all the above-mentioned features. In this paper, we propose RoLA, a real-time online lightweight anomaly detection system for multivariate time series based on a divide-and-conquer strategy, parallel processing, and the majority rule. RoLA employs multiple lightweight anomaly detectors to monitor multivariate time series in parallel, determine the correlations between variables dynamically on the fly, and then jointly detect anomalies based on the majority rule in real time. To demonstrate the performance of RoLA, we conducted an experiment based on a public dataset provided by the FerryBox of the One Ocean Expedition. The results show that RoLA provides satisfactory detection accuracy and lightweight performance. | ['Jia-Chun Lin', 'Ming-Chang Lee'] | 2023-05-25 | null | null | null | null | ['time-series-anomaly-detection'] | ['time-series'] | [-1.37402713e-01 -9.10943329e-01 3.57390344e-01 -2.93485641e-01
-1.97605416e-01 -3.86497319e-01 3.92463475e-01 7.05321431e-01
-2.42016241e-01 1.65143475e-01 -4.07663524e-01 -3.90068203e-01
-4.01897162e-01 -7.93187499e-01 -1.68930873e-01 -6.95773721e-01
-7.32048571e-01 3.87120217e-01 4.97697949e-01 -1.14927255e-01
3.11114520e-01 1.08402729e+00 -1.74687123e+00 -3.91570032e-02
7.67763793e-01 1.43931913e+00 -4.14577574e-01 7.80422866e-01
-2.66893566e-01 4.34455723e-01 -9.18708444e-01 4.61880118e-01
3.58638763e-01 -2.89381921e-01 -1.33440152e-01 3.87771167e-02
-1.88749179e-01 -1.42724544e-01 6.49337322e-02 7.52260923e-01
2.26426899e-01 3.24784577e-01 3.76360178e-01 -1.63391852e+00
1.94123179e-01 1.88758105e-01 -6.17040634e-01 8.78349662e-01
2.84198791e-01 2.12734520e-01 6.46568775e-01 -4.57665682e-01
-1.55542428e-02 9.63595033e-01 5.16095400e-01 1.70184411e-02
-1.08336973e+00 -5.54515958e-01 4.02676046e-01 3.90356898e-01
-1.20247424e+00 -1.10748447e-01 7.50583291e-01 -3.49041224e-01
1.17087507e+00 5.41694641e-01 6.23061776e-01 4.37789619e-01
6.19703829e-01 3.45514864e-01 5.73380113e-01 -3.35242778e-01
5.72070241e-01 -3.07668597e-01 4.00404148e-02 3.85302275e-01
2.00932488e-01 -9.11746547e-02 -3.98225933e-01 -6.38434291e-01
1.91514894e-01 6.67546511e-01 5.28441332e-02 5.64666912e-02
-9.13824439e-01 4.12274420e-01 -1.10277735e-01 4.18352038e-01
-7.80512810e-01 -4.09300208e-01 7.24179149e-01 8.49116325e-01
4.93803918e-01 4.01583284e-01 -5.83386362e-01 -4.01870310e-01
-6.49117291e-01 1.08004704e-01 7.90148795e-01 3.58223736e-01
4.53719676e-01 4.96637613e-01 1.87196046e-01 4.68520731e-01
2.43806288e-01 7.17460036e-01 8.09241474e-01 -3.81143540e-01
1.23940632e-01 8.15151215e-01 -1.46958441e-01 -1.28376627e+00
-6.88554466e-01 -2.08199501e-01 -8.65131259e-01 4.65517104e-01
2.90544987e-01 -9.70782265e-02 -5.14968574e-01 1.30484498e+00
5.39622366e-01 2.65787214e-01 7.90947080e-02 5.80271780e-01
1.31467544e-02 6.96175158e-01 -2.19181076e-01 -6.38213396e-01
9.78430867e-01 -1.75430030e-01 -8.73337567e-01 2.18582958e-01
7.33444810e-01 -7.42159784e-01 6.95682287e-01 7.71634340e-01
-5.90484679e-01 -3.78457427e-01 -1.00087988e+00 8.88473094e-01
-5.67539632e-01 -2.84833342e-01 4.09457564e-01 4.08325016e-01
-5.59078574e-01 6.03210449e-01 -1.23837698e+00 -4.94303763e-01
-2.05262795e-01 2.53762096e-01 -1.39603436e-01 3.08737785e-01
-9.72706497e-01 4.94810402e-01 3.62047434e-01 2.95175374e-01
-5.06062090e-01 -3.22575957e-01 -7.11162984e-01 -9.08016507e-03
3.19763333e-01 4.43317629e-02 1.08150303e+00 -8.04554164e-01
-1.19703853e+00 3.06624293e-01 -1.64255604e-01 -6.02750957e-01
4.12720203e-01 -2.04771996e-01 -1.35080159e+00 -2.65549999e-02
-2.86716837e-02 -6.26152754e-01 6.99649334e-01 -3.47020358e-01
-1.01339507e+00 -6.06652856e-01 -5.39279580e-01 -3.95312041e-01
-4.24563766e-01 3.71532172e-01 9.94923115e-02 -5.62517345e-01
3.70493203e-01 -6.94235027e-01 -1.68761358e-01 -2.92040586e-01
-5.34142144e-02 -3.43847305e-01 1.45352650e+00 -5.00526428e-01
1.55237246e+00 -2.52474236e+00 -3.65082204e-01 9.57187951e-01
-7.40084648e-02 2.78605312e-01 2.12434411e-01 6.60755336e-01
-3.41591537e-01 -2.75295794e-01 -4.05195385e-01 -8.69304873e-03
-3.30632180e-01 5.11686563e-01 -6.08454108e-01 7.72087872e-01
1.82881489e-01 -4.76158671e-02 -7.27731645e-01 -9.07062441e-02
3.21401328e-01 -1.54793471e-01 -2.25762337e-01 5.89098155e-01
-1.21639982e-01 4.72667545e-01 -4.11996067e-01 8.81955385e-01
4.00884837e-01 2.01364737e-02 -1.20594874e-01 1.71999291e-01
-4.74378943e-01 -2.65539587e-01 -1.44357252e+00 9.61878061e-01
-1.19497739e-01 5.85362971e-01 -1.55765384e-01 -1.31174076e+00
1.32307827e+00 5.80776095e-01 9.89527047e-01 -9.67412353e-01
-1.59371505e-03 5.70837855e-01 1.17557548e-01 -8.41397762e-01
2.44393319e-01 3.27485442e-01 -2.46350188e-02 7.90860474e-01
-3.96629244e-01 2.92708814e-01 4.56630945e-01 2.68420801e-02
1.41980469e+00 -3.34354728e-01 4.23683286e-01 7.15278368e-03
7.68775880e-01 -1.02414377e-02 7.64289737e-01 5.62623739e-01
-1.63644046e-01 1.17435768e-01 5.08094668e-01 -8.56398284e-01
-6.75971925e-01 -9.67348218e-01 9.27748755e-02 9.01035428e-01
-1.49318412e-01 -3.60397160e-01 3.82378921e-02 -4.96047556e-01
1.35522768e-01 5.90935469e-01 -4.56502855e-01 -3.64940703e-01
-6.17527962e-01 -7.92135000e-01 3.55003506e-01 3.43356758e-01
3.28033715e-01 -1.12206745e+00 -9.75669682e-01 5.12868464e-01
1.42192557e-01 -7.46429443e-01 -9.58614126e-02 2.36949727e-01
-9.85212624e-01 -1.23186457e+00 7.88231492e-02 -2.07566440e-01
5.81909359e-01 2.42539391e-01 6.93473637e-01 1.58185348e-01
-5.37790418e-01 4.44486439e-01 -5.19323111e-01 -7.75497258e-01
-3.31985891e-01 -3.85022640e-01 5.46484828e-01 5.72295666e-01
4.21918392e-01 -8.73763740e-01 -2.90723443e-01 4.13686156e-01
-1.12665939e+00 -8.11292470e-01 1.84926271e-01 4.32072967e-01
6.60340607e-01 5.47832429e-01 6.66416764e-01 -5.25093973e-01
6.59600139e-01 -8.53281736e-01 -1.01791537e+00 8.52167979e-02
-8.73977602e-01 -6.92210495e-02 1.04834020e+00 -4.60542679e-01
-5.55714011e-01 -1.60934657e-01 1.68390814e-02 -5.05909026e-01
-2.51269847e-01 6.90158010e-01 1.23844326e-01 2.16930792e-01
7.24310756e-01 4.01620239e-01 2.19395101e-01 -5.85779428e-01
-3.57302994e-01 8.28242183e-01 5.61592638e-01 -4.65942353e-01
7.01534033e-01 6.24684453e-01 1.32281438e-01 -9.61830139e-01
-2.86895812e-01 -8.98955584e-01 -4.64055330e-01 -5.01891613e-01
2.57292956e-01 -5.31390786e-01 -6.90775037e-01 7.16272116e-01
-7.68028736e-01 5.34027256e-02 -2.60826141e-01 6.00746572e-01
-1.43119112e-01 3.83699000e-01 -2.37152159e-01 -1.10854125e+00
-4.60480571e-01 -5.31188130e-01 5.20186067e-01 3.22106302e-01
-4.91246879e-01 -9.63284910e-01 2.02759653e-01 -5.29999256e-01
7.94664204e-01 7.82957196e-01 6.44757330e-01 -1.23415720e+00
-9.89251584e-02 -8.20620418e-01 2.48011738e-01 1.90687418e-01
5.93932152e-01 5.85515380e-01 -6.18620157e-01 -4.80940819e-01
2.43263602e-01 3.68340880e-01 2.14236319e-01 -9.15127154e-03
1.38768733e+00 -3.34326535e-01 -2.83381585e-02 5.33093750e-01
1.05473399e+00 7.23446369e-01 2.96562940e-01 6.36159360e-01
3.42990249e-01 4.08522546e-01 8.31858754e-01 1.02367926e+00
-5.48474723e-03 3.34874481e-01 5.55226088e-01 4.62220795e-02
8.68565798e-01 3.27574193e-01 6.16736889e-01 9.75500166e-01
-9.61997639e-03 -2.74298824e-02 -1.01644957e+00 5.36847234e-01
-1.89034784e+00 -1.12638569e+00 -2.70805627e-01 2.46121645e+00
2.85888940e-01 3.56326371e-01 4.45568055e-01 6.46658421e-01
5.07483304e-01 -1.54982274e-02 -7.44804144e-01 -7.14908123e-01
-7.48068765e-02 -1.03994809e-01 2.59678781e-01 -4.04283218e-02
-9.90464747e-01 1.28015533e-01 6.00403929e+00 2.13584796e-01
-1.68214476e+00 -2.50637501e-01 1.19174898e-01 -1.72011033e-01
8.17278847e-02 -1.89260632e-01 -1.90139681e-01 5.88798583e-01
1.22914803e+00 -5.04829764e-01 2.42384642e-01 1.02158892e+00
5.55739641e-01 -1.42008558e-01 -8.54252517e-01 8.85713995e-01
-6.59263879e-02 -7.53209114e-01 -1.48616597e-01 1.48766804e-02
3.03747773e-01 1.93529740e-01 -3.45583886e-01 1.72324240e-01
-1.23206429e-01 -5.49335659e-01 3.02984923e-01 7.31006920e-01
3.89173962e-02 -7.45819926e-01 7.69414783e-01 4.15467471e-01
-1.44198096e+00 -4.19129074e-01 1.54156223e-01 -3.03788483e-01
1.37120917e-01 8.46578300e-01 -5.58341205e-01 5.53868532e-01
9.78641272e-01 6.27135098e-01 -5.52425861e-01 1.34244549e+00
1.72639623e-01 8.71889949e-01 -7.41508305e-01 6.70574233e-02
1.26142651e-01 -1.58119619e-01 8.29939485e-01 8.90005827e-01
5.06074548e-01 8.38878378e-03 4.08957899e-01 2.63530403e-01
6.26283467e-01 3.18451047e-01 -4.72802907e-01 -6.89988434e-02
4.52272236e-01 1.03892803e+00 -7.22958505e-01 -3.50227177e-01
-5.71051598e-01 4.96052653e-01 -2.73048848e-01 3.10993642e-01
-7.10150898e-01 -6.90489829e-01 7.17375755e-01 1.04391329e-01
-3.05964351e-02 -4.15123492e-01 -3.72988097e-02 -1.04566562e+00
3.66958410e-01 -9.97982383e-01 1.03001547e+00 -3.16417575e-01
-1.41973042e+00 5.61096191e-01 -1.29437804e-01 -1.71122563e+00
-5.80675125e-01 -4.61407304e-01 -1.09980011e+00 8.16590250e-01
-9.98906612e-01 -3.06561023e-01 -3.41255218e-01 1.03017092e+00
4.04666841e-01 -3.92898172e-01 9.55698550e-01 2.46432975e-01
-9.85908031e-01 2.85367280e-01 3.46819252e-01 9.12913028e-03
5.91191173e-01 -1.10777712e+00 4.47609007e-01 1.35463989e+00
-1.38647616e-01 3.69965136e-01 8.83903623e-01 -6.69553280e-01
-1.39417243e+00 -1.00711787e+00 5.43242931e-01 1.29994258e-01
8.37006211e-01 4.07927632e-02 -1.54654229e+00 5.59707582e-01
-2.89142609e-01 3.23404759e-01 8.05744827e-01 1.00705318e-01
-1.66305631e-01 -6.64814770e-01 -1.19983876e+00 3.36359322e-01
3.47093076e-01 -3.48101318e-01 -4.93462890e-01 2.01160252e-01
2.25951806e-01 -1.36517152e-01 -8.91777039e-01 4.22259629e-01
3.54230046e-01 -9.64412153e-01 3.78114522e-01 -5.00365376e-01
-3.89830351e-01 -6.77646637e-01 -2.11224273e-01 -1.26159024e+00
-8.52391273e-02 -7.17542708e-01 -4.76064324e-01 1.17216551e+00
1.57368377e-01 -1.33182204e+00 2.08779618e-01 6.67442679e-01
-3.04077491e-02 -4.53192681e-01 -1.05941713e+00 -1.08533514e+00
-8.60224247e-01 -7.76980519e-01 9.31932747e-01 1.01504886e+00
8.98256153e-02 -3.10653478e-01 -1.97899297e-01 6.48265898e-01
5.20464540e-01 2.81445265e-01 8.94554496e-01 -1.49994469e+00
-2.07895800e-01 -4.43602681e-01 -8.10284317e-01 -1.50993466e-01
-3.91114712e-01 -3.45797688e-01 -4.65813689e-02 -8.68801475e-01
-5.84962726e-01 -2.56727964e-01 -7.03188360e-01 5.76285660e-01
2.94605251e-02 -2.30145976e-01 -3.92539531e-01 4.36161667e-01
-5.50695181e-01 3.83686095e-01 1.93423972e-01 1.35772482e-01
-5.34469128e-01 3.89906287e-01 2.83164368e-03 7.07655489e-01
1.05427456e+00 -5.19567728e-01 -2.01544568e-01 -1.36692151e-01
6.32303655e-02 6.49177730e-02 8.24674591e-02 -1.29225552e+00
4.42792743e-01 -4.25607562e-01 3.58101517e-01 -6.15115225e-01
-1.04061663e-01 -1.11264122e+00 2.42915183e-01 6.87275708e-01
1.58699900e-01 9.47766483e-01 4.06654090e-01 6.66127503e-01
-5.43570578e-01 2.38421276e-01 5.05893946e-01 3.02479804e-01
-1.00421929e+00 4.61734563e-01 -5.83511293e-01 -1.50773942e-01
1.26070416e+00 -9.94813368e-02 -6.57641217e-02 -1.30528554e-01
-6.67869329e-01 5.26915908e-01 1.03156835e-01 6.42508924e-01
5.00732422e-01 -1.15655041e+00 -5.84853888e-01 7.97870755e-01
4.67655629e-01 -1.52887344e-01 1.90975517e-01 9.47646558e-01
-6.23848379e-01 9.77935642e-03 -3.77407283e-01 -8.81641090e-01
-1.34912109e+00 5.06870508e-01 3.53367716e-01 -4.09438387e-02
-7.37712681e-01 1.05298609e-01 -6.58125699e-01 -8.56921002e-02
2.97331870e-01 -2.15691328e-01 -2.60701627e-01 1.66349083e-01
1.10188949e+00 5.39665997e-01 3.94051880e-01 -3.44598085e-01
-5.31361878e-01 2.90405631e-01 6.04017004e-02 5.69361076e-02
1.29797447e+00 -1.29213497e-01 -3.39438289e-01 1.03484178e+00
7.49145269e-01 9.06888172e-02 -7.79671609e-01 -4.90772456e-01
3.20733607e-01 -6.18495822e-01 -2.17105269e-01 -3.50563616e-01
-1.03742480e+00 5.27626514e-01 7.77543008e-01 9.89785314e-01
1.63575888e+00 -4.16939557e-01 6.10202730e-01 3.84943634e-01
2.20025957e-01 -1.11978197e+00 -1.67464212e-01 5.38938522e-01
6.84148729e-01 -9.85691309e-01 -1.40946969e-01 2.54610002e-01
-4.37809378e-01 1.41108084e+00 6.37573481e-01 -9.86890420e-02
8.17732453e-01 2.67460972e-01 2.22782671e-01 -8.00217688e-02
-8.69077742e-01 -1.67679656e-02 -2.32532974e-02 3.98045570e-01
9.91983246e-03 3.29625160e-02 -2.22585037e-01 2.02958331e-01
7.21690580e-02 -3.56305063e-01 2.83340514e-01 1.14247274e+00
-5.62389553e-01 -8.55919540e-01 -7.75784254e-01 6.37512326e-01
-5.11122227e-01 5.38521171e-01 -6.90315478e-03 5.90315700e-01
-1.19991109e-01 1.12094116e+00 4.54932243e-01 -4.19589520e-01
8.10534656e-01 2.29903758e-01 -4.12187546e-01 -1.81270719e-01
-5.07843971e-01 -3.61984260e-02 -2.66450703e-01 -8.30138624e-01
-1.08628176e-01 -9.50322390e-01 -1.46431398e+00 -4.75454837e-01
-7.68629974e-03 4.20412481e-01 7.68242598e-01 1.25377619e+00
4.73921001e-01 6.79274559e-01 1.17329752e+00 -4.30101573e-01
-5.46177447e-01 -9.27717388e-01 -8.14445436e-01 4.45791930e-01
5.38209021e-01 -2.57005632e-01 -7.39273071e-01 -2.72088200e-01] | [7.265730857849121, 2.7464351654052734] |
288e7e27-f1e4-4706-92a2-07b5269713c8 | semi-supervised-community-detection-via | 2306.01089 | null | https://arxiv.org/abs/2306.01089v1 | https://arxiv.org/pdf/2306.01089v1.pdf | Semi-supervised Community Detection via Structural Similarity Metrics | Motivated by social network analysis and network-based recommendation systems, we study a semi-supervised community detection problem in which the objective is to estimate the community label of a new node using the network topology and partially observed community labels of existing nodes. The network is modeled using a degree-corrected stochastic block model, which allows for severe degree heterogeneity and potentially non-assortative communities. We propose an algorithm that computes a `structural similarity metric' between the new node and each of the $K$ communities by aggregating labeled and unlabeled data. The estimated label of the new node corresponds to the value of $k$ that maximizes this similarity metric. Our method is fast and numerically outperforms existing semi-supervised algorithms. Theoretically, we derive explicit bounds for the misclassification error and show the efficiency of our method by comparing it with an ideal classifier. Our findings highlight, to the best of our knowledge, the first semi-supervised community detection algorithm that offers theoretical guarantees. | ['Tracy Ke', 'Yicong Jiang'] | 2023-06-01 | null | null | null | null | ['stochastic-block-model', 'community-detection'] | ['graphs', 'graphs'] | [ 5.81343994e-02 5.06704748e-01 -3.80919158e-01 -2.35110059e-01
-2.86001027e-01 -7.56182730e-01 3.19648325e-01 6.61917627e-01
-1.42306566e-01 5.00059843e-01 -2.60263234e-01 -7.82345608e-02
-4.89894867e-01 -9.10186231e-01 -3.05398405e-01 -7.09678888e-01
-6.24970376e-01 1.02501714e+00 2.52750427e-01 4.01018590e-01
2.18651056e-01 2.39996493e-01 -9.79132593e-01 -1.09376386e-01
7.54248083e-01 6.00794315e-01 3.93699966e-02 5.69497108e-01
9.81354937e-02 7.59359360e-01 -3.16112965e-01 -4.96309012e-01
3.28892231e-01 -3.82732779e-01 -1.11540496e+00 6.51615858e-01
-2.93151960e-02 1.69753000e-01 -3.75248581e-01 1.22821736e+00
4.74316366e-02 -3.66603196e-01 8.36882114e-01 -1.66720080e+00
-2.29901463e-01 1.14570725e+00 -8.85362923e-01 1.80108547e-01
2.60529310e-01 -5.44124722e-01 1.30652976e+00 -4.98292983e-01
9.95998204e-01 8.51671040e-01 8.82223547e-01 1.68290794e-01
-1.58747625e+00 -7.50757575e-01 5.97072095e-02 -1.10966645e-01
-1.93064678e+00 -1.28529266e-01 9.29097891e-01 -7.03875959e-01
3.19702059e-01 -8.91043097e-02 5.38843989e-01 4.14173096e-01
-5.19240797e-01 4.85593826e-01 1.08146489e+00 -5.52069545e-01
4.69707459e-01 4.05050844e-01 3.75810027e-01 8.83713841e-01
6.93911076e-01 -2.70702690e-01 -2.89824873e-01 -7.22572148e-01
3.70674878e-01 1.62207648e-01 2.70454790e-02 -1.02327824e+00
-1.10856128e+00 8.81546021e-01 3.76057565e-01 2.39315227e-01
-1.12567097e-01 2.24732786e-01 2.11050600e-01 4.59205866e-01
5.56103647e-01 4.92110290e-02 -1.91679731e-01 5.04256129e-01
-1.27942669e+00 -5.07920921e-01 1.18312109e+00 1.10726249e+00
1.08315372e+00 -4.35845435e-01 5.98783433e-01 5.92709124e-01
5.31858742e-01 3.53646755e-01 -2.52271622e-01 -1.34081972e+00
1.34165376e-01 9.93122995e-01 2.36248039e-02 -9.79111373e-01
-2.78928995e-01 -6.93294227e-01 -9.21360612e-01 -3.54106799e-02
5.76320767e-01 -1.75022244e-01 -3.01593959e-01 1.63730562e+00
4.07710135e-01 3.50346982e-01 -1.08854547e-01 5.07349133e-01
2.96024412e-01 -4.48985249e-02 -1.77502587e-01 -7.14868128e-01
8.25657964e-01 -8.69622171e-01 -2.58125454e-01 1.31542817e-01
7.32737601e-01 -4.85952705e-01 8.40718672e-02 3.31283808e-01
-7.84133136e-01 -1.35984141e-02 -9.90635693e-01 7.65133679e-01
3.38683315e-02 8.96968879e-03 5.34939051e-01 8.30969095e-01
-1.25909245e+00 6.57437921e-01 -7.02095270e-01 -6.62330091e-01
2.72568464e-01 3.42138737e-01 -3.55946898e-01 -3.01248342e-01
-6.59621954e-01 3.46603006e-01 3.11166793e-01 -2.47544631e-01
-8.62413526e-01 -2.32673153e-01 -5.38328409e-01 -5.98837659e-02
7.36769140e-01 -4.02220488e-01 8.49163711e-01 -1.20907509e+00
-5.65846860e-01 1.12239361e+00 -2.89741695e-01 -5.45539737e-01
3.65616232e-01 8.16845894e-01 -2.52103776e-01 6.41239822e-01
3.99109870e-01 2.49256447e-01 5.94188869e-01 -1.61838400e+00
-5.66679120e-01 -3.77169102e-01 -3.55060510e-02 4.35198545e-02
-3.81310344e-01 -1.42017333e-02 -1.60722837e-01 -3.89569849e-01
6.89351916e-01 -1.12316644e+00 -4.22834724e-01 2.20261812e-01
-5.92026770e-01 -3.65167290e-01 6.14693403e-01 -1.71076179e-01
1.21515942e+00 -2.06593251e+00 -1.17941037e-01 1.00100374e+00
1.00910592e+00 -1.61853522e-01 -2.80160997e-02 8.31453621e-01
2.94734277e-02 3.27119172e-01 -4.43613529e-01 -5.01468599e-01
-2.53364593e-01 4.19929065e-02 1.65468708e-01 9.02625382e-01
-2.66150951e-01 2.46824890e-01 -1.30416834e+00 -7.70363390e-01
-2.69879222e-01 3.55633497e-02 -3.37485313e-01 -1.50978386e-01
1.50513992e-01 4.95903753e-03 -2.99415678e-01 6.13506138e-01
5.93492508e-01 -9.10151362e-01 1.17879844e+00 2.84342974e-01
2.84153372e-01 9.76080354e-03 -1.60580289e+00 1.15733051e+00
9.33880135e-02 6.34846091e-01 6.86758459e-01 -1.36920607e+00
1.03484929e+00 4.67195183e-01 8.46011341e-01 4.80568200e-01
-1.93354376e-02 3.49091679e-01 4.68187593e-02 -8.32437500e-02
3.82804126e-02 -9.13833827e-03 9.44761485e-02 1.25970840e+00
-7.25126592e-03 3.70372683e-01 3.39047372e-01 9.57486808e-01
1.42493248e+00 -6.11249924e-01 3.78300458e-01 -4.98841047e-01
3.25413227e-01 2.39481293e-02 4.44823831e-01 1.00616658e+00
-4.26016331e-01 1.66017950e-01 6.82230592e-01 1.20724134e-01
-1.03398585e+00 -1.15037549e+00 1.06850185e-01 8.18629801e-01
3.16466153e-01 -4.58971977e-01 -8.31517160e-01 -7.56947398e-01
3.00398588e-01 -1.67190313e-01 -5.47057927e-01 1.47008568e-01
-2.37073135e-02 -5.16113997e-01 2.95582086e-01 2.32146397e-01
1.91662699e-01 -5.43186843e-01 3.27113211e-01 2.67688513e-01
-4.32335049e-01 -1.01899564e+00 -5.67324460e-01 2.39448678e-02
-1.05732274e+00 -1.72890139e+00 -2.23126680e-01 -1.20511472e+00
1.28220606e+00 6.13729835e-01 9.76529717e-01 7.14985847e-01
-5.71326260e-03 5.06007969e-01 -4.88371849e-01 3.22437257e-01
-7.07273662e-01 1.12312838e-01 2.57525384e-01 2.98597872e-01
4.53366220e-01 -7.30372846e-01 -7.09043682e-01 5.31036794e-01
-4.10039365e-01 -3.17144573e-01 2.30846241e-01 5.72747231e-01
3.00128549e-01 5.62861443e-01 6.11487925e-01 -1.12922704e+00
4.70856667e-01 -8.28355312e-01 -5.65681934e-01 3.53934258e-01
-1.10623121e+00 -1.46087915e-01 3.41931015e-01 -3.14813524e-01
-5.47641456e-01 3.57689142e-01 7.15988219e-01 -9.71060023e-02
1.17110431e-01 7.27702916e-01 1.16454609e-01 -3.01230282e-01
6.18848562e-01 -6.07882142e-02 2.17642009e-01 -4.15380031e-01
2.73427486e-01 1.04793549e+00 2.52214164e-01 -4.60406601e-01
9.57330048e-01 9.00271952e-01 2.35366940e-01 -6.51271284e-01
-5.88559866e-01 -1.03632808e+00 -9.44103599e-01 -4.36223924e-01
-6.75407276e-02 -1.12603819e+00 -9.35947061e-01 2.79822022e-01
-7.92732477e-01 -2.30565965e-01 -1.27286881e-01 4.50102270e-01
-2.56489158e-01 9.27802026e-01 -7.42337406e-01 -1.22663832e+00
-1.14403285e-01 -5.38321614e-01 3.59967351e-01 -3.90192240e-01
-1.87278375e-01 -1.16995537e+00 2.01667309e-01 4.36990887e-01
-1.03806272e-01 1.84789225e-01 4.01031017e-01 -8.52191389e-01
-5.93201578e-01 -5.15120745e-01 -4.88685995e-01 2.02958241e-01
8.40852708e-02 2.46038109e-01 -6.09182596e-01 -6.82511151e-01
-4.55843925e-01 -1.16585121e-01 8.82794082e-01 2.38115966e-01
5.88771403e-01 -4.10971373e-01 -9.81571794e-01 8.47933739e-02
1.47430289e+00 -2.16403976e-01 8.97699445e-02 -2.10075825e-01
5.49603343e-01 7.65975177e-01 2.57444680e-01 5.83823383e-01
5.60501397e-01 1.56598285e-01 2.81590194e-01 1.72162011e-01
1.08834118e-01 -2.72073090e-01 3.33547778e-02 9.89710987e-01
-1.07251881e-02 -3.15688819e-01 -1.00924969e+00 8.83359551e-01
-1.81405973e+00 -1.02088964e+00 -4.15391654e-01 2.37833738e+00
7.10067213e-01 2.62705624e-01 4.98712838e-01 5.04433155e-01
1.35876906e+00 -2.25122496e-01 -3.13240141e-01 1.01425901e-01
-4.66747172e-02 6.50077686e-02 5.77675164e-01 7.44856477e-01
-1.04653060e+00 4.88422900e-01 6.88315964e+00 5.88759601e-01
-2.60134697e-01 2.25331888e-01 5.14277339e-01 1.02399848e-01
-1.50195569e-01 4.66291368e-01 -4.53164726e-01 3.29009980e-01
8.63982201e-01 -4.01043534e-01 5.48811555e-01 9.47035730e-01
1.58027336e-01 -1.45003885e-01 -1.09940159e+00 5.24596691e-01
6.65636137e-02 -1.21564949e+00 -4.84720498e-01 4.88134682e-01
1.08735538e+00 1.50787652e-01 -5.29522598e-01 -2.44750157e-01
1.04949403e+00 -5.12921035e-01 3.16546261e-01 3.08428645e-01
8.25263977e-01 -6.51707113e-01 4.31020409e-01 6.25160933e-01
-1.42726147e+00 -2.93250144e-01 -7.32949078e-02 -2.81388044e-01
-5.38189895e-02 9.70304966e-01 -1.02117097e+00 9.31869000e-02
4.77629125e-01 9.42399561e-01 -4.38887596e-01 1.16035366e+00
-3.82713079e-02 8.50380242e-01 -5.18388033e-01 -4.57099639e-02
-8.62882957e-02 -4.12482500e-01 6.00129128e-01 1.08438659e+00
1.00858435e-01 1.13061905e-01 4.71366018e-01 5.64182818e-01
-4.75878030e-01 5.49432859e-02 -5.69064319e-01 -1.91968471e-01
1.11426413e+00 1.31014585e+00 -1.32678413e+00 -5.42939544e-01
-2.58902758e-01 7.86157370e-01 6.69181585e-01 3.14678192e-01
-1.18590854e-01 -2.84998029e-01 2.48317480e-01 6.49281442e-01
4.48750615e-01 -2.08516017e-01 -1.40412539e-01 -1.00248671e+00
-2.35495105e-01 -4.15807992e-01 6.86963797e-01 -4.27521467e-01
-1.57238507e+00 2.02624306e-01 -2.25515261e-01 -1.16028702e+00
-8.01729336e-02 -8.15983936e-02 -4.92700964e-01 3.90436709e-01
-9.64254379e-01 -8.36975157e-01 -1.64521620e-01 3.73706728e-01
-1.86183110e-01 -1.39998600e-01 6.48621023e-01 1.72143981e-01
-4.40733522e-01 5.11561453e-01 5.73018670e-01 4.87720698e-01
5.36198914e-01 -1.33267939e+00 -6.90175593e-03 8.79856467e-01
-1.76742487e-02 6.06748283e-01 5.92756212e-01 -8.88045967e-01
-7.33321369e-01 -1.11581767e+00 1.22971308e+00 -2.11505502e-01
9.59360182e-01 -4.86601144e-01 -6.33612573e-01 7.35951364e-01
-1.03498161e-01 3.06977808e-01 9.04576957e-01 1.97948381e-01
-6.85282350e-01 -1.99807823e-01 -1.43598211e+00 1.73988581e-01
1.32367182e+00 -7.21215367e-01 3.87366675e-02 6.21897459e-01
3.95406127e-01 4.18724597e-01 -1.12127328e+00 1.77663535e-01
4.47639853e-01 -8.30513239e-01 8.30469966e-01 -2.17885911e-01
9.43859816e-02 -4.13075089e-01 -3.51997167e-02 -1.08034849e+00
-4.59230572e-01 -6.38364971e-01 -2.17827231e-01 1.09306538e+00
4.80444014e-01 -6.40656412e-01 1.34427583e+00 2.68288761e-01
6.56979859e-01 -3.63003284e-01 -8.42358589e-01 -8.37612808e-01
-1.92714557e-01 -5.21951765e-02 2.64779866e-01 1.39447463e+00
5.69294095e-01 2.73714125e-01 -1.45415127e-01 2.65830517e-01
1.57699549e+00 3.28299314e-01 3.86298180e-01 -1.95239985e+00
-3.78036648e-01 -2.65662193e-01 -2.77586967e-01 -7.68176198e-01
2.98423201e-01 -1.07264721e+00 -5.96405603e-02 -1.27863836e+00
8.03209305e-01 -9.19360816e-01 -1.81642771e-01 1.63085744e-01
1.34824365e-01 5.42153180e-01 -2.10462272e-01 5.40358245e-01
-1.03874564e+00 -3.09882283e-01 6.84249640e-01 -2.21954569e-01
-1.92254648e-01 2.28239566e-01 -7.30386317e-01 7.09716618e-01
9.13823843e-01 -8.45984221e-01 -3.74511123e-01 1.38598576e-01
4.51659292e-01 1.70615479e-01 2.19925851e-01 -5.68499386e-01
5.99391520e-01 -6.07305504e-02 -1.85851738e-01 -4.50621188e-01
-9.24575031e-02 -7.35594749e-01 4.35959935e-01 7.77634859e-01
-5.25527537e-01 -4.20677632e-01 -8.03677738e-01 1.28811502e+00
1.05918527e-01 -3.54869813e-01 7.17449069e-01 -2.15099260e-01
-1.19143374e-01 3.42224389e-01 -6.04506671e-01 1.80637278e-02
1.17096794e+00 -4.13210601e-01 -2.89080530e-01 -5.72153389e-01
-9.79075789e-01 3.90426248e-01 6.91783786e-01 -1.73017263e-01
2.79947579e-01 -1.03669357e+00 -8.89682174e-01 -1.72696322e-01
3.87519360e-01 -1.35433927e-01 -1.73852697e-01 8.24360669e-01
-3.79406869e-01 -2.40219459e-02 4.56880540e-01 -5.25100350e-01
-1.68083847e+00 7.74420381e-01 2.78303176e-01 -3.37058574e-01
-1.39902532e-01 5.63468874e-01 -3.48213047e-01 -5.48288405e-01
2.23812029e-01 3.87205154e-01 -1.66725330e-02 6.49541169e-02
1.54193357e-01 7.93766379e-01 -2.33089194e-01 -7.74067342e-01
-4.83673245e-01 1.74554810e-01 -9.91210341e-02 -1.03623755e-01
1.05404270e+00 -4.96192724e-01 -6.39156759e-01 2.78541535e-01
1.20252490e+00 -1.42970169e-02 -9.08019960e-01 -9.52867806e-01
2.22255841e-01 -3.33553195e-01 -9.76953804e-02 -3.68584216e-01
-1.14715815e+00 3.74953598e-01 1.25045717e-01 6.77831769e-01
6.85164750e-01 4.37882334e-01 1.64878651e-01 4.49900806e-01
6.44303262e-01 -1.13943565e+00 2.68502921e-01 6.33541942e-02
5.20911906e-03 -1.19037354e+00 3.18012893e-01 -9.37386811e-01
-1.72878370e-01 7.33285189e-01 1.96665630e-01 -3.68279219e-01
1.23292971e+00 1.40338898e-01 -2.33312622e-01 -3.49414438e-01
-7.24872887e-01 -2.01691955e-01 -1.79929003e-01 8.13658595e-01
1.62161604e-01 4.24205691e-01 -3.65199775e-01 1.92226917e-01
1.65723994e-01 -1.41913056e-01 8.55018139e-01 7.50437796e-01
-7.02893674e-01 -1.09510136e+00 -3.33530873e-01 8.12436461e-01
-2.30883375e-01 -5.11248782e-02 -7.60609210e-01 3.52477163e-01
2.78154202e-02 1.42252338e+00 1.28786594e-01 -3.22274745e-01
-2.82452822e-01 -9.07923728e-02 3.00065964e-01 -8.05424631e-01
-2.54506677e-01 9.84455347e-02 1.37291819e-01 6.91170171e-02
-8.56005669e-01 -8.60920727e-01 -1.07381201e+00 -7.80113995e-01
-8.26243460e-01 5.83897531e-01 4.17290121e-01 7.75987744e-01
2.86372364e-01 -2.51649797e-01 1.33095586e+00 -2.26462007e-01
-4.69667107e-01 -8.30122828e-01 -1.17895913e+00 3.99480432e-01
1.44549370e-01 -5.08743584e-01 -9.67275381e-01 1.32909447e-01] | [6.909959316253662, 5.176337242126465] |
abea5a3f-8744-440c-b8dd-2563345991eb | fairness-index-measures-to-evaluate-bias-in | 2306.10919 | null | https://arxiv.org/abs/2306.10919v1 | https://arxiv.org/pdf/2306.10919v1.pdf | Fairness Index Measures to Evaluate Bias in Biometric Recognition | The demographic disparity of biometric systems has led to serious concerns regarding their societal impact as well as applicability of such systems in private and public domains. A quantitative evaluation of demographic fairness is an important step towards understanding, assessment, and mitigation of demographic bias in biometric applications. While few, existing fairness measures are based on post-decision data (such as verification accuracy) of biometric systems, we discuss how pre-decision data (score distributions) provide useful insights towards demographic fairness. In this paper, we introduce multiple measures, based on the statistical characteristics of score distributions, for the evaluation of demographic fairness of a generic biometric verification system. We also propose different variants for each fairness measure depending on how the contribution from constituent demographic groups needs to be combined towards the final measure. In each case, the behavior of the measure has been illustrated numerically and graphically on synthetic data. The demographic imbalance in benchmarking datasets is often overlooked during fairness assessment. We provide a novel weighing strategy to reduce the effect of such imbalance through a non-linear function of sample sizes of demographic groups. The proposed measures are independent of the biometric modality, and thus, applicable across commonly used biometric modalities (e.g., face, fingerprint, etc.). | ['Sebastien Marcel', 'Ketan Kotwal'] | 2023-06-19 | null | null | null | null | ['benchmarking', 'benchmarking'] | ['miscellaneous', 'robots'] | [ 2.99677312e-01 -9.84318182e-02 -9.55931395e-02 -8.75715196e-01
-1.72084495e-01 -6.06929362e-01 6.50725245e-01 6.52113080e-01
-6.34296238e-01 8.85460377e-01 2.25759000e-01 -2.78729290e-01
-3.08205515e-01 -8.59078944e-01 -7.04751238e-02 -7.53944874e-01
3.32248390e-01 2.32176855e-01 -3.74559611e-01 -7.72811696e-02
7.35610187e-01 6.85811937e-01 -1.66410112e+00 -2.50573307e-01
1.19436967e+00 1.02519298e+00 -5.93256354e-01 4.38181490e-01
-9.59927961e-02 1.62023511e-02 -8.06270123e-01 -9.89978671e-01
3.62052947e-01 -4.84445840e-01 -4.72685486e-01 -2.89735913e-01
5.24796665e-01 -3.49127650e-01 3.29937428e-01 1.07196641e+00
9.16760087e-01 -1.59491226e-01 1.04261625e+00 -1.30605149e+00
-4.64669824e-01 6.06888294e-01 -6.89710140e-01 -1.36766523e-01
3.56964171e-01 1.11550391e-01 7.92554677e-01 -6.63548529e-01
2.43154719e-01 1.24906802e+00 6.47908211e-01 5.00945091e-01
-1.23103940e+00 -7.88863182e-01 -1.82724118e-01 -4.00233269e-02
-1.31193972e+00 -7.18188703e-01 4.97320443e-01 -6.07137442e-01
1.66232184e-01 3.98121387e-01 4.05781180e-01 5.29267609e-01
5.41409589e-02 6.11150265e-02 1.48319554e+00 -5.26336670e-01
2.40786105e-01 4.22698587e-01 3.71536762e-01 2.28485346e-01
8.36656153e-01 6.08687885e-02 -5.95094860e-01 -5.79877973e-01
4.17409599e-01 -2.46141031e-01 7.87190646e-02 -3.75436753e-01
-6.67006314e-01 6.95539355e-01 -7.79480562e-02 1.97232783e-01
-2.45360091e-01 -1.78818285e-01 4.42516387e-01 3.16479117e-01
4.32345718e-01 2.37085760e-01 -2.41146833e-01 -2.69776046e-01
-9.99108970e-01 3.39071006e-01 8.09111416e-01 4.78034109e-01
6.30267680e-01 -1.90228358e-01 -4.60117847e-01 9.16323245e-01
1.68444842e-01 6.69404209e-01 1.26680478e-01 -8.02067995e-01
3.15655887e-01 6.42982423e-01 3.68407398e-01 -1.15528870e+00
-2.94274241e-01 -1.58465594e-01 -1.02371466e+00 2.77729183e-01
1.07477105e+00 -1.56076565e-01 -3.40191752e-01 1.78803837e+00
3.10383469e-01 -4.77782100e-01 -2.14126468e-01 5.91535211e-01
4.98043239e-01 1.33161135e-02 3.35575521e-01 -4.22501624e-01
1.33106458e+00 6.31713718e-02 -5.23828804e-01 4.16395962e-01
1.07910283e-01 -8.23711574e-01 8.78500521e-01 2.12548494e-01
-1.06974220e+00 -4.20566171e-01 -6.75498188e-01 2.64700592e-01
-3.83612275e-01 4.05845493e-02 2.28727028e-01 1.72447157e+00
-9.37664568e-01 6.82339132e-01 -3.13295096e-01 -6.07719004e-01
4.09402877e-01 6.21474624e-01 -2.08460480e-01 1.82765618e-01
-1.01628053e+00 9.85020399e-01 -1.15866303e-01 1.86461806e-01
2.93168910e-02 -6.58772051e-01 -4.80314553e-01 1.68024912e-01
-1.03156835e-01 -5.31084776e-01 5.83994746e-01 -9.06109810e-01
-1.33103824e+00 9.90224957e-01 -2.47582048e-01 -1.89196378e-01
7.49321163e-01 6.66776225e-02 -2.70341814e-01 -3.49401176e-01
-1.81031957e-01 3.08861971e-01 6.55753255e-01 -1.17905176e+00
-4.02004689e-01 -7.18769670e-01 1.31935170e-02 4.38464843e-02
-5.99200368e-01 4.08356220e-01 3.51620466e-01 -4.74516988e-01
-2.23897710e-01 -6.47005558e-01 -4.13163826e-02 -4.14833911e-02
-2.42655173e-01 -3.99876684e-02 3.69271874e-01 -6.92620635e-01
1.45192480e+00 -1.88484895e+00 -2.47969687e-01 7.35904276e-01
5.92211671e-02 1.94589555e-01 9.86619890e-02 3.72324914e-01
1.97601870e-01 2.61897504e-01 -2.25554824e-01 -2.34313644e-02
7.82394633e-02 -1.88708007e-01 1.35022693e-03 7.27654576e-01
2.18904883e-01 3.65096271e-01 -6.05906606e-01 -3.97722304e-01
2.88690686e-01 3.99655730e-01 -5.05776465e-01 1.11857481e-01
5.31847119e-01 5.11436701e-01 -2.08465368e-01 7.80475676e-01
1.05265737e+00 3.57211620e-01 2.73465842e-01 -2.94263750e-01
-3.00519437e-01 9.95613821e-03 -1.18856692e+00 6.93069577e-01
-2.72400022e-01 1.93957984e-01 -1.03942513e-01 -9.06651139e-01
1.31270802e+00 3.54923122e-02 6.67066455e-01 -6.22039914e-01
4.48062271e-01 4.16740060e-01 3.33182037e-01 -9.69503298e-02
7.19008207e-01 -3.60950261e-01 -7.01563284e-02 6.41751170e-01
-1.26577154e-01 1.46666288e-01 2.21551865e-01 -3.96470994e-01
3.94729435e-01 -2.56125450e-01 6.64946198e-01 -6.35311067e-01
8.79104257e-01 -5.99388361e-01 5.11251390e-01 7.74287820e-01
-6.06992781e-01 5.76263607e-01 7.12430537e-01 -3.10751736e-01
-1.12075865e+00 -8.69573414e-01 -3.91013354e-01 8.00082207e-01
2.50256777e-01 -5.64633943e-02 -7.94150472e-01 -4.62504536e-01
4.32914495e-01 4.20287251e-01 -5.88009715e-01 -1.79010615e-01
-2.72892088e-01 -1.12326396e+00 7.74914980e-01 1.98197156e-01
3.24448705e-01 -6.67859018e-01 -7.20544279e-01 -1.44777566e-01
5.88614829e-02 -7.40778267e-01 -1.91431478e-01 -4.29663599e-01
-9.11003113e-01 -1.28098321e+00 -1.03059709e+00 3.85478474e-02
6.70676947e-01 -9.89784002e-02 1.03669786e+00 2.09105164e-01
-1.29724562e-01 5.09484470e-01 -1.11100353e-01 -6.31866574e-01
-4.23713058e-01 1.48412988e-01 3.37181479e-01 3.10933650e-01
5.20734012e-01 -4.42353338e-01 -8.61317396e-01 5.98048747e-01
-6.02805078e-01 -4.15631980e-01 2.56232113e-01 8.25054109e-01
6.55192062e-02 -2.60792613e-01 6.97234750e-01 -9.57698345e-01
9.92578626e-01 -2.48377740e-01 -4.82272446e-01 4.91362482e-01
-1.09399390e+00 -8.80805776e-03 4.31936860e-01 -1.72828987e-01
-1.13543606e+00 -3.38082284e-01 5.97602986e-02 2.86268175e-01
-3.41729634e-02 3.29804599e-01 -3.12582880e-01 -3.67244929e-01
6.25632644e-01 -1.96392924e-01 2.07532808e-01 -3.01317126e-01
8.42904449e-02 8.59925508e-01 3.44854444e-01 -1.03614545e+00
6.22033358e-01 2.34305650e-01 2.58623928e-01 -6.66798472e-01
-2.72275716e-01 -2.51626283e-01 -6.58921778e-01 -5.54335654e-01
1.68408915e-01 -3.56496066e-01 -1.22127843e+00 7.68864930e-01
-8.31413031e-01 2.57741094e-01 -1.90486982e-01 2.86001235e-01
-3.38851988e-01 6.35913551e-01 -2.05785349e-01 -1.49191189e+00
-5.46146095e-01 -9.38020885e-01 7.26427794e-01 6.46712899e-01
-4.32840198e-01 -8.53508234e-01 -2.37565823e-02 3.16897959e-01
6.90504611e-01 4.29866821e-01 9.63277340e-01 -4.71979767e-01
1.40643254e-01 -3.33499104e-01 -4.57007617e-01 3.89406711e-01
4.77721155e-01 3.98561120e-01 -1.11541128e+00 -3.11583400e-01
-2.66732246e-01 -6.20007999e-02 4.68873322e-01 5.23703516e-01
1.09138834e+00 -3.00647646e-01 1.59868240e-01 1.46458328e-01
1.41223490e+00 -2.13806704e-02 7.72749782e-01 5.78591339e-02
3.26009095e-01 1.26798451e+00 6.25373960e-01 9.21307027e-01
4.61670041e-01 5.92758060e-01 1.86604530e-01 -1.24256238e-02
2.77418554e-01 1.50115266e-01 1.60160765e-01 4.13759947e-01
-7.56535292e-01 1.01282476e-02 -1.04532433e+00 2.45695233e-01
-1.58909130e+00 -8.92024875e-01 -1.51573896e-01 2.92525935e+00
7.03698337e-01 -1.04050227e-01 6.13289058e-01 3.51846486e-01
1.12207723e+00 -1.09469041e-01 -5.10435820e-01 -7.79783010e-01
-2.46796116e-01 2.64407605e-01 7.43738830e-01 4.58601058e-01
-6.45529091e-01 4.30083424e-01 6.48261595e+00 5.62574625e-01
-1.07733607e+00 -1.86714038e-01 1.08085084e+00 1.02050252e-01
-2.79659897e-01 -1.79122575e-02 -6.79993451e-01 6.08298719e-01
8.63887608e-01 -6.92050993e-01 1.82883337e-01 3.46829712e-01
4.63000029e-01 -4.75178987e-01 -9.14401889e-01 9.89934444e-01
-1.18814297e-01 -6.90272748e-01 -7.84044154e-03 3.21062118e-01
5.58229804e-01 -7.34690547e-01 2.46533200e-01 -2.26378962e-01
-1.00422919e-01 -1.12078452e+00 5.44284582e-01 7.03081965e-01
9.25358653e-01 -9.96419430e-01 1.05921352e+00 -5.38798841e-03
-8.63522291e-01 -3.47651616e-02 -3.77786994e-01 -2.38160804e-01
-4.85278517e-02 9.11156416e-01 -4.44291800e-01 6.63084567e-01
3.44497740e-01 1.57657668e-01 -6.04347110e-01 1.05514777e+00
4.16521221e-01 3.63799840e-01 -1.73214212e-01 -5.22627570e-02
-4.28315520e-01 -3.74469578e-01 1.70647994e-01 1.05251348e+00
5.90691090e-01 -5.12681529e-02 -5.02051651e-01 8.55853856e-01
2.24681254e-02 6.17567003e-01 -4.39317733e-01 -5.06206229e-03
6.44424677e-01 1.13643372e+00 -5.36141574e-01 -8.47264752e-02
-2.84411579e-01 3.00823867e-01 -7.73534775e-02 3.94986756e-02
-5.01482785e-01 -1.99998036e-01 8.80798936e-01 2.40529194e-01
-3.43118757e-01 9.38212574e-02 -1.06642747e+00 -9.12190437e-01
5.26200235e-02 -9.36960697e-01 4.05302674e-01 3.29283737e-02
-1.56283522e+00 1.70301840e-01 1.22231543e-01 -1.22689998e+00
-2.17767745e-01 -4.86706585e-01 -5.72925746e-01 1.35084164e+00
-1.26673460e+00 -8.59445453e-01 -4.58326727e-01 4.31512594e-01
-3.57459158e-01 -3.07574332e-01 8.48804176e-01 4.72012758e-01
-3.86333674e-01 1.12630892e+00 1.05858482e-02 -1.12251580e-01
9.52189088e-01 -1.02087855e+00 1.91436812e-01 5.96454442e-01
-5.42865694e-01 8.01383972e-01 8.42614293e-01 -4.71814036e-01
-8.33881915e-01 -4.53347892e-01 9.51940656e-01 -4.73114192e-01
1.27966419e-01 -1.08266488e-01 -5.79223752e-01 -2.63850421e-01
-1.30459219e-01 -3.27789009e-01 1.13684762e+00 4.49139386e-01
-4.45229471e-01 -5.12247860e-01 -1.80123425e+00 5.01577497e-01
8.86130929e-01 -4.05999899e-01 -4.71859537e-02 -3.96003872e-01
-2.27624014e-01 -1.14899939e-02 -1.26140583e+00 5.93377173e-01
1.38632572e+00 -1.42898357e+00 7.91875899e-01 -3.97075206e-01
3.93526703e-01 -1.38584182e-01 -9.54452083e-02 -9.95353341e-01
-2.90347394e-02 -3.33046138e-01 1.43773183e-01 1.65888405e+00
3.20475519e-01 -8.63379657e-01 8.39122355e-01 1.16238105e+00
7.69258201e-01 -6.44391596e-01 -8.73592496e-01 -5.84114194e-01
2.57186145e-01 3.33759487e-02 9.05681431e-01 9.16509509e-01
-7.17987865e-03 -1.38284892e-01 -5.94384015e-01 -1.72778711e-01
8.96365762e-01 6.07992709e-02 9.60512400e-01 -1.57134008e+00
2.77222898e-02 -7.80235887e-01 -6.80054069e-01 -2.46754199e-01
-2.12479979e-01 -2.64297634e-01 -4.04725611e-01 -8.46893907e-01
5.01016915e-01 -7.92666256e-01 -5.06650746e-01 1.91931769e-01
-3.03739697e-01 3.22452366e-01 3.14788252e-01 6.56039193e-02
2.43542716e-01 2.89194137e-01 9.80477989e-01 -5.78522421e-02
-1.79022893e-01 3.23077649e-01 -1.07332373e+00 4.05641079e-01
9.23016548e-01 -2.77888834e-01 -1.12316191e-01 6.88941106e-02
7.30086416e-02 -3.68346944e-02 1.83925018e-01 -7.61155665e-01
7.33252838e-02 -5.33473551e-01 3.92143995e-01 -3.23377252e-01
-3.49571463e-03 -5.84595978e-01 8.66141692e-02 5.76646745e-01
-2.23322764e-01 1.23017155e-01 -6.07086346e-02 1.50501683e-01
-1.50328651e-01 -2.28455514e-01 9.41778541e-01 2.19296724e-01
-2.30712611e-02 1.67022422e-01 3.72080617e-02 -8.93875733e-02
6.82240248e-01 -5.74774563e-01 -3.86151284e-01 -4.79575247e-01
-3.67509186e-01 7.59949610e-02 6.66073561e-01 2.40898982e-01
2.52133727e-01 -1.21101260e+00 -9.11284924e-01 -2.18637884e-02
3.17517340e-01 -6.82116091e-01 2.27006689e-01 9.41062331e-01
-3.36422652e-01 2.19323963e-01 -7.37427115e-01 -3.97672057e-01
-1.68147159e+00 4.14271206e-02 4.02327359e-01 -7.09028095e-02
3.45564812e-01 3.96810830e-01 4.36887220e-02 -3.54613453e-01
2.62733102e-01 -4.14107069e-02 -4.01069522e-01 3.63504559e-01
4.48037565e-01 7.72499859e-01 5.83461449e-02 -7.59028196e-01
-5.84192693e-01 6.88978732e-01 1.94205821e-01 -1.62812307e-01
1.12302506e+00 -2.72147954e-01 -3.02275896e-01 3.47531885e-01
6.05253100e-01 3.77917886e-01 -7.68075168e-01 5.15444316e-02
1.24412894e-01 -9.31088805e-01 -6.48378551e-01 -7.08994806e-01
-9.34539437e-01 6.73865259e-01 7.63826191e-01 3.40637505e-01
1.20159066e+00 -6.20945454e-01 2.04801053e-01 -6.03611581e-02
5.48192501e-01 -1.17149675e+00 -4.90671307e-01 2.33199790e-01
6.33675396e-01 -1.27678168e+00 4.09015000e-01 -5.24612784e-01
-4.89945769e-01 1.19169986e+00 4.55365092e-01 1.71508610e-01
6.79571331e-01 1.08395494e-03 1.41172171e-01 2.79412329e-01
-1.19224280e-01 5.35956770e-02 3.89516205e-01 7.43635118e-01
9.08093512e-01 3.54234159e-01 -1.24936509e+00 7.01210082e-01
-3.83834362e-01 5.73324673e-02 5.34312785e-01 7.02493310e-01
-1.15567267e-01 -1.57794166e+00 -5.02918243e-01 7.18578756e-01
-5.40811300e-01 4.09111828e-02 -6.19692802e-01 4.82933313e-01
3.19605350e-01 8.84640455e-01 -1.10929519e-01 -4.00101393e-01
4.02899951e-01 5.23836315e-02 7.12308288e-01 -2.17885375e-01
-8.72357547e-01 -4.58637863e-01 1.63210288e-01 -2.02425733e-01
-7.76071131e-01 -8.90585661e-01 -4.77576464e-01 -1.01871157e+00
-4.10465747e-01 7.22088367e-02 8.05879414e-01 7.64845490e-01
1.62120268e-01 -2.23053858e-01 8.79029930e-01 -6.20671332e-01
-7.50455081e-01 -8.69949818e-01 -7.11993039e-01 4.34519410e-01
-1.76723171e-02 -7.52391398e-01 -1.63302094e-01 -3.49605173e-01] | [13.032125473022461, 1.301426887512207] |
ed83896c-3643-4204-bcf7-c87abf43e487 | recent-developments-in-machine-learning | 2303.10257 | null | https://arxiv.org/abs/2303.10257v1 | https://arxiv.org/pdf/2303.10257v1.pdf | Recent Developments in Machine Learning Methods for Stochastic Control and Games | Stochastic optimal control and games have found a wide range of applications, from finance and economics to social sciences, robotics and energy management. Many real-world applications involve complex models which have driven the development of sophisticated numerical methods. Recently, computational methods based on machine learning have been developed for stochastic control problems and games. We review such methods, with a focus on deep learning algorithms that have unlocked the possibility to solve such problems even when the dimension is high or when the structure is very complex, beyond what is feasible with traditional numerical methods. Here, we consider mostly the continuous time and continuous space setting. Many of the new approaches build on recent neural-network based methods for high-dimensional partial differential equations or backward stochastic differential equations, or on model-free reinforcement learning for Markov decision processes that have led to breakthrough results. In this paper we provide an introduction to these methods and summarize state-of-the-art works on machine learning for stochastic control and games. | ['Mathieu Laurière', 'Ruimeng Hu'] | 2023-03-17 | null | null | null | null | ['energy-management'] | ['time-series'] | [-9.78554338e-02 -8.59414190e-02 -1.78098470e-01 4.98531669e-01
-4.64946568e-01 -1.98381469e-01 6.45862281e-01 -1.03043132e-01
-6.08093679e-01 1.08500946e+00 -3.01286578e-01 -4.38916445e-01
-4.74155992e-01 -8.84434938e-01 -2.64962852e-01 -1.11409867e+00
-2.46154666e-01 6.12889886e-01 1.40767574e-01 -4.29470599e-01
2.46968076e-01 3.01373184e-01 -1.52406287e+00 -2.82331795e-01
6.85697734e-01 1.25467122e+00 2.26242855e-01 9.37236786e-01
-2.83260942e-01 9.63713944e-01 -1.67834565e-01 -2.51063854e-01
1.47653341e-01 -5.86315930e-01 -4.83442426e-01 -1.04583809e-02
-4.84412938e-01 5.10555925e-03 -4.53865021e-01 1.09769619e+00
6.23517811e-01 5.63775837e-01 1.07793593e+00 -1.10631216e+00
-6.08975232e-01 2.69101501e-01 -4.25494075e-01 1.26529604e-01
-5.66985160e-02 4.23848122e-01 9.08216476e-01 -6.09126091e-01
3.17319691e-01 1.17216599e+00 8.06008458e-01 7.82147944e-01
-1.06715715e+00 -3.26102257e-01 1.50169745e-01 4.74092066e-01
-8.14281762e-01 1.28344730e-01 8.01551104e-01 -6.67336583e-01
8.78015578e-01 -1.38520434e-01 9.46441472e-01 9.93625879e-01
5.59077740e-01 9.07569528e-01 1.27014017e+00 -5.87555647e-01
8.08915675e-01 -2.29945391e-01 -3.67217511e-01 7.10556090e-01
1.04296647e-01 4.79120016e-01 2.64854908e-01 -1.48517732e-02
8.11198413e-01 5.74592724e-02 1.12588093e-01 -7.87107646e-01
-9.69764113e-01 1.64035118e+00 4.18300815e-02 3.04888874e-01
-6.89148784e-01 3.28664541e-01 3.60612690e-01 3.08034062e-01
4.15991247e-01 6.04658544e-01 -6.15474582e-01 -2.38081202e-01
-6.52603745e-01 7.34104276e-01 1.15646434e+00 5.07580042e-01
4.93683726e-01 3.43670428e-01 -2.55504921e-02 6.21284425e-01
2.87749529e-01 6.33029461e-01 3.86241347e-01 -1.49070418e+00
9.03157517e-02 1.14469476e-01 4.50904250e-01 -8.79835367e-01
-6.23162866e-01 -1.75424237e-02 -1.59643745e+00 7.10996866e-01
6.06498122e-01 -7.25320816e-01 -5.64352989e-01 1.60845041e+00
3.65751773e-01 2.14059442e-01 1.55187652e-01 8.26448798e-01
6.84897527e-02 6.99018955e-01 -1.53248638e-01 -4.20708537e-01
9.76157427e-01 -6.92621291e-01 -8.03056538e-01 3.50907780e-02
4.98949051e-01 -3.27000380e-01 7.46149302e-01 5.28903842e-01
-1.13895047e+00 -1.07402153e-01 -8.34244072e-01 4.22754616e-01
-5.78504503e-01 -4.82724011e-01 7.08624423e-01 6.41243100e-01
-1.10780287e+00 1.14821780e+00 -9.27446365e-01 -2.60195285e-01
4.32130694e-01 3.65669280e-01 2.60353297e-01 2.19541788e-01
-1.65006948e+00 1.03965962e+00 -2.23245528e-02 1.99071050e-01
-9.43375349e-01 -4.20414031e-01 -8.02885950e-01 3.07622328e-02
7.88567185e-01 -9.42744553e-01 1.68049383e+00 -8.09024215e-01
-2.11769366e+00 6.49849772e-01 3.24662894e-01 -6.48247182e-01
9.63495791e-01 8.11787546e-02 1.23567455e-01 -1.45767154e-02
-1.82751622e-02 1.81859955e-01 8.83725464e-01 -6.09274983e-01
-5.24446785e-01 -1.69762239e-01 2.99414426e-01 4.99060452e-02
-5.45798242e-02 4.65619192e-02 3.74696106e-01 -5.66326141e-01
-5.11692524e-01 -1.05436492e+00 -1.05773962e+00 -1.74347516e-02
-1.68523580e-01 -3.61404419e-01 5.32930553e-01 -1.89385906e-01
9.72009301e-01 -1.41616774e+00 8.53512347e-01 -2.56991267e-01
1.04095377e-01 4.86879081e-01 1.41131908e-01 6.94956064e-01
1.32856565e-02 2.26255000e-01 -3.71109277e-01 -3.66891325e-01
2.68314004e-01 1.51281089e-01 -4.65580933e-02 4.81824577e-01
2.20093399e-01 8.69392335e-01 -9.02504086e-01 -3.50134283e-01
5.32512844e-01 2.19344899e-01 -6.27668381e-01 5.75303249e-02
-5.85697293e-01 5.57300091e-01 -6.92201436e-01 1.39771830e-02
3.85968029e-01 -2.41807893e-01 -7.94185624e-02 6.40886366e-01
-2.53859133e-01 -2.26732995e-02 -1.58768904e+00 1.32191992e+00
-7.80316174e-01 4.08873886e-01 4.43917602e-01 -1.72850215e+00
3.30276638e-01 3.88154566e-01 8.90474975e-01 -5.39292812e-01
4.09886509e-01 1.12563990e-01 -6.75878003e-02 -6.00977719e-01
2.61329561e-01 -8.40801001e-01 -2.29141369e-01 4.92507905e-01
-1.46182626e-01 -7.48761714e-01 2.65692085e-01 -1.68875396e-01
1.06529176e+00 -1.13886297e-01 4.19659883e-01 -3.66966218e-01
7.42605984e-01 -6.95847347e-02 4.33419406e-01 9.21529293e-01
-3.25619161e-01 2.90175438e-01 7.86142528e-01 -4.35153514e-01
-1.09504056e+00 -6.79428935e-01 3.05909663e-01 7.75785923e-01
-8.98276493e-02 9.60475057e-02 -8.29693079e-01 -8.86890590e-02
3.05096898e-02 3.86552125e-01 -8.39351058e-01 -3.79441082e-01
-5.57880998e-01 -9.22896147e-01 3.43874283e-02 3.20895433e-01
3.52454513e-01 -1.52047634e+00 -5.21466434e-01 5.93396008e-01
3.59226286e-01 -9.46248949e-01 -5.99060580e-02 2.26191521e-01
-8.38746667e-01 -1.04649782e+00 -1.18605518e+00 -5.48145533e-01
-2.38650039e-01 -3.08277905e-01 9.68477130e-01 -2.70141959e-01
-5.03878534e-01 4.45693076e-01 4.69122343e-02 -7.49762595e-01
-5.20352602e-01 3.30473512e-01 4.51932460e-01 2.45380793e-02
-7.30130300e-02 -5.85326254e-01 -3.31973821e-01 6.98915124e-02
-7.30923831e-01 -7.71122351e-02 2.15018913e-01 9.71045256e-01
3.96967143e-01 1.01541981e-01 7.68043637e-01 -5.81995726e-01
8.83161366e-01 -5.14594972e-01 -1.12881446e+00 -2.34253913e-01
-4.25421059e-01 3.56254548e-01 8.80551398e-01 -6.35677636e-01
-8.23414326e-01 -9.92532670e-02 -2.17031524e-01 -4.72871900e-01
1.06596239e-01 6.19085014e-01 2.53753066e-01 9.25083980e-02
3.55496377e-01 2.39617601e-01 3.85614693e-01 -2.56857872e-01
2.72594482e-01 4.47131276e-01 -1.70007378e-01 -5.75490892e-01
5.39791882e-01 4.45313662e-01 5.06867170e-01 -8.72027576e-01
-8.78358424e-01 3.91422398e-03 -5.98595500e-01 -1.94056317e-01
1.18276489e+00 -4.65841740e-01 -1.23536348e+00 9.72800493e-01
-1.06568110e+00 -1.04534805e+00 -6.60625517e-01 4.89682943e-01
-1.45915389e+00 -2.02759467e-02 -9.70092475e-01 -1.50649107e+00
1.45656019e-01 -1.13627148e+00 6.44874036e-01 4.03164744e-01
2.38359585e-01 -1.52254891e+00 4.33016390e-01 -3.16403270e-01
6.67709589e-01 2.99619645e-01 8.30969393e-01 -1.20418176e-01
-4.30832654e-01 -8.71827751e-02 2.39469260e-01 3.55730057e-01
-9.89037082e-02 -4.57670167e-02 -5.07955074e-01 -6.78609535e-02
2.37148330e-01 -3.04866433e-01 9.76843715e-01 9.76647615e-01
9.86674309e-01 -1.59067333e-01 -2.03683883e-01 4.53848660e-01
1.48009789e+00 3.53085160e-01 2.18737677e-01 2.60646075e-01
4.56195235e-01 6.89067364e-01 3.60791862e-01 6.55232131e-01
2.55940586e-01 4.48559493e-01 6.99701488e-01 -3.75006236e-02
5.39900243e-01 9.48953852e-02 1.99504063e-01 5.41274846e-01
-2.98544556e-01 -1.81595221e-01 -9.06470835e-01 4.52240348e-01
-2.18794799e+00 -1.27688062e+00 -1.20024486e-02 1.97349524e+00
6.17969990e-01 5.05525656e-02 2.81735539e-01 1.94200873e-01
8.42677653e-01 1.69900596e-01 -7.36668348e-01 -6.88671410e-01
-1.42476493e-02 2.33820766e-01 4.34821695e-01 5.69496453e-01
-1.21613050e+00 9.02629852e-01 6.81922340e+00 9.19237971e-01
-1.02061069e+00 7.57004917e-02 8.03907871e-01 -8.51588249e-02
2.97017038e-01 -5.19457281e-01 -5.87675571e-01 4.02223438e-01
1.02064335e+00 -2.36739919e-01 7.35871673e-01 8.39646697e-01
6.80315673e-01 -3.63477260e-01 -7.27797151e-01 1.03557813e+00
-6.15276873e-01 -1.61529446e+00 -4.57206726e-01 3.83406341e-01
1.15802598e+00 9.59261209e-02 3.00440818e-01 6.77781343e-01
8.16468596e-01 -1.16116345e+00 4.38939333e-01 7.01258957e-01
5.17508268e-01 -9.88999069e-01 7.99158990e-01 6.39303029e-01
-9.61705744e-01 -4.87574011e-01 -3.91597897e-01 -8.03912580e-01
2.81486720e-01 6.11397028e-01 3.93579341e-02 1.06824741e-01
6.13656938e-01 8.63547921e-01 1.57848790e-01 9.11353767e-01
5.06416410e-02 3.58090907e-01 -3.74162257e-01 -8.92984688e-01
6.12475753e-01 -6.60132766e-01 6.25077248e-01 8.91724169e-01
2.64366835e-01 2.38307014e-01 2.59281069e-01 9.12979066e-01
2.41632298e-01 2.30041742e-02 -8.48871291e-01 -2.40482122e-01
-3.04413468e-01 9.80429769e-01 -7.74462521e-01 -2.24232361e-01
-5.99447489e-01 7.69338906e-01 2.99357772e-01 3.71363878e-01
-8.58389854e-01 -3.64325732e-01 1.07512844e+00 5.26642054e-02
4.78425860e-01 -5.55362999e-01 -1.26773700e-01 -1.20262873e+00
-1.96959123e-01 -5.52497327e-01 3.72831106e-01 -2.42519438e-01
-1.44684982e+00 9.12094191e-02 -4.83738892e-02 -8.26824784e-01
-7.82023728e-01 -1.03466284e+00 -7.43517458e-01 6.97296202e-01
-1.27321625e+00 -4.82478112e-01 4.42588270e-01 5.78621805e-01
4.98246729e-01 -3.76992762e-01 7.42674351e-01 2.13958789e-02
-6.68288946e-01 -1.55064389e-01 8.68969738e-01 1.17597058e-01
6.14757128e-02 -1.53966141e+00 1.87052786e-01 4.67449486e-01
-2.56954044e-01 -8.60340595e-02 8.49940896e-01 -1.67850882e-01
-1.49843156e+00 -7.19463408e-01 5.01722276e-01 -3.88541013e-01
1.23835325e+00 -5.47463298e-01 -6.61100984e-01 1.43027708e-01
1.47154033e-01 1.54575109e-01 6.80611357e-02 -1.58224076e-01
4.26962972e-01 6.71969950e-02 -1.07995999e+00 6.78364933e-01
7.51965463e-01 -3.16093445e-01 -1.73099890e-01 3.23153973e-01
5.47239900e-01 -5.62699437e-02 -4.29644585e-01 8.38570595e-02
4.22635317e-01 -8.89566839e-01 8.86282086e-01 -9.99446630e-01
5.57990074e-01 1.89380258e-01 1.23087704e-01 -1.57261240e+00
-2.84484059e-01 -1.18020487e+00 -4.97581214e-01 6.23197973e-01
1.83604091e-01 -6.85378611e-01 8.24008942e-01 4.44865495e-01
9.24294069e-02 -1.17731881e+00 -1.25944197e+00 -8.00565124e-01
9.54707682e-01 -6.19219959e-01 1.21380702e-01 6.06274903e-01
-4.29724455e-02 2.13686123e-01 -3.76705468e-01 -3.96541178e-01
6.36860073e-01 7.73038045e-02 2.92389333e-01 -1.34282029e+00
-4.24380660e-01 -1.03286886e+00 -3.19752723e-01 -8.32777619e-01
5.48962295e-01 -4.06818807e-01 1.30465791e-01 -1.53893125e+00
-8.60185176e-02 -2.85235494e-01 -2.71932900e-01 -1.76084504e-01
-1.23655036e-01 4.80411239e-02 1.36800691e-01 -3.30647558e-01
-7.70916641e-01 9.98759866e-01 1.29972780e+00 -2.77730316e-01
-3.02571625e-01 6.42584503e-01 -4.21540827e-01 9.00384426e-01
1.02789414e+00 -3.93602252e-01 -1.32543683e-01 -3.39903161e-02
4.52062488e-01 4.23894167e-01 2.42664352e-01 -1.14492416e+00
1.94404289e-01 -6.52040005e-01 -3.90379131e-02 -8.19811970e-02
3.66017967e-01 -5.93216479e-01 -4.15711701e-01 8.88709068e-01
-3.03373992e-01 1.97852269e-01 -1.98489334e-02 8.41185212e-01
-1.78524330e-01 -3.10321689e-01 1.27947807e+00 -5.40749311e-01
-3.95240933e-01 7.15041459e-01 -1.21490860e+00 4.85038519e-01
1.16620135e+00 1.75714016e-01 6.17467701e-01 -9.17879105e-01
-1.17770720e+00 3.10749590e-01 1.34361893e-01 1.92682132e-01
2.25288957e-01 -1.13631082e+00 -7.15155840e-01 -2.44470194e-01
-4.56541181e-01 -1.87361464e-01 3.14801216e-01 8.89559448e-01
-3.69114995e-01 3.88721794e-01 -6.52278736e-02 -3.29890043e-01
-5.77342391e-01 1.09997714e+00 7.82124817e-01 -8.88206065e-01
-3.64432633e-01 5.87362528e-01 1.15735590e-01 -3.69191408e-01
-2.78631854e-03 -4.48849052e-01 -2.15639278e-01 2.77018666e-01
4.52814341e-01 5.97787082e-01 -2.84587592e-01 -2.62424052e-01
-5.60112335e-02 6.66018307e-01 5.07720888e-01 -5.92916608e-01
1.49284148e+00 -1.74661189e-01 8.90981704e-02 8.37333500e-01
1.03069937e+00 -5.64253688e-01 -1.55016804e+00 -2.47692600e-01
1.72177106e-01 3.53708833e-01 6.32094592e-02 -2.14471847e-01
-1.15770531e+00 1.44809425e+00 5.20496666e-01 7.03510761e-01
8.67448807e-01 -2.13179126e-01 6.60288274e-01 6.34557962e-01
6.88252091e-01 -1.53882146e+00 2.60319799e-01 1.21670568e+00
7.26477385e-01 -1.47382891e+00 -3.26228708e-01 8.15113783e-02
-4.53214914e-01 1.21410441e+00 2.71836907e-01 -7.40656316e-01
1.29561436e+00 4.96917397e-01 -3.73446941e-01 2.32655406e-01
-8.98552239e-01 -6.11738682e-01 -1.92421064e-01 8.12953174e-01
4.54263687e-02 -2.54884129e-04 -2.27459386e-01 5.99639058e-01
6.91326857e-02 1.99125826e-01 6.66129231e-01 7.85509288e-01
-4.21706170e-01 -9.60506976e-01 -1.66381136e-01 4.38469291e-01
-7.27789879e-01 -6.75219372e-02 -1.12668000e-01 7.81752706e-01
-2.35862747e-01 8.72419536e-01 1.12184301e-01 5.18146232e-02
4.97337282e-02 1.18112482e-01 6.19017482e-01 -5.35656989e-01
-2.90987223e-01 -2.36220807e-01 -2.90831536e-01 -4.64885980e-01
-4.05814886e-01 -9.20287669e-01 -1.04981899e+00 -3.51747394e-01
-2.78927684e-01 1.87546849e-01 6.21712565e-01 1.22502577e+00
9.40084979e-02 5.55183113e-01 5.61488926e-01 -1.33104980e+00
-9.16988134e-01 -9.69965518e-01 -8.32880318e-01 -1.34181365e-01
5.82458675e-01 -1.04110336e+00 -3.76726121e-01 -3.68347526e-01] | [4.1397223472595215, 2.4264612197875977] |
270419e0-25fb-41c9-9428-d1f313712cba | time-series-continuous-modeling-for | 2306.05880 | null | https://arxiv.org/abs/2306.05880v2 | https://arxiv.org/pdf/2306.05880v2.pdf | Time Series Continuous Modeling for Imputation and Forecasting with Implicit Neural Representations | Although widely explored, time series modeling continues to encounter significant challenges when confronted with real-world data. We propose a novel modeling approach leveraging Implicit Neural Representations (INR). This approach enables us to effectively capture the continuous aspect of time series and provides a natural solution to recurring modeling issues such as handling missing data, dealing with irregular sampling, or unaligned observations from multiple sensors. By introducing conditional modulation of INR parameters and leveraging meta-learning techniques, we address the issue of generalization to both unseen samples and time window shifts. Through extensive experimentation, our model demonstrates state-of-the-art performance in forecasting and imputation tasks, while exhibiting flexibility in handling a wide range of challenging scenarios that competing models cannot. | ['Vincent Guigue', 'Nicolas Baskiotis', 'Ghislain Agoua', 'Patrick Gallinari', 'Yuan Yin', 'Léon Migus', 'Louis Serrano', 'Etienne Le Naour'] | 2023-06-09 | null | null | null | null | ['imputation', 'meta-learning', 'imputation', 'imputation'] | ['computer-vision', 'methodology', 'miscellaneous', 'time-series'] | [ 5.11368811e-01 -3.96766037e-01 -5.12336195e-01 -3.43744665e-01
-9.21221256e-01 -3.74726057e-01 7.53632724e-01 -3.05411890e-02
-2.04118043e-01 8.42385113e-01 2.50219345e-01 -4.29153711e-01
-4.55584526e-01 -5.45583129e-01 -8.14854741e-01 -4.60125506e-01
-4.06680435e-01 1.71268687e-01 -6.05758190e-01 -5.35519831e-02
7.97885954e-02 4.54647303e-01 -1.59403062e+00 1.49833083e-01
7.70927131e-01 1.03153348e+00 -1.06168158e-01 2.65181690e-01
-6.03484921e-02 9.15272653e-01 -6.65338755e-01 8.30954239e-02
1.90605521e-01 1.69470146e-01 -6.98034987e-02 1.48637472e-02
3.76914628e-02 -2.97751009e-01 -3.76693279e-01 3.90305579e-01
2.30215833e-01 2.72559464e-01 5.88421702e-01 -1.43631804e+00
-5.84807098e-01 3.87834013e-01 -7.33131528e-01 2.80312210e-01
1.46199629e-01 4.49040812e-03 4.91786480e-01 -6.18975520e-01
1.83662437e-02 1.03176475e+00 1.27493906e+00 3.36947143e-01
-1.66726255e+00 -7.19932318e-01 5.33931434e-01 3.67107689e-02
-1.05285239e+00 -4.83636945e-01 9.13460314e-01 -4.23891813e-01
9.81193006e-01 2.48241499e-01 2.12728798e-01 1.69262099e+00
3.31724495e-01 8.50136280e-01 1.00752115e+00 -1.25830010e-01
3.70371431e-01 -4.06030923e-01 1.60181046e-01 -4.56412256e-01
-7.77221343e-04 3.46608758e-01 -4.53120798e-01 -5.81522465e-01
7.93773949e-01 7.28131592e-01 -1.64940003e-02 -9.80653614e-02
-1.33289003e+00 6.05498970e-01 -4.33871988e-03 4.13568653e-02
-7.24772692e-01 3.18359673e-01 3.53008807e-01 4.75239128e-01
7.09934235e-01 1.98846161e-01 -9.59751844e-01 -3.64210546e-01
-1.17286718e+00 3.28320175e-01 6.40083790e-01 9.31157649e-01
3.91886115e-01 6.24069214e-01 -8.70308653e-03 7.64472425e-01
1.98943820e-03 4.54315096e-01 6.05212629e-01 -9.03900981e-01
6.66301429e-01 3.65236223e-01 4.67849016e-01 -7.84358382e-01
-5.48360348e-01 -7.18016565e-01 -1.27649987e+00 -1.57037571e-01
3.89290571e-01 -2.21497685e-01 -1.18906093e+00 2.17641139e+00
7.33577386e-02 7.55360007e-01 6.22083284e-02 2.43437588e-01
2.40678623e-01 7.83934653e-01 3.33888769e-01 -5.52746475e-01
8.39118063e-01 -5.27106822e-01 -9.10174906e-01 -4.82202321e-01
1.76954046e-01 -3.70658994e-01 5.16290426e-01 4.47398901e-01
-9.56357181e-01 -5.90220511e-01 -8.63398015e-01 8.03002045e-02
-3.86222333e-01 -1.66714206e-01 7.56536126e-01 2.16420829e-01
-6.47951365e-01 6.88605368e-01 -1.33506441e+00 3.21737118e-02
2.07300738e-01 4.56700563e-01 -2.13530973e-01 -1.18586317e-01
-9.70157266e-01 7.38738775e-01 1.53044060e-01 2.57219106e-01
-6.64156795e-01 -1.20485449e+00 -8.14994037e-01 2.29163114e-02
3.05597842e-01 -6.35445952e-01 1.23012424e+00 -6.67114437e-01
-1.03102326e+00 2.23134696e-01 -5.10369956e-01 -7.41204143e-01
5.57235181e-01 -5.03921688e-01 -1.10899675e+00 -5.22659421e-01
-1.16817817e-01 7.52896070e-02 1.09538019e+00 -8.16603124e-01
-1.19611293e-01 -5.43209493e-01 -4.82853740e-01 -2.76105493e-01
-1.06239825e-01 -2.34966457e-01 1.29760757e-01 -1.23748755e+00
3.58080178e-01 -7.81954885e-01 -4.96216267e-01 -1.97801724e-01
-1.95760921e-01 -3.21901329e-02 1.02893269e+00 -7.45135069e-01
1.27304530e+00 -2.09773898e+00 5.95651157e-02 1.33711338e-01
-4.03005034e-02 6.26525059e-02 -1.95079163e-01 9.48507011e-01
-3.49368602e-01 -1.45813435e-01 -3.71699691e-01 -6.91904902e-01
2.33138055e-02 4.30423915e-01 -1.04715395e+00 2.58311391e-01
2.11058810e-01 8.56902421e-01 -6.13632619e-01 2.01292723e-01
2.39269644e-01 6.79017961e-01 -8.89813825e-02 1.94657683e-01
-2.98618585e-01 8.12257588e-01 -2.06195608e-01 8.90797257e-01
6.59790635e-01 -5.33716798e-01 2.05185622e-01 1.14151396e-01
-3.05437911e-02 1.07729435e-01 -1.26681209e+00 1.62076056e+00
-5.52658737e-01 3.85239094e-01 -9.83335152e-02 -1.12067771e+00
9.14377332e-01 4.34934109e-01 7.04571664e-01 -7.75490642e-01
-6.87891170e-02 2.34176852e-02 -3.11270863e-01 -4.33072746e-01
4.74879414e-01 -9.99632850e-02 -1.62780046e-01 5.15604019e-01
-4.32381965e-02 4.41477835e-01 -2.63813704e-01 -4.32997674e-01
1.06138337e+00 3.92772198e-01 2.55361557e-01 -1.85256097e-02
-1.16078846e-01 -3.40283304e-01 8.49888682e-01 1.02218390e+00
1.17295794e-01 5.22166431e-01 8.33620653e-02 -8.52334499e-01
-9.23465014e-01 -1.12613451e+00 -3.02506983e-01 1.11302710e+00
-5.13818204e-01 -1.45710126e-01 -4.79710177e-02 -5.41482382e-02
4.57691580e-01 8.61753941e-01 -9.47789013e-01 -1.55206680e-01
-5.21570683e-01 -1.23150527e+00 2.91809350e-01 9.35519457e-01
-7.37290876e-03 -8.45651269e-01 -6.08773112e-01 8.51931155e-01
-2.71343321e-01 -1.08418047e+00 -1.41095314e-02 4.68058348e-01
-1.35145605e+00 -6.95727944e-01 -4.22097057e-01 -2.19158605e-01
2.28560552e-01 3.10985029e-01 1.31040680e+00 -2.95620292e-01
-2.83281237e-01 4.36181903e-01 -7.82352537e-02 -7.34363139e-01
-4.74296957e-02 -1.63080152e-02 8.77951682e-02 1.93314463e-01
4.80411410e-01 -1.09905267e+00 -4.88732755e-01 1.48516268e-01
-1.10256183e+00 -1.31481618e-01 3.41990173e-01 8.83020520e-01
6.19470000e-01 -1.59348235e-01 1.30332220e+00 -5.84757984e-01
5.52108943e-01 -1.20603073e+00 -4.96597707e-01 3.01794171e-01
-5.85900009e-01 -9.79674682e-02 6.60831451e-01 -1.15565133e+00
-7.93654323e-01 -2.33569100e-01 9.50935557e-02 -6.87817931e-01
-2.51515985e-01 8.67653131e-01 1.95233390e-01 3.94635081e-01
3.34683746e-01 4.65928346e-01 9.57999304e-02 -7.92839408e-01
1.28149629e-01 5.92417955e-01 7.20190465e-01 -7.14664340e-01
6.47079468e-01 6.93575144e-01 7.22484365e-02 -8.36602569e-01
-9.09059465e-01 -3.03227454e-01 -4.67884392e-01 1.85262635e-01
2.32965946e-01 -1.39529192e+00 -5.14108539e-01 5.79137146e-01
-8.18918586e-01 -2.83142239e-01 -4.45486337e-01 5.27518570e-01
-7.07561731e-01 1.20304391e-01 -5.58607340e-01 -1.24635589e+00
-1.80634961e-01 -6.85134590e-01 1.11714983e+00 -1.33070335e-01
-5.00297129e-01 -1.20503390e+00 1.91131487e-01 -5.03024645e-02
7.47122467e-01 8.51442575e-01 9.51550961e-01 -6.39036834e-01
-4.84999180e-01 -5.86331129e-01 2.55734801e-01 6.66303635e-02
3.15711886e-01 -3.17484826e-01 -1.07904744e+00 -4.27376151e-01
2.47659415e-01 -1.80090964e-01 7.94490516e-01 5.95346332e-01
1.67960370e+00 -2.90170103e-01 -1.69279218e-01 6.77821338e-01
1.23139191e+00 2.45721877e-01 6.63719356e-01 3.18186909e-01
3.60354543e-01 4.12710428e-01 3.80237699e-01 9.10481393e-01
5.98850071e-01 5.05505919e-01 4.24057394e-01 -1.61101073e-02
5.33051074e-01 -4.57504988e-01 3.97041161e-03 6.31276488e-01
9.99998823e-02 -1.87839970e-01 -7.66421854e-01 8.67577255e-01
-2.25236464e+00 -1.34239721e+00 1.10793404e-01 2.44841886e+00
5.62430859e-01 -1.60174277e-02 1.89534202e-01 4.37552422e-01
3.79928738e-01 2.49474496e-01 -1.16574132e+00 -4.31222888e-03
-2.37255424e-01 2.02049702e-01 3.22298050e-01 -4.92959805e-02
-1.25698340e+00 2.72532612e-01 7.78934288e+00 3.70574743e-01
-1.10765028e+00 -8.97130743e-02 5.80181956e-01 -2.49221087e-01
-2.29067966e-01 -1.78080380e-01 -5.07817805e-01 7.72589386e-01
1.33209610e+00 -1.09842606e-01 5.64703524e-01 5.11285126e-01
3.26848030e-01 3.18419755e-01 -1.33555555e+00 1.03486907e+00
-1.23995081e-01 -1.28598416e+00 -1.65769711e-01 5.32570593e-02
7.80220449e-01 3.79606247e-01 2.79313684e-01 6.13531053e-01
2.11154893e-01 -1.24281347e+00 5.16901314e-01 9.74679530e-01
5.99786460e-01 -5.20726979e-01 4.16743666e-01 7.45926797e-01
-9.73798215e-01 -5.69606423e-01 -5.57126384e-03 -5.86512864e-01
2.46746957e-01 7.25784242e-01 -2.36365557e-01 7.07333684e-01
6.80717826e-01 1.04400694e+00 -1.14810377e-01 1.06793284e+00
5.89600623e-01 7.34077036e-01 -5.39789498e-01 6.49283826e-01
-6.27807602e-02 -6.57450333e-02 3.33838254e-01 7.87363529e-01
6.78816915e-01 -6.78294599e-02 2.55867392e-01 6.47473037e-01
8.67696479e-02 -4.42517221e-01 -7.68729985e-01 -4.43681292e-02
6.70567572e-01 7.16227591e-01 -1.18234336e-01 -2.38299072e-01
-5.44072628e-01 4.30899590e-01 3.36890161e-01 8.74139905e-01
-7.19895482e-01 7.90795088e-02 7.59995222e-01 4.08158600e-02
2.73933679e-01 -6.17637038e-01 -3.49882811e-01 -1.43314195e+00
2.18175054e-01 -9.26780164e-01 5.37894428e-01 -7.09004402e-01
-1.72798860e+00 3.71877044e-01 1.60062164e-01 -1.54147947e+00
-6.88850164e-01 -3.01470160e-01 -7.47778356e-01 8.70460749e-01
-1.55342638e+00 -1.28878474e+00 -6.77096993e-02 6.13551378e-01
5.43210447e-01 -3.30212489e-02 8.90453160e-01 3.58619899e-01
-4.53488350e-01 3.38939220e-01 6.35760009e-01 -2.83761352e-01
4.90129828e-01 -9.10952806e-01 5.82872033e-01 6.42925441e-01
-3.74401212e-02 9.07766461e-01 7.85243094e-01 -6.39908373e-01
-1.70607293e+00 -1.36894631e+00 6.56011462e-01 -5.03904283e-01
7.33202040e-01 -3.05148751e-01 -1.35765171e+00 1.23701727e+00
-1.23468287e-01 -7.53684491e-02 1.10330975e+00 4.65745866e-01
-6.95575297e-01 -1.72260031e-01 -1.12822604e+00 2.66804188e-01
8.19558501e-01 -6.53737903e-01 -7.72904694e-01 2.50802994e-01
5.38189173e-01 -3.62421960e-01 -1.09862554e+00 8.30267668e-01
6.45394027e-01 -4.86061513e-01 1.25579464e+00 -9.07930732e-01
1.70587718e-01 -8.23271945e-02 -4.90775257e-01 -1.21032858e+00
-3.54444981e-01 -8.01809609e-01 -9.83469188e-01 1.16072559e+00
1.19931445e-01 -7.65833795e-01 4.77550328e-01 9.10434723e-01
1.71807762e-02 -5.41254818e-01 -1.03783464e+00 -9.99838591e-01
1.49707511e-01 -8.21634054e-01 1.11645436e+00 1.27831626e+00
-2.03807324e-01 -1.47440717e-01 -1.18145251e+00 2.34459892e-01
9.59795654e-01 3.20867747e-01 6.07537568e-01 -1.68736291e+00
-4.30470467e-01 -3.81313749e-02 -1.84210166e-01 -9.36404347e-01
2.88039356e-01 -3.76626611e-01 -2.70632893e-01 -1.13344502e+00
-1.74794197e-01 -4.21062887e-01 -7.48333395e-01 5.17335176e-01
-4.33962792e-02 1.22140497e-01 -5.63497469e-03 3.75976205e-01
-3.08516622e-01 7.53963113e-01 6.03302181e-01 -1.18034206e-01
-2.37792566e-01 4.49479133e-01 -7.04561174e-01 5.49534619e-01
9.19266462e-01 -3.32208425e-01 -4.65426862e-01 -7.75387943e-01
1.09532274e-01 7.22121656e-01 6.49780214e-01 -9.24170375e-01
1.67872325e-01 -4.82056677e-01 8.28584373e-01 -1.01245511e+00
6.99025095e-01 -1.16002655e+00 7.68637002e-01 1.18364602e-01
-3.98883253e-01 6.22376263e-01 6.07917547e-01 1.05192947e+00
-1.70001447e-01 4.02568340e-01 2.46491194e-01 1.01417184e-01
-4.61780787e-01 3.87379795e-01 -3.51422876e-01 2.39317175e-02
8.05393577e-01 -2.38024279e-01 -3.30763847e-01 -6.08249962e-01
-1.07364380e+00 3.72716099e-01 2.55810648e-01 8.51778626e-01
3.63646388e-01 -1.33068991e+00 -5.75070143e-01 6.51794970e-01
1.65811121e-01 -2.04224274e-01 6.58488452e-01 7.16446161e-01
4.93435949e-01 1.35735959e-01 -7.21877664e-02 -6.72771573e-01
-4.79762495e-01 7.08792388e-01 4.32629623e-02 -2.08352327e-01
-7.76944697e-01 9.72761214e-02 -1.22681551e-01 -6.45034969e-01
4.34479415e-01 -4.83162135e-01 4.41125333e-02 4.26683985e-02
7.54599154e-01 5.43354750e-01 1.29567057e-01 -4.02198017e-01
-5.74329868e-02 2.73779064e-01 -2.93317921e-02 1.94713727e-01
1.70185900e+00 -1.48661479e-01 1.25595093e-01 1.12789929e+00
8.65641236e-01 -4.86533046e-01 -1.65463603e+00 -4.65560704e-01
1.70932457e-01 -3.26454431e-01 -1.96962774e-01 -1.10129881e+00
-5.55647254e-01 8.42460275e-01 6.17160320e-01 3.63642454e-01
1.17364085e+00 -5.04263222e-01 6.70096993e-01 2.82344699e-01
6.03364646e-01 -8.38801861e-01 -3.60562563e-01 5.43161452e-01
7.79610932e-01 -1.27386928e+00 -7.58004263e-02 1.60837948e-01
-3.05545628e-01 8.80300820e-01 2.96109080e-01 -9.40720364e-02
8.03918898e-01 4.63939279e-01 3.67841348e-02 3.58720720e-01
-1.31505132e+00 3.02075505e-01 2.16091409e-01 7.07764447e-01
2.92133301e-01 1.46651059e-01 2.45543107e-01 6.73216701e-01
1.11068405e-01 3.32771212e-01 4.73836362e-01 1.19574273e+00
8.66435617e-02 -1.06406128e+00 -5.31194389e-01 7.67373383e-01
-6.96669400e-01 9.75111499e-02 2.98594177e-01 7.53977478e-01
-3.38726163e-01 1.06091511e+00 3.13665420e-01 -1.33803993e-01
2.71652102e-01 3.30138952e-01 9.85765532e-02 -1.84508145e-01
-4.57741857e-01 3.12963277e-01 -2.56460577e-01 -4.52874422e-01
-5.20712197e-01 -9.33777988e-01 -7.05473959e-01 -2.59618402e-01
8.68797526e-02 -9.72262397e-02 6.22401714e-01 1.08376598e+00
8.07175994e-01 6.21062517e-01 5.38258195e-01 -1.10600078e+00
-1.00727880e+00 -9.44800794e-01 -4.98119175e-01 4.38647479e-01
1.03324199e+00 -7.62023866e-01 -3.71736705e-01 -1.02906168e-01] | [7.029759407043457, 3.142416000366211] |
396c8a7f-ae4c-4109-9837-734e9ce02f61 | abstractive-meeting-summarization | 1609.07035 | null | http://arxiv.org/abs/1609.07035v1 | http://arxiv.org/pdf/1609.07035v1.pdf | Abstractive Meeting Summarization UsingDependency Graph Fusion | Automatic summarization techniques on meeting conversations developed so far
have been primarily extractive, resulting in poor summaries. To improve this,
we propose an approach to generate abstractive summaries by fusing important
content from several utterances. Any meeting is generally comprised of several
discussion topic segments. For each topic segment within a meeting
conversation, we aim to generate a one sentence summary from the most important
utterances using an integer linear programming-based sentence fusion approach.
Experimental results show that our method can generate more informative
summaries than the baselines. | ['Siddhartha Banerjee', 'Kazunari Sugiyama', 'Prasenjit Mitra'] | 2016-09-22 | null | null | null | null | ['meeting-summarization'] | ['natural-language-processing'] | [ 5.51073194e-01 7.35880196e-01 -8.46543089e-02 -5.61401427e-01
-1.72712028e+00 -3.39870483e-01 6.93885863e-01 7.99365163e-01
4.38801870e-02 1.31438017e+00 1.36219978e+00 1.87024727e-01
2.88232118e-01 -3.62616360e-01 -4.20687497e-01 -4.84944671e-01
1.85585663e-01 3.82970333e-01 -1.43454492e-01 -2.32354015e-01
5.83941340e-01 -8.69892687e-02 -1.23196340e+00 7.97236681e-01
1.23601782e+00 3.36241335e-01 3.42504531e-01 1.16261005e+00
-5.35238862e-01 8.58549297e-01 -1.59293854e+00 -1.92212075e-01
-3.33347470e-01 -8.83158088e-01 -9.73845184e-01 4.04215455e-01
3.82833540e-01 -3.17984015e-01 6.40167892e-02 7.66743422e-01
6.48993909e-01 4.19736445e-01 7.86018610e-01 -1.01992524e+00
-1.15668558e-01 1.59464490e+00 -3.97860587e-01 1.37712657e-01
1.01458740e+00 -3.79360378e-01 1.08858728e+00 -6.33782446e-01
5.02867579e-01 1.67242122e+00 2.35538483e-01 5.41275561e-01
-1.19494843e+00 -2.54811883e-01 5.19851983e-01 -3.87032926e-01
-8.70548964e-01 -8.38265121e-01 9.06992793e-01 -8.60464722e-02
1.10690069e+00 7.51573384e-01 5.62281966e-01 9.19802010e-01
6.61847055e-01 1.21491301e+00 5.03442943e-01 -5.40165365e-01
2.39154965e-01 2.72926182e-01 6.50611162e-01 1.85700521e-01
2.32859671e-01 -1.01783371e+00 -6.81816161e-01 -6.76908374e-01
-5.17720804e-02 -2.44664878e-01 -3.45725656e-01 5.89212954e-01
-1.34232986e+00 9.04686332e-01 -1.93421721e-01 2.41809011e-01
-7.70927072e-01 -9.17394087e-02 6.85301065e-01 3.36644858e-01
9.45749462e-01 7.58067966e-01 -2.83741131e-02 -1.41583160e-01
-1.21862245e+00 8.37106466e-01 1.42667580e+00 9.96589243e-01
5.41669130e-01 3.39872390e-02 -8.80696058e-01 8.57801914e-01
1.05720736e-01 2.95015723e-01 4.35546011e-01 -1.07430434e+00
8.68261337e-01 4.74306315e-01 2.96021789e-01 -9.13858950e-01
-2.92020679e-01 6.72619417e-02 -7.17997253e-01 -5.25181532e-01
-2.00823382e-01 -6.65346920e-01 -4.20093715e-01 1.28143227e+00
1.30231157e-02 -3.22554082e-01 5.56901395e-01 3.32754076e-01
1.34031534e+00 1.26189280e+00 -9.28680003e-02 -9.85632122e-01
1.19980514e+00 -1.39326870e+00 -1.25999320e+00 -2.11836457e-01
4.14736509e-01 -8.54232728e-01 6.93699121e-01 3.77267778e-01
-1.76787245e+00 -4.39358950e-01 -1.13856685e+00 2.88946964e-02
1.18409403e-01 5.36690205e-02 1.38385236e-01 3.43483359e-01
-1.17662048e+00 3.19299459e-01 -6.26449704e-01 -3.93227100e-01
-1.14967674e-01 3.06274623e-01 -3.37477028e-02 2.34495759e-01
-8.77098858e-01 6.83785558e-01 5.95535278e-01 -2.98093170e-01
-4.44203705e-01 -5.35543799e-01 -1.08298588e+00 3.02163482e-01
2.67681748e-01 -9.90185857e-01 1.82433724e+00 -5.85963666e-01
-1.62181878e+00 2.30734751e-01 -9.03043330e-01 -5.45750856e-01
1.27541274e-01 -4.19590324e-01 1.02455150e-02 3.89833808e-01
1.58098057e-01 6.99117005e-01 5.95114231e-01 -1.43647635e+00
-7.83369601e-01 6.48398623e-02 -3.76341864e-02 7.28571892e-01
-2.67963022e-01 3.58591497e-01 -1.94718346e-01 -7.59685636e-01
-1.11211069e-01 -6.52422607e-01 -5.64190328e-01 -1.17072594e+00
-9.44160879e-01 -4.97830391e-01 7.92135358e-01 -9.28137064e-01
1.59234309e+00 -1.58674300e+00 3.54661912e-01 -3.87001753e-01
2.34861374e-01 8.19252580e-02 -3.35590839e-02 9.70069945e-01
2.39630088e-01 3.27999949e-01 -1.42303020e-01 -9.70848739e-01
-1.98856201e-02 -1.54245973e-01 -7.19799161e-01 -5.80851771e-02
1.42702788e-01 6.54645503e-01 -9.68739390e-01 -8.43794167e-01
-7.16050490e-05 2.01264694e-01 -3.99627090e-01 5.55706263e-01
-4.87823546e-01 2.95473218e-01 -6.96166277e-01 2.71353424e-01
4.13753569e-01 1.34848222e-01 -6.89760447e-02 1.38039351e-01
-2.84103543e-01 8.22730064e-01 -6.91595912e-01 1.81341732e+00
-2.51448721e-01 9.36225295e-01 4.31370512e-02 -7.18604505e-01
1.03604782e+00 6.44376934e-01 4.19265300e-01 1.39080644e-01
9.17890966e-02 -7.49983862e-02 -3.62484753e-01 -3.63601476e-01
1.40651214e+00 -1.16975240e-01 -6.05472326e-01 8.41984212e-01
-2.50798404e-01 -7.15001583e-01 5.99713385e-01 7.28180826e-01
1.07820547e+00 -3.91199887e-01 4.89843488e-01 -1.12483829e-01
6.19884253e-01 1.96153998e-01 5.65873146e-01 1.01830113e+00
-2.01756507e-01 1.01361442e+00 7.57509053e-01 2.17080414e-01
-8.19799662e-01 -7.69875586e-01 3.84572238e-01 9.48393345e-01
-1.18888080e-01 -8.68790567e-01 -1.00825560e+00 -5.39537728e-01
-4.04708803e-01 1.22294772e+00 -1.49833694e-01 -2.52172556e-02
-9.00689781e-01 -5.35844803e-01 3.33746344e-01 2.28250593e-01
1.92459181e-01 -1.15389419e+00 -4.20559317e-01 5.85683525e-01
-9.45518494e-01 -9.52130318e-01 -9.09719288e-01 -3.03827673e-02
-7.44787574e-01 -3.94583702e-01 -9.02784169e-01 -8.74103487e-01
5.96157968e-01 5.82091033e-01 1.16690409e+00 -2.32793078e-01
3.43092769e-01 3.38721365e-01 -4.48924839e-01 -9.58604455e-01
-1.14867520e+00 5.31911671e-01 -1.85178425e-02 -3.26168716e-01
9.30588692e-02 -2.64462113e-01 -2.36531898e-01 -2.31167167e-01
-8.33185971e-01 2.95195222e-01 3.69365960e-01 6.12187862e-01
6.56818077e-02 -1.16289899e-01 1.19242716e+00 -9.39042151e-01
1.54080486e+00 -3.57109874e-01 1.63741559e-01 3.41241419e-01
-3.12877484e-02 1.17545433e-01 5.25434017e-01 -2.40730867e-01
-1.48107624e+00 -3.13914865e-01 -1.07717086e-02 3.23255926e-01
-1.63485304e-01 7.35187471e-01 -1.72689527e-01 6.29098296e-01
5.57939410e-01 3.39600414e-01 -1.09313875e-02 -1.97622404e-01
1.33796439e-01 1.15304422e+00 3.15679669e-01 -4.77801323e-01
2.44835228e-01 -8.03843290e-02 -8.54589701e-01 -1.37691951e+00
-1.05050468e+00 -7.29061425e-01 -4.07148480e-01 -3.38244051e-01
7.17648983e-01 -9.33358490e-01 -3.26004744e-01 9.64336693e-02
-1.68159103e+00 3.56210135e-02 -3.03241462e-01 3.96065205e-01
-5.63281059e-01 5.79148233e-01 -7.38682568e-01 -1.03108883e+00
-8.99339974e-01 -1.11433756e+00 1.31614995e+00 6.87596202e-01
-1.01737463e+00 -8.49962711e-01 3.27565789e-01 4.44031417e-01
1.72782719e-01 3.74073476e-01 5.56541741e-01 -1.06214964e+00
-1.15450099e-01 -2.86835194e-01 1.16055645e-01 1.97233588e-01
4.60513175e-01 2.16945499e-01 -7.01061308e-01 -2.76245862e-01
2.46909082e-01 -2.73683071e-01 1.11014104e+00 7.47594357e-01
5.49403012e-01 -7.95202613e-01 -5.81797004e-01 -2.67516345e-01
7.32771814e-01 2.25889593e-01 2.87538081e-01 3.22532654e-02
4.87177461e-01 8.93663406e-01 8.99743557e-01 6.11349761e-01
6.31442964e-01 2.96894580e-01 -1.93888977e-01 2.53142029e-01
6.79750144e-02 -1.71748884e-02 7.75020540e-01 1.44703686e+00
3.91675085e-01 -8.32703650e-01 -4.26615983e-01 8.62660646e-01
-2.09268570e+00 -1.08662665e+00 -1.09280765e-01 1.72528601e+00
1.03454435e+00 1.91986427e-01 1.39468595e-01 -2.46113967e-02
8.41869235e-01 7.18027234e-01 -2.42684469e-01 -1.02633178e+00
-1.68479428e-01 -3.83354366e-01 -9.61457640e-02 6.21716440e-01
-9.56327438e-01 7.25043535e-01 7.07647514e+00 4.28854764e-01
-7.08875537e-01 -3.22961569e-01 5.39464474e-01 -3.68948221e-01
-6.49582505e-01 -4.81772237e-02 -1.11604059e+00 3.14456910e-01
1.25149119e+00 -1.00562906e+00 -3.05921614e-01 5.22697389e-01
8.72017443e-01 -4.37868536e-01 -1.06811178e+00 5.89385569e-01
5.39110065e-01 -1.51582909e+00 2.98117459e-01 -9.78353843e-02
1.12804639e+00 -4.43481982e-01 -4.16932940e-01 2.67595410e-01
5.15302658e-01 -6.61731780e-01 4.91684318e-01 4.66522157e-01
1.46441115e-03 -9.59464788e-01 8.07550490e-01 6.83471978e-01
-8.91040504e-01 3.08027834e-01 -4.34392869e-01 -2.16409847e-01
7.42882788e-01 6.45975173e-01 -1.36652124e+00 7.09458292e-01
1.22855611e-01 5.97581685e-01 -1.61188141e-01 9.57951725e-01
-1.85889721e-01 7.94045806e-01 -2.19916608e-02 -3.59071136e-01
2.83765018e-01 -1.67396650e-01 1.11765075e+00 1.67054427e+00
2.17307121e-01 3.05363506e-01 6.46204531e-01 4.14378434e-01
-2.57248819e-01 3.28610152e-01 -7.18942702e-01 -1.98214035e-02
5.26178062e-01 1.13541138e+00 -6.17796838e-01 -8.44456255e-01
-3.68278287e-02 9.32319880e-01 7.81493336e-02 2.43449256e-01
-4.76470679e-01 -5.91721177e-01 2.94632763e-01 -3.76607656e-01
2.02081762e-02 -5.38649783e-02 -4.17803407e-01 -1.07046294e+00
1.14433073e-01 -9.62185860e-01 2.80494958e-01 -5.77692509e-01
-8.68256271e-01 7.17217445e-01 3.91972423e-01 -8.98407519e-01
-7.16303349e-01 4.54108238e-01 -1.32662642e+00 7.48023391e-01
-1.11984861e+00 -8.64897728e-01 -1.00638688e-01 -2.36489773e-01
1.51448166e+00 -6.30431473e-02 9.05494690e-01 -3.22565317e-01
-7.97441840e-01 3.98225844e-01 6.20484501e-02 -4.13509846e-01
9.66431320e-01 -1.51565707e+00 5.74724019e-01 7.23297060e-01
-3.11575502e-01 5.22030652e-01 1.43544888e+00 -8.77413929e-01
-9.73197699e-01 -1.06757998e+00 1.51263607e+00 -2.35871449e-01
2.07440913e-01 -1.46953270e-01 -9.62937117e-01 6.83674812e-01
1.17871666e+00 -1.24210668e+00 8.82046878e-01 -9.73699465e-02
5.45772672e-01 1.65869251e-01 -9.19381380e-01 8.38075638e-01
3.73135805e-01 4.93539646e-02 -1.24734318e+00 5.53110540e-01
1.29331863e+00 -4.47731167e-01 -7.14790881e-01 7.27318451e-02
1.80110022e-01 -5.88321745e-01 5.22047579e-01 -3.21066529e-01
6.75325572e-01 9.73148365e-03 1.19324446e-01 -1.77571654e+00
7.50547051e-02 -1.24426675e+00 -1.44429475e-01 1.80932236e+00
5.84085464e-01 -3.13973427e-01 6.50880575e-01 7.79938757e-01
-5.53809404e-01 -2.97156930e-01 -3.75385404e-01 -2.83720762e-01
-6.67783543e-02 5.05041219e-02 5.35159528e-01 3.14162999e-01
8.04302275e-01 1.04001701e+00 -4.49446857e-01 -1.32087797e-01
4.38849717e-01 1.80422276e-01 1.04354608e+00 -1.03103161e+00
2.10249767e-01 -5.98856866e-01 2.85717964e-01 -1.16456199e+00
4.48365569e-01 -3.15633088e-01 7.52862990e-01 -2.15175176e+00
4.71993595e-01 1.28164619e-01 2.32287601e-01 2.83971317e-02
-7.90716469e-01 -3.14535022e-01 2.36183330e-01 1.45895183e-01
-8.16631258e-01 7.26410449e-01 9.82154310e-01 -5.05893588e-01
-8.51975083e-01 3.71224046e-01 -1.22019827e+00 7.25336552e-01
7.83573270e-01 -4.98485744e-01 -3.45626652e-01 -5.07418513e-02
-1.49821013e-01 6.35779142e-01 -5.05753815e-01 -7.31545448e-01
4.66174066e-01 -2.09084123e-01 -1.60487801e-01 -1.41111279e+00
3.47781658e-01 1.14922295e-04 -1.79829627e-01 2.44968057e-01
-1.02125311e+00 -3.83601524e-02 2.64651746e-01 5.55799544e-01
-6.83360934e-01 -4.07636017e-01 2.64810622e-01 -1.59487873e-01
4.20164578e-02 -3.51192445e-01 -9.88544881e-01 -6.30112514e-02
8.24608266e-01 -7.87217394e-02 -2.11985365e-01 -8.94284904e-01
-3.67860854e-01 6.21690035e-01 2.42160112e-01 3.89447570e-01
6.60806417e-01 -1.04227030e+00 -1.40019381e+00 -5.07663012e-01
-1.55813143e-01 3.20268661e-01 1.31999254e-01 5.36669075e-01
-2.88824052e-01 7.47106910e-01 1.15113929e-01 -3.88247937e-01
-1.70458639e+00 1.41700640e-01 -3.88854980e-01 -7.48719990e-01
-7.09538221e-01 5.78751624e-01 2.94638485e-01 -1.08117402e-01
4.24734801e-01 -5.01496375e-01 -7.85317004e-01 5.24182558e-01
1.07162297e+00 6.17260158e-01 -9.26459301e-03 -5.34449041e-01
-2.43700985e-02 4.29170653e-02 -5.13822019e-01 -4.38423932e-01
1.24818039e+00 -6.38300002e-01 -1.69946805e-01 7.75016248e-01
1.07241809e+00 3.50310773e-01 -9.95061696e-01 -5.64738326e-02
4.20712829e-02 -8.47787876e-03 -1.14414997e-01 -4.14839238e-01
-2.13789403e-01 2.68699050e-01 -6.09119594e-01 7.82333076e-01
9.51475441e-01 -6.03884412e-03 1.13145781e+00 5.71738541e-01
-2.72397131e-01 -9.55713034e-01 1.03991844e-01 5.96711516e-01
1.33656383e+00 -1.27898645e+00 3.35040599e-01 -4.42292780e-01
-9.82830226e-01 1.07468963e+00 3.02574962e-01 2.25674696e-02
1.15029633e-01 2.36730382e-01 -3.18091246e-03 -1.14880845e-01
-1.28001356e+00 2.46647358e-01 4.14001465e-01 1.92774415e-01
7.22663760e-01 7.54868910e-02 -7.42569864e-01 5.97408533e-01
-5.55208862e-01 -3.41643840e-01 1.21683872e+00 1.08599055e+00
-9.00202155e-01 -9.94738162e-01 -6.51051044e-01 5.59202135e-01
-7.19374716e-01 1.73327245e-03 -9.09293354e-01 2.80233532e-01
-9.30697262e-01 1.68965161e+00 5.97732849e-02 -3.16428617e-02
4.79019970e-01 2.00227886e-01 4.17870469e-02 -1.16609383e+00
-1.08860946e+00 4.70424920e-01 7.70511866e-01 5.18479757e-02
-5.33406913e-01 -9.78412867e-01 -1.24352419e+00 -3.79649341e-01
-5.15981436e-01 7.91866124e-01 5.60702443e-01 1.03426278e+00
2.72302300e-01 9.54571366e-01 8.06964755e-01 -1.07909477e+00
-4.98063654e-01 -1.41695881e+00 -2.88279295e-01 -1.13509692e-01
5.50151229e-01 1.89310655e-01 -2.73747206e-01 3.14649820e-01] | [12.61026382446289, 9.381625175476074] |
5fca7d4a-24fd-4593-a694-5441e13a053d | circular-antenna-array-design-for-breast | 1801.05068 | null | http://arxiv.org/abs/1801.05068v1 | http://arxiv.org/pdf/1801.05068v1.pdf | Circular Antenna Array Design for Breast Cancer Detection | Microwave imaging for breast cancer detection is based on the contrast in the
electrical properties of healthy fatty breast tissues. This paper presents an
industrial, scientific and medical (ISM) bands comparative study of five
microstrip patch antennas for microwave imaging at a frequency of 2.45 GHz. The
choice of one antenna is made for an antenna array composed of 8 antennas for a
microwave breast imaging system. Each antenna element is arranged in a circular
configuration so that it can be directly faced to the breast phantom for better
tumor detection. This choice is made by putting each antenna alone on the
Breast skin to study the electric field, magnetic fields and current density in
the healthy tissue of the breast phantom designed and simulated in Ansoft High
Frequency Simulation Software (HFSS). | ['Kalthoum Ouerghi', 'Najib Fadlallah', 'Amor Smida', 'Jaouhar Fattahi', 'Ridha Ghayoula', 'Noureddine Boulejfen'] | 2018-01-15 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 2.02239037e-01 4.37403798e-01 1.00988992e-01 -1.14125282e-01
-1.72195032e-01 -1.12821236e-01 -8.19636509e-02 4.80525708e-03
-2.14764494e-02 3.06932122e-01 -1.41057326e-02 -9.24274802e-01
-2.65555441e-01 -8.57584715e-01 -8.66579264e-02 -1.01750660e+00
-3.88041079e-01 9.51477513e-02 2.39438608e-01 2.50963569e-01
-2.48159349e-01 7.84462214e-01 -4.62319791e-01 4.38823313e-01
4.73955035e-01 7.49211550e-01 3.25719059e-01 5.53187430e-01
3.69930714e-01 5.85405052e-01 -2.49023587e-01 -9.93072707e-03
3.40910405e-02 -6.05887890e-01 -6.22627497e-01 2.97913373e-01
-3.03248703e-01 -3.60368401e-01 -1.87143162e-01 8.73674631e-01
3.54621232e-01 -3.89058053e-01 1.14590526e+00 -7.38690436e-01
1.90186143e-01 4.74623919e-01 -9.48098421e-01 3.13713491e-01
3.95781696e-02 -4.57011372e-01 -1.03370309e-01 -3.39561760e-01
5.19919515e-01 8.27782512e-01 7.43165910e-01 2.51496226e-01
-9.06701386e-01 -5.17180502e-01 -7.96914518e-01 2.99861133e-01
-9.91199374e-01 -2.03654900e-01 7.83739150e-01 -3.47199976e-01
6.15122437e-01 7.41133034e-01 6.46043718e-01 6.70262277e-01
1.17408693e+00 -1.50098518e-01 1.17757344e+00 -9.66611624e-01
3.14964801e-01 6.73292875e-01 2.86187947e-01 6.49096906e-01
7.32095897e-01 8.17516670e-02 4.22535896e-01 -6.52297556e-01
8.01490009e-01 -3.48272383e-01 -3.46167415e-01 -3.44424605e-01
-8.72544646e-01 4.35358077e-01 5.40271759e-01 1.17240775e+00
-7.00474083e-01 3.91998664e-02 3.56905341e-01 -1.02752214e-02
1.43322691e-01 1.26473621e-01 -4.78680804e-02 6.39847219e-01
-3.55449408e-01 -2.44858354e-01 8.60437155e-01 3.26405585e-01
-1.23764254e-01 -1.05745129e-01 1.11270107e-01 6.59778893e-01
7.90851355e-01 7.36885846e-01 2.02436596e-01 -2.91735798e-01
-3.00662875e-01 -1.87083095e-01 6.77500898e-03 -1.08579493e+00
-6.17328703e-01 -9.17375803e-01 -1.06858850e+00 -3.21363583e-02
3.01938444e-01 -3.49651426e-01 -6.08762205e-01 9.47550952e-01
7.57183671e-01 -1.68977417e-02 3.98636520e-01 6.94839418e-01
9.93334472e-01 8.42365623e-01 2.46921718e-01 -4.16350573e-01
1.85563707e+00 -4.29245293e-01 -6.44685268e-01 -1.80991232e-01
7.32025683e-01 -9.23177540e-01 -8.77032727e-02 2.49524876e-01
-6.05472326e-01 -2.70474732e-01 -1.13601708e+00 9.61337209e-01
2.28640616e-01 3.68307054e-01 7.54498661e-01 1.11688411e+00
-4.85021561e-01 1.64375022e-01 -1.12014878e+00 -6.95836365e-01
2.17262298e-01 1.68745637e-01 -4.71135080e-01 -2.10611776e-01
-5.35578072e-01 8.64059031e-01 -3.98631632e-01 1.54432744e-01
-1.39949948e-01 -7.80822098e-01 -6.73789501e-01 -3.73334080e-01
-6.74308240e-01 -8.30456257e-01 1.05107045e+00 -1.00288892e+00
-1.24408066e+00 8.65784526e-01 -2.40186788e-02 -3.04041296e-01
2.09798783e-01 7.32873917e-01 -6.57266855e-01 5.64458489e-01
7.64093846e-02 -5.29181287e-02 3.60753357e-01 -1.37973130e+00
-9.74365920e-02 -3.95517677e-01 -6.69809401e-01 -3.42479438e-01
1.95435300e-01 2.60353833e-01 2.90241152e-01 -4.36724365e-01
8.03488314e-01 -7.58352935e-01 -5.13103366e-01 -1.84054106e-01
-1.58833966e-01 2.89375663e-01 8.50728512e-01 -1.12412846e+00
7.84817159e-01 -2.38343954e+00 -5.79516113e-01 8.11846614e-01
-4.00155842e-01 -1.67649105e-01 3.34032290e-02 5.55139184e-01
-4.02474880e-01 -3.04850161e-01 7.57124051e-02 8.33117902e-01
-6.20685339e-01 -1.81885436e-01 5.47615945e-01 1.27590919e+00
-3.85404438e-01 2.43606701e-01 -4.50719386e-01 -4.11338478e-01
-9.06794965e-02 5.73046684e-01 -4.31026071e-01 -3.12156677e-02
7.28115737e-01 4.65502322e-01 -1.05388284e+00 7.01379240e-01
1.42015970e+00 1.15598105e-01 6.01611912e-01 -6.95476890e-01
-1.10397495e-01 -3.05071265e-01 -1.14312840e+00 8.49202156e-01
-5.06697476e-01 2.15448335e-01 6.10879123e-01 -1.50805962e+00
9.38104928e-01 7.71297634e-01 7.29036808e-01 -9.74098444e-01
3.61803114e-01 2.58733869e-01 5.72419465e-01 -1.00507414e+00
-6.43138170e-01 -6.03485107e-01 4.24334735e-01 2.66875029e-01
-2.63816714e-01 -2.18308368e-03 -3.06442589e-01 -1.42191812e-01
1.11147022e+00 -6.27572715e-01 2.53531337e-01 -1.01500189e+00
4.30809945e-01 1.60770938e-01 2.30236411e-01 3.56250226e-01
-8.47387537e-02 -1.09322570e-01 2.62875825e-01 -3.99290532e-01
-8.54463637e-01 -1.11439240e+00 -9.73672152e-01 4.99097928e-02
2.79480338e-01 2.69479275e-01 -7.39215732e-01 -2.66834795e-01
-1.32918164e-01 5.45496464e-01 -5.26969552e-01 -5.09055741e-02
-3.85647416e-01 -1.17030478e+00 6.62818477e-02 1.12863168e-01
5.41137755e-01 -6.74321711e-01 -1.14780259e+00 2.45242000e-01
-6.87481835e-02 -7.61479855e-01 4.79815274e-01 4.62322146e-01
-1.01146293e+00 -9.34716523e-01 -6.77890062e-01 -1.04726398e+00
9.82500196e-01 1.40865162e-01 9.28352416e-01 -4.09236811e-02
-1.01232159e+00 7.91598201e-01 -6.19296372e-01 -7.47250974e-01
-8.50302160e-01 -5.05444288e-01 -3.76624614e-01 -1.88805193e-01
1.20359898e-01 -5.74676812e-01 -7.77423561e-01 5.07082880e-01
-5.41301847e-01 -1.16343930e-01 8.44921827e-01 7.87779510e-01
7.84708932e-02 4.44002390e-01 6.00394130e-01 -5.21633506e-01
5.98442974e-03 -7.47235656e-01 -2.04616800e-01 -1.36133716e-01
2.94905752e-01 -4.74490374e-01 3.05805681e-03 -2.38820508e-01
-1.17774355e+00 -2.43883431e-01 -3.75663996e-01 5.64784229e-01
-3.04508537e-01 3.43006402e-01 -2.70206720e-01 -5.76660335e-01
8.62538874e-01 -5.45933358e-02 3.18683743e-01 -2.19593018e-01
-6.78249478e-01 8.07751954e-01 3.34821582e-01 -1.73075601e-01
3.48316163e-01 1.96112916e-01 5.05949318e-01 -1.42801523e+00
-6.65147379e-02 -4.43403363e-01 -2.26036012e-02 -4.99108642e-01
7.16634750e-01 -6.37416959e-01 -2.34303981e-01 3.02883536e-01
-8.75302911e-01 -2.43388638e-01 3.96584064e-01 1.12961078e+00
-1.38599873e-01 2.53152847e-01 -6.56030536e-01 -7.76410997e-01
-4.28604782e-01 -7.08495259e-01 2.85289854e-01 2.55535007e-01
-2.72386134e-01 -1.04988897e+00 -1.38864204e-01 1.50335193e-01
5.94013870e-01 7.08761871e-01 1.37520611e+00 3.02899051e-02
2.45260564e-03 -4.45507169e-01 -6.70387149e-02 4.64196615e-02
6.57165721e-02 -1.38840795e-01 -6.52641296e-01 -3.81089538e-01
5.93416154e-01 1.90516740e-01 4.88223404e-01 1.26387262e+00
1.02684665e+00 -9.34962109e-02 -1.19495547e+00 4.50327426e-01
1.87786925e+00 8.13115954e-01 9.74761128e-01 2.81126380e-01
-2.93963403e-01 5.46379447e-01 6.72792912e-01 3.80891025e-01
-4.73069608e-01 9.11739245e-02 5.16989589e-01 -6.48106515e-01
-7.13363364e-02 6.78210855e-01 -2.86089003e-01 2.40176290e-01
1.35368422e-01 -3.29402477e-01 -8.62800777e-01 3.71633798e-01
-8.04075420e-01 -6.93730116e-01 -7.39610851e-01 1.98869789e+00
3.56862098e-01 -4.73742276e-01 -2.47646838e-01 2.51752883e-01
4.17758048e-01 -6.04849994e-01 2.87290663e-01 -6.11570358e-01
5.31674504e-01 4.69427109e-01 6.57740235e-01 3.86494070e-01
-9.41454411e-01 -4.27278757e-01 6.40608931e+00 5.75034499e-01
-1.27315652e+00 2.51136214e-01 9.93740916e-01 6.41463399e-01
-3.46139997e-01 -1.47541910e-01 -2.82557964e-01 1.09773919e-01
6.58842742e-01 3.61028463e-01 -5.09629905e-01 4.08705324e-01
-5.42739108e-02 -1.06010532e+00 -5.61506808e-01 6.49729252e-01
-6.09014630e-01 -8.77049267e-01 -2.42139935e-01 9.95922312e-02
2.37380266e-01 -5.03551424e-01 -2.37905458e-01 -3.81780118e-01
-6.95871949e-01 -9.10124838e-01 3.88718881e-02 3.32032382e-01
5.93738496e-01 -8.79417598e-01 9.89517808e-01 4.24832851e-01
-6.07695282e-01 7.21490532e-02 -5.04179657e-01 2.26768628e-01
-2.07496852e-01 1.07794130e+00 -1.05374491e+00 6.94414258e-01
7.38380015e-01 -1.90648437e-01 -6.48728684e-02 1.30111158e+00
8.71657610e-01 1.01023746e+00 -4.43844289e-01 -1.22175694e-01
2.41205916e-01 -2.23825499e-01 4.20590073e-01 1.44308209e+00
8.65545094e-01 6.23879731e-01 -4.08613861e-01 7.29077399e-01
7.28055000e-01 6.50516093e-01 -5.94987273e-01 4.97835398e-01
2.48077795e-01 1.49597931e+00 -1.15897715e+00 -1.41005516e-02
-5.15129149e-01 5.37355304e-01 -6.96343005e-01 2.39940241e-01
-8.61626327e-01 -4.36121434e-01 1.14495426e-01 6.96021497e-01
1.79613337e-01 4.37999181e-02 -4.20596916e-03 -1.92724288e-01
-1.87355891e-01 -5.86973488e-01 2.42840037e-01 -4.66943324e-01
-1.06362414e+00 4.50487316e-01 2.42733330e-01 -8.43661368e-01
3.59677613e-01 -6.73965633e-01 -8.69554579e-01 1.07843709e+00
-8.71957242e-01 -1.00043714e+00 -4.50430810e-01 9.63603482e-02
-2.69315600e-01 3.36286575e-02 1.08082306e+00 1.77533537e-01
-1.26075491e-01 3.48690480e-01 4.10388678e-01 4.55143768e-03
4.55312848e-01 -6.66149676e-01 -5.33977151e-01 1.42862231e-01
-6.20709181e-01 9.83102769e-02 8.93193901e-01 -5.18970251e-01
-1.57775307e+00 -8.50137055e-01 6.36805415e-01 5.16564310e-01
4.79219928e-02 -2.04100817e-01 -4.54920202e-01 7.52956718e-02
6.36354148e-01 2.37124369e-01 1.20037532e+00 -3.04336786e-01
3.38400155e-01 -3.45846146e-01 -1.84427619e+00 2.52389908e-01
2.60319948e-01 5.85704803e-01 5.38347894e-03 6.05252683e-01
-4.55719978e-01 -2.34672859e-01 -1.14709365e+00 9.17584717e-01
6.50718033e-01 -8.67818356e-01 1.00504589e+00 1.03715924e-03
1.37055412e-01 5.20241261e-02 -1.00844778e-01 -1.13119328e+00
-7.31727242e-01 4.79998477e-02 1.07236946e+00 7.54726768e-01
3.91780466e-01 -8.61669540e-01 7.12854207e-01 -2.02534318e-01
1.67928755e-01 -9.07349885e-01 -9.53132391e-01 -3.46856445e-01
1.29732728e-01 2.45562658e-01 -1.37833223e-01 6.83532298e-01
4.24929976e-01 4.17594798e-02 3.92463446e-01 5.22764146e-01
7.34055936e-01 -2.13348586e-02 1.61757067e-01 -1.06635714e+00
-5.16198575e-01 8.01234990e-02 -6.67863488e-01 -2.23023385e-01
-3.35062832e-01 -6.17912114e-01 -2.08324328e-01 -1.75370550e+00
3.35941702e-01 -8.76898825e-01 1.79491013e-01 -8.30974951e-02
3.78191501e-01 3.33280534e-01 -5.28694630e-01 -5.24524450e-01
6.14209771e-01 -3.01763535e-01 1.41363537e+00 -1.88060403e-01
2.20714539e-01 2.09598467e-01 -3.25550348e-01 5.84165752e-01
8.07563841e-01 -3.76884013e-01 -2.66164035e-01 1.70549899e-01
-2.58259207e-01 7.09771574e-01 4.30052429e-01 -1.41294527e+00
1.43074859e-02 8.80443901e-02 1.10743308e+00 -4.26111728e-01
-1.48813665e-01 -1.23875403e+00 9.12509203e-01 1.36239302e+00
3.94168913e-01 -5.15312493e-01 4.04940128e-01 1.72397792e-01
-4.25186418e-02 -5.90399325e-01 1.40254831e+00 -3.12625229e-01
-3.42376567e-02 -4.35239792e-01 -1.21119666e+00 -8.81481171e-01
1.53414667e+00 -2.33731344e-01 1.84560642e-01 -6.12165667e-02
-8.74395072e-01 -2.90815055e-01 -1.14048950e-01 -5.79026520e-01
2.56688356e-01 -1.19867563e+00 -9.82163608e-01 4.19643134e-01
-9.42369774e-02 -6.39133632e-01 7.32130349e-01 1.28341091e+00
-1.02453244e+00 3.78997713e-01 -3.83637697e-01 -8.34209561e-01
-1.65962780e+00 3.13815564e-01 1.23543119e+00 -2.71935850e-01
-6.16800010e-01 9.29640889e-01 3.18519533e-01 1.66511282e-01
-1.33698493e-01 -2.58655995e-01 -2.61610508e-01 -5.09318709e-01
3.68293256e-01 2.12825298e-01 2.86477566e-01 -1.96304366e-01
-4.43644285e-01 6.35635257e-01 4.77074474e-01 2.71911472e-01
1.17776930e+00 1.98089629e-01 -5.51232994e-01 -1.70283809e-01
1.07594824e+00 1.86206669e-01 -1.99913338e-01 2.19190359e-01
-4.94015634e-01 -2.81380564e-01 3.37040305e-01 -1.03717923e+00
-8.74554634e-01 2.66723335e-01 7.77076781e-01 3.02543223e-01
1.28223312e+00 1.50130734e-01 9.15345699e-02 1.52870134e-01
3.56430054e-01 -5.23081779e-01 -1.55585989e-01 -2.16811314e-01
8.54911983e-01 -6.55047417e-01 1.23447046e-01 -1.23325717e+00
-3.81868370e-02 1.57866657e+00 3.12071949e-01 -5.07806122e-01
1.44733036e+00 9.99791384e-01 1.86514944e-01 -1.94454417e-01
-1.92156434e-01 6.83699548e-01 -4.06523841e-03 7.01754808e-01
9.21076834e-01 1.71412125e-01 -8.68319750e-01 7.62547076e-01
2.35382155e-01 -9.42325741e-02 3.05268377e-01 1.11781693e+00
-6.49329841e-01 -8.13405156e-01 -8.10125113e-01 4.57718879e-01
-9.25384700e-01 4.48833913e-01 1.17590576e-01 7.16940641e-01
2.29541749e-01 8.91078949e-01 2.44233355e-01 4.33609873e-01
2.34826267e-01 -3.12953651e-01 1.21280527e+00 -1.19872443e-01
-4.13063318e-01 4.19753402e-01 2.59494722e-01 -1.23895757e-01
-2.20475703e-01 -5.54515004e-01 -1.12542605e+00 7.86147043e-02
-3.85302424e-01 6.77476645e-01 1.03758490e+00 4.72296804e-01
-2.36363739e-01 8.77244473e-01 7.93911338e-01 -3.04082066e-01
-3.35623741e-01 -1.14462566e+00 -1.32870507e+00 -2.87131131e-01
-2.21596304e-02 -4.05884773e-01 -1.73012286e-01 -1.26266316e-01] | [15.187472343444824, -2.776240348815918] |
76b38ebb-7cbb-4c3d-9c83-ed4c4ca05942 | deep-reflection-prior | 1912.03623 | null | https://arxiv.org/abs/1912.03623v2 | https://arxiv.org/pdf/1912.03623v2.pdf | Deep Reflection Prior | Reflections are very common phenomena in our daily photography, which distract people's attention from the scene behind the glass. The problem of removing reflection artifacts is important but challenging due to its ill-posed nature. Recent learning-based approaches have demonstrated a significant improvement in removing reflections. However, these methods are limited as they require a large number of synthetic reflection/clean image pairs for supervision, at the risk of overfitting in the synthetic image domain. In this paper, we propose a learning-based approach that captures the reflection statistical prior for single image reflection removal. Our algorithm is driven by optimizing the target with joint constraints enhanced between multiple input images during the training stage, but is able to eliminate reflections only from a single input for evaluation. Our framework allows to predict both background and reflection via a one-branch deep neural network, which is implemented by the controllable latent code that indicates either the background or reflection output. We demonstrate superior performance over the state-of-the-art methods on a large range of real-world images. We further provide insightful analysis behind the learned latent code, which may inspire more future work. | ['Dong-Dong Chen', 'Carola-Bibiane Schnlieb', 'Angelica Aviles-Rivero', 'Qingnan Fan', 'Yingda Yin', 'Yujie Wang', 'Ruoteng Li', 'Dani Lischinski', 'Baoquan Chen'] | 2019-12-08 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 7.72563100e-01 -1.68314159e-01 5.36615014e-01 -3.51342231e-01
-1.09066391e+00 -1.59083799e-01 4.45360065e-01 -6.84028387e-01
-1.18806921e-01 4.63690490e-01 2.24223301e-01 -1.05776362e-01
3.51594478e-01 -7.43080020e-01 -8.41575146e-01 -1.05336618e+00
4.45333540e-01 -1.37412727e-01 1.70308128e-01 -2.37074737e-02
1.74230248e-01 2.94829160e-01 -1.74101830e+00 6.45576119e-01
9.67399180e-01 7.50543416e-01 2.14160636e-01 5.55678785e-01
2.42159292e-01 9.43369448e-01 -6.09828651e-01 -2.84182400e-01
6.04892194e-01 -5.07027984e-01 -8.99152979e-02 4.35391307e-01
8.99808347e-01 -5.35611808e-01 -4.10958916e-01 1.09818959e+00
5.53070724e-01 8.52169693e-02 5.48867047e-01 -8.83335292e-01
-6.46250844e-01 -1.82070734e-03 -6.53225958e-01 -2.15038180e-01
2.32674375e-01 4.63559747e-01 7.80962229e-01 -9.62497473e-01
9.82201844e-02 1.08543611e+00 7.07379937e-01 3.70996028e-01
-1.19754148e+00 -6.71981692e-01 2.12550059e-01 4.84023131e-02
-1.37614024e+00 -6.87173367e-01 9.63749766e-01 -4.81465310e-01
7.61356592e-01 4.08977002e-01 6.13839626e-01 1.27150536e+00
-5.09859025e-02 7.64417946e-01 1.39487779e+00 -5.69973767e-01
-7.20398650e-02 3.20334315e-01 -4.16209809e-02 7.24604726e-01
2.65101105e-01 2.25842819e-01 -5.07657707e-01 7.51079842e-02
7.21091807e-01 2.76528031e-01 -6.40787423e-01 -1.40955284e-01
-7.42860436e-01 6.06479287e-01 4.33258653e-01 -1.38127923e-01
-2.41150022e-01 7.47436509e-02 -2.81058460e-01 1.92187548e-01
6.71120167e-01 2.64789701e-01 -1.37998983e-01 3.48411381e-01
-9.85492826e-01 1.22371905e-01 6.40566409e-01 6.81720257e-01
8.18656862e-01 5.67611642e-02 -2.42930666e-01 1.00000310e+00
3.80600631e-01 8.15919042e-01 -1.00933693e-01 -8.85768235e-01
4.62655932e-01 6.20673418e-01 3.95782351e-01 -9.07032132e-01
-1.48090988e-01 -5.07847488e-01 -1.00106120e+00 7.58489430e-01
2.53245831e-01 7.89321400e-03 -1.07374358e+00 1.39309144e+00
1.49784282e-01 3.17941397e-01 -2.23287437e-02 1.12551212e+00
7.28306949e-01 7.36172795e-01 -5.29845059e-01 -1.19246140e-01
1.03878641e+00 -1.29754257e+00 -4.98182386e-01 -4.90299284e-01
8.01972076e-02 -1.04174352e+00 1.25721931e+00 7.65507400e-01
-9.80301440e-01 -5.18186748e-01 -1.02986073e+00 -1.35979295e-01
1.44743532e-01 6.01709902e-01 4.04086769e-01 6.90556288e-01
-7.28744030e-01 2.64153779e-01 -7.18332052e-01 -9.70495790e-02
3.62081319e-01 5.82694896e-02 -1.23499610e-01 -3.95994127e-01
-8.47680509e-01 8.84026289e-01 -5.10764837e-01 5.33759236e-01
-9.16985333e-01 -4.63831902e-01 -8.06962252e-01 4.58259881e-02
5.11763573e-01 -6.78473413e-01 9.74880934e-01 -1.10102248e+00
-1.73713696e+00 9.61846411e-01 -2.16543093e-01 -1.28726929e-01
7.58727670e-01 -6.54803872e-01 -4.10526454e-01 1.57896817e-01
-6.67822585e-02 4.61241268e-02 1.36028862e+00 -1.71790373e+00
-3.62408787e-01 -6.85886294e-02 4.39525470e-02 2.71365792e-01
-1.99188173e-01 -1.31014613e-02 -7.14583039e-01 -5.09270728e-01
3.29148658e-02 -1.01963365e+00 -2.60816336e-01 2.66564459e-01
-4.49307203e-01 4.68030572e-01 6.90081656e-01 -5.66218555e-01
9.04136717e-01 -2.23927736e+00 -2.46597156e-01 -1.44175319e-02
1.61421165e-01 5.05276144e-01 -2.66722798e-01 3.44676882e-01
-6.46519959e-02 -2.65817225e-01 -4.55108047e-01 -6.26779437e-01
-1.52549103e-01 -2.29535639e-01 -6.05570674e-01 7.44473517e-01
2.87999302e-01 5.76339543e-01 -5.98649681e-01 -6.78423420e-02
4.79998171e-01 8.40012074e-01 -4.37582821e-01 5.69240689e-01
-1.18742652e-01 5.47932684e-01 -2.09247261e-01 4.07902449e-01
1.02788723e+00 -1.98920503e-01 -7.45756999e-02 -4.06228513e-01
-2.72415608e-01 2.79205441e-01 -1.19276857e+00 1.33720410e+00
-7.83748865e-01 7.96017945e-01 1.06214352e-01 -6.81271613e-01
1.09246242e+00 -3.15853693e-02 3.51284534e-01 -9.39083040e-01
-9.10481140e-02 8.72689933e-02 -2.47906327e-01 -8.32369447e-01
2.73325384e-01 -2.97023803e-01 1.37094095e-01 6.39291942e-01
-6.07336581e-01 -1.42162114e-01 -2.74387777e-01 -4.37722821e-03
1.05343115e+00 2.78664619e-01 -1.92343816e-02 2.53240466e-01
6.02336764e-01 -3.70805889e-01 5.81966698e-01 8.72335792e-01
2.07442418e-01 1.27107942e+00 3.54109295e-02 -5.16448855e-01
-8.54744971e-01 -1.30248737e+00 1.66400343e-01 6.85363293e-01
3.14641863e-01 -1.12433329e-01 -5.51952004e-01 -3.58165383e-01
-2.16023117e-01 7.14828372e-01 -4.32400584e-01 -2.37003013e-01
-8.45065832e-01 -8.49364042e-01 1.81265652e-01 2.56907165e-01
7.23909557e-01 -9.63432133e-01 -7.25954056e-01 -1.19073406e-01
-5.12123942e-01 -1.20859396e+00 -3.78047556e-01 -1.57018974e-01
-5.76783419e-01 -1.10241151e+00 -7.40718722e-01 -4.29859340e-01
9.38429177e-01 9.04926956e-01 1.10313582e+00 3.78524750e-01
-6.31230712e-01 3.67281705e-01 -1.71601012e-01 -4.03506905e-01
-2.52370596e-01 -3.58892262e-01 -3.48456353e-01 4.77112383e-01
2.85885364e-01 -4.93119657e-01 -9.57566142e-01 5.01614451e-01
-9.12468791e-01 3.88616592e-01 7.80644834e-01 7.97840178e-01
2.30547369e-01 2.05259904e-01 -1.38886198e-01 -8.96226227e-01
3.16402644e-01 9.12331883e-03 -8.01758587e-01 1.79855824e-01
-2.23987862e-01 -5.24904057e-02 5.86373210e-01 -1.72232524e-01
-1.57014310e+00 8.59308913e-02 1.24094546e-01 -4.36618090e-01
-9.13875327e-02 1.16941072e-02 -2.67556578e-01 -5.70051782e-02
6.92117333e-01 2.22451642e-01 -1.36250421e-01 -4.48273003e-01
1.47521064e-01 6.90822959e-01 5.35185933e-01 -2.54616499e-01
1.03319657e+00 9.72653329e-01 -4.89920676e-02 -1.15678477e+00
-1.46946490e+00 -6.31130695e-01 -4.01388288e-01 -3.65470231e-01
6.08698428e-01 -1.17757070e+00 -6.19727671e-01 7.40663171e-01
-1.13916981e+00 -5.86932063e-01 1.98826101e-02 3.83934736e-01
-3.14938784e-01 3.91626447e-01 -4.67441082e-01 -1.16292834e+00
-1.74441501e-01 -1.09125805e+00 1.33705616e+00 1.57610118e-01
1.15842685e-01 -6.41235352e-01 7.68956263e-03 7.03596592e-01
3.99669558e-01 6.11581057e-02 5.10634601e-01 3.23784709e-01
-1.31806612e+00 -1.73481449e-01 -5.01070023e-01 7.64210999e-01
2.36676320e-01 -2.86628045e-02 -1.42627943e+00 -4.17514563e-01
4.65341985e-01 -1.87105358e-01 1.38135517e+00 3.33953828e-01
1.02327538e+00 -1.97554812e-01 -1.58819035e-01 9.43247318e-01
1.56311524e+00 -2.69707769e-01 1.08622420e+00 3.58168215e-01
8.88900340e-01 6.80293679e-01 4.96037096e-01 3.44308883e-01
1.20079502e-01 7.47732401e-01 5.47821999e-01 -7.01072872e-01
-4.53205734e-01 2.00706888e-02 5.97709417e-01 4.84311908e-01
-1.52068391e-01 -4.16594416e-01 -5.80140948e-01 3.11725348e-01
-1.81551325e+00 -1.09580779e+00 -3.67687285e-01 2.48506904e+00
5.98494828e-01 7.59235993e-02 -2.04072163e-01 3.32140215e-02
5.49609363e-01 3.54688287e-01 -3.37375879e-01 6.19007759e-02
-2.13636488e-01 2.05735385e-01 4.65693653e-01 7.91732132e-01
-1.07391238e+00 7.47572243e-01 6.28352070e+00 3.08196485e-01
-1.31827390e+00 -1.28153920e-01 4.13093358e-01 -3.80500585e-01
-3.77063841e-01 -1.99581664e-02 -7.56260872e-01 4.08746064e-01
4.93360639e-01 4.31785464e-01 5.46008527e-01 4.60931957e-01
5.65768301e-01 -3.11901480e-01 -8.80045533e-01 1.19359791e+00
5.52175999e-01 -9.95634437e-01 -1.20577984e-01 -1.52970523e-01
9.23567951e-01 6.94371164e-02 2.57024258e-01 -6.08958118e-02
2.29995623e-01 -1.11244881e+00 4.68920320e-01 9.02146399e-01
7.73920357e-01 -4.04027611e-01 3.94713849e-01 2.80050904e-01
-7.71789789e-01 -3.90399136e-02 -4.17834878e-01 -1.76300287e-01
2.87010521e-01 9.60119724e-01 -5.27251720e-01 3.53785813e-01
7.93021500e-01 8.04189682e-01 -4.20435190e-01 1.05543399e+00
-8.06264699e-01 6.15076423e-01 -3.08283359e-01 4.41888601e-01
-9.62236524e-02 -6.21066153e-01 5.54696023e-01 1.21332610e+00
3.20481688e-01 5.76027222e-02 1.68891594e-01 1.03723466e+00
1.55383795e-01 -3.18444967e-01 -5.70999146e-01 3.88081670e-01
-4.05247025e-02 1.22305870e+00 -4.59327757e-01 -9.46570262e-02
-7.10777104e-01 1.20193517e+00 1.82217062e-01 7.07317889e-01
-8.77339959e-01 -1.62359521e-01 5.93934834e-01 4.23230976e-01
3.71590942e-01 -2.11802974e-01 -3.57441783e-01 -1.43290150e+00
4.13439542e-01 -9.56583321e-01 -1.22332443e-02 -1.02669954e+00
-1.34178233e+00 5.11290550e-01 -4.00294125e-01 -1.52494776e+00
2.90258139e-01 -7.92965174e-01 -7.66717851e-01 1.00752580e+00
-1.94167936e+00 -1.19075930e+00 -8.92018199e-01 6.14336252e-01
4.82354015e-01 -1.87404212e-02 6.80271983e-01 4.93055522e-01
-5.62510610e-01 3.30265820e-01 2.40680829e-01 5.34797162e-02
9.93489921e-01 -9.06906009e-01 3.62461776e-01 1.21103120e+00
1.36242077e-01 5.23305297e-01 8.66269290e-01 -4.02628362e-01
-1.29402065e+00 -1.16035604e+00 5.65219223e-01 -4.34921265e-01
1.60022974e-01 -5.72314858e-01 -8.75101745e-01 6.84715748e-01
6.52660429e-02 -3.59187759e-02 5.52566111e-01 -9.32142884e-03
-4.57903773e-01 -3.27608645e-01 -7.17130423e-01 8.57376873e-01
9.92214680e-01 -4.75810111e-01 -2.92832881e-01 3.10750157e-01
3.51249158e-01 -2.96126246e-01 -7.19366968e-02 2.02862248e-01
5.71074367e-01 -1.51638949e+00 1.21198654e+00 7.77027681e-02
6.27612293e-01 -4.74782139e-01 -1.26786837e-02 -1.20583475e+00
-2.05481470e-01 -7.50947833e-01 8.85229930e-02 9.69918132e-01
2.95176685e-01 -7.31278718e-01 7.79343486e-01 5.48228383e-01
-1.38456747e-01 -3.68090719e-01 -3.85385782e-01 -6.42020345e-01
-3.78215015e-01 -4.57652211e-01 1.98429301e-01 5.59583783e-01
-7.38689363e-01 4.11746681e-01 -9.80127752e-01 4.87683296e-01
1.03247893e+00 5.78296900e-01 1.28201449e+00 -9.78242099e-01
-5.74492455e-01 -9.55354273e-02 -4.12208848e-02 -1.23256814e+00
-3.82180624e-02 -4.68159080e-01 4.88539517e-01 -1.57783318e+00
2.96525925e-01 -2.83910125e-01 -1.27121419e-01 2.82352358e-01
-3.49556953e-01 5.84109724e-01 1.23791859e-01 2.00448409e-01
-5.56719661e-01 7.63803720e-01 1.26467657e+00 -2.14064062e-01
-2.04568446e-01 2.70822436e-01 -7.85000980e-01 1.09254599e+00
6.24061763e-01 -5.78855157e-01 -3.78977597e-01 -7.86193311e-01
2.97731817e-01 -2.95075178e-01 7.42978275e-01 -9.59413886e-01
9.55079868e-02 -1.65970281e-01 4.06916022e-01 -4.87861425e-01
7.64201820e-01 -8.88175786e-01 1.27496898e-01 1.80622652e-01
-1.64984629e-01 -5.77148855e-01 -7.51996487e-02 5.50804138e-01
-1.57746747e-01 -7.30024353e-02 1.02851248e+00 -3.41614366e-01
-2.94693857e-01 1.68073863e-01 -2.20348269e-01 -5.15836477e-02
5.86499214e-01 -2.11000800e-01 -4.44572836e-01 -5.63216209e-01
-4.02034074e-01 -9.53843072e-02 6.80122972e-01 1.52448833e-01
7.35252440e-01 -9.65222120e-01 -9.01255310e-01 4.34279144e-01
9.29281414e-02 1.96599096e-01 3.86471301e-01 7.44386971e-01
-4.91219401e-01 -4.62273508e-02 9.75869782e-03 -6.40345156e-01
-1.50162911e+00 3.75645012e-01 4.40595686e-01 -1.68910295e-01
-1.13002634e+00 7.94361830e-01 7.09487140e-01 -1.56060964e-01
1.52907729e-01 -1.41574159e-01 1.26021445e-01 -3.50026250e-01
7.16087937e-01 3.92479002e-01 -1.59144178e-02 -4.12060350e-01
-4.63095717e-02 8.28306377e-01 -8.27272683e-02 -8.18784386e-02
1.56348479e+00 -2.70494521e-01 -5.90780452e-02 3.39143842e-01
1.20330489e+00 3.80373657e-01 -1.75069177e+00 -4.70121443e-01
-5.70962787e-01 -1.05020368e+00 8.39627981e-02 -6.85277998e-01
-1.20240831e+00 1.01537228e+00 4.43554163e-01 -5.72515763e-02
1.34454298e+00 -3.57559234e-01 7.09182501e-01 3.97711128e-01
1.20636716e-01 -9.62364733e-01 4.57411140e-01 2.84276694e-01
1.09211934e+00 -1.48523402e+00 3.87164742e-01 -7.76748419e-01
-5.78694105e-01 1.10829663e+00 6.05285525e-01 -3.25446904e-01
4.48596299e-01 4.33437914e-01 6.07442498e-01 -2.47670546e-01
-6.46091998e-01 -1.89529538e-01 3.20027351e-01 5.18959105e-01
3.34810615e-01 -2.06596375e-01 2.48588413e-01 1.30365267e-01
-8.85652155e-02 -1.17980048e-01 6.15267158e-01 7.17309415e-01
-4.56637979e-01 -9.54524875e-01 -6.60575092e-01 2.61815757e-01
-2.70739108e-01 -2.31303379e-01 -3.23561043e-01 4.16124493e-01
4.77949419e-04 9.74738061e-01 -1.80156574e-01 8.58877748e-02
3.02954793e-01 -3.50551903e-01 6.73791528e-01 -7.16647148e-01
-3.10179591e-01 3.26690406e-01 -6.95205182e-02 -6.51140511e-01
-6.33770108e-01 -5.90478241e-01 -7.01395392e-01 3.06743309e-02
-4.43098128e-01 -2.75675625e-01 4.29445416e-01 7.70405531e-01
1.03116080e-01 5.64769864e-01 8.17769170e-01 -9.62785959e-01
-4.20334727e-01 -8.16131651e-01 -5.62988102e-01 7.02734649e-01
6.12092316e-01 -6.26672208e-01 -7.02816844e-01 2.83464581e-01] | [10.544488906860352, -2.77262020111084] |
9d401349-9d65-440d-a6a9-da01284708a9 | multi-layer-personalized-federated-learning | 2212.02985 | null | https://arxiv.org/abs/2212.02985v1 | https://arxiv.org/pdf/2212.02985v1.pdf | Multi-Layer Personalized Federated Learning for Mitigating Biases in Student Predictive Analytics | Traditional learning-based approaches to student modeling (e.g., predicting grades based on measured activities) generalize poorly to underrepresented/minority student groups due to biases in data availability. In this paper, we propose a Multi-Layer Personalized Federated Learning (MLPFL) methodology which optimizes inference accuracy over different layers of student grouping criteria, such as by course and by demographic subgroups within each course. In our approach, personalized models for individual student subgroups are derived from a global model, which is trained in a distributed fashion via meta-gradient updates that account for subgroup heterogeneity while preserving modeling commonalities that exist across the full dataset. To evaluate our methodology, we consider case studies of two popular downstream student modeling tasks, knowledge tracing and outcome prediction, which leverage multiple modalities of student behavior (e.g., visits to lecture videos and participation on forums) in model training. Experiments on three real-world datasets from online courses demonstrate that our approach obtains substantial improvements over existing student modeling baselines in terms of increasing the average and decreasing the variance of prediction quality across different student subgroups. Visual analysis of the resulting students' knowledge state embeddings confirm that our personalization methodology extracts activity patterns which cluster into different student subgroups, consistent with the performance enhancements we obtain over the baselines. | ['Christopher Brinton', 'Andrew Lan', 'Kerrie Douglas', 'Laura Cruz', 'Elizabeth Tenorio', 'Seyyedali Hosseinalipour', 'Yun-Wei Chu'] | 2022-12-05 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-4.60925838e-03 1.48855999e-01 -7.79597342e-01 -7.52219737e-01
-8.39796782e-01 -8.36030483e-01 4.73300785e-01 7.46608078e-01
-2.82301068e-01 4.88005072e-01 6.00246549e-01 -6.09579027e-01
-5.52914381e-01 -1.03687882e+00 -8.02284062e-01 -2.92689115e-01
1.82858407e-01 2.80380189e-01 8.94173980e-02 1.25918275e-04
2.86905199e-01 6.42118812e-01 -1.81801951e+00 5.00750184e-01
9.81879711e-01 6.12475991e-01 -3.92376870e-01 7.50950038e-01
-2.87576497e-01 9.88729000e-01 -3.88108969e-01 -4.77942377e-01
-7.20735341e-02 -7.20663816e-02 -9.39399660e-01 -1.69701025e-01
1.21258354e+00 -2.45110616e-01 -3.03753644e-01 7.08800673e-01
3.96475703e-01 7.79036999e-01 5.80602527e-01 -1.21710014e+00
-8.59105110e-01 8.92273903e-01 -1.41244665e-01 1.21325199e-02
1.67980716e-01 -3.22316401e-02 1.15743947e+00 -6.27479613e-01
4.62679923e-01 1.04537702e+00 9.66289282e-01 6.16884112e-01
-1.78417683e+00 -5.55947065e-01 6.25343263e-01 1.70282066e-01
-1.17668104e+00 -2.26117224e-01 6.04571342e-01 -7.64103591e-01
5.28753698e-01 4.76629913e-01 4.71641034e-01 1.04136109e+00
5.75301610e-02 1.06431401e+00 9.60319340e-01 -2.67286956e-01
1.25600204e-01 6.90224111e-01 8.56939971e-01 8.49659622e-01
3.00241083e-01 -2.94791937e-01 -9.67945278e-01 -3.86786222e-01
9.55761746e-02 3.03999007e-01 -4.70957197e-02 -6.34237587e-01
-7.85772204e-01 7.09266067e-01 7.89930224e-02 -8.77364576e-02
-1.72781110e-01 -1.68703005e-01 1.06475383e-01 2.84370363e-01
6.06905103e-01 4.29728597e-01 -8.36441219e-01 -2.59283036e-01
-1.05118513e+00 5.53911924e-01 9.84759927e-01 8.07783604e-01
1.07743740e+00 -1.40239149e-02 -5.96581876e-01 7.95159638e-01
3.49627882e-01 -1.22024961e-01 8.64759445e-01 -9.15844262e-01
2.80374557e-01 1.18776894e+00 -4.99585234e-02 -6.54913127e-01
-2.58226991e-01 -6.24637902e-01 6.93612769e-02 3.42757022e-03
5.24809837e-01 -3.90008569e-01 -8.24632525e-01 2.02276659e+00
5.21373689e-01 6.67060435e-01 -3.86774540e-02 1.18708782e-01
1.00894022e+00 3.08044255e-01 4.98985589e-01 4.53929096e-01
1.27393472e+00 -8.61754775e-01 -3.25346768e-01 -6.95330426e-02
1.05447686e+00 -3.60809863e-01 1.12983549e+00 4.80399668e-01
-1.08978450e+00 -6.36814296e-01 -5.94012380e-01 -8.26683044e-02
-7.42031336e-01 -1.93772197e-01 4.86500412e-01 1.15108609e+00
-9.69491899e-01 9.54138935e-01 -8.56686771e-01 -9.57246497e-02
5.37395418e-01 4.68768537e-01 -7.73678571e-02 5.09804189e-02
-6.38548732e-01 5.47160447e-01 -5.45729604e-03 -6.75662637e-01
-8.56384158e-01 -1.82926214e+00 -7.20218301e-01 7.73793876e-01
-1.37843207e-01 -4.96613324e-01 1.41119969e+00 -6.67510390e-01
-1.56185985e+00 6.18766844e-01 -1.00690641e-01 -2.18512490e-01
3.43923122e-01 -2.60435879e-01 -1.40843511e-01 -5.17520189e-01
-2.32796475e-01 3.47377270e-01 1.17463857e-01 -1.12411594e+00
-8.74064386e-01 -5.36568224e-01 -9.08906758e-02 1.53677076e-01
-1.12975526e+00 -3.19251835e-01 -2.10117519e-01 -1.42740235e-01
-1.27295613e-01 -7.68111587e-01 -2.55894929e-01 -4.94152874e-01
-8.42534304e-02 -6.48472548e-01 8.38849247e-01 -6.48714662e-01
1.81816959e+00 -1.70951712e+00 -1.04608715e-01 5.56220770e-01
2.02816606e-01 6.69508800e-02 -1.76338822e-01 2.63857484e-01
-1.71601936e-01 1.66504949e-01 3.52054089e-01 -6.39301717e-01
3.74629229e-01 2.39284872e-03 -2.13386506e-01 3.24257672e-01
-2.90858865e-01 7.04547286e-01 -8.49488378e-01 -5.05615473e-01
6.79757223e-02 3.93442541e-01 -9.80673373e-01 3.47004682e-01
-5.23282364e-02 1.61458030e-01 -4.18804407e-01 6.48010373e-01
2.59856910e-01 -1.74294934e-01 2.95289099e-01 2.12440923e-01
-2.62962222e-01 6.57495677e-01 -1.17774451e+00 1.59402585e+00
-7.12033510e-01 4.91699457e-01 -5.50526828e-02 -9.67479408e-01
8.17472339e-01 2.54838437e-01 6.75481260e-01 -1.84729844e-01
-2.43931860e-01 -1.47526771e-01 -1.89715773e-01 -3.45516920e-01
7.61572182e-01 2.57288367e-01 -5.83168417e-02 7.56332695e-01
4.19557810e-01 3.78487140e-01 -3.00592575e-02 2.63335496e-01
1.15889001e+00 3.11221302e-01 -2.03842580e-01 -4.28607196e-01
3.43926162e-01 -8.99505541e-02 6.29460871e-01 9.73153889e-01
-3.48071307e-01 2.83276588e-01 3.46046150e-01 -4.27730352e-01
-5.27146280e-01 -1.16414571e+00 -1.88732713e-01 2.04262376e+00
-3.76413941e-01 -5.37585139e-01 -5.73744297e-01 -1.04727411e+00
5.61648786e-01 8.98465455e-01 -6.79751575e-01 -6.75230861e-01
-4.76498276e-01 -7.74706304e-01 4.45117086e-01 6.83028877e-01
-2.20040187e-01 -6.14800394e-01 -2.55098611e-01 2.68707365e-01
1.51150748e-01 -6.78534985e-01 -3.57064635e-01 2.00110838e-01
-9.65757012e-01 -9.31629658e-01 -1.42643929e-01 -7.79133677e-01
6.64164782e-01 6.42968342e-02 1.47756696e+00 -2.91963350e-02
-1.43069267e-01 1.16890013e+00 1.77809447e-01 -6.60641670e-01
-1.93097487e-01 4.57345366e-01 1.24515086e-01 1.72740802e-01
9.25726354e-01 -6.49034083e-01 -5.67531586e-01 -1.20479025e-01
-7.06231654e-01 -2.73536026e-01 1.32094622e-01 5.46130121e-01
3.32422882e-01 -2.81278521e-01 7.57676065e-01 -1.43559968e+00
7.41151810e-01 -8.77993464e-01 -4.77683246e-01 4.90150273e-01
-1.20731378e+00 2.67267674e-02 5.51386893e-01 -6.40316367e-01
-1.31115150e+00 -4.06974293e-02 2.63215214e-01 -2.77768344e-01
-4.29687709e-01 5.02140939e-01 -7.83040002e-02 -1.43707097e-01
7.48248577e-01 8.26792195e-02 -3.11263084e-01 -8.13527167e-01
1.89814016e-01 7.20868349e-01 3.65552336e-01 -1.04668558e+00
5.58462322e-01 2.31755803e-05 -1.73417225e-01 -5.03428757e-01
-1.10159802e+00 -5.17978430e-01 -5.17563105e-01 -2.39056632e-01
4.42304969e-01 -1.08007693e+00 -1.23059714e+00 1.10070869e-01
-5.24716973e-01 -5.22481799e-01 -5.85276544e-01 6.92642808e-01
-2.39074171e-01 -8.06306824e-02 -6.47648513e-01 -7.79264688e-01
1.80640884e-04 -1.06256509e+00 5.25155127e-01 6.63368642e-01
-3.83420080e-01 -1.60826969e+00 3.81270021e-01 7.39352286e-01
5.02739370e-01 4.74191271e-03 1.21853817e+00 -1.31537747e+00
-5.15850306e-01 -1.81370497e-01 4.42403257e-01 2.53255904e-01
1.58728302e-01 8.48544687e-02 -1.27442968e+00 -2.79276937e-01
-7.52329767e-01 -3.07724237e-01 7.98856854e-01 2.44036838e-01
1.63714933e+00 -3.35877329e-01 -3.81493211e-01 5.66142917e-01
1.28251016e+00 -1.38854057e-01 -8.11562240e-02 4.26672876e-01
1.05844307e+00 8.07007730e-01 3.73470001e-02 4.84007418e-01
8.06631804e-01 4.09450501e-01 2.19751224e-01 3.72526348e-01
-5.53721972e-02 -6.16270304e-01 5.88351250e-01 6.81597531e-01
1.67465448e-01 -6.10496104e-02 -1.01740217e+00 1.01361370e+00
-1.76100886e+00 -9.13488686e-01 -6.57436112e-03 2.47569060e+00
1.22447765e+00 -2.37466052e-01 1.91070348e-01 -4.75190997e-01
3.57036620e-01 1.17912740e-01 -6.00048780e-01 -8.48364174e-01
3.66482407e-01 4.97003257e-01 7.00957417e-01 7.70588219e-01
-8.13563883e-01 7.28710771e-01 6.29290247e+00 5.34639359e-01
-8.32513690e-01 4.55361009e-02 6.96465492e-01 -4.98109818e-01
-6.99843645e-01 -1.00116864e-01 -1.44884491e+00 3.57570887e-01
1.64278531e+00 -4.69179720e-01 1.94855243e-01 9.35234368e-01
9.93472189e-02 4.37903106e-01 -1.45483065e+00 4.03694957e-01
-1.57153923e-02 -1.48476326e+00 2.01938391e-01 1.68607071e-01
1.33336699e+00 -1.06994629e-01 3.77299756e-01 9.44497526e-01
1.24018133e+00 -1.02844632e+00 3.23302865e-01 7.51767993e-01
4.93279956e-02 -1.00251496e+00 1.75128743e-01 4.19966906e-01
-8.05921674e-01 -4.76150185e-01 3.25476192e-02 3.56480069e-02
-6.49688601e-01 2.79406399e-01 -1.01883280e+00 2.00556457e-01
8.39096904e-01 5.40420711e-01 -8.72436166e-01 8.39086533e-01
1.38298914e-01 1.27324688e+00 -2.18108565e-01 2.07843438e-01
8.49113837e-02 -1.52407452e-01 1.04642123e-01 1.26790369e+00
3.49374592e-01 -1.62194565e-01 3.37837458e-01 7.72535980e-01
-5.58549702e-01 2.11115375e-01 -5.01949787e-01 1.51396260e-01
6.83670759e-01 1.33680892e+00 -1.27418842e-02 -4.30696756e-01
-7.40394413e-01 2.49893665e-01 5.86941004e-01 4.27494258e-01
-4.05644149e-01 -3.37907553e-01 1.29255486e+00 3.13194066e-01
2.53271043e-01 1.21412009e-01 -3.85605544e-01 -1.16672862e+00
-4.06287074e-01 -8.89876962e-01 7.19796002e-01 -6.95551857e-02
-1.25880849e+00 -1.37120038e-01 -7.28406534e-02 -8.25463951e-01
-2.98049867e-01 -5.47524154e-01 -9.50262487e-01 1.07120848e+00
-1.50791132e+00 -1.02290678e+00 -3.03790361e-01 5.71418345e-01
2.11821437e-01 -2.13487089e-01 8.36643934e-01 3.48910362e-01
-1.00776529e+00 1.20947409e+00 4.99073178e-01 5.69923185e-02
1.01818705e+00 -1.50703657e+00 -6.60054013e-02 6.66842997e-01
1.67247355e-01 8.18386316e-01 5.14699280e-01 -4.70890582e-01
-1.27652204e+00 -1.22733378e+00 1.13884354e+00 -1.02730823e+00
5.95694363e-01 -1.77394226e-01 -1.17621589e+00 1.33737707e+00
1.58067301e-01 -7.11747557e-02 1.48982775e+00 8.26870680e-01
-4.06304955e-01 -2.09837988e-01 -1.09279776e+00 6.35742307e-01
6.42771304e-01 -5.09229779e-01 -3.89476508e-01 3.02501380e-01
4.88759726e-01 -3.90070707e-01 -1.53837359e+00 9.45581868e-02
5.89711130e-01 -7.83793330e-01 8.98690343e-01 -1.64088392e+00
4.57956642e-01 2.85307586e-01 -5.67921204e-03 -1.45294511e+00
-5.09113967e-01 -4.86145198e-01 -5.08511245e-01 1.62734509e+00
3.99349719e-01 -3.70953888e-01 1.53211582e+00 1.44883764e+00
-2.93476313e-01 -9.72437561e-01 -4.76207107e-01 -3.58526617e-01
4.83514428e-01 -1.72772780e-01 1.00580263e+00 1.34762216e+00
6.05376959e-02 6.53523952e-02 1.43903732e-01 3.33995879e-01
7.95022488e-01 4.52802628e-01 8.52454007e-01 -1.74425590e+00
-2.23567456e-01 -6.71749353e-01 -1.95334464e-01 -8.85529876e-01
5.87624252e-01 -1.01343274e+00 -4.95439112e-01 -1.25858617e+00
2.59460449e-01 -5.47672927e-01 -8.85776758e-01 6.95154727e-01
-4.47435588e-01 -2.37696573e-01 -1.91162497e-01 -3.10192674e-01
-7.13419199e-01 3.06955338e-01 8.61906648e-01 7.97578245e-02
-5.28844059e-01 1.11338466e-01 -1.10497880e+00 8.40002060e-01
7.37305284e-01 -4.65344816e-01 -3.99171501e-01 -4.07341242e-01
6.11048192e-02 -3.07887703e-01 1.36922643e-01 -9.34664369e-01
4.54051375e-01 -4.80577230e-01 6.65682197e-01 -2.51129627e-01
1.63298696e-01 -7.43574321e-01 -3.77015889e-01 1.20752424e-01
-9.24065888e-01 -1.71100479e-02 3.54531854e-01 4.71996874e-01
-7.46679530e-02 -3.69843483e-01 6.56240702e-01 7.00841397e-02
-5.00353575e-01 3.57830733e-01 -3.67201209e-01 8.30711275e-02
8.83655429e-01 -1.27403796e-01 -4.69894439e-01 -2.46410549e-01
-8.69975269e-01 5.31043828e-01 2.39099488e-01 5.42854369e-01
2.68264830e-01 -1.33408189e+00 -6.44367218e-01 3.02498817e-01
9.80371460e-02 -7.68239498e-02 4.96499896e-01 6.28714383e-01
6.98030069e-02 4.08550382e-01 -2.08330043e-02 -3.96329075e-01
-1.63786447e+00 6.07034303e-02 3.68769437e-01 -5.83511293e-01
-8.05732086e-02 9.35637414e-01 -7.98216835e-02 -1.25027263e+00
4.97296423e-01 -3.99779499e-01 -3.95507336e-01 3.54528785e-01
3.85934979e-01 7.82460272e-01 2.53989428e-01 -2.11339876e-01
-8.13979954e-02 5.91965988e-02 -3.74351770e-01 2.59063721e-01
1.64426553e+00 4.44940776e-02 5.15816569e-01 4.83496606e-01
1.25954914e+00 4.46798444e-01 -1.32606757e+00 -5.36302388e-01
1.78888649e-01 -2.92096734e-01 2.31616408e-01 -8.80137384e-01
-1.03538430e+00 6.61367297e-01 6.52778506e-01 -4.11971100e-02
6.44400001e-01 -2.58673519e-01 5.17108202e-01 3.96611750e-01
-1.91316098e-01 -1.09507453e+00 -2.45441452e-01 4.22957987e-01
-2.79381067e-01 -1.21539831e+00 4.28105108e-02 2.16741294e-01
-2.51590341e-01 8.96367252e-01 1.11668384e+00 1.79063305e-01
7.94559836e-01 -1.39486626e-01 1.16355143e-01 9.44340900e-02
-1.20255983e+00 2.83612996e-01 5.06053627e-01 1.95671618e-01
9.05305266e-01 2.46064320e-01 1.64968953e-01 1.02168489e+00
-2.92503804e-01 -1.91625774e-01 6.52909875e-01 1.15968823e+00
-5.28748393e-01 -1.25189340e+00 -3.93694341e-01 7.65677869e-01
-6.38929963e-01 4.00002068e-03 -1.29139334e-01 5.50536990e-01
1.53692469e-01 7.20262110e-01 2.18718529e-01 -4.57258016e-01
6.08277857e-01 9.05981064e-01 4.06528026e-01 -9.88138497e-01
-1.38562405e+00 -7.35436201e-01 -1.56421989e-01 -5.97888589e-01
-1.16798179e-02 -8.70442748e-01 -9.62929428e-01 -5.02612948e-01
2.65728123e-03 4.19706881e-01 5.20170629e-01 6.41814709e-01
7.07644403e-01 7.93351889e-01 4.21040028e-01 -4.54626799e-01
-9.76264536e-01 -7.09957957e-01 -5.70733786e-01 6.27673209e-01
3.38840872e-01 -4.14677173e-01 -5.14102697e-01 1.47937983e-01] | [10.127290725708008, 7.177610874176025] |
92c27560-8367-4601-9bb3-12a38720a218 | masked-conditional-neural-networks-for-1 | 1805.10004 | null | http://arxiv.org/abs/1805.10004v2 | http://arxiv.org/pdf/1805.10004v2.pdf | Masked Conditional Neural Networks for Environmental Sound Classification | The ConditionaL Neural Network (CLNN) exploits the nature of the temporal
sequencing of the sound signal represented in a spectrogram, and its variant
the Masked ConditionaL Neural Network (MCLNN) induces the network to learn in
frequency bands by embedding a filterbank-like sparseness over the network's
links using a binary mask. Additionally, the masking automates the exploration
of different feature combinations concurrently analogous to handcrafting the
optimum combination of features for a recognition task. We have evaluated the
MCLNN performance using the Urbansound8k dataset of environmental sounds.
Additionally, we present a collection of manually recorded sounds for rail and
road traffic, YorNoise, to investigate the confusion rates among machine
generated sounds possessing low-frequency components. MCLNN has achieved
competitive results without augmentation and using 12% of the trainable
parameters utilized by an equivalent model based on state-of-the-art
Convolutional Neural Networks on the Urbansound8k. We extended the Urbansound8k
dataset with YorNoise, where experiments have shown that common tonal
properties affect the classification performance. | ['John Robinson', 'Fady Medhat', 'David Chesmore'] | 2018-05-25 | null | null | null | null | ['environmental-sound-classification', 'sound-classification'] | ['audio', 'audio'] | [ 3.64064157e-01 6.69070520e-03 3.27436954e-01 -2.37908170e-01
-6.06946409e-01 -2.99656391e-01 3.14349204e-01 -4.54625815e-01
-5.61737120e-01 5.36195755e-01 1.61601797e-01 -2.29945630e-01
-2.75310367e-01 -4.80347931e-01 -6.20108366e-01 -7.46277869e-01
-7.71884084e-01 -2.21211433e-01 3.07933748e-01 9.67659131e-02
-7.36941025e-02 2.73250252e-01 -1.93121755e+00 6.87163949e-01
1.69183806e-01 1.11779499e+00 4.66751039e-01 1.17026770e+00
2.12274715e-01 7.84674227e-01 -7.94479728e-01 1.12610590e-02
3.52524906e-01 -4.90404457e-01 -7.17583179e-01 -2.01082498e-01
5.20007789e-01 2.10773900e-01 -3.82374346e-01 7.93965876e-01
4.59757745e-01 3.84155005e-01 5.05365610e-01 -9.26854312e-01
-3.73195171e-01 1.07487011e+00 2.42858782e-01 6.21785581e-01
6.26394451e-02 3.73048395e-01 1.27353752e+00 -1.02113724e+00
7.86546543e-02 1.04541016e+00 1.11515057e+00 2.41129726e-01
-1.43234789e+00 -8.79898250e-01 -1.88770413e-01 3.33598971e-01
-1.47658956e+00 -7.25237548e-01 7.76001513e-01 -5.46105802e-01
1.16852140e+00 3.05494130e-01 6.86654508e-01 1.01953435e+00
-4.17678095e-02 4.13856894e-01 9.55827653e-01 -7.01892972e-01
1.85911030e-01 -5.64923249e-02 -1.82213902e-01 5.69340467e-01
-2.62161851e-01 5.42024612e-01 -9.56095338e-01 -1.48738116e-01
6.93216741e-01 -4.24053073e-01 -5.39985359e-01 2.24171802e-01
-1.19746470e+00 7.70008087e-01 2.62796164e-01 4.51278716e-01
-3.10453951e-01 5.84617138e-01 2.92022496e-01 4.68299836e-01
3.03252310e-01 7.02072024e-01 -7.34392703e-01 -2.90922612e-01
-1.18360341e+00 -5.52930199e-02 1.02168787e+00 4.74313736e-01
7.13304043e-01 7.64704347e-01 2.42904902e-01 1.02770364e+00
1.86137795e-01 2.55833179e-01 6.49126709e-01 -1.03522277e+00
3.49510431e-01 -1.36626348e-01 -1.57688290e-01 -1.00110531e+00
-4.22331661e-01 -5.70725620e-01 -6.02050245e-01 1.09730870e-01
4.21222657e-01 -3.76776725e-01 -8.43670726e-01 1.73409259e+00
-3.37471694e-01 4.83148932e-01 7.30715096e-02 8.01526368e-01
6.37639880e-01 6.89466178e-01 -1.00213803e-01 8.55877697e-02
1.14595723e+00 -7.77761877e-01 -6.58399463e-01 -3.07948917e-01
3.14526767e-01 -8.59295189e-01 1.12894225e+00 6.38697088e-01
-7.89781690e-01 -1.05109322e+00 -1.06941998e+00 5.03838062e-01
-5.02469897e-01 4.01772410e-01 7.66120195e-01 8.82169306e-01
-9.23967540e-01 8.56929243e-01 -7.06016839e-01 7.01119229e-02
1.75341517e-01 4.01916742e-01 -3.91049713e-01 3.28783244e-01
-1.66123152e+00 7.56363511e-01 5.55877268e-01 4.78726238e-01
-1.07734716e+00 -8.95526588e-01 -9.16693032e-01 2.74783283e-01
7.23435953e-02 1.25289559e-01 1.36966610e+00 -9.05813992e-01
-1.70470214e+00 3.36466461e-01 3.79551083e-01 -7.85612345e-01
2.46483907e-02 -3.43764603e-01 -9.91284728e-01 3.43517125e-01
-1.37628973e-01 6.63745522e-01 9.74881411e-01 -8.30582559e-01
-4.02171850e-01 5.50932705e-01 -2.22772747e-01 -4.02213126e-01
-2.83942789e-01 -1.18073404e-01 2.83229500e-01 -8.71286750e-01
7.43485913e-02 -9.08712208e-01 -2.11300910e-01 -3.73082340e-01
-4.13458854e-01 -7.15889633e-02 7.27619588e-01 -4.41472054e-01
1.25819707e+00 -2.61057901e+00 -3.14432830e-01 2.05456287e-01
-2.02280864e-01 1.63429126e-01 -3.28355402e-01 5.13281345e-01
-5.38329363e-01 1.64399091e-02 -5.22586524e-01 -2.81772256e-01
8.82080942e-02 2.79812813e-01 -5.13090670e-01 4.69501883e-01
4.81519401e-01 3.73963445e-01 -7.27433264e-01 -4.07478921e-02
7.66902864e-02 4.92912889e-01 -7.54861593e-01 2.58468419e-01
-1.05892420e-01 1.57352567e-01 2.80398130e-01 2.93635309e-01
3.11369240e-01 1.48022577e-01 3.92108560e-02 -2.46602640e-01
-3.20074230e-01 7.27781117e-01 -1.30193281e+00 1.77037692e+00
-7.90357232e-01 1.08462667e+00 2.09281296e-02 -6.92221999e-01
9.36116695e-01 6.65426791e-01 1.78611830e-01 -1.57176986e-01
-4.44681868e-02 2.72339433e-01 4.28155959e-01 -5.24488211e-01
3.88290256e-01 -3.64452720e-01 -3.61833498e-02 3.39119673e-01
4.78883803e-01 -3.00084859e-01 2.00115573e-02 -1.62976846e-01
1.10748255e+00 -3.12473029e-02 -2.70793382e-02 -4.09233361e-01
1.90799385e-01 -3.32737088e-01 3.66221011e-01 8.02225530e-01
-7.31469095e-02 6.77202582e-01 2.22501591e-01 -3.90214354e-01
-6.12509072e-01 -1.12319827e+00 -4.47312236e-01 1.14761341e+00
-6.97528541e-01 -3.06410164e-01 -5.20118415e-01 -3.50602776e-01
-8.30870941e-02 7.22616196e-01 -7.30501771e-01 -2.57050693e-01
-6.09014153e-01 -6.72040582e-01 1.16610825e+00 4.34715092e-01
3.33640963e-01 -1.53883839e+00 -8.90027642e-01 4.43373948e-01
-5.37673198e-02 -1.22904027e+00 -2.89116412e-01 1.20475543e+00
-4.56336826e-01 -9.14555311e-01 -1.95358485e-01 -6.68139160e-01
6.33648261e-02 -1.96593568e-01 1.17427385e+00 -3.27316850e-01
-4.42832321e-01 2.52321213e-01 -4.16493863e-01 -4.82959181e-01
-4.16596562e-01 7.88708329e-02 2.24378183e-01 1.92298457e-01
2.98668921e-01 -1.11527252e+00 -4.03509289e-01 1.05258144e-01
-8.88895929e-01 -3.04688841e-01 5.41597784e-01 8.48988593e-01
1.66049778e-01 4.16453421e-01 7.00792909e-01 -5.92809260e-01
3.62673938e-01 -4.31335449e-01 -2.52904624e-01 -4.94178772e-01
5.96431382e-02 1.94261726e-02 7.41706967e-01 -7.09759593e-01
-7.55563319e-01 2.80357629e-01 -3.47156435e-01 -4.68939394e-01
-3.86059225e-01 4.30018306e-01 2.30529100e-01 1.73122603e-02
8.84886444e-01 1.09583355e-01 -3.29791307e-01 -8.69484663e-01
4.07458246e-01 6.84204519e-01 8.69956434e-01 -3.39732230e-01
6.47913516e-01 2.63350576e-01 -1.62825122e-01 -1.18361866e+00
-8.61976922e-01 -3.60341012e-01 -7.15533078e-01 -1.73112735e-01
9.27244425e-01 -8.45527470e-01 -5.57964444e-01 4.39593434e-01
-1.09115517e+00 -4.05494988e-01 -7.14159250e-01 9.06790316e-01
-5.13661444e-01 -9.71489549e-02 -5.41342080e-01 -1.06701696e+00
1.29746690e-01 -8.31896901e-01 4.49922681e-01 -2.14913502e-01
-4.67970967e-01 -7.28141308e-01 1.41878337e-01 -2.31911182e-01
5.22429109e-01 1.95244730e-01 9.74791586e-01 -6.82913244e-01
-2.11927235e-01 -1.21279798e-01 1.64659292e-01 9.06051815e-01
2.02081114e-01 -1.40914947e-01 -1.62293947e+00 -4.96524312e-02
8.34665149e-02 -4.05196041e-01 1.13924909e+00 2.35749915e-01
1.17928147e+00 -3.15395236e-01 2.93036133e-01 6.51238501e-01
1.13290894e+00 2.89312422e-01 4.74236697e-01 9.67768356e-02
4.56315547e-01 5.36356866e-01 -9.65103060e-02 3.75487387e-01
-7.35702962e-02 2.84980267e-01 5.66838145e-01 -1.04072936e-01
-4.20648903e-01 -1.86965421e-01 5.02379596e-01 1.42023087e+00
8.14642683e-02 7.96407834e-02 -7.08087623e-01 8.10767829e-01
-1.18011177e+00 -9.14259374e-01 -4.44536768e-02 1.77535784e+00
1.02186048e+00 3.43814611e-01 -1.58496708e-01 8.56924891e-01
6.05877340e-01 3.44233185e-01 -1.16274536e-01 -5.58807731e-01
-3.13458219e-02 7.56683648e-01 3.30851883e-01 4.41865146e-01
-1.24577081e+00 6.88348830e-01 7.40488338e+00 8.36656332e-01
-1.23784494e+00 1.25490859e-01 2.06075683e-01 -1.18286356e-01
7.85349160e-02 -2.68461108e-01 -5.17099380e-01 4.65475380e-01
1.63860750e+00 6.29179001e-01 6.23305261e-01 6.39134049e-01
1.01381280e-01 1.77639700e-03 -1.10378003e+00 7.65391886e-01
-2.14205712e-01 -1.32366168e+00 -2.99537241e-01 -3.18618149e-01
5.39259791e-01 2.69824415e-01 2.50996917e-01 4.33619440e-01
3.98586005e-01 -1.19862473e+00 9.76887703e-01 6.28270745e-01
9.94909525e-01 -8.70449245e-01 5.50528228e-01 1.25399485e-01
-1.31076384e+00 -2.56759256e-01 -2.42187709e-01 -4.11611348e-01
-2.46019319e-01 7.01861978e-01 -1.14978266e+00 3.25044632e-01
8.95941019e-01 4.10490453e-01 -3.88048530e-01 8.63774896e-01
-2.48512611e-01 1.30595016e+00 -4.74335164e-01 1.17499039e-01
4.68899608e-01 3.61171156e-01 5.98004878e-01 1.84605420e+00
7.92734101e-02 -3.59759718e-01 5.08594625e-02 8.22485924e-01
1.25075027e-01 -1.75399780e-01 -3.62097412e-01 -2.15330705e-01
4.36920196e-01 9.16724980e-01 -5.62463939e-01 6.99084029e-02
-3.26547295e-01 5.65622747e-01 8.04212615e-02 4.54101950e-01
-5.60264409e-01 -7.99246669e-01 7.30234265e-01 -1.37205333e-01
8.34861994e-01 -3.36219728e-01 -1.53969511e-01 -7.10948944e-01
-2.37304211e-01 -7.27041602e-01 1.63270935e-01 -7.00354457e-01
-1.22232962e+00 1.06421757e+00 -1.06415242e-01 -1.41096520e+00
-3.50879103e-01 -6.11465871e-01 -5.97896218e-01 1.05855525e+00
-1.55886662e+00 -6.21099889e-01 1.17583022e-01 4.54022229e-01
4.82423604e-01 -2.92203605e-01 1.25187254e+00 3.68795961e-01
-3.20220649e-01 3.17776889e-01 -1.90251619e-01 3.38432878e-01
4.79000658e-01 -1.28635943e+00 5.41095495e-01 8.04944754e-01
6.11632466e-01 5.23928702e-01 7.09542513e-01 -2.24057324e-02
-8.34909856e-01 -1.22467947e+00 9.57456231e-01 -8.88862684e-02
1.01970375e+00 -5.21774292e-01 -7.96520889e-01 4.22535866e-01
1.84558868e-01 3.06298524e-01 9.93387401e-01 -2.50788331e-02
-6.27880991e-01 -1.26523495e-01 -7.18780875e-01 3.10509354e-01
7.87405372e-01 -1.04404593e+00 -7.60027707e-01 -2.92835990e-03
1.00802207e+00 -1.04984455e-01 -5.61273515e-01 3.31435770e-01
5.76541960e-01 -1.12053847e+00 8.30438972e-01 -5.10325849e-01
4.19986695e-01 -4.55042511e-01 -5.14307201e-01 -1.58850014e+00
-4.24394161e-01 -7.41321385e-01 -1.20302208e-01 9.74969864e-01
5.30193627e-01 -4.62136716e-01 5.18075705e-01 -1.92458272e-01
-5.40925264e-01 -6.04867280e-01 -1.24573553e+00 -9.04354453e-01
-1.83949128e-01 -1.20725501e+00 3.60313803e-01 8.42266023e-01
-1.89823300e-01 -6.40309462e-03 -4.78743047e-01 2.90213853e-01
2.02541262e-01 -2.81043611e-02 3.47977132e-01 -1.20737720e+00
-6.93923354e-01 -2.99254566e-01 -5.54506600e-01 -9.94149745e-01
1.79882377e-01 -1.02543533e+00 5.97144365e-01 -7.23679304e-01
-6.05154693e-01 -5.07757962e-01 -6.92607760e-01 6.62721157e-01
1.76354453e-01 6.55850351e-01 2.73630321e-01 -8.77173766e-02
-1.04715399e-01 3.17602992e-01 9.26890492e-01 6.74509928e-02
-1.26316696e-01 8.57800944e-04 -3.31429183e-01 9.01091814e-01
8.03345442e-01 -8.02528322e-01 -3.15713376e-01 -3.29884529e-01
5.55427819e-02 8.82631466e-02 5.46269238e-01 -1.47515082e+00
2.04589233e-01 1.54028952e-01 3.42294246e-01 -4.69187945e-01
5.92069447e-01 -8.67984533e-01 1.89081073e-01 5.44024408e-01
-6.13753855e-01 -1.37833595e-01 4.03454334e-01 5.67998767e-01
-4.34841663e-01 -2.89268464e-01 9.09429014e-01 -2.06917018e-01
-6.31753623e-01 -1.34334147e-01 -9.84335244e-01 6.02410175e-02
2.91987360e-01 -6.29220307e-02 1.83431134e-01 -2.80089289e-01
-9.60586727e-01 -5.00779331e-01 -5.61310649e-01 2.78055966e-01
5.35456061e-01 -1.27262509e+00 -5.74765086e-01 7.02812731e-01
-2.84631401e-01 -3.31926644e-01 1.66991755e-01 5.27164817e-01
-3.47217649e-01 4.45831299e-01 -2.06170410e-01 -5.37211716e-01
-8.09521735e-01 1.56179786e-01 4.91953224e-01 1.66168343e-02
-5.30324042e-01 1.37610638e+00 -4.72273715e-02 -4.32848960e-01
3.88796628e-01 -7.83911586e-01 -2.40297273e-01 1.36251763e-01
3.70151997e-01 2.76180923e-01 2.98163444e-01 -4.84466553e-01
-4.20358300e-01 6.49232194e-02 4.05991554e-01 -3.04288149e-01
1.41848350e+00 2.08134338e-01 1.73865557e-02 8.88563037e-01
1.42043138e+00 1.78668469e-01 -1.27642906e+00 -2.01960087e-01
1.10467456e-01 -1.88320726e-01 2.86462396e-01 -8.26804578e-01
-1.03946173e+00 8.48825037e-01 5.29430211e-01 6.80582047e-01
1.15488052e+00 -3.10402215e-01 7.00159907e-01 6.31626785e-01
1.14387736e-01 -8.41777205e-01 -4.21421975e-03 8.64020109e-01
7.93848038e-01 -7.22779095e-01 -3.14560890e-01 -2.54073232e-01
-4.32406873e-01 1.46904683e+00 1.75440073e-01 -3.93982142e-01
1.15060723e+00 6.45210385e-01 3.58907700e-01 -7.02088922e-02
-9.03783143e-01 -3.24764758e-01 3.56940001e-01 5.63411832e-01
3.93199325e-01 1.59471601e-01 2.21945375e-01 7.96171188e-01
-9.07599151e-01 -4.09181923e-01 5.19038260e-01 7.03349888e-01
-4.40829605e-01 -6.40762389e-01 -4.19228107e-01 3.49436074e-01
-7.87258089e-01 -5.63119948e-01 8.49009678e-03 6.81401193e-01
5.74638009e-01 1.20757043e+00 2.27813065e-01 -6.72469318e-01
3.61545652e-01 5.61203897e-01 8.92226845e-02 -7.38701999e-01
-9.34136450e-01 1.60070434e-01 3.71893317e-01 -4.18192983e-01
-4.97977972e-01 -6.42244041e-01 -1.08508587e+00 3.88107657e-01
-4.31572407e-01 5.58569729e-01 6.19947076e-01 8.31051528e-01
-1.18521936e-01 1.11827028e+00 7.78051317e-01 -1.10803390e+00
-3.36954385e-01 -1.23600304e+00 -7.79063106e-01 7.22262114e-02
8.44320297e-01 -4.87257183e-01 -7.85713315e-01 4.68474448e-01] | [15.3378324508667, 5.326955318450928] |
be32bf41-366c-44ec-bbc4-e5c7ad75ca58 | a-slot-is-not-built-in-one-utterance-spoken-1 | 2203.10759 | null | https://arxiv.org/abs/2203.10759v1 | https://arxiv.org/pdf/2203.10759v1.pdf | A Slot Is Not Built in One Utterance: Spoken Language Dialogs with Sub-Slots | A slot value might be provided segment by segment over multiple-turn interactions in a dialog, especially for some important information such as phone numbers and names. It is a common phenomenon in daily life, but little attention has been paid to it in previous work. To fill the gap, this paper defines a new task named Sub-Slot based Task-Oriented Dialog (SSTOD) and builds a Chinese dialog dataset SSD for boosting research on SSTOD. The dataset includes a total of 40K dialogs and 500K utterances from four different domains: Chinese names, phone numbers, ID numbers and license plate numbers. The data is well annotated with sub-slot values, slot values, dialog states and actions. We find some new linguistic phenomena and interactive manners in SSTOD which raise critical challenges of building dialog agents for the task. We test three state-of-the-art dialog models on SSTOD and find they cannot handle the task well on any of the four domains. We also investigate an improved model by involving slot knowledge in a plug-in manner. More work should be done to meet the new challenges raised from SSTOD which widely exists in real-life applications. The dataset and code are publicly available via https://github.com/shunjiu/SSTOD. | ['Xiaojie Wang', 'Caixia Yuan', 'Jian Sun', 'Yongbin Li', 'Jiaman Wu', 'Yuchuan Wu', 'Yuwei Hu', 'Sai Zhang'] | 2022-03-21 | null | https://aclanthology.org/2022.findings-acl.27 | https://aclanthology.org/2022.findings-acl.27.pdf | findings-acl-2022-5 | ['sstod'] | ['natural-language-processing'] | [-5.05787849e-01 2.29647115e-01 -2.18035206e-01 -8.33180010e-01
-4.04462725e-01 -9.52910662e-01 8.20905983e-01 -3.71692181e-01
-3.03744644e-01 1.24936354e+00 6.81230724e-01 -5.39948225e-01
2.33657137e-01 -4.23984796e-01 1.72696307e-01 -2.77012527e-01
3.34329009e-01 1.12947798e+00 6.72251821e-01 -9.90665138e-01
1.65236607e-01 1.55499205e-01 -8.93371880e-01 4.21619773e-01
7.38291800e-01 6.97449088e-01 3.20434272e-01 4.09792840e-01
-1.03616643e+00 9.74289536e-01 -9.68493938e-01 -6.13959491e-01
2.67490763e-02 -6.14164352e-01 -1.45431292e+00 1.40992895e-01
-1.75901800e-01 -4.97164428e-01 -2.18101382e-01 7.76475132e-01
5.27120709e-01 3.38179499e-01 4.95534778e-01 -1.60815644e+00
-4.96349186e-01 1.10537207e+00 4.58628349e-02 6.43328577e-02
4.23197746e-01 1.76707543e-02 9.75532711e-01 -6.16225719e-01
6.85941041e-01 1.60693550e+00 3.99188101e-01 9.79139149e-01
-7.43696213e-01 -5.56001008e-01 1.65121406e-01 -6.57719001e-02
-8.48752677e-01 -4.76019502e-01 5.88287592e-01 -3.24163198e-01
1.03626585e+00 3.82730633e-01 9.41170603e-02 1.33566952e+00
-2.12976411e-01 8.05414617e-01 1.17055130e+00 -3.22846144e-01
2.30150089e-01 7.87370384e-01 5.68326294e-01 2.05696389e-01
-5.15798509e-01 -2.64020503e-01 -5.11849701e-01 -2.81066298e-01
7.45554149e-01 -1.67050719e-01 1.27879560e-01 3.70426103e-02
-1.38522017e+00 1.07870626e+00 1.14737123e-01 5.83995342e-01
2.77130753e-01 -5.45516312e-01 6.77954912e-01 5.58898091e-01
4.61265028e-01 5.27098119e-01 -1.11569023e+00 -6.94716811e-01
-1.85606666e-02 5.33768713e-01 1.47225702e+00 1.48929322e+00
6.68047547e-01 -3.92620564e-01 -1.18424587e-01 1.37270403e+00
1.83375821e-01 3.14197987e-01 6.00025177e-01 -9.56587553e-01
7.07803488e-01 8.35206449e-01 4.18606997e-01 -3.63461107e-01
-6.55307829e-01 5.27741492e-01 -5.14086306e-01 -2.89523453e-01
8.50005865e-01 -7.05148160e-01 -4.70173061e-01 1.70307541e+00
5.76098382e-01 -2.73263127e-01 3.74048293e-01 8.75614822e-01
1.47910643e+00 6.84128404e-01 2.30025351e-01 -8.92848670e-02
2.00076199e+00 -1.21114457e+00 -1.18028700e+00 -3.15219849e-01
7.52330661e-01 -9.64110136e-01 1.34954226e+00 -2.78315917e-02
-5.40244877e-01 -4.78090227e-01 -4.68815923e-01 -2.92142630e-01
-8.44922662e-01 2.81775147e-02 7.80455112e-01 6.76611662e-01
-9.05684531e-01 -5.60870767e-02 -2.83126384e-01 -7.29627907e-01
-3.92341971e-01 4.10734653e-01 -1.89620927e-01 4.68166977e-01
-1.71516836e+00 1.07027757e+00 3.85492325e-01 -1.59194857e-01
-4.28044438e-01 -2.24816561e-01 -8.49936128e-01 -1.90765887e-01
8.55500996e-01 -1.40455201e-01 1.97723925e+00 -4.83765066e-01
-1.99462140e+00 6.95427895e-01 -2.12259889e-01 -1.68384790e-01
3.37274224e-01 -1.82397246e-01 -5.65601349e-01 -4.00777549e-01
1.59526080e-01 7.91859150e-01 2.05291733e-01 -9.66446877e-01
-7.33532250e-01 -2.23159209e-01 5.64527810e-01 4.62115228e-01
-2.37661213e-01 4.35291439e-01 -3.27802300e-01 -4.21927124e-01
-2.43517086e-01 -1.17808306e+00 -2.59755015e-01 -4.78232265e-01
-6.76235497e-01 -8.30253780e-01 9.37191248e-01 -5.27366698e-01
1.20873725e+00 -2.05519772e+00 -9.40245241e-02 -4.85794663e-01
4.32251580e-02 3.08397263e-01 2.55903184e-01 1.00297678e+00
4.88438755e-01 -7.29944929e-02 -1.50113463e-01 -4.34645146e-01
3.82092834e-01 4.24718618e-01 -4.52824056e-01 -2.87919194e-01
-7.38959983e-02 8.97615433e-01 -5.61171770e-01 -5.65094829e-01
3.18660587e-01 -1.11952201e-01 -1.97956219e-01 5.94402850e-01
-8.08565855e-01 7.42119968e-01 -6.94788337e-01 6.10609889e-01
7.29183376e-01 -1.59676790e-01 3.23682368e-01 1.83295816e-01
-2.90064752e-01 5.70874631e-01 -9.73417401e-01 1.70199883e+00
-3.05401921e-01 4.18639898e-01 1.97396800e-01 -6.95718110e-01
1.21664548e+00 4.71858829e-01 8.23863149e-02 -4.31551546e-01
3.38576078e-01 2.81755865e-01 5.15273288e-02 -5.47833562e-01
9.47805882e-01 -1.50803581e-01 -7.63580561e-01 3.62756461e-01
1.13041736e-01 -3.28154624e-01 7.63351396e-02 3.41953844e-01
7.85828531e-01 -3.18581969e-01 4.48211551e-01 -3.50188106e-01
8.34400177e-01 3.77263308e-01 3.67333740e-01 5.81632793e-01
-5.28011262e-01 1.68359131e-01 8.60836744e-01 -2.71770805e-01
-7.68126726e-01 -7.29856908e-01 -2.98271537e-01 1.74361312e+00
3.76113594e-01 -3.33500713e-01 -7.21387804e-01 -1.03385782e+00
-9.15402547e-02 6.11376643e-01 -3.68888915e-01 4.45688367e-01
-4.89936769e-01 -3.95540297e-01 6.71946645e-01 3.18593204e-01
1.17245340e+00 -1.52375245e+00 1.59107596e-02 3.54511619e-01
-3.94921362e-01 -1.56008601e+00 -6.73066735e-01 2.79064804e-01
-4.66786712e-01 -7.96294689e-01 -7.96700716e-01 -1.20020366e+00
1.48612604e-01 1.67528987e-01 1.11485231e+00 -2.76280075e-01
3.43000859e-01 1.90623894e-01 -8.14029515e-01 -4.56491143e-01
-6.11676395e-01 3.70915025e-01 -3.91391814e-02 -4.09812242e-01
9.80087876e-01 -7.60930628e-02 -1.30023345e-01 1.09766996e+00
-4.44860578e-01 4.46934775e-02 8.51289555e-02 8.94079268e-01
-3.42450619e-01 -2.19155610e-01 9.12395775e-01 -1.20256293e+00
1.11367917e+00 -6.36594415e-01 -5.59617996e-01 8.51057321e-02
7.68436417e-02 -3.51562425e-02 5.42472184e-01 -4.77338225e-01
-1.55849874e+00 -3.64113785e-02 -4.51912135e-01 4.33816254e-01
-6.68568015e-01 3.23913157e-01 -4.03112561e-01 3.76748651e-01
4.34867620e-01 1.46181673e-01 1.76767305e-01 -8.77768636e-01
3.78630251e-01 1.47537017e+00 2.50739515e-01 -7.68560171e-01
3.68350595e-01 -2.16481974e-03 -8.13117266e-01 -1.00912225e+00
-7.00684547e-01 -9.11411703e-01 -6.27906203e-01 -2.22583190e-02
9.67254460e-01 -8.33844483e-01 -9.99146461e-01 7.28564858e-01
-1.25135815e+00 -8.00214231e-01 1.34729624e-01 2.72489816e-01
-3.04644078e-01 2.74475843e-01 -9.66494024e-01 -8.48842680e-01
-2.04668432e-01 -1.25481510e+00 9.99412119e-01 4.54954922e-01
-3.54118168e-01 -1.19402802e+00 9.11295190e-02 5.95882893e-01
6.01982892e-01 -4.26636070e-01 8.58122826e-01 -1.42027831e+00
-2.51382858e-01 3.68049592e-01 -1.23655990e-01 1.79691225e-01
4.57700551e-01 -5.76014757e-01 -8.92925262e-01 1.42363161e-01
6.51605129e-02 -7.42312074e-01 2.14059234e-01 3.55402380e-03
5.89409888e-01 -3.05950910e-01 -2.61636168e-01 -3.38510722e-01
7.41193354e-01 8.29530239e-01 3.69144946e-01 3.26717138e-01
3.76004964e-01 1.02156830e+00 1.11313033e+00 5.56653082e-01
7.38385439e-01 1.16467929e+00 -2.34758575e-02 1.80530529e-02
1.22925870e-01 -7.75723457e-02 3.48699540e-01 8.26051414e-01
5.04544616e-01 -3.07772905e-01 -9.41904545e-01 5.72932601e-01
-2.05533099e+00 -6.32266581e-01 -1.91392303e-01 1.55430877e+00
1.36279047e+00 3.49841453e-02 5.74224293e-01 -3.41876835e-01
7.86552727e-01 2.52703816e-01 -5.51743567e-01 -5.54928839e-01
-1.31352460e-02 -2.88326412e-01 -8.20403844e-02 7.49033451e-01
-1.21445310e+00 1.64276755e+00 5.56423950e+00 8.59299123e-01
-8.27288210e-01 5.92965856e-02 7.32984424e-01 7.29694068e-01
6.64076582e-02 1.61764637e-01 -1.32987010e+00 5.54815769e-01
8.84039283e-01 -1.13959126e-01 3.28905135e-01 1.15884805e+00
3.35545018e-02 -3.18214297e-01 -9.55954373e-01 7.36854851e-01
-2.53529608e-01 -1.08973575e+00 -3.63626480e-01 4.69048098e-02
2.83473849e-01 -2.05845773e-01 -1.79357916e-01 8.03904891e-01
8.08879316e-01 -8.53401721e-01 1.97997287e-01 -1.52135551e-01
5.51200688e-01 -2.40700290e-01 1.01552546e+00 5.46468854e-01
-1.12843382e+00 8.84448737e-02 -5.28907180e-01 -2.36301020e-01
3.06667179e-01 5.47630005e-02 -1.40305126e+00 2.78698474e-01
5.63348114e-01 5.15297711e-01 -2.52353340e-01 4.57749814e-01
-1.82316359e-02 3.84520084e-01 -2.35911161e-01 -5.87159574e-01
4.41899210e-01 -2.81024128e-01 2.70486385e-01 1.21353221e+00
-3.29880673e-03 3.03277284e-01 3.47333640e-01 5.77417493e-01
1.09624133e-01 2.15863615e-01 -6.36455357e-01 9.65984985e-02
9.24326479e-01 1.27795136e+00 -7.76440799e-01 -4.15484488e-01
-5.96547008e-01 8.64380121e-01 4.18943614e-02 2.04290375e-01
-7.94320166e-01 -3.68644834e-01 8.08144867e-01 -2.69546747e-01
9.26251262e-02 -3.52320999e-01 -2.31332570e-01 -1.08161569e+00
-1.77859768e-01 -1.12552881e+00 4.60380495e-01 -5.13973892e-01
-1.65469241e+00 1.02961135e+00 3.74551982e-01 -1.23148477e+00
-5.02792537e-01 -7.51113534e-01 -4.65019584e-01 7.73496926e-01
-1.11105216e+00 -1.03725958e+00 -3.24487001e-01 8.71497750e-01
1.27375412e+00 -6.37020171e-01 1.00490034e+00 2.03707948e-01
-3.61154824e-01 4.06491935e-01 3.01930923e-02 5.17549157e-01
1.23143828e+00 -1.42579377e+00 6.95211649e-01 9.36741158e-02
-1.94917813e-01 7.60488629e-01 7.54136860e-01 -5.92514038e-01
-9.18128788e-01 -6.35830879e-01 1.17539310e+00 -8.22179914e-01
7.60594308e-01 -8.72095466e-01 -8.01152170e-01 8.72650027e-01
7.37842143e-01 -5.78058362e-01 7.29614139e-01 3.70263696e-01
3.49290040e-03 2.24075347e-01 -1.27725577e+00 6.34436131e-01
9.59619761e-01 -3.30058157e-01 -9.87334967e-01 5.68898797e-01
1.15758979e+00 -7.63901114e-01 -6.73367321e-01 1.50783863e-02
1.04032405e-01 -9.47676122e-01 7.11471438e-01 -5.90267837e-01
1.40664279e-01 -1.77331362e-02 -1.21518612e-01 -1.21689188e+00
3.19136769e-01 -7.35205531e-01 3.14837664e-01 1.88140333e+00
6.14008904e-01 -8.64620268e-01 8.76833737e-01 1.07947195e+00
-2.47156218e-01 -2.94603884e-01 -1.02703953e+00 -7.31055379e-01
2.53751576e-01 -1.28996849e-01 9.70479071e-01 1.02101493e+00
7.12839842e-01 8.41557384e-01 -6.96728587e-01 -3.23050559e-01
-1.40846103e-01 3.63768488e-02 1.13151419e+00 -1.19154906e+00
-1.19801397e-02 9.86343436e-03 1.06526859e-01 -1.70337689e+00
2.91432470e-01 -3.63800973e-01 8.99596512e-02 -1.27936542e+00
-8.51850361e-02 -8.90812993e-01 4.93457228e-01 7.15380728e-01
-5.34213334e-02 -3.16492021e-01 3.18821192e-01 2.95420945e-01
-7.28275537e-01 7.35073566e-01 1.19099391e+00 -7.91255478e-03
-6.38457775e-01 3.64923686e-01 -7.72659898e-01 4.67592537e-01
1.09706783e+00 -2.29671225e-01 -6.41540587e-01 -1.25896960e-01
-3.95738602e-01 6.04954481e-01 -1.01275593e-01 -5.76398075e-01
3.45706284e-01 -4.17995542e-01 -3.39433402e-01 -5.41689336e-01
7.35443532e-01 -9.03941512e-01 -2.34801039e-01 6.69818595e-02
-6.18206441e-01 -4.58241887e-02 2.61820167e-01 2.00654775e-01
-3.95148605e-01 -3.73078078e-01 4.60753798e-01 -4.63205010e-01
-1.11996508e+00 -1.42405510e-01 -6.34366691e-01 2.54078925e-01
1.02570724e+00 9.88546461e-02 -8.52006972e-01 -7.72909164e-01
-7.91515231e-01 5.73871195e-01 1.18506953e-01 8.35437059e-01
1.41324282e-01 -1.16146517e+00 -3.84383589e-01 -3.54474522e-02
1.72542542e-01 -1.78889200e-01 3.55628937e-01 3.70763868e-01
-3.98211598e-01 8.62846553e-01 -3.71945918e-01 -3.78006309e-01
-1.42372561e+00 2.30718330e-01 2.24344164e-01 -4.14600611e-01
-3.37505907e-01 8.26586843e-01 1.83932960e-01 -1.26774347e+00
4.91344124e-01 -2.73109287e-01 -6.09224439e-01 2.76122898e-01
4.59535211e-01 -8.71899910e-03 -2.47176692e-01 -6.51162446e-01
-4.28756714e-01 1.58228666e-01 -3.36682111e-01 -4.61816639e-01
9.35711205e-01 -5.14824390e-01 -8.35577920e-02 6.59427285e-01
8.98810029e-01 -5.10240346e-03 -1.01630342e+00 -5.60192883e-01
3.24355185e-01 -3.11855793e-01 -8.28048944e-01 -1.14703560e+00
-5.48205614e-01 6.11440659e-01 3.15257013e-01 7.14522481e-01
6.07856333e-01 4.57854539e-01 9.95012701e-01 6.87037170e-01
5.87730885e-01 -1.23207116e+00 2.79428661e-01 1.29576671e+00
9.31754768e-01 -1.62420166e+00 -5.75641453e-01 -6.96375787e-01
-1.56586337e+00 8.99520636e-01 1.16258776e+00 4.64047134e-01
6.98283434e-01 3.59539032e-01 8.38253617e-01 -1.86068207e-01
-7.08329678e-01 -2.68286675e-01 -2.72204310e-01 7.38538444e-01
6.52168691e-01 8.16666558e-02 -5.28811157e-01 8.87496233e-01
-5.66815853e-01 -3.14895123e-01 5.87987959e-01 1.04602718e+00
-5.15277863e-01 -1.64524174e+00 -3.15276891e-01 2.41747305e-01
-1.82431430e-01 -2.50809431e-01 -9.45357382e-01 9.35303271e-01
-2.90108442e-01 1.49087012e+00 -1.64426208e-01 -4.95840162e-01
3.12020183e-01 4.25679415e-01 -4.10693556e-01 -7.88154423e-01
-9.12907302e-01 -5.28300032e-02 7.26668954e-01 -1.58454433e-01
-4.30466264e-01 -4.14867252e-01 -1.27329195e+00 -5.28148949e-01
-3.18901122e-01 6.84072852e-01 3.52512926e-01 8.26754391e-01
2.30827302e-01 9.88023132e-02 5.10995746e-01 -5.80144405e-01
-6.26897573e-01 -1.42831421e+00 -7.78193235e-01 4.08435881e-01
8.36352855e-02 -8.14247787e-01 -2.97835678e-01 -2.24900618e-01] | [12.780228614807129, 7.928560256958008] |
ae9f64fa-e62a-494e-b9fa-11cea7c06f17 | learnable-gated-temporal-shift-module-for | 1907.01131 | null | https://arxiv.org/abs/1907.01131v2 | https://arxiv.org/pdf/1907.01131v2.pdf | Learnable Gated Temporal Shift Module for Deep Video Inpainting | How to efficiently utilize temporal information to recover videos in a consistent way is the main issue for video inpainting problems. Conventional 2D CNNs have achieved good performance on image inpainting but often lead to temporally inconsistent results where frames will flicker when applied to videos (see https://www.youtube.com/watch?v=87Vh1HDBjD0&list=PLPoVtv-xp_dL5uckIzz1PKwNjg1yI0I94&index=1); 3D CNNs can capture temporal information but are computationally intensive and hard to train. In this paper, we present a novel component termed Learnable Gated Temporal Shift Module (LGTSM) for video inpainting models that could effectively tackle arbitrary video masks without additional parameters from 3D convolutions. LGTSM is designed to let 2D convolutions make use of neighboring frames more efficiently, which is crucial for video inpainting. Specifically, in each layer, LGTSM learns to shift some channels to its temporal neighbors so that 2D convolutions could be enhanced to handle temporal information. Meanwhile, a gated convolution is applied to the layer to identify the masked areas that are poisoning for conventional convolutions. On the FaceForensics and Free-form Video Inpainting (FVI) dataset, our model achieves state-of-the-art results with simply 33% of parameters and inference time. | ['Kuan-Ying Lee', 'Ya-Liang Chang', 'Winston Hsu', 'Zhe Yu Liu'] | 2019-07-02 | null | null | null | null | ['video-inpainting'] | ['computer-vision'] | [ 1.04981087e-01 -7.70944357e-02 -3.95149142e-01 -2.58864313e-01
-4.08651114e-01 -2.37969398e-01 1.81744710e-01 -7.13907838e-01
-2.56449789e-01 7.88558066e-01 1.00678265e-01 -3.88367712e-01
1.65161744e-01 -6.73197329e-01 -1.16391861e+00 -5.40376008e-01
-3.16050723e-02 -1.61824033e-01 7.38320798e-02 -1.18334077e-01
-1.72803879e-01 5.61409354e-01 -1.34755862e+00 5.69055200e-01
6.02572799e-01 1.07674325e+00 4.48621750e-01 6.10167921e-01
-1.95970550e-01 9.07239735e-01 -5.68698108e-01 -3.03274244e-01
4.41234767e-01 -4.68795359e-01 -7.73817241e-01 9.32792649e-02
7.14971602e-01 -9.69033182e-01 -9.13690686e-01 8.27047050e-01
4.45304930e-01 2.54630953e-01 3.98048192e-01 -1.18464565e+00
-6.90631688e-01 3.19365054e-01 -9.27439570e-01 4.36506420e-01
2.56267935e-01 2.70238072e-01 4.57230747e-01 -9.35855985e-01
7.61296153e-01 1.33763301e+00 5.56836545e-01 8.66060138e-01
-1.22606456e+00 -1.02082598e+00 3.50847453e-01 5.50126731e-01
-1.18387854e+00 -5.63190997e-01 7.71075368e-01 -1.44433081e-01
1.03740025e+00 2.84497380e-01 7.48036921e-01 1.26658595e+00
1.81315601e-01 8.89248073e-01 7.37450838e-01 -1.25153529e-04
2.11401172e-02 -5.48453271e-01 -4.12707210e-01 8.15422833e-01
-1.90956414e-01 1.55821905e-01 -8.79741311e-01 2.32681736e-01
1.30945218e+00 4.07641321e-01 -4.81302112e-01 4.70199376e-01
-9.73252177e-01 7.77554095e-01 3.62014353e-01 1.79182574e-01
-4.62960809e-01 4.33348894e-01 2.58236229e-01 6.00892901e-01
7.02612460e-01 1.48004040e-01 -4.50252146e-01 -4.95788381e-02
-1.35238612e+00 4.92231578e-01 1.65126830e-01 1.13876534e+00
8.95708382e-01 1.87959716e-01 -4.73378718e-01 7.75266707e-01
-9.64950398e-02 3.12354803e-01 2.34953433e-01 -1.36526775e+00
4.92808342e-01 2.59067088e-01 6.39568046e-02 -9.21932936e-01
-5.51904030e-02 3.45123187e-02 -9.08678949e-01 1.70724332e-01
2.01879844e-01 -1.98306829e-01 -1.18298280e+00 1.58648431e+00
2.00389847e-01 8.48285317e-01 -3.16575944e-01 1.00205469e+00
1.11931288e+00 9.06359494e-01 -4.30288399e-03 -4.23787862e-01
1.01454031e+00 -1.19497406e+00 -9.56139624e-01 -2.59910434e-01
2.81511724e-01 -9.87130225e-01 1.01355994e+00 2.03587353e-01
-1.42944014e+00 -7.73894429e-01 -7.65426576e-01 -4.44418460e-01
7.95922428e-02 -1.08108670e-01 5.91223657e-01 3.98849174e-02
-1.29401982e+00 8.53189290e-01 -9.09398913e-01 -1.36812329e-01
6.95558846e-01 2.92502224e-01 -4.77223337e-01 -3.23145002e-01
-1.30635571e+00 6.89982116e-01 2.28961706e-01 3.47601235e-01
-1.14978766e+00 -7.49834359e-01 -9.03345942e-01 -1.34940654e-01
4.73253936e-01 -5.23060083e-01 1.23782516e+00 -1.07153845e+00
-1.42322588e+00 7.48596430e-01 -4.44550693e-01 -5.44485033e-01
6.32145941e-01 -2.46211827e-01 -4.68584448e-01 6.07706070e-01
1.82198927e-01 1.19834745e+00 1.44082737e+00 -7.91533232e-01
-4.68370974e-01 -5.90850748e-02 6.44909739e-02 1.11079961e-01
-2.28606761e-01 8.20422769e-02 -9.74274278e-01 -1.23342431e+00
-2.00781375e-01 -7.52656460e-01 -3.53127718e-02 4.25508350e-01
-2.30958074e-01 1.10261396e-01 1.29858553e+00 -1.23935759e+00
1.32167900e+00 -2.11107445e+00 1.18360609e-01 -2.70828754e-01
1.93378270e-01 6.30830646e-01 -4.81931716e-01 2.93986052e-01
-1.42353997e-01 9.36805084e-02 -2.42557004e-01 -5.59633255e-01
-2.92933196e-01 4.23272103e-01 -4.10783142e-01 2.67447323e-01
5.53910553e-01 1.13397050e+00 -6.91361845e-01 -5.60566306e-01
3.64414334e-01 7.50286162e-01 -7.60634720e-01 4.10963625e-01
-3.65406692e-01 5.55776775e-01 -2.02727601e-01 7.80546308e-01
1.04523349e+00 -1.06788158e-01 -4.25620005e-02 -3.93509805e-01
-1.02686338e-01 2.76630372e-02 -8.99335086e-01 1.95358276e+00
-3.73309374e-01 7.36861169e-01 2.16393888e-01 -8.59628201e-01
4.79373842e-01 4.30096567e-01 6.62171662e-01 -8.60528052e-01
1.55045629e-01 -1.62088275e-01 -3.94142240e-01 -6.96404278e-01
3.27696621e-01 1.19037889e-01 3.51394862e-01 2.66341060e-01
1.61020264e-01 -1.06797770e-01 8.22987705e-02 6.07174709e-02
1.21439576e+00 4.33592170e-01 -1.22673996e-01 2.14106604e-01
1.65045425e-01 -2.90714055e-01 6.73422039e-01 5.10057330e-01
-1.59634292e-01 8.43784094e-01 4.05166209e-01 -6.33360684e-01
-9.30931151e-01 -7.94002950e-01 5.92398010e-02 8.09674203e-01
1.49902448e-01 -3.18367690e-01 -9.11160648e-01 -6.39687598e-01
-2.23758351e-02 5.70026636e-01 -6.51825249e-01 -1.44506961e-01
-7.88283706e-01 -2.64702618e-01 4.70973104e-01 5.35813928e-01
7.03838110e-01 -1.29832435e+00 -5.37132561e-01 3.63249183e-01
-2.42609844e-01 -1.19986939e+00 -9.50578034e-01 1.13648526e-01
-9.62442398e-01 -1.02952397e+00 -8.87812376e-01 -8.27616870e-01
6.76478803e-01 5.75810432e-01 8.72873306e-01 2.20501021e-01
-3.00723732e-01 -4.13864888e-02 -4.18831706e-01 -3.97270545e-02
-1.00185245e-01 -1.11135706e-01 -1.72059536e-01 -7.29843453e-02
1.49769813e-01 -7.90605009e-01 -8.09870005e-01 2.88769662e-01
-1.21067154e+00 3.58100146e-01 3.61353725e-01 1.02923167e+00
5.25863111e-01 -1.03801452e-01 1.82181060e-01 -6.49026990e-01
2.35037401e-01 -3.31350356e-01 -5.04606485e-01 6.76195323e-02
-1.08071148e-01 -5.62243499e-02 6.71539962e-01 -7.51455784e-01
-9.66882825e-01 8.14782307e-02 -4.88138258e-01 -1.34308827e+00
-4.23920825e-02 1.76320553e-01 -3.26290764e-02 -1.09768897e-01
2.25239322e-01 2.25438282e-01 2.73148507e-01 -5.44580698e-01
2.64845818e-01 2.86182314e-01 4.46232229e-01 -2.98877686e-01
8.38243127e-01 6.11820042e-01 -2.03422144e-01 -8.09646070e-01
-7.06036389e-01 1.13789206e-02 -3.88634145e-01 -2.16528133e-01
9.06891823e-01 -1.07340634e+00 -5.80521584e-01 5.89050412e-01
-1.29240680e+00 -7.00653255e-01 -2.38580510e-01 4.01894242e-01
-4.52297181e-01 4.29283470e-01 -1.01641142e+00 -3.08984756e-01
-3.21228504e-01 -1.21638775e+00 1.02263403e+00 2.83360481e-01
-5.59040494e-02 -7.48384833e-01 -4.20893252e-01 2.24727556e-01
3.81041914e-01 2.82501340e-01 6.93973899e-01 -8.71621445e-02
-5.65096200e-01 3.55287522e-01 -2.61057585e-01 5.89993238e-01
2.97144920e-01 1.79635510e-02 -8.92583191e-01 -2.73334712e-01
5.45135364e-02 -1.27064094e-01 1.15060461e+00 7.27561474e-01
1.54723001e+00 -6.62225425e-01 -1.29316866e-01 1.06306362e+00
1.03165936e+00 3.68832141e-01 8.94120514e-01 1.40987635e-01
8.41093123e-01 2.23247871e-01 6.73408449e-01 5.28621554e-01
-6.53453991e-02 7.20226586e-01 5.83874583e-01 -3.27242911e-01
-4.04431701e-01 -4.12840247e-01 5.85235655e-01 4.58078504e-01
-1.95066426e-02 -2.21429572e-01 -1.57778367e-01 3.37232500e-01
-1.88960862e+00 -1.12915230e+00 1.30452499e-01 1.81403625e+00
1.04436803e+00 -9.06738341e-02 -8.53966773e-02 -5.54338954e-02
7.44415283e-01 5.27124107e-01 -5.30542195e-01 -3.76444250e-01
-8.13321844e-02 5.96097827e-01 5.22574067e-01 6.38908207e-01
-1.02274454e+00 1.14300251e+00 5.34486151e+00 1.03591883e+00
-1.22291470e+00 3.41886461e-01 8.84615541e-01 -5.26103020e-01
-2.72017539e-01 -2.32096370e-02 -6.46211863e-01 8.85367036e-01
6.59223318e-01 4.60422635e-01 8.50167394e-01 3.52798164e-01
6.00803733e-01 -1.50510654e-01 -9.48529303e-01 1.23731005e+00
-7.40451664e-02 -1.73297226e+00 1.59782410e-01 -1.42467543e-01
7.63320744e-01 -3.14108461e-01 2.27204353e-01 9.79279503e-02
-1.73309118e-01 -1.16588318e+00 8.31786990e-01 4.75224614e-01
1.17214811e+00 -8.23725283e-01 4.23774868e-01 8.99047703e-02
-1.12393367e+00 4.07583872e-03 -4.21435416e-01 -1.26741439e-01
3.26499164e-01 5.00373244e-01 -6.34904742e-01 2.68174767e-01
1.10806441e+00 9.01119947e-01 -1.70556486e-01 8.27953517e-01
-3.89258027e-01 4.80071872e-01 -1.92413688e-01 4.69966382e-01
3.32358062e-01 -1.70762748e-01 4.37001795e-01 9.18325603e-01
7.23783195e-01 2.60891706e-01 -2.01231748e-01 8.37949157e-01
-3.69072258e-01 -3.35052252e-01 -5.93373418e-01 9.20552202e-03
4.74782258e-01 1.01667798e+00 -4.71188545e-01 -1.62271246e-01
-6.27019227e-01 1.44136894e+00 2.67906159e-01 5.92163444e-01
-1.32038593e+00 -1.80186242e-01 1.08872557e+00 4.60064530e-01
8.00301790e-01 -2.51805723e-01 3.62233520e-01 -1.20956957e+00
7.99998119e-02 -8.61404300e-01 2.94508606e-01 -8.25033903e-01
-9.76700783e-01 6.26307249e-01 -6.40969276e-02 -1.26563537e+00
-1.61848888e-01 -5.36536813e-01 -4.63776141e-01 8.14075351e-01
-1.53504002e+00 -1.22904623e+00 -3.17843586e-01 1.07627165e+00
7.77471840e-01 -2.80616973e-02 4.37652171e-01 6.52382910e-01
-5.85458755e-01 6.72701359e-01 -2.30671182e-01 1.36930630e-01
9.01075184e-01 -5.52283466e-01 4.91827816e-01 1.03072798e+00
-1.05705909e-01 3.95081341e-01 5.10873914e-01 -7.83938468e-01
-1.50003982e+00 -1.55588293e+00 7.58014381e-01 6.02577031e-02
3.25108379e-01 -2.01539591e-01 -8.70046377e-01 7.84581304e-01
2.48796687e-01 3.57776463e-01 1.41567677e-01 -6.01254940e-01
-3.02442163e-01 -1.68771341e-01 -1.19053233e+00 7.79490530e-01
1.43195260e+00 -5.39238989e-01 -1.74796179e-01 2.08130479e-01
9.45128560e-01 -7.18311548e-01 -6.79314017e-01 2.58003861e-01
3.38479966e-01 -9.71199393e-01 1.02359593e+00 -5.07888258e-01
6.73776746e-01 -4.92694348e-01 -1.33392783e-02 -8.90588880e-01
-2.20892072e-01 -1.00365531e+00 -5.37325144e-01 1.07170713e+00
-5.38167506e-02 -3.60147029e-01 8.27334225e-01 3.73566270e-01
-2.82666773e-01 -7.60664880e-01 -1.05373871e+00 -5.98297656e-01
-3.44125509e-01 -5.19642115e-01 5.64596415e-01 8.78335953e-01
-5.43988168e-01 -2.19488904e-01 -8.77791345e-01 9.46698934e-02
3.12589556e-01 4.50547459e-03 5.20391583e-01 -4.42966133e-01
-2.86777854e-01 -2.85293877e-01 -1.00409798e-01 -1.28365254e+00
2.71016866e-01 -6.20782554e-01 -2.03907296e-01 -1.22367156e+00
-2.02594940e-02 -6.24352619e-02 -1.03544876e-01 8.55425775e-01
-1.85246140e-01 5.73714197e-01 2.48010546e-01 2.52999477e-02
-2.52710372e-01 5.24611235e-01 1.65551877e+00 -3.00723791e-01
-1.78470030e-01 -1.22839764e-01 -5.42610645e-01 4.39597368e-01
5.60138106e-01 -3.80803317e-01 -5.20938218e-01 -8.13694417e-01
-1.62949458e-01 4.82629061e-01 7.82643616e-01 -7.25272834e-01
1.06813222e-01 -3.43541533e-01 7.26251125e-01 -6.96509719e-01
6.48194373e-01 -7.55506814e-01 5.33554494e-01 4.51823443e-01
-5.57406545e-02 3.81948918e-01 3.72226685e-01 2.97605604e-01
-4.89595264e-01 -3.74532044e-02 7.30441272e-01 -2.22395524e-01
-1.03975677e+00 8.35836530e-01 -2.66228855e-01 -1.11171231e-01
1.03708506e+00 -3.44696999e-01 -8.32284316e-02 -5.22877574e-01
-7.33258009e-01 2.14829147e-01 5.53957641e-01 5.07491827e-01
9.87743795e-01 -1.43812966e+00 -5.95430315e-01 3.72340769e-01
-3.67155850e-01 2.50623226e-01 8.94430757e-01 6.59065902e-01
-7.38465786e-01 8.38388950e-02 -5.97735047e-01 -4.87794727e-01
-1.17636609e+00 6.15589499e-01 2.04134628e-01 -1.22644462e-01
-6.91207886e-01 1.22413599e+00 1.43443316e-01 1.71710119e-01
3.29397649e-01 -5.38123488e-01 1.31036028e-01 1.81205228e-01
6.63697958e-01 1.33132577e-01 2.60407869e-02 -4.21090186e-01
-3.51007462e-01 2.95935422e-01 -2.77731031e-01 5.74379563e-02
1.54495680e+00 5.97699434e-02 -1.14729822e-01 -2.11275458e-01
1.40914667e+00 -2.91599780e-01 -2.03321433e+00 -1.04966871e-01
-5.87590396e-01 -7.70183861e-01 -8.24355707e-02 -3.73046517e-01
-1.63353002e+00 8.70472193e-01 5.89449644e-01 -3.06604028e-01
1.51911521e+00 -1.47373140e-01 1.24837387e+00 -1.16147123e-01
1.94915608e-01 -1.04844677e+00 2.50470519e-01 3.90577644e-01
1.02886641e+00 -1.00873983e+00 -2.41572242e-02 -3.32029253e-01
-5.45971632e-01 1.17855537e+00 7.24233687e-01 -1.00153044e-01
6.19580269e-01 4.16107118e-01 -1.45553350e-01 6.62622675e-02
-6.46320045e-01 7.33952299e-02 7.63153732e-02 5.19175410e-01
2.02333689e-01 -2.78009087e-01 -1.61091298e-01 2.93626696e-01
7.61790872e-02 8.31166357e-02 2.36601770e-01 9.81977284e-01
1.01349264e-01 -1.10592246e+00 -2.31635839e-01 3.46450537e-01
-4.24694180e-01 -3.11881006e-01 5.05206697e-02 7.94113040e-01
4.33195472e-01 7.66135991e-01 2.49332353e-01 -4.56031889e-01
1.47621557e-01 -1.64349884e-01 6.48820221e-01 -3.53814185e-01
-5.28623283e-01 3.85241657e-01 -1.19159415e-01 -9.26441848e-01
-5.28223336e-01 -4.71214861e-01 -1.12717462e+00 -6.60237253e-01
1.42645955e-01 -1.83548868e-01 1.39309078e-01 7.23865807e-01
6.16698086e-01 5.82634449e-01 5.33149600e-01 -1.22098184e+00
-9.73391756e-02 -8.15297127e-01 -4.45922554e-01 2.34851599e-01
5.86935461e-01 -5.99956393e-01 -1.32044792e-01 3.25067312e-01] | [10.907159805297852, -1.2936288118362427] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.