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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2b3e21b2-0eaa-436f-a875-807c7ff7137c | video-chatcaptioner-towards-the-enriched | 2304.04227 | null | https://arxiv.org/abs/2304.04227v3 | https://arxiv.org/pdf/2304.04227v3.pdf | Video ChatCaptioner: Towards Enriched Spatiotemporal Descriptions | Video captioning aims to convey dynamic scenes from videos using natural language, facilitating the understanding of spatiotemporal information within our environment. Although there have been recent advances, generating detailed and enriched video descriptions continues to be a substantial challenge. In this work, we introduce Video ChatCaptioner, an innovative approach for creating more comprehensive spatiotemporal video descriptions. Our method employs a ChatGPT model as a controller, specifically designed to select frames for posing video content-driven questions. Subsequently, a robust algorithm is utilized to answer these visual queries. This question-answer framework effectively uncovers intricate video details and shows promise as a method for enhancing video content. Following multiple conversational rounds, ChatGPT can summarize enriched video content based on previous conversations. We qualitatively demonstrate that our Video ChatCaptioner can generate captions containing more visual details about the videos. The code is publicly available at https://github.com/Vision-CAIR/ChatCaptioner | ['Mohamed Elhoseiny', 'Xiang Li', 'Kilichbek Haydarov', 'Deyao Zhu', 'Jun Chen'] | 2023-04-09 | null | null | null | null | ['video-captioning'] | ['computer-vision'] | [ 2.09110573e-01 3.68396305e-02 -1.24736652e-01 -2.82884389e-01
-1.15858567e+00 -6.84101760e-01 5.66940010e-01 -2.45172366e-01
3.12045664e-02 7.71778941e-01 9.55438793e-01 1.32065594e-01
5.09773731e-01 -2.17615873e-01 -7.42456436e-01 -4.07990128e-01
2.28130911e-02 8.69581942e-03 3.28315496e-01 -2.02117145e-01
5.14971435e-01 1.09165542e-01 -1.58083797e+00 8.41898739e-01
2.76724696e-01 6.24642193e-01 7.38971353e-01 1.09522855e+00
-3.88263047e-01 1.54660511e+00 -6.81173682e-01 -2.58333504e-01
-1.60098717e-01 -6.70694828e-01 -1.01833951e+00 4.30063546e-01
5.12188137e-01 -8.60350609e-01 -8.44489694e-01 6.82922959e-01
2.77397841e-01 4.38655972e-01 2.18688160e-01 -1.61458611e+00
-7.67915308e-01 5.53283513e-01 -2.51181066e-01 4.37298328e-01
1.13189805e+00 5.42198539e-01 8.61089528e-01 -1.08796227e+00
1.06931162e+00 1.20158410e+00 2.86280960e-01 8.79990578e-01
-9.54937875e-01 -5.01338780e-01 2.59124398e-01 4.61575896e-01
-1.20258045e+00 -7.00233281e-01 8.82509530e-01 -4.00016934e-01
6.63434565e-01 4.77955163e-01 9.20851052e-01 1.48648381e+00
-2.00432792e-01 1.19249594e+00 4.53342944e-01 -5.50587066e-02
9.26672816e-02 8.97261500e-02 -2.48551339e-01 6.12910748e-01
-4.32939380e-01 -3.17448020e-01 -1.04697418e+00 6.07531182e-02
1.02650642e+00 2.41548955e-01 -5.61150670e-01 -1.95733741e-01
-1.42744875e+00 5.58101237e-01 2.33848587e-01 2.70462066e-01
-4.17723268e-01 6.01506948e-01 3.98624986e-01 -3.15092988e-02
3.16612810e-01 4.97232020e-01 9.94994417e-02 -8.83225083e-01
-8.27786744e-01 3.97608697e-01 6.03943706e-01 1.25923276e+00
8.34138870e-01 -1.23852625e-01 -7.27880597e-01 6.17054939e-01
-4.33945656e-03 3.58304262e-01 2.93565188e-02 -1.85200298e+00
4.80498850e-01 4.15616959e-01 3.08937222e-01 -1.17636275e+00
1.19311541e-01 4.21279430e-01 -2.32321918e-01 -3.46568555e-01
2.46447310e-01 -2.64486820e-01 -5.37055790e-01 1.53096604e+00
2.22755373e-01 4.19308662e-01 -1.14789002e-01 1.24019659e+00
1.15106428e+00 1.19064057e+00 1.29153624e-01 -2.63899751e-02
1.46638012e+00 -1.14276910e+00 -1.03568494e+00 -1.26551703e-01
3.75250071e-01 -7.46277392e-01 1.24156976e+00 4.45189507e-04
-1.36099887e+00 -3.51994038e-01 -4.10193771e-01 -3.37746829e-01
-6.29034713e-02 -1.92936644e-01 3.36863846e-01 6.21483214e-02
-1.34346366e+00 3.98301110e-02 -6.52024567e-01 -5.35481632e-01
4.04322892e-01 -2.15063438e-01 -4.45423633e-01 -3.10892671e-01
-1.03436863e+00 4.35310751e-01 2.02359974e-01 -3.99974994e-02
-1.29391229e+00 -7.09917307e-01 -9.93434250e-01 1.45162642e-01
6.65858269e-01 -5.21215558e-01 1.76080370e+00 -8.82006347e-01
-1.40902674e+00 6.33958817e-01 -4.54852760e-01 -2.89388955e-01
4.37376320e-01 -1.85289666e-01 4.34772409e-02 1.16755271e+00
2.40242466e-01 1.34968460e+00 6.96676075e-01 -1.37471533e+00
-6.54665887e-01 2.31844887e-01 4.29548591e-01 3.72803420e-01
-3.47566396e-01 2.84792304e-01 -9.92610455e-01 -6.88725650e-01
-4.40942317e-01 -7.16522813e-01 -9.04911477e-03 3.26451093e-01
-2.50617802e-01 4.64912951e-02 1.19925070e+00 -7.39443064e-01
1.08329391e+00 -2.11015558e+00 6.22352846e-02 -3.61593425e-01
3.62641990e-01 3.25002894e-02 -2.70033985e-01 9.07930493e-01
2.85149217e-01 1.29771844e-01 -1.14535227e-01 -4.17969584e-01
-2.00770214e-01 5.70056327e-02 -2.31451869e-01 -4.83921671e-04
3.34548950e-01 1.13786757e+00 -1.22943020e+00 -7.51550972e-01
4.61650103e-01 7.90870905e-01 -7.12804794e-01 6.51149809e-01
-5.61360061e-01 7.14451194e-01 -5.66312671e-01 6.03738427e-01
3.30326051e-01 -5.58624566e-01 -6.00991445e-03 -1.70074418e-01
-1.30053639e-01 -6.45472854e-02 -6.32260025e-01 1.78674757e+00
-2.98010409e-01 1.46479440e+00 2.62244493e-01 -6.26521051e-01
6.88548207e-01 5.03845215e-01 6.43468976e-01 -5.04768372e-01
2.83303354e-02 -2.69976497e-01 -8.16352844e-01 -1.17109907e+00
8.09321463e-01 2.45766953e-01 -8.38227421e-02 5.61024189e-01
-1.36295766e-01 -2.07015067e-01 3.91991317e-01 6.70787632e-01
1.30538082e+00 3.41156393e-01 2.93543609e-03 3.57723624e-01
3.32196593e-01 4.58159894e-01 2.04564989e-01 7.17012525e-01
-5.33083200e-01 1.04056728e+00 5.00359058e-01 -3.05671006e-01
-1.05768144e+00 -1.02732587e+00 4.93163973e-01 1.18916070e+00
4.17468756e-01 -5.64647317e-01 -8.10943782e-01 -3.65293920e-01
-3.40316534e-01 5.44679403e-01 -5.67830622e-01 1.55537456e-01
-6.40767097e-01 1.90717638e-01 3.33214968e-01 5.00337839e-01
5.60732007e-01 -1.51899564e+00 -6.31548703e-01 9.38812196e-02
-1.06409383e+00 -1.54691660e+00 -1.02528811e+00 -5.21921396e-01
-3.83995563e-01 -1.10471773e+00 -1.18889654e+00 -1.00487077e+00
6.42216027e-01 9.61960912e-01 1.14706516e+00 2.57101327e-01
-2.81708002e-01 1.11354899e+00 -8.10570836e-01 3.68200392e-02
-4.24390435e-01 -2.52307594e-01 -4.75718558e-01 8.33367109e-02
2.59192705e-01 -3.72983247e-01 -8.84387851e-01 3.72594774e-01
-9.92565334e-01 7.00009525e-01 1.13885179e-01 4.59193558e-01
3.33748668e-01 -7.08061755e-01 4.17705059e-01 -5.30547380e-01
6.51251435e-01 -7.25513339e-01 -1.18611500e-01 2.08592899e-02
2.92972386e-01 -3.17606777e-01 4.66535807e-01 -3.91045183e-01
-1.28314185e+00 1.88335076e-01 4.62359525e-02 -8.19353521e-01
-4.64348406e-01 1.91685036e-01 1.64115444e-01 1.16266094e-01
3.31528753e-01 4.84748155e-01 1.19835332e-01 -1.21987469e-01
4.41338062e-01 7.30474710e-01 8.36640775e-01 -6.19492710e-01
6.02923393e-01 6.85740173e-01 -6.07636929e-01 -1.05297565e+00
-6.04262352e-01 -7.18326271e-01 -2.48963758e-01 -1.02332032e+00
1.13799369e+00 -1.05842054e+00 -1.02668345e+00 2.27304958e-02
-1.43668687e+00 -6.44499481e-01 5.33656850e-02 2.41320357e-01
-9.36287165e-01 4.12711620e-01 -7.85381138e-01 -7.02518523e-01
-1.62539467e-01 -1.16995811e+00 1.31170452e+00 3.67979228e-01
-3.93407136e-01 -7.81136155e-01 -1.91776529e-02 8.42893600e-01
3.21085662e-01 4.23231632e-01 9.36461911e-02 -2.49346301e-01
-1.15252280e+00 1.46109508e-02 -4.01944011e-01 -2.63785087e-02
4.76203300e-02 1.48494586e-01 -7.38630891e-01 1.11818817e-02
-3.86594594e-01 -4.67303336e-01 4.80380505e-01 2.59501636e-01
1.33032429e+00 -6.67290688e-01 -2.22793624e-01 3.41083169e-01
1.13440120e+00 3.38723689e-01 6.96662962e-01 2.57951409e-01
8.34041715e-01 8.18734288e-01 7.91795731e-01 5.77030778e-01
7.08603740e-01 6.42453253e-01 6.32915735e-01 -4.50148955e-02
-1.63049519e-01 -4.67925280e-01 5.74642479e-01 5.30383050e-01
2.47081369e-02 -4.67186600e-01 -8.67285967e-01 7.49892890e-01
-2.02575564e+00 -1.56175649e+00 -4.23567230e-03 1.49270427e+00
5.94335794e-01 -4.03115988e-01 2.29023501e-01 -2.90329337e-01
1.04595339e+00 3.59281957e-01 -4.66377258e-01 -1.74445137e-01
8.52991343e-02 -5.91032863e-01 4.24599461e-02 4.07271564e-01
-6.33336723e-01 1.01855612e+00 6.13056707e+00 3.19792479e-01
-9.00907934e-01 2.02924130e-03 6.66694582e-01 -3.60111803e-01
-4.23773885e-01 -1.45736039e-01 -2.50682175e-01 5.63381910e-01
7.71252871e-01 -2.94820309e-01 4.88339961e-01 7.01575935e-01
9.20886695e-01 -2.43400335e-01 -9.50724840e-01 1.23896384e+00
4.21595305e-01 -1.99230623e+00 2.99607933e-01 -3.15054774e-01
7.47796595e-01 -2.47424811e-01 1.84579089e-03 1.53494209e-01
-2.07671132e-02 -6.68746471e-01 8.26536179e-01 4.75482792e-01
7.68094778e-01 -4.80815470e-01 4.01250899e-01 -5.37593896e-03
-1.30638802e+00 -2.06919108e-02 1.90086365e-02 -1.49233758e-01
7.34131217e-01 -8.48182887e-02 -9.82792854e-01 2.92030978e-03
9.72966611e-01 9.85416055e-01 -4.48817164e-01 1.11553824e+00
-6.44012094e-02 3.88992935e-01 1.82748273e-01 -2.62527227e-01
3.16822767e-01 2.09634453e-02 7.63307810e-01 1.53384030e+00
3.65852267e-01 4.76434380e-01 1.54670641e-01 7.85708964e-01
-2.42308304e-01 -5.03094979e-02 -7.46125042e-01 -3.76303941e-01
6.77400589e-01 1.24145031e+00 -7.05891848e-01 -5.52454948e-01
-5.36118448e-01 1.19138765e+00 1.84303194e-01 6.08145714e-01
-1.05142224e+00 -2.49421835e-01 7.55302727e-01 1.80629402e-01
2.31864005e-01 -1.94749281e-01 2.41504923e-01 -1.39591920e+00
6.57044202e-02 -8.89923275e-01 2.19422087e-01 -1.73186743e+00
-8.01786304e-01 5.03448546e-01 2.32705757e-01 -1.35350561e+00
-3.59021038e-01 -2.97698945e-01 -7.47241616e-01 1.47633180e-01
-1.09794092e+00 -9.06388044e-01 -9.88224089e-01 7.24886358e-01
1.20206523e+00 2.44347289e-01 4.40461010e-01 2.05385521e-01
-4.91330683e-01 4.70561944e-02 -1.64332956e-01 2.61866003e-01
8.85939598e-01 -9.42610979e-01 2.71953195e-01 6.80867195e-01
1.06410787e-01 2.83132643e-01 1.08760738e+00 -5.06320119e-01
-1.61995995e+00 -1.03206754e+00 5.56449533e-01 -6.16493404e-01
6.40826285e-01 -4.50749189e-01 -9.30293977e-01 7.64342487e-01
7.43121266e-01 -2.48216346e-01 6.21776402e-01 -6.49067283e-01
-6.31330013e-02 1.71109363e-01 -8.56636882e-01 9.74724948e-01
1.10185194e+00 -7.12558925e-01 -5.13168812e-01 4.01201755e-01
9.61965442e-01 -4.47269529e-01 -5.54039001e-01 -2.86604285e-01
5.47569573e-01 -1.04182923e+00 9.27307308e-01 -2.36710072e-01
9.68331695e-01 -4.56100345e-01 -2.26242952e-02 -1.04114008e+00
5.75806238e-02 -1.15891647e+00 -2.45295122e-01 1.35478747e+00
1.32378832e-01 1.83291151e-03 9.85197484e-01 7.98228383e-01
-1.77180648e-01 -3.44556630e-01 -4.47767496e-01 -3.60752255e-01
-6.75792396e-01 -3.28823239e-01 3.59846264e-01 7.69353330e-01
4.18556988e-01 1.49771214e-01 -6.60476685e-01 -6.47796094e-02
3.87959450e-01 -5.83203360e-02 1.08834994e+00 -5.86889684e-01
5.18899783e-02 -2.34528631e-01 -2.21768722e-01 -1.25294912e+00
1.15544617e-01 -5.10901928e-01 3.54602635e-01 -1.94792187e+00
5.04447699e-01 3.87537807e-01 2.93356478e-01 3.10048670e-01
-1.60306290e-01 6.18891060e-01 5.72904825e-01 2.40031600e-01
-1.21615744e+00 5.50016105e-01 1.57833505e+00 -1.57030508e-01
-2.16566026e-01 -5.09433568e-01 -6.77517414e-01 4.05137002e-01
9.21597123e-01 -2.57019192e-01 -4.21715885e-01 -4.37083364e-01
1.15433126e-03 6.96180642e-01 6.75335944e-01 -9.84138250e-01
3.59236896e-01 -3.97080809e-01 2.31388435e-01 -6.86795115e-01
7.79178798e-01 -6.34295046e-01 2.05896929e-01 1.81305125e-01
-7.42344022e-01 2.99771786e-01 2.64296830e-01 6.86408281e-01
-5.07280588e-01 2.74322946e-02 3.01786721e-01 -3.67151022e-01
-1.05656004e+00 3.69271904e-01 -9.91530538e-01 2.12439060e-01
1.11462104e+00 -4.11670327e-01 -5.98973513e-01 -1.25963342e+00
-7.48569131e-01 6.15184546e-01 6.34688497e-01 6.29884064e-01
1.08240056e+00 -1.34332240e+00 -7.43697643e-01 -3.03926319e-01
3.99049550e-01 -1.34314090e-01 7.17404306e-01 4.94568408e-01
-8.84399176e-01 3.62815440e-01 -3.55336040e-01 -6.57292783e-01
-1.48766041e+00 5.17862439e-01 -5.72210737e-03 4.38045323e-01
-8.03527474e-01 7.01175272e-01 4.43462491e-01 1.76254034e-01
3.46030921e-01 -1.17863923e-01 -1.97464108e-01 -3.46836671e-02
9.57341373e-01 3.56304139e-01 -7.01327085e-01 -7.50168622e-01
-1.48814052e-01 3.32960188e-01 -5.91649907e-03 -2.52008587e-01
1.21508741e+00 -9.05036330e-01 2.49855816e-01 2.22199410e-01
1.42895937e+00 -2.38964751e-01 -1.83046234e+00 2.23004259e-02
-3.04106772e-01 -6.83938980e-01 -2.99114525e-01 -4.93149281e-01
-9.05007005e-01 8.35485876e-01 2.63646487e-02 2.30354041e-01
1.02500856e+00 2.38117829e-01 1.03235638e+00 3.69346052e-01
2.99661845e-01 -9.16557550e-01 8.35840344e-01 3.88430268e-01
1.18895400e+00 -1.40195131e+00 -4.19255406e-01 -4.64874566e-01
-1.10109675e+00 1.20746863e+00 9.05516744e-01 3.02297566e-02
3.73337716e-02 1.52984828e-01 2.60712564e-01 -2.47801527e-01
-1.12622797e+00 -2.28430912e-01 -9.35711712e-02 7.60605216e-01
2.45726436e-01 -3.21712077e-01 1.66416973e-01 1.13742694e-01
8.78847465e-02 9.10549760e-02 1.06408763e+00 7.60746062e-01
-5.00217319e-01 -6.43295646e-01 -5.84431648e-01 7.98692480e-02
-3.76875907e-01 1.72879230e-02 -4.12736267e-01 6.19563878e-01
-5.39682865e-01 1.23249614e+00 3.31424892e-01 -2.88548619e-01
2.80119525e-03 -1.23434722e-01 2.22865716e-01 -6.49042130e-01
-4.45543855e-01 -2.96131149e-02 2.30718434e-01 -9.65277612e-01
-6.23564065e-01 -6.37798488e-01 -1.30832529e+00 -2.92907298e-01
3.64346057e-01 3.84181708e-01 4.81958956e-01 6.43760622e-01
6.32967055e-01 4.31035370e-01 5.47471404e-01 -1.34514737e+00
3.84098202e-01 -6.52960241e-01 8.86062086e-02 6.12950921e-01
6.19360387e-01 -3.87333721e-01 -3.77164930e-01 6.73052728e-01] | [10.537740707397461, 0.8803253173828125] |
41542ba7-80f5-4634-8c62-38c3edab361b | lmu-munichas-neural-machine-translation | null | null | https://aclanthology.org/W18-6446 | https://aclanthology.org/W18-6446.pdf | LMU Munich's Neural Machine Translation Systems at WMT 2018 | We present the LMU Munich machine translation systems for the English{--}German language pair. We have built neural machine translation systems for both translation directions (English→German and German→English) and for two different domains (the biomedical domain and the news domain). The systems were used for our participation in the WMT18 biomedical translation task and in the shared task on machine translation of news. The main focus of our recent system development efforts has been on achieving improvements in the biomedical domain over last year{'}s strong biomedical translation engine for English→German (Huck et al., 2017a). Considerable progress has been made in the latter task, which we report on in this paper. | ['er', 'Viktor Hangya', 'Alex Fraser', 'Matthias Huck', 'Dario Stojanovski'] | 2018-10-01 | null | null | null | ws-2018-10 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 4.02652085e-01 3.95562619e-01 -4.44846123e-01 -5.08051157e-01
-1.35334492e+00 -5.18659770e-01 8.22574437e-01 3.77935544e-02
-7.55907714e-01 1.39803207e+00 5.68814635e-01 -1.00066090e+00
3.26352745e-01 -3.95974219e-01 -5.79310000e-01 -2.40914240e-01
4.53737438e-01 9.77614820e-01 -3.41656864e-01 -5.29894590e-01
-2.24733934e-01 1.28193960e-01 -2.90415615e-01 8.21977854e-01
6.38563633e-01 2.53242403e-01 3.37411021e-03 7.69803703e-01
1.56057209e-01 4.55674738e-01 -1.97792053e-01 -9.73049223e-01
1.35649472e-01 -9.36566412e-01 -1.56560695e+00 -5.67756176e-01
2.47169495e-01 -6.21494949e-02 -2.91942477e-01 6.98387742e-01
8.61140728e-01 -2.53798723e-01 7.26829290e-01 -5.85523665e-01
-9.69021142e-01 9.54941809e-01 -2.15043977e-01 4.90916103e-01
3.47809345e-01 -4.36674766e-02 9.78709519e-01 -1.05365586e+00
1.16212285e+00 9.67375875e-01 8.11586738e-01 9.70476449e-01
-1.08357251e+00 -4.78909850e-01 -3.16774189e-01 3.18722688e-02
-1.08123982e+00 -8.81796479e-01 1.78106010e-01 -2.39129201e-01
1.66852558e+00 1.03352323e-01 3.81775409e-01 1.66221380e+00
7.94491649e-01 6.73385859e-01 1.22103894e+00 -5.95799446e-01
-1.94559187e-01 9.30032656e-02 -2.83248544e-01 4.66086715e-01
6.65550455e-02 1.85343638e-01 -5.87105513e-01 -2.85502493e-01
4.26493168e-01 -6.76343381e-01 -1.15267798e-01 5.81061721e-01
-1.83297956e+00 6.19759679e-01 1.07185483e-01 7.28174329e-01
-4.01054382e-01 1.53049082e-01 6.29253626e-01 8.99754643e-01
1.04155481e+00 5.74926555e-01 -9.15248334e-01 -2.89766878e-01
-8.52189124e-01 1.07859904e-02 1.05969059e+00 1.14890981e+00
7.09938109e-02 -1.70664474e-01 -1.91500172e-01 9.82479811e-01
7.14639127e-02 6.32877827e-01 7.97836483e-01 -4.02701467e-01
9.74500060e-01 -1.26830563e-01 4.50688452e-02 -4.85359818e-01
-6.06196404e-01 -4.13029790e-01 -8.35366309e-01 -6.00511074e-01
4.41072375e-01 -6.90630257e-01 -7.38426387e-01 1.88084733e+00
3.88347246e-02 -1.94922104e-01 5.69160759e-01 6.73202753e-01
9.83140588e-01 4.84213471e-01 2.80676901e-01 -1.43941298e-01
1.50299287e+00 -9.51433241e-01 -7.36792505e-01 -4.28121179e-01
1.12055838e+00 -1.36320448e+00 4.17920679e-01 7.24031636e-03
-1.54276478e+00 -2.87746608e-01 -8.77866983e-01 -4.41169947e-01
-4.30480033e-01 4.59396869e-01 4.42391127e-01 2.80392855e-01
-1.51331174e+00 6.59222782e-01 -9.07015324e-01 -1.05013180e+00
-3.70758539e-03 3.86276335e-01 -7.67247558e-01 -1.00621022e-01
-1.69753695e+00 1.63077033e+00 2.50745773e-01 1.90008238e-01
-3.05767566e-01 -3.82200062e-01 -4.49646473e-01 -4.20327008e-01
-3.39848429e-01 -1.26715958e+00 1.56990814e+00 -1.02445078e+00
-1.55154943e+00 1.58601022e+00 4.57675429e-03 -5.49700856e-01
7.80422628e-01 1.88844744e-02 -7.07881451e-01 -1.87351644e-01
3.14289957e-01 6.69969857e-01 1.35249481e-01 -2.69368798e-01
-5.80320835e-01 -2.59929061e-01 -4.95479554e-01 1.51094183e-01
7.93448985e-02 5.12871802e-01 -1.28211714e-02 -8.66775870e-01
2.10157130e-03 -1.19462514e+00 -4.90252711e-02 -4.50583279e-01
-4.17856604e-01 -2.17745572e-01 -1.10944606e-01 -1.19162464e+00
7.87829340e-01 -1.67192233e+00 4.83189702e-01 -2.38661364e-01
-2.25410745e-01 6.78164512e-02 -3.66106451e-01 8.88368666e-01
-3.49074483e-01 8.73285457e-02 -2.63418317e-01 -5.12123346e-01
-1.70315623e-01 8.82706791e-02 -3.05195361e-01 5.36994576e-01
4.11137223e-01 1.27819347e+00 -9.02315915e-01 -3.84760499e-01
-4.55651045e-01 1.93392769e-01 -2.24415481e-01 -8.64631832e-02
-2.38398444e-02 6.72461510e-01 -1.94256008e-01 5.35403073e-01
2.86753803e-01 3.07509173e-02 3.52029830e-01 -4.33998071e-02
-1.21532388e-01 9.21625376e-01 5.07355630e-02 2.09914231e+00
-5.60361922e-01 4.91642863e-01 4.73124124e-02 -7.35102654e-01
6.40009284e-01 9.36536729e-01 4.93239433e-01 -4.84611243e-01
4.32689697e-01 9.43886876e-01 4.22253758e-01 -3.48067731e-01
2.92357892e-01 -6.38215721e-01 -1.07210450e-01 6.03053451e-01
4.01686072e-01 -1.03166766e-01 1.80914924e-01 -2.90095657e-01
1.00601113e+00 6.36272192e-01 4.97557551e-01 -7.05753386e-01
2.47216672e-01 4.82661664e-01 4.19764966e-01 2.94849634e-01
-3.09472084e-01 6.69384599e-01 9.16643292e-02 -4.79719162e-01
-1.15083241e+00 -7.04234362e-01 -2.71421939e-01 9.96204376e-01
-7.07547545e-01 -2.81580299e-01 -7.94107199e-01 -6.42354786e-01
-4.06316131e-01 7.44199157e-01 -6.04956627e-01 -3.35575610e-01
-1.04152763e+00 -1.08025146e+00 1.30774081e+00 4.88409758e-01
4.12019342e-01 -9.05414701e-01 -1.98386163e-02 6.96127236e-01
-7.40527689e-01 -1.41352189e+00 -7.10421979e-01 2.21135929e-01
-1.00071371e+00 -6.36922538e-01 -1.12831604e+00 -1.12515330e+00
1.72834516e-01 -2.95527309e-01 1.19693434e+00 -4.48380649e-01
-4.81070578e-02 -2.80055907e-02 -1.85046151e-01 -8.06791604e-01
-8.47770333e-01 6.01460397e-01 2.27893591e-01 -5.69983721e-01
6.58486664e-01 -4.27291721e-01 -3.73544902e-01 3.56666058e-01
-3.43624026e-01 5.03009200e-01 8.12700629e-01 1.10554719e+00
3.75942737e-01 -1.22331107e+00 9.92293596e-01 -9.14949894e-01
9.20203567e-01 -6.74049973e-01 -7.01060370e-02 2.67446637e-01
-6.59440279e-01 -1.45861000e-01 4.95669603e-01 -1.54152960e-01
-8.49140823e-01 -2.21868664e-01 -5.91692269e-01 2.52081513e-01
9.36580822e-02 8.96740973e-01 1.70504093e-01 3.41282904e-01
7.30671763e-01 2.24594161e-01 -7.54866451e-02 -4.68535751e-01
1.84607491e-01 8.72454405e-01 5.24680555e-01 -6.02474511e-01
2.53041804e-01 -1.36237845e-01 -2.47705616e-02 -4.05298740e-01
-4.05127585e-01 -1.20513968e-01 -7.66539693e-01 4.02441174e-01
1.25697005e+00 -1.10539639e+00 1.06201597e-01 3.37303102e-01
-1.57317352e+00 -4.06933159e-01 2.22611744e-02 9.10984397e-01
-6.46778464e-01 -8.08608904e-02 -1.26583827e+00 -4.76735309e-02
-8.16357255e-01 -1.09234095e+00 1.12342572e+00 -5.54578960e-01
-7.64548659e-01 -1.28571582e+00 5.56433618e-01 2.79891670e-01
5.19121766e-01 7.45026171e-02 1.13260221e+00 -9.07852471e-01
8.31809416e-02 -1.56453341e-01 -4.28660810e-02 1.28764942e-01
2.00012714e-01 -2.32803449e-01 -6.53109193e-01 -2.94840276e-01
-4.73131724e-02 -2.44296789e-01 8.90749574e-01 1.00790530e-01
1.85375050e-01 -1.09017663e-01 -4.21426088e-01 3.48035246e-01
8.84198844e-01 8.87667313e-02 3.78259838e-01 3.44902754e-01
2.82047302e-01 5.94651222e-01 4.37855780e-01 -3.55310977e-01
4.77511555e-01 7.35423207e-01 -4.06201899e-01 -5.17067194e-01
-2.89340883e-01 -1.81659251e-01 6.20813727e-01 1.34799337e+00
-2.04212070e-01 -3.99888754e-01 -1.10487044e+00 8.08628082e-01
-2.01741600e+00 -5.79103589e-01 -3.06043681e-02 1.77671051e+00
1.27643669e+00 -1.91313013e-01 -5.84365577e-02 -6.11158848e-01
4.88482922e-01 -3.90800983e-01 -1.31985664e-01 -1.00280309e+00
-1.27746105e-01 4.20954585e-01 4.34321493e-01 6.78541839e-01
-9.82736826e-01 1.17936587e+00 7.08549404e+00 4.55577940e-01
-1.12512410e+00 4.15068179e-01 8.19695532e-01 -2.15549525e-02
-1.37414441e-01 -1.74860373e-01 -7.10656285e-01 2.03148127e-01
1.63479996e+00 -3.21383715e-01 3.36861163e-01 3.48513961e-01
1.75415009e-01 3.41877222e-01 -1.58278298e+00 6.52336717e-01
-6.81057125e-02 -1.32054269e+00 -1.87550023e-01 1.47514701e-01
6.86278701e-01 7.80228078e-01 -6.04474125e-03 3.38114232e-01
4.65033084e-01 -1.15776908e+00 5.48767507e-01 3.14219922e-01
1.29481304e+00 -5.72145343e-01 1.07596993e+00 3.36915523e-01
-5.93349099e-01 5.62811196e-01 -4.39959049e-01 1.57590568e-01
1.00044340e-01 5.37411451e-01 -1.19462907e+00 7.48974860e-01
4.34061676e-01 7.28029013e-01 -8.07685331e-02 2.20755026e-01
-2.35223621e-01 7.95071661e-01 -1.96687236e-01 -1.10796783e-02
3.39000642e-01 -3.44954193e-01 5.70136905e-01 1.65462494e+00
4.47760135e-01 -4.41334471e-02 2.24652868e-02 5.82336962e-01
-6.39301538e-01 5.00179410e-01 -6.51491463e-01 -3.89307618e-01
-5.63149154e-02 1.09224975e+00 -4.64297652e-01 -4.80070710e-01
-3.05605322e-01 1.25557601e+00 3.76137346e-01 3.29546899e-01
-6.74836338e-01 -2.88481802e-01 6.01576805e-01 -2.60245681e-01
-2.94131339e-01 -1.08523242e-01 -4.51873928e-01 -1.35884750e+00
-2.05258802e-01 -1.05544305e+00 2.87728459e-01 -6.56436980e-01
-1.40663326e+00 1.20469069e+00 -6.79089576e-02 -1.05550003e+00
-7.73331285e-01 -5.98075449e-01 -3.58106852e-01 1.44682920e+00
-1.11510706e+00 -1.66279626e+00 8.27890396e-01 3.49913359e-01
5.24610519e-01 -4.62358534e-01 1.37396038e+00 5.83349228e-01
-3.57629746e-01 7.17717171e-01 3.49208862e-01 1.84015751e-01
1.54154742e+00 -9.80823338e-01 1.05768275e+00 5.95745444e-01
-3.41677554e-02 9.68719304e-01 6.93156600e-01 -7.18150020e-01
-1.24212360e+00 -1.15037870e+00 2.10587716e+00 -5.99082887e-01
7.22813308e-01 -3.79925102e-01 -2.48093739e-01 1.06342876e+00
8.74946237e-01 -5.79414845e-01 9.65485454e-01 1.27551571e-01
-7.31069073e-02 3.55640173e-01 -1.28900802e+00 6.01915777e-01
9.79631901e-01 -7.62282491e-01 -7.04230547e-01 8.65691066e-01
7.25782514e-01 -4.46805418e-01 -1.21475971e+00 2.37856761e-01
5.50106406e-01 -1.95023403e-01 6.74818933e-01 -1.39691126e+00
9.41492379e-01 1.45046636e-01 -1.61480188e-01 -1.82982552e+00
-4.32325870e-01 -7.36270547e-01 3.58764827e-01 9.33655798e-01
1.05355906e+00 -8.92786562e-01 2.40323693e-01 1.53805435e-01
-3.58926564e-01 -9.21315730e-01 -1.42248106e+00 -6.00785851e-01
9.82025146e-01 -6.28808886e-02 3.84825975e-01 1.10917413e+00
4.84276295e-01 7.79038489e-01 -4.36538070e-01 -5.07867098e-01
-9.25131291e-02 -1.98742524e-02 3.99947762e-01 -7.85730660e-01
-5.67182064e-01 -3.19789469e-01 -1.08908571e-01 -6.24453485e-01
9.93358567e-02 -1.64426589e+00 -4.59731892e-02 -1.69156945e+00
2.61514366e-01 3.35782170e-01 -2.20347479e-01 8.47154737e-01
-1.39857829e-01 4.58583862e-01 -1.19961873e-01 2.75020897e-01
1.81235641e-01 1.06860168e-01 1.09740925e+00 -1.29577458e-01
3.00765276e-01 6.70682713e-02 -8.73754978e-01 2.37996086e-01
9.36924756e-01 -5.08189142e-01 1.15802381e-02 -9.34768379e-01
2.25920439e-01 7.44306967e-02 -1.48351446e-01 -4.71251398e-01
-1.50731578e-01 4.04064655e-02 4.11554664e-01 -1.52115509e-01
1.27488121e-01 -3.69428664e-01 1.23685554e-01 7.11130857e-01
-5.30680716e-01 8.24195385e-01 4.32180613e-01 5.92514612e-02
-1.05664454e-01 1.46781668e-01 8.41790974e-01 -3.43500525e-01
1.13002777e-01 9.79285762e-02 -7.34861434e-01 -3.38889919e-02
5.36732137e-01 2.57838398e-01 -4.35866117e-01 -3.45921189e-01
-9.63165879e-01 6.37712702e-02 1.91729724e-01 4.65781748e-01
2.25308418e-01 -1.30735874e+00 -1.52140045e+00 -1.32305905e-01
1.42921761e-01 -8.60439718e-01 -3.57557654e-01 1.50270295e+00
-5.21388471e-01 8.64932716e-01 -3.77066463e-01 -1.33563951e-01
-1.11577058e+00 3.23242486e-01 4.60522801e-01 -7.01959908e-01
-3.25990021e-01 8.60194266e-01 -2.34977052e-01 -8.53592932e-01
-4.97684658e-01 -3.55413526e-01 1.74413234e-01 -1.22054279e-01
4.16092962e-01 2.26988271e-01 5.06257594e-01 -6.40952587e-01
-4.46243912e-01 1.87622711e-01 -2.99965680e-01 -4.41710740e-01
1.31186056e+00 1.77930426e-02 -4.53618675e-01 4.87962216e-01
1.40385795e+00 -6.92628101e-02 3.24098058e-02 -3.17978591e-01
1.39808550e-01 4.15140033e-01 -2.13447720e-01 -1.41642749e+00
-4.79348570e-01 9.00357187e-01 4.91347432e-01 -4.37523603e-01
7.67644227e-01 -1.34958982e-01 1.16519010e+00 4.76288319e-01
5.23886263e-01 -1.14377725e+00 -8.82032931e-01 9.04932141e-01
9.88122046e-01 -1.11679780e+00 -1.47365302e-01 -1.35322049e-01
-4.51308101e-01 1.36961114e+00 1.96886823e-01 2.04291135e-01
3.87913078e-01 1.89510688e-01 3.66963625e-01 -4.20588329e-02
-8.11693609e-01 1.26484156e-01 4.84635115e-01 3.66252393e-01
1.28993392e+00 1.82771564e-01 -1.15395212e+00 5.54693222e-01
-3.81026506e-01 2.16867924e-01 2.27366254e-01 1.05760801e+00
-2.32907338e-03 -1.57783294e+00 -1.69850439e-01 3.72342527e-01
-8.54043782e-01 -5.45646727e-01 -1.14985883e+00 6.13609374e-01
-3.81802283e-02 1.02258563e+00 -4.15371299e-01 -3.74762595e-01
2.36117825e-01 5.20012677e-01 7.69197226e-01 -7.11353242e-01
-9.94132757e-01 2.18137830e-01 7.59966016e-01 -3.22144866e-01
-1.92288101e-01 -7.79800117e-01 -1.05475724e+00 -3.52659881e-01
-4.03034464e-02 3.18652779e-01 7.70197392e-01 1.06207514e+00
4.13068682e-01 6.42112792e-01 -1.94585137e-02 -4.64324653e-01
-7.55885780e-01 -1.50228012e+00 -8.78445134e-02 -6.51200637e-02
5.19582536e-03 6.42294958e-02 1.75401747e-01 2.61193030e-02] | [11.547478675842285, 10.41109848022461] |
324d6fcb-8349-4e38-b98e-cff3d840acb9 | dsvae-interpretable-disentangled | 2304.03323 | null | https://arxiv.org/abs/2304.03323v1 | https://arxiv.org/pdf/2304.03323v1.pdf | DSVAE: Interpretable Disentangled Representation for Synthetic Speech Detection | Tools to generate high quality synthetic speech signal that is perceptually indistinguishable from speech recorded from human speakers are easily available. Several approaches have been proposed for detecting synthetic speech. Many of these approaches use deep learning methods as a black box without providing reasoning for the decisions they make. This limits the interpretability of these approaches. In this paper, we propose Disentangled Spectrogram Variational Auto Encoder (DSVAE) which is a two staged trained variational autoencoder that processes spectrograms of speech using disentangled representation learning to generate interpretable representations of a speech signal for detecting synthetic speech. DSVAE also creates an activation map to highlight the spectrogram regions that discriminate synthetic and bona fide human speech signals. We evaluated the representations obtained from DSVAE using the ASVspoof2019 dataset. Our experimental results show high accuracy (>98%) on detecting synthetic speech from 6 known and 10 out of 11 unknown speech synthesizers. We also visualize the representation obtained from DSVAE for 17 different speech synthesizers and verify that they are indeed interpretable and discriminate bona fide and synthetic speech from each of the synthesizers. | ['Edward J. Delp', 'Stefano Tubaro', 'Paolo Bestagini', 'Ziyue Xiang', 'Kratika Bhagtani', 'Amit Kumar Singh Yadav'] | 2023-04-06 | null | null | null | null | ['synthetic-speech-detection'] | ['audio'] | [ 2.11273685e-01 6.20698750e-01 4.34593707e-01 -1.97437197e-01
-1.09112060e+00 -9.69471455e-01 7.58305490e-01 -4.30011272e-01
1.83673546e-01 5.78513801e-01 4.95030016e-01 -4.32132661e-01
3.18306565e-01 -3.71234000e-01 -6.27082586e-01 -7.49072015e-01
2.10401952e-01 4.93290454e-01 -6.33533020e-03 -1.49894983e-01
-2.91554451e-01 4.33357418e-01 -1.75006187e+00 8.07845354e-01
6.52179539e-01 6.47705495e-01 3.70610386e-01 1.49001849e+00
2.35790133e-01 6.19955301e-01 -1.57643080e+00 -3.47171098e-01
2.22862706e-01 -7.61806250e-01 -3.40477347e-01 2.18521178e-01
2.73139894e-01 -3.11237693e-01 -4.39029872e-01 1.03515160e+00
6.39565229e-01 -6.62575513e-02 9.08528626e-01 -1.28578877e+00
-1.03864086e+00 1.08732510e+00 -5.64185977e-02 3.85020316e-01
3.96846712e-01 4.99115050e-01 9.94498551e-01 -9.98860955e-01
3.14921081e-01 1.58585501e+00 9.59436819e-02 7.89506078e-01
-1.20148206e+00 -8.09904575e-01 -3.09940279e-01 -5.51850721e-02
-9.26411390e-01 -1.00007629e+00 8.29043627e-01 -6.66949928e-01
1.12977779e+00 7.97335088e-01 4.27062571e-01 1.91446662e+00
-3.63235772e-01 9.25909579e-01 1.15311837e+00 -3.79781932e-01
2.24844992e-01 3.63185197e-01 -1.85538650e-01 5.64355910e-01
7.64793903e-02 6.91059649e-01 -6.95099056e-01 -2.39895165e-01
4.70881224e-01 -6.39979303e-01 -5.75930893e-01 2.09858611e-01
-1.49633574e+00 9.34479713e-01 6.78896084e-02 2.28346452e-01
-2.48433709e-01 -2.04510331e-01 3.70088130e-01 2.92898625e-01
4.00359094e-01 7.02921510e-01 -1.83033600e-01 -6.10821955e-02
-9.39263046e-01 -4.26990874e-02 6.80806220e-01 6.04625881e-01
2.25761160e-01 1.15591192e+00 -1.66085109e-01 8.20316076e-01
4.79735196e-01 8.66217852e-01 7.64227688e-01 -8.27100754e-01
4.78782892e-01 1.74329743e-01 -4.47460264e-02 -7.47475922e-01
1.43615350e-01 -3.05428565e-01 -5.38761079e-01 6.66790664e-01
2.57446796e-01 -3.45317036e-01 -1.06616116e+00 1.43633318e+00
-1.46481723e-01 3.36433172e-01 8.05798292e-01 1.04373324e+00
1.31632543e+00 1.11435616e+00 -4.63540554e-01 7.25337788e-02
1.31729770e+00 -8.55251193e-01 -8.85590255e-01 -2.22289294e-01
8.16586148e-03 -9.08985615e-01 1.12337291e+00 4.90687698e-01
-8.41076672e-01 -7.06444442e-01 -1.60083294e+00 1.46305859e-01
-1.78501070e-01 6.51008725e-01 4.97451052e-02 9.61252034e-01
-8.07804525e-01 2.88886279e-01 -7.40374029e-01 2.85051286e-01
2.00796410e-01 4.19018269e-02 -3.70747179e-01 8.26312661e-01
-1.29677927e+00 7.80573130e-01 4.93971646e-01 1.54675385e-02
-1.68029225e+00 -3.70162934e-01 -9.20302987e-01 2.04919189e-01
-6.48963973e-02 -3.00401360e-01 1.44297493e+00 -9.83175933e-01
-1.78387177e+00 8.37149620e-01 1.86321765e-01 -6.48811638e-01
5.05582094e-01 2.72924781e-01 -1.07276762e+00 2.43329406e-01
-7.69783324e-03 4.79415894e-01 1.48370254e+00 -1.35066330e+00
-2.75947630e-01 -1.73435673e-01 -3.09275627e-01 -7.63884932e-02
7.34440461e-02 5.31620346e-02 6.02749228e-01 -8.61690104e-01
-3.20386402e-02 -8.60769033e-01 3.29061270e-01 -2.78417736e-01
-8.39842141e-01 -1.46141648e-01 1.20182836e+00 -1.07381892e+00
6.06591940e-01 -2.34014177e+00 8.18235502e-02 -3.73288602e-01
5.36114931e-01 6.45284534e-01 -5.33159636e-02 2.20137760e-01
-2.31800959e-01 2.70014614e-01 -2.46866539e-01 -5.60808539e-01
1.69289663e-01 2.00413331e-01 -9.87545133e-01 4.65596378e-01
4.70043629e-01 3.84303778e-01 -9.17788327e-01 8.51808302e-03
3.09665442e-01 7.82795191e-01 -2.45863438e-01 6.04663789e-01
-3.27770919e-01 6.63721323e-01 -3.53819155e-03 1.87153712e-01
5.11533797e-01 1.43865258e-01 6.02292567e-02 1.23905465e-02
-2.03407425e-02 6.55668437e-01 -8.77750874e-01 1.14829874e+00
-4.37333286e-01 1.39548481e+00 1.80885836e-01 -5.78923166e-01
1.30616796e+00 9.65130389e-01 -4.58489209e-01 -2.42485777e-01
8.70509371e-02 2.15132162e-01 2.98524678e-01 -5.67668796e-01
5.01744330e-01 -5.00891618e-02 3.46279354e-03 5.40556908e-01
3.65473419e-01 -4.43982333e-01 -2.51454979e-01 1.92981720e-01
8.89948487e-01 -2.16673478e-01 2.62770295e-01 -3.37298736e-02
4.22520310e-01 -4.65606660e-01 4.68370914e-01 6.35598898e-01
-3.57247144e-01 9.64212418e-01 6.12074673e-01 -2.04631716e-01
-1.27023542e+00 -1.72257304e+00 3.16252321e-01 8.36198986e-01
-4.49970186e-01 -1.33486450e-01 -7.92763472e-01 -4.10513073e-01
-3.26436222e-01 1.14199412e+00 -5.06973922e-01 -1.74242854e-01
-3.42312694e-01 -2.29218349e-01 1.15999281e+00 4.15000081e-01
9.98556763e-02 -1.18647254e+00 -5.91650426e-01 -1.33351505e-01
-2.60585934e-01 -1.31866848e+00 -1.72598109e-01 2.28642896e-01
-2.17677698e-01 -8.24473798e-01 -6.01359725e-01 -6.46805823e-01
3.88218343e-01 9.59768593e-02 9.31872904e-01 -4.81623590e-01
-8.89887661e-02 -2.06081659e-01 -3.94086838e-01 -6.81967318e-01
-1.62076688e+00 -4.25336897e-01 5.10861218e-01 2.00285181e-01
-7.19451811e-04 -3.65382403e-01 -1.81841955e-01 2.06993252e-01
-6.58690095e-01 1.05436318e-01 2.23710895e-01 7.51285374e-01
1.58835560e-01 -2.71408349e-01 5.84761202e-01 -4.80281889e-01
1.02709150e+00 -4.70983267e-01 -7.48380601e-01 1.46891668e-01
-5.11469552e-03 3.74344677e-01 9.18314636e-01 -5.61270058e-01
-1.06480241e+00 -2.50275396e-02 -2.65311897e-01 -7.63513684e-01
-5.10244608e-01 -2.11082160e-01 -1.88177347e-01 6.68458879e-01
1.26994407e+00 3.69013935e-01 1.37160756e-02 -3.43911469e-01
5.10289848e-01 1.44629323e+00 7.65576243e-01 -1.12238631e-01
6.83453977e-01 2.67639518e-01 -8.10994208e-01 -1.25000787e+00
-3.75025660e-01 1.60561413e-01 -1.18034996e-01 -2.44027346e-01
1.03899682e+00 -9.40147698e-01 -7.26335466e-01 3.20529193e-01
-1.65397823e+00 4.53580618e-02 -2.36584052e-01 6.59196615e-01
-5.07939398e-01 2.15550810e-01 -4.40105706e-01 -1.24656618e+00
-3.35783601e-01 -1.62364101e+00 1.20977998e+00 -7.12215304e-02
-5.38746178e-01 -5.11673033e-01 1.14096969e-01 6.10432565e-01
1.78578615e-01 3.91054064e-01 7.25101888e-01 -1.16253138e+00
-3.03686261e-01 1.36239499e-01 2.85797983e-01 6.83656216e-01
2.79733300e-01 5.23844600e-01 -1.89611256e+00 -5.57387136e-02
1.81608364e-01 -3.02695125e-01 8.18673670e-01 1.69448644e-01
5.64740837e-01 -7.26234138e-01 1.84489354e-01 3.39087456e-01
5.15089035e-01 5.05972862e-01 3.38131368e-01 -4.47749257e-01
4.22454149e-01 6.75627470e-01 -1.48180708e-01 1.52916968e-01
-3.03190023e-01 4.97044146e-01 4.55321699e-01 1.11992851e-01
-5.51925242e-01 -5.17588973e-01 9.30769801e-01 1.26940274e+00
2.43990183e-01 -7.57215381e-01 -6.63505912e-01 4.79231507e-01
-1.24535298e+00 -1.15763819e+00 -6.87827542e-02 1.92862225e+00
4.14052784e-01 1.88998625e-01 6.70135766e-02 7.04344690e-01
1.00758505e+00 5.40695667e-01 -5.18342912e-01 -9.26612735e-01
-3.97771299e-01 2.36312732e-01 -7.77597651e-02 6.69300616e-01
-9.19089556e-01 7.79663861e-01 6.26955700e+00 6.00640237e-01
-1.26411653e+00 1.25507236e-01 4.04338300e-01 3.46768997e-03
-5.40567160e-01 -4.07435030e-01 -2.60941416e-01 4.79759961e-01
1.32994866e+00 -1.86034039e-01 5.19261241e-01 9.04871285e-01
2.74690211e-01 4.69491214e-01 -1.15220320e+00 1.13587821e+00
1.10695347e-01 -1.44530964e+00 6.33507743e-02 -1.91251785e-01
5.71351171e-01 1.06088519e-01 5.52120805e-01 1.21259518e-01
6.95395112e-01 -1.25081193e+00 1.13192177e+00 7.06441253e-02
8.05088460e-01 -5.24563074e-01 2.56252021e-01 4.97734636e-01
-8.09632301e-01 2.46824980e-01 -1.88768730e-01 -1.00892782e-01
-1.46786526e-01 2.13377386e-01 -1.87410271e+00 1.19959086e-01
3.85909021e-01 9.04558077e-02 -3.65978628e-01 3.82429332e-01
-8.08180034e-01 1.01983345e+00 -3.20623666e-02 -2.44092315e-01
8.57595652e-02 3.02452117e-01 1.17111063e+00 1.13841605e+00
3.94944489e-01 -3.83664489e-01 -4.89502370e-01 1.40659690e+00
-2.51872651e-02 -3.08264256e-01 -8.99449348e-01 -8.53235602e-01
6.13688290e-01 9.85471547e-01 -5.88453174e-01 -3.83779794e-01
2.38416307e-02 9.87636447e-01 8.02915469e-02 5.14182091e-01
-8.62486482e-01 -4.32308972e-01 8.94468844e-01 -3.37697953e-01
3.07887077e-01 -2.92575717e-01 6.48982674e-02 -1.24133801e+00
-1.91232607e-01 -1.21321094e+00 -2.23086234e-02 -1.14455402e+00
-1.12076664e+00 1.54643905e+00 -2.77297348e-01 -1.30858040e+00
-9.28771913e-01 -6.46975756e-01 -5.50543964e-01 1.06657493e+00
-6.96563542e-01 -7.76376545e-01 -1.50574580e-01 2.32042044e-01
1.04871178e+00 -8.87172282e-01 1.23954368e+00 -2.38256410e-01
-5.30648530e-01 4.01283979e-01 6.09848276e-02 4.51245457e-01
1.92520320e-01 -1.28392291e+00 1.14023924e+00 1.20765853e+00
9.03596461e-01 3.55840713e-01 1.20419610e+00 -3.37455004e-01
-8.46209884e-01 -8.14397991e-01 5.16399264e-01 -3.19283813e-01
4.47606564e-01 -9.54817951e-01 -8.41552734e-01 5.71423411e-01
6.94921732e-01 -1.41920929e-03 8.80258143e-01 -4.11195040e-01
-6.62568748e-01 3.67622733e-01 -1.09381020e+00 6.25389874e-01
4.23542559e-01 -9.58675504e-01 -1.15133286e+00 1.99514911e-01
1.20486665e+00 -3.28328371e-01 -3.09342921e-01 1.25557423e-01
4.06275898e-01 -1.24904752e+00 8.10498655e-01 -7.72886455e-01
2.85977751e-01 -5.71822941e-01 -2.56119996e-01 -1.91925895e+00
1.33524835e-01 -9.31278348e-01 -3.18449199e-01 1.04933727e+00
7.62365937e-01 -6.29385352e-01 5.32955229e-01 -4.22023460e-02
-3.52285653e-01 -1.18948333e-01 -8.27986300e-01 -8.41338038e-01
-3.44400376e-01 -5.96668065e-01 8.21848035e-01 9.41516876e-01
-2.79732049e-01 5.21569371e-01 -4.24746990e-01 6.26349628e-01
4.20119435e-01 -2.22683307e-02 6.74738526e-01 -8.90900075e-01
-6.44210041e-01 -1.55390978e-01 -6.68194771e-01 -6.56243443e-01
3.46388727e-01 -1.00292599e+00 1.18802615e-01 -1.26020372e+00
-5.77068627e-01 2.17621326e-01 6.62538335e-02 1.12411581e-01
1.00965276e-01 8.43512639e-02 2.09909379e-01 -9.63711515e-02
1.77517384e-01 6.46213174e-01 1.01587105e+00 -4.51175660e-01
6.83995485e-02 1.74186274e-01 -5.13675690e-01 8.20595682e-01
9.32270467e-01 -6.30609035e-01 -6.75455093e-01 -5.50725162e-01
-2.37351015e-01 3.69711548e-01 5.01169026e-01 -9.66432750e-01
-1.51953369e-01 2.91006953e-01 4.40472603e-01 -5.60351372e-01
7.18875527e-01 -3.97928298e-01 2.75667310e-01 4.33863252e-01
-4.59819734e-01 -6.25195801e-02 1.18336432e-01 4.47143018e-01
-3.68511111e-01 -1.44073024e-01 8.29508662e-01 -1.72510281e-01
-4.83770579e-01 -2.81329215e-01 -7.38499284e-01 7.39906356e-02
8.31585646e-01 -1.13350525e-01 -4.87620026e-01 -6.86822116e-01
-1.06949985e+00 -4.66377079e-01 -1.45563439e-01 7.31392741e-01
9.73931909e-01 -1.04823411e+00 -1.04966962e+00 6.36690855e-01
3.48706022e-02 -4.85231400e-01 9.03788507e-02 -4.16391809e-03
-6.41481102e-01 4.38418180e-01 -2.28153527e-01 -5.46066999e-01
-1.38144040e+00 5.11775970e-01 4.56043869e-01 3.91108572e-01
-3.72749716e-01 1.11729479e+00 2.91642934e-01 -6.41767740e-01
2.15910658e-01 -5.43207109e-01 -1.68404698e-01 3.57173420e-02
5.54025829e-01 3.82978559e-01 2.61810757e-02 -1.00335193e+00
-4.13898289e-01 -3.29166651e-01 2.45182276e-01 -6.35121942e-01
1.13596201e+00 2.84146100e-01 3.45671207e-01 7.36399114e-01
1.33400285e+00 4.60544795e-01 -1.34185767e+00 2.00442359e-01
-3.30583900e-01 -3.86667550e-01 -1.12067796e-01 -8.07900190e-01
-9.21584606e-01 1.33992302e+00 6.11851037e-01 7.21289575e-01
6.92200363e-01 8.26502368e-02 6.16158068e-01 4.77670021e-02
6.22683465e-02 -6.65356576e-01 1.46787837e-01 2.58736968e-01
1.29931962e+00 -1.19453907e+00 -5.72880089e-01 -2.46697173e-01
-1.07821929e+00 1.27787066e+00 3.76223415e-01 -1.58972204e-01
1.36495426e-01 5.20924926e-01 4.70803320e-01 -4.41289954e-02
-8.10927808e-01 -4.07919548e-02 3.32095623e-01 9.30543184e-01
3.40959609e-01 5.59090316e-01 4.10829186e-01 5.95181167e-01
-9.43097472e-01 -8.36963773e-01 6.47054076e-01 1.77698150e-01
-4.04038310e-01 -6.11356974e-01 -5.92038512e-01 8.84607509e-02
-3.28143835e-01 -7.77939931e-02 -8.94212723e-01 3.60811770e-01
-4.02704589e-02 1.47988605e+00 8.68571028e-02 -7.99295664e-01
1.14728600e-01 1.27293438e-01 2.13019699e-01 -6.39241159e-01
-4.94782686e-01 1.45088499e-02 2.06663966e-01 -1.14401929e-01
-6.60046488e-02 -4.48510826e-01 -1.26197410e+00 1.54638782e-01
-2.43296608e-01 3.33494186e-01 8.74827504e-01 7.59404838e-01
2.41040647e-01 8.20014179e-01 7.31392443e-01 -6.43726289e-01
-7.23705888e-01 -1.08687520e+00 -4.02182579e-01 1.79472595e-01
1.13069534e+00 -3.59840423e-01 -8.62968028e-01 6.20272085e-02] | [15.040329933166504, 6.481487274169922] |
88ea59cc-daad-4367-9fd2-79e941ca015d | robust-classification-of-high-dimensional-1 | 2306.13985 | null | https://arxiv.org/abs/2306.13985v1 | https://arxiv.org/pdf/2306.13985v1.pdf | Robust Classification of High-Dimensional Data using Data-Adaptive Energy Distance | Classification of high-dimensional low sample size (HDLSS) data poses a challenge in a variety of real-world situations, such as gene expression studies, cancer research, and medical imaging. This article presents the development and analysis of some classifiers that are specifically designed for HDLSS data. These classifiers are free of tuning parameters and are robust, in the sense that they are devoid of any moment conditions of the underlying data distributions. It is shown that they yield perfect classification in the HDLSS asymptotic regime, under some fairly general conditions. The comparative performance of the proposed classifiers is also investigated. Our theoretical results are supported by extensive simulation studies and real data analysis, which demonstrate promising advantages of the proposed classification techniques over several widely recognized methods. | ['Subhajit Dutta', 'Sarbojit Roy', 'Aytijhya Saha', 'Jyotishka Ray Choudhury'] | 2023-06-24 | null | null | null | null | ['classification-1'] | ['methodology'] | [ 3.59549999e-01 -2.51862139e-01 -6.80515110e-01 -4.92023051e-01
-4.96963084e-01 -1.76680431e-01 5.87435603e-01 3.94855261e-01
-2.73514062e-01 9.60471392e-01 -3.84913892e-01 -3.02152187e-01
-4.90704089e-01 -4.51138794e-01 -2.53949463e-01 -1.26979184e+00
-2.38014266e-01 3.36474121e-01 -1.09670922e-01 -5.62509149e-02
3.24330211e-01 6.80952311e-01 -1.67294824e+00 -3.20131958e-01
7.01696873e-01 1.24964643e+00 -2.71349996e-01 6.20254934e-01
1.56227544e-01 8.59307274e-02 -6.40923917e-01 -8.23249146e-02
1.07601874e-01 -1.96546763e-01 -4.29426402e-01 3.53248894e-01
1.86784416e-01 3.68478864e-01 1.07909381e-01 1.08873355e+00
4.97206658e-01 8.98721814e-02 9.70933437e-01 -1.41183162e+00
-2.84139425e-01 -6.51478246e-02 -6.47698164e-01 3.59030426e-01
7.06486851e-02 -2.32947424e-01 8.09898257e-01 -8.22558105e-01
3.23229015e-01 1.06225204e+00 7.62406647e-01 3.24778259e-01
-1.07435048e+00 -6.70662642e-01 -1.76707372e-01 -6.41357675e-02
-1.67711365e+00 -3.91814262e-01 4.93852198e-01 -5.24349034e-01
4.63734716e-01 4.83527929e-01 3.04556936e-01 7.65655398e-01
5.71684062e-01 6.57298923e-01 1.42101002e+00 -6.14319384e-01
4.92451578e-01 2.64332026e-01 5.62375069e-01 6.83052361e-01
4.75966364e-01 4.29035984e-02 -8.51633251e-02 -6.01591170e-01
4.52997416e-01 -1.22072492e-02 -2.37893537e-01 -4.94760036e-01
-9.05512452e-01 8.42617989e-01 -1.03352182e-01 4.76221502e-01
-1.91268280e-01 -4.65006143e-01 4.15566981e-01 1.57407373e-01
5.54523528e-01 6.96404278e-02 -3.60841841e-01 1.91582009e-01
-6.82916820e-01 1.67381793e-01 7.34538078e-01 1.23121619e+00
2.04823628e-01 -1.03231691e-01 9.94620174e-02 9.30216491e-01
-3.97136286e-02 4.91325796e-01 7.36793339e-01 -3.63648862e-01
1.98238268e-01 3.16700846e-01 1.16055682e-01 -9.60715950e-01
-8.07561874e-01 -7.38273323e-01 -1.40674210e+00 -2.76693285e-01
4.59156334e-01 1.09438382e-01 -5.62628448e-01 1.55021286e+00
3.51162821e-01 2.45429859e-01 3.38342041e-01 5.64112186e-01
5.09918034e-01 5.97050250e-01 5.90012074e-02 -8.01681757e-01
1.14925277e+00 -2.37300485e-01 -7.88137615e-01 2.01021835e-01
5.88927209e-01 -5.36930919e-01 8.95286858e-01 5.87652326e-01
-6.61324263e-01 -4.64558572e-01 -1.00615036e+00 3.11475545e-01
-2.77072966e-01 4.19358850e-01 7.17229903e-01 9.42701578e-01
-4.82739419e-01 2.84984440e-01 -6.66060686e-01 -6.54985785e-01
4.18974221e-01 5.63911021e-01 -3.68048191e-01 8.35161656e-02
-1.05221415e+00 4.76867259e-01 2.05513723e-02 4.20530468e-01
-2.58352757e-01 -3.84025007e-01 -5.90996325e-01 -5.25972992e-03
-6.22646213e-02 -4.01334792e-01 1.13882470e+00 -5.73599100e-01
-1.17302811e+00 8.85628462e-01 -2.42846817e-01 -2.61113495e-01
4.35639769e-01 2.40298763e-01 -5.62838972e-01 -5.07582426e-02
-7.52073973e-02 -1.47034213e-01 6.59105659e-01 -6.60372138e-01
-6.77519441e-01 -5.72310448e-01 -4.16715562e-01 -7.83986077e-02
-5.28698385e-01 -1.58569068e-01 -6.67713508e-02 -6.08906925e-01
1.66627631e-01 -1.03909409e+00 -4.31049019e-01 -1.09705023e-01
-5.02435625e-01 -1.65779889e-01 8.07113707e-01 9.99317691e-02
1.05067027e+00 -2.10972071e+00 -8.54448006e-02 6.43484235e-01
-1.59847021e-01 3.56519848e-01 6.08259849e-02 4.41320568e-01
-2.12456375e-01 8.12509805e-02 -2.00147778e-01 9.18514878e-02
-2.78606772e-01 1.00544907e-01 -1.42620102e-01 9.83825803e-01
-4.65746522e-02 2.97912329e-01 -5.04902840e-01 -4.69372302e-01
1.75063744e-01 7.49601871e-02 -1.48068100e-01 1.13812268e-01
2.29948759e-01 4.04727876e-01 -7.51358271e-01 6.16929054e-01
5.64716637e-01 -3.73620659e-01 2.69611806e-01 -2.26353109e-01
6.22551180e-02 -3.57755870e-01 -1.14689219e+00 8.35251689e-01
-3.26199442e-01 4.91376817e-01 -4.25202958e-02 -1.56682527e+00
1.07956839e+00 2.17349619e-01 5.60681224e-01 -3.54589164e-01
4.91532087e-01 1.68737382e-01 2.06493497e-01 -4.26524252e-01
-2.86948355e-03 -4.35802460e-01 -3.30705345e-01 1.46272138e-01
-1.02702565e-01 8.56122598e-02 2.63184309e-01 -1.96273431e-01
8.76193702e-01 -8.08690548e-01 1.04238379e+00 -7.09532678e-01
9.51330364e-01 -2.85025746e-01 6.37055218e-01 6.97168171e-01
-2.97922373e-01 2.47671649e-01 6.36379778e-01 -1.43974528e-01
-8.08997929e-01 -7.91871667e-01 -8.19179595e-01 4.98083830e-01
2.31867969e-01 3.32593806e-02 -6.92292988e-01 -4.19844329e-01
1.52715042e-01 1.75074622e-01 -6.00029588e-01 -1.30278334e-01
-1.79188803e-01 -1.55700040e+00 5.27266204e-01 3.41888219e-01
3.07134748e-01 -2.24228472e-01 -2.59313077e-01 -1.05748273e-01
2.36479059e-01 -1.34920716e+00 -1.42153725e-02 1.64678857e-01
-1.06603825e+00 -1.41888797e+00 -7.54744232e-01 -8.78086388e-01
8.34704399e-01 1.32361948e-01 4.33438063e-01 1.51340634e-01
-5.50128400e-01 1.40010789e-01 -1.23696521e-01 -4.43078429e-01
-5.41899085e-01 1.61372676e-01 3.99549127e-01 1.43751293e-01
4.33741391e-01 -2.47023404e-01 -3.63387018e-01 6.71095848e-01
-6.82571113e-01 -2.54832596e-01 7.56494999e-01 1.27539480e+00
6.80010617e-01 5.14939129e-01 1.05180478e+00 -1.18324077e+00
7.30177760e-01 -5.85957587e-01 -7.91807115e-01 2.57725030e-01
-7.65897155e-01 -1.84910953e-01 8.47939730e-01 -3.75139982e-01
-8.54873240e-01 2.54111961e-02 -2.26580337e-01 -7.09857941e-02
-3.65968704e-01 4.72199917e-01 -3.00615430e-01 -2.22302541e-01
5.62005043e-01 1.91517338e-01 2.95795381e-01 -6.06021464e-01
-2.58213520e-01 1.18597221e+00 3.27314734e-01 -4.52687800e-01
4.32025105e-01 3.74767035e-01 6.28034234e-01 -1.36033785e+00
-8.63137126e-01 -6.53870940e-01 -5.47286391e-01 1.38946861e-01
2.04603523e-01 -5.89801669e-01 -8.51956725e-01 6.66201413e-01
-3.86212498e-01 1.49356917e-01 2.18723625e-01 5.77072322e-01
-6.41489148e-01 3.36906135e-01 -6.07884169e-01 -9.48247313e-01
-2.25655839e-01 -9.89697039e-01 8.09825599e-01 2.96322227e-01
-8.89890119e-02 -1.25732076e+00 -2.16070846e-01 6.43861741e-02
4.89825942e-02 3.46061379e-01 1.28913140e+00 -9.89662766e-01
-2.43919473e-02 -7.80462861e-01 -3.69883589e-02 3.23019713e-01
4.79497969e-01 1.03219524e-01 -8.70298445e-01 -4.79331046e-01
6.97800284e-03 -2.57319156e-02 3.89177978e-01 5.58910966e-01
1.46221900e+00 -2.25903764e-01 -7.02031076e-01 3.46225709e-01
1.30055809e+00 3.89580011e-01 2.72788197e-01 -1.34747505e-01
7.30207488e-02 4.97040272e-01 1.00002694e+00 6.59437835e-01
-1.53638333e-01 5.86174011e-01 9.43519846e-02 -3.20408821e-01
4.17724043e-01 2.13906959e-01 -1.90107673e-01 8.18895400e-01
2.35109121e-01 -5.19396126e-01 -7.20033228e-01 4.06081021e-01
-1.70293772e+00 -7.45582759e-01 -3.10033888e-01 2.35865307e+00
6.78663731e-01 1.28272735e-02 1.81780562e-01 6.47695839e-01
9.14856970e-01 -2.08714664e-01 -5.65006435e-01 -4.27285582e-01
-3.00590873e-01 2.40475729e-01 5.10235369e-01 2.25676715e-01
-1.23123980e+00 2.25834250e-01 7.30832815e+00 9.03917372e-01
-1.01907659e+00 -1.77642122e-01 9.11451519e-01 2.58420110e-01
4.03117716e-01 -4.29757655e-01 -9.09586847e-01 4.40542579e-01
9.03766632e-01 -3.64049196e-01 -3.85488331e-01 8.32107306e-01
1.53561443e-01 -3.93312454e-01 -9.84294355e-01 1.00959682e+00
-1.61174312e-02 -1.21384394e+00 -1.59180939e-01 5.62963225e-02
5.69835782e-01 -5.04947901e-01 3.00761759e-01 5.03661856e-02
-2.58436590e-01 -8.81469846e-01 1.02912739e-01 3.47004026e-01
7.64972746e-01 -9.33380723e-01 1.14687800e+00 5.69718421e-01
-7.34162509e-01 -3.51665139e-01 -4.16044503e-01 -1.05724566e-01
-2.55276710e-01 9.78281200e-01 -8.21011603e-01 5.59967339e-01
2.76803762e-01 6.04925632e-01 -4.78338718e-01 1.13959384e+00
4.41126019e-01 5.66833317e-01 -2.64510661e-01 -2.86546081e-01
-7.62413368e-02 -5.66209517e-02 2.23981231e-01 1.14820278e+00
5.23169041e-01 2.70600766e-01 5.79114780e-02 1.09754100e-01
2.58388370e-01 5.62503636e-01 -5.34195542e-01 8.65461398e-03
3.91387522e-01 1.09495282e+00 -7.41121888e-01 -2.71667689e-01
-4.45064902e-01 2.91361660e-01 -1.04751512e-01 1.97902322e-01
-4.32769567e-01 -6.39210939e-01 7.46856153e-01 8.89095366e-02
1.11176439e-01 6.37465045e-02 -3.01286578e-01 -1.08929729e+00
-1.97403103e-01 -6.99243963e-01 7.42433071e-01 -1.27284229e-01
-1.51330137e+00 3.61651123e-01 2.64975011e-01 -1.39657295e+00
-3.50247920e-01 -9.32451069e-01 -3.13798755e-01 5.18747509e-01
-1.31642187e+00 -6.61569834e-01 -1.57351926e-01 5.15163422e-01
2.19428778e-01 -5.14451802e-01 9.04084504e-01 3.92982274e-01
-8.54009986e-01 7.59209394e-01 6.99458122e-01 -5.87235726e-02
4.63132948e-01 -8.37116241e-01 -1.52510554e-01 4.43614960e-01
-1.96998209e-01 6.83407903e-01 1.00622702e+00 -3.07992905e-01
-1.36502922e+00 -9.47356462e-01 7.37797379e-01 2.63988197e-01
4.80712980e-01 -4.18413937e-01 -7.27865756e-01 1.49480864e-01
-4.00404513e-01 4.25813615e-01 1.07851577e+00 2.83202529e-01
7.10848048e-02 -3.95132929e-01 -1.34554982e+00 3.63645375e-01
5.57517707e-01 -9.12076458e-02 -2.71957964e-01 5.57050109e-01
-2.70381212e-01 -2.79244244e-01 -1.16145289e+00 6.50471091e-01
7.10252821e-01 -8.39678943e-01 9.01911736e-01 -6.74566686e-01
-2.32977241e-01 -1.68790556e-02 -1.48888215e-01 -1.11581314e+00
-2.73723472e-02 -4.62735295e-01 1.62861869e-01 1.00284410e+00
1.72637552e-01 -8.44236851e-01 7.98407912e-01 3.09950352e-01
1.46242529e-01 -9.95126247e-01 -1.06437659e+00 -1.08680153e+00
-6.90396177e-03 -1.52114004e-01 4.92331743e-01 7.96561718e-01
3.54688883e-01 2.70375907e-01 -2.49094158e-01 6.18348233e-02
6.64898634e-01 3.78784955e-01 6.94420040e-01 -1.43539977e+00
-5.20449467e-02 -2.49532104e-01 -9.41758752e-01 -5.67118287e-01
4.06975746e-01 -6.66793108e-01 2.69251522e-02 -8.64967525e-01
2.48210192e-01 -8.56778502e-01 -5.21742642e-01 9.64285284e-02
-1.18917599e-02 2.47420385e-01 -3.96675467e-01 1.18151769e-01
-2.20655054e-01 4.07645822e-01 8.85073066e-01 1.65844992e-01
4.96927015e-02 6.28987968e-01 -6.13477945e-01 8.13225746e-01
9.37357306e-01 -3.63870472e-01 -6.31948113e-01 4.75750804e-01
-2.62941301e-01 2.66682118e-01 1.00746766e-01 -9.82830167e-01
-1.49960160e-01 -3.99371535e-01 4.81225610e-01 -4.52670127e-01
3.37563545e-01 -6.98435843e-01 6.91414112e-03 5.28183699e-01
-4.36283976e-01 -1.50798187e-01 2.34543495e-02 6.87414229e-01
-2.39910379e-01 -1.88841566e-01 1.23778868e+00 3.80540937e-01
-4.55390573e-01 2.28179023e-01 -4.57880497e-01 -1.09659746e-01
1.35327494e+00 -1.62356064e-01 -2.69463927e-01 -4.19536322e-01
-6.41654015e-01 -1.05197476e-02 2.04717144e-01 7.53118396e-02
3.01463991e-01 -1.15816236e+00 -2.85501987e-01 5.37917495e-01
4.25759107e-01 -4.82767463e-01 1.65342405e-01 1.10576773e+00
-7.64662027e-02 8.98057282e-01 -2.45565951e-01 -7.68331647e-01
-1.44790089e+00 5.34541488e-01 2.00501725e-01 -5.26256599e-02
-4.21984553e-01 3.17286730e-01 3.59914184e-01 -1.59180209e-01
1.14063211e-01 -4.94042516e-01 -1.55776918e-01 -3.55889685e-02
4.68027472e-01 3.45710099e-01 3.80084991e-01 -6.80174232e-01
-3.85852963e-01 6.11631811e-01 -5.82018085e-02 3.08589995e-01
9.94783223e-01 -2.43754417e-01 4.68386197e-03 6.04729652e-01
1.31678283e+00 -2.72774309e-01 -6.14276886e-01 -2.48366997e-01
2.28824571e-01 -5.57881415e-01 -1.21383741e-01 -4.63145584e-01
-7.22496629e-01 8.84595096e-01 7.20652223e-01 3.58838469e-01
1.27316749e+00 -1.17283128e-01 4.87743944e-01 4.60084260e-01
3.90813619e-01 -1.10672975e+00 -3.06979656e-01 1.78275093e-01
5.04182637e-01 -1.12310517e+00 1.83278188e-01 -8.15087259e-01
-3.23133975e-01 1.14971864e+00 3.98851871e-01 -4.25697714e-02
9.14249003e-01 3.58998388e-01 -1.44560769e-01 1.78250313e-01
-9.68086421e-01 -5.46046793e-02 1.88493729e-01 6.32667005e-01
5.07616997e-01 -4.09907438e-02 -7.64139831e-01 8.12928379e-01
5.65606579e-02 1.64487809e-01 6.11063898e-01 8.42554450e-01
-4.20015395e-01 -9.23056304e-01 -3.86487514e-01 7.48883545e-01
-6.89969063e-01 4.30640668e-01 -4.24629860e-02 1.08585131e+00
-2.71723360e-01 7.77181804e-01 1.66570842e-01 -8.46080035e-02
1.12368174e-01 4.99070017e-03 4.57605839e-01 -3.21232140e-01
7.49621615e-02 2.47402862e-01 9.45702791e-02 -1.98216558e-01
-4.07717437e-01 -9.08334255e-01 -1.07362914e+00 -1.45683348e-01
-5.01001656e-01 4.70031530e-01 6.15777731e-01 9.85429525e-01
2.17405349e-01 6.43603727e-02 9.23178375e-01 -3.94657969e-01
-7.33502269e-01 -7.05469906e-01 -9.29767489e-01 1.74866959e-01
4.15810764e-01 -8.86287570e-01 -3.04592967e-01 -8.66471455e-02] | [8.057577133178711, 4.238565444946289] |
23731197-1bd5-47a6-8df3-31038ef347e6 | dialoguebert-a-self-supervised-learning-based | 2109.10480 | null | https://arxiv.org/abs/2109.10480v1 | https://arxiv.org/pdf/2109.10480v1.pdf | DialogueBERT: A Self-Supervised Learning based Dialogue Pre-training Encoder | With the rapid development of artificial intelligence, conversational bots have became prevalent in mainstream E-commerce platforms, which can provide convenient customer service timely. To satisfy the user, the conversational bots need to understand the user's intention, detect the user's emotion, and extract the key entities from the conversational utterances. However, understanding dialogues is regarded as a very challenging task. Different from common language understanding, utterances in dialogues appear alternately from different roles and are usually organized as hierarchical structures. To facilitate the understanding of dialogues, in this paper, we propose a novel contextual dialogue encoder (i.e. DialogueBERT) based on the popular pre-trained language model BERT. Five self-supervised learning pre-training tasks are devised for learning the particularity of dialouge utterances. Four different input embeddings are integrated to catch the relationship between utterances, including turn embedding, role embedding, token embedding and position embedding. DialogueBERT was pre-trained with 70 million dialogues in real scenario, and then fine-tuned in three different downstream dialogue understanding tasks. Experimental results show that DialogueBERT achieves exciting results with 88.63% accuracy for intent recognition, 94.25% accuracy for emotion recognition and 97.04% F1 score for named entity recognition, which outperforms several strong baselines by a large margin. | ['Meng Chen', 'Tao Guo', 'Zhenyu Zhang'] | 2021-09-22 | null | null | null | null | ['role-embedding', 'intent-recognition', 'dialogue-understanding'] | ['graphs', 'natural-language-processing', 'natural-language-processing'] | [-2.44993091e-01 3.33610058e-01 -2.30924189e-02 -5.57816625e-01
-2.18881354e-01 -6.80175245e-01 7.41434693e-01 -1.14153624e-01
-5.63757837e-01 5.97579658e-01 6.29661500e-01 -2.07276717e-01
6.38373733e-01 -6.16776705e-01 6.25656247e-02 -4.65133369e-01
8.53065923e-02 5.87086022e-01 -3.12561989e-02 -7.58042932e-01
2.61835754e-01 -5.37026115e-02 -7.80169070e-01 4.74334717e-01
8.94173622e-01 1.01498151e+00 1.47513524e-01 9.30643260e-01
-6.44482017e-01 1.17573607e+00 -1.05417395e+00 -6.68387413e-01
-2.63469458e-01 -5.49878538e-01 -1.28142595e+00 1.21516943e-01
-7.70209491e-01 -5.76615214e-01 -3.03967983e-01 7.53681004e-01
5.20073593e-01 2.53659457e-01 6.57866001e-01 -1.33847213e+00
-6.63658559e-01 6.95598006e-01 -3.09255689e-01 1.82383478e-01
5.67284644e-01 2.54060179e-01 1.28212535e+00 -6.62930787e-01
4.69279557e-01 1.58055091e+00 3.91746432e-01 9.08192456e-01
-7.78240144e-01 -6.57629132e-01 1.52873755e-01 1.74918488e-01
-6.68027520e-01 -4.22516197e-01 7.51018703e-01 -2.12573409e-01
1.33196342e+00 1.16022497e-01 4.37821418e-01 1.45150876e+00
4.87433709e-02 1.11449635e+00 8.12595546e-01 -1.88389808e-01
-1.08590715e-01 4.79963243e-01 4.49325383e-01 4.76490080e-01
-6.57972693e-01 -3.62094641e-01 -3.16474497e-01 -2.44513199e-01
4.70566034e-01 -1.84164166e-01 -1.16631217e-01 2.70092517e-01
-1.03570926e+00 1.37896216e+00 3.24905545e-01 4.36216265e-01
-3.47811610e-01 -2.10484698e-01 1.23520827e+00 4.91587669e-01
4.87010151e-01 4.17568386e-01 -6.34950340e-01 -8.03344011e-01
2.96230346e-01 1.19570382e-02 1.45446467e+00 8.39935899e-01
4.84896511e-01 2.97565181e-02 -1.85353547e-01 1.24215686e+00
3.94473493e-01 1.71600863e-01 8.58137965e-01 -6.87567353e-01
5.72127700e-01 9.62830722e-01 -1.79320090e-02 -1.08841085e+00
-4.07732308e-01 2.07065955e-01 -1.04829299e+00 -6.29339099e-01
2.41114482e-01 -7.60577142e-01 -3.07015508e-01 1.54622555e+00
4.84997839e-01 -4.98890616e-02 6.40370011e-01 7.26413250e-01
1.21493113e+00 1.18890178e+00 2.55525947e-01 -2.46405765e-01
1.57725286e+00 -1.26627469e+00 -1.07161951e+00 -1.33438751e-01
7.42686689e-01 -6.76976740e-01 1.04580140e+00 1.03386685e-01
-4.67508584e-01 -4.59757924e-01 -6.08605146e-01 -2.82811463e-01
-5.33186853e-01 5.00612371e-02 6.55111909e-01 6.43066525e-01
-4.89809275e-01 2.47723889e-02 -3.56533587e-01 -4.08129632e-01
2.18396988e-02 2.97456861e-01 -2.97221839e-01 3.98122519e-01
-1.78888714e+00 1.04627430e+00 5.28572261e-01 3.06919336e-01
-6.24512851e-01 -1.70359343e-01 -1.04026806e+00 9.35219899e-02
3.14163268e-01 -1.78303286e-01 1.54632425e+00 -8.93488765e-01
-2.17311406e+00 8.10998201e-01 -2.05031022e-01 -4.59836334e-01
2.15346724e-01 -2.38533691e-01 -5.14516175e-01 -6.60344511e-02
5.13495021e-02 4.64518428e-01 6.36893809e-01 -9.64405894e-01
-8.37370694e-01 -1.93734765e-01 3.60603333e-01 5.71722746e-01
-4.19924676e-01 3.74172837e-01 -1.79506481e-01 -1.32849842e-01
-3.65485340e-01 -7.09256530e-01 -8.16488862e-02 -7.50639856e-01
-5.86117566e-01 -9.28474426e-01 1.26699889e+00 -8.49802732e-01
1.06040609e+00 -2.14584017e+00 9.52856094e-02 -3.45984191e-01
2.85249263e-01 4.69035119e-01 -6.20696284e-02 7.45970011e-01
2.17525363e-01 6.44554570e-02 -7.62673244e-02 -2.50814229e-01
3.57382119e-01 4.16476429e-01 -4.19963628e-01 8.21772739e-02
4.22596931e-01 9.78054583e-01 -1.01840699e+00 -2.61926413e-01
2.66338825e-01 3.25348347e-01 -4.68271911e-01 8.54998767e-01
-1.79862902e-01 6.95968568e-01 -6.22242332e-01 3.09094727e-01
2.31944740e-01 -7.56535912e-03 2.53967404e-01 -2.07474098e-01
7.73302093e-02 6.15632534e-01 -6.76104426e-01 1.33060181e+00
-8.88450444e-01 7.52190769e-01 3.28168511e-01 -1.14767587e+00
1.04871845e+00 6.73404574e-01 1.42602369e-01 -5.01418471e-01
5.46819925e-01 -5.37549891e-02 3.12968791e-01 -8.60515594e-01
6.45369232e-01 -2.77457088e-01 -5.67962885e-01 6.63181424e-01
2.34027594e-01 -6.94451183e-02 -9.18260813e-02 2.62831062e-01
9.80360627e-01 -4.03923005e-01 5.96678674e-01 1.14462726e-01
8.56175840e-01 -9.01516229e-02 4.93866473e-01 2.05310807e-01
-5.77898681e-01 -8.25456679e-02 8.90358984e-01 -4.38238114e-01
-5.82407475e-01 -4.99063909e-01 2.06442207e-01 1.47692347e+00
2.76531100e-01 -3.52948993e-01 -7.18875110e-01 -1.10062277e+00
-2.57620454e-01 6.72778904e-01 -4.29418147e-01 -1.87715054e-01
-6.25505745e-01 -6.89629674e-01 7.86973417e-01 2.50899762e-01
1.02097130e+00 -1.65342593e+00 -3.06819499e-01 4.55740869e-01
-5.60184240e-01 -1.40743780e+00 -4.33935642e-01 1.62070036e-01
-2.44400918e-01 -1.11589277e+00 -3.60702336e-01 -1.00870585e+00
2.35362872e-01 1.02325074e-01 9.92263258e-01 -1.97977405e-02
-1.80013888e-02 3.66811037e-01 -6.81066394e-01 -3.77493769e-01
-7.14602411e-01 3.50058138e-01 -1.17265619e-02 1.04463570e-01
7.10492253e-01 -3.23565781e-01 -3.41875821e-01 3.72616976e-01
-6.48767054e-01 -2.52083614e-02 3.28344434e-01 1.27593052e+00
-4.55397993e-01 5.08711673e-02 9.85794663e-01 -1.07684052e+00
1.38867891e+00 -6.82731688e-01 1.17785200e-01 1.20809861e-01
-1.15587970e-03 -1.92306206e-01 9.13166821e-01 -4.88698035e-01
-1.41636670e+00 -2.69040555e-01 -5.66472113e-01 2.48361334e-01
-4.52552885e-01 4.33050811e-01 -3.63334328e-01 5.14805615e-01
2.41583422e-01 3.29186261e-01 1.17308185e-01 -3.01065832e-01
6.29447043e-01 1.48960757e+00 2.15138927e-01 -5.41760862e-01
3.00633013e-01 -6.49851933e-02 -1.03482020e+00 -1.30528188e+00
-6.32715344e-01 -8.33902955e-01 -5.47186255e-01 -1.44388914e-01
1.10335541e+00 -7.08878100e-01 -1.29544735e+00 6.46274686e-01
-1.79745853e+00 -3.00017238e-01 1.60883456e-01 2.14249417e-01
-2.84187257e-01 4.57746506e-01 -1.12580156e+00 -9.78227794e-01
-6.17619932e-01 -1.12477171e+00 8.56628895e-01 4.19283032e-01
-4.56357896e-01 -1.12868643e+00 1.05383374e-01 5.56973159e-01
4.43299979e-01 -2.05205873e-01 9.79328334e-01 -1.39152956e+00
1.64117470e-01 -9.27658528e-02 -4.07761931e-01 5.39770842e-01
4.45498675e-01 -2.52500862e-01 -1.06971133e+00 1.38094157e-01
2.72834152e-01 -7.18337417e-01 3.96657109e-01 -2.48006031e-01
6.28572643e-01 -7.07238972e-01 -1.80471435e-01 -6.42536655e-02
7.07323790e-01 7.44211674e-01 4.93215740e-01 1.86705478e-02
7.00652719e-01 1.02545369e+00 4.87665772e-01 4.88057762e-01
7.88249135e-01 4.29600775e-01 2.55561680e-01 1.29691973e-01
5.03694773e-01 -2.00077415e-01 5.66115260e-01 1.30060935e+00
2.88586497e-01 -2.91613370e-01 -7.16630220e-01 4.65042174e-01
-1.90134895e+00 -8.69240642e-01 3.69611615e-03 1.32799506e+00
1.11961687e+00 3.19810323e-02 2.80508161e-01 -2.28257567e-01
8.99307966e-01 3.81954104e-01 -7.32503057e-01 -8.85346591e-01
2.73318648e-01 -1.67445257e-01 -7.28188232e-02 6.18203461e-01
-1.13500702e+00 1.32388413e+00 5.33977890e+00 6.65386617e-01
-1.08009100e+00 2.65525430e-01 8.00742269e-01 6.43590629e-01
1.26647249e-01 -2.08027348e-01 -8.37579966e-01 4.88095492e-01
9.91820753e-01 -2.53444761e-01 3.42048973e-01 9.50717509e-01
1.98041663e-01 1.51128694e-01 -8.97363722e-01 9.72406685e-01
5.20318467e-03 -8.84443045e-01 -2.86361724e-01 -1.86417401e-01
2.58469015e-01 -8.47420543e-02 -3.42917889e-01 1.05463469e+00
6.50039971e-01 -1.04779792e+00 -8.92220810e-02 -8.03790689e-02
3.69216621e-01 -7.93871582e-01 9.96993721e-01 7.39126086e-01
-1.10736287e+00 -1.44548178e-01 -3.70542616e-01 -2.52929300e-01
3.45329195e-01 1.39048651e-01 -1.20299959e+00 3.52029681e-01
4.29971874e-01 8.58226717e-01 9.89216715e-02 1.62212804e-01
-2.94996381e-01 6.91713333e-01 -1.56315789e-01 -7.03845918e-01
5.27704954e-01 -3.62807959e-01 4.36915666e-01 1.55450952e+00
-3.68448079e-01 4.47091699e-01 2.29903221e-01 6.36354029e-01
-2.75521636e-01 3.52535725e-01 -6.95281267e-01 -4.02239799e-01
4.39251065e-01 1.50133085e+00 -4.14194405e-01 -3.13780457e-01
-2.82496780e-01 1.40416086e+00 3.78156215e-01 1.86162710e-01
-8.61257970e-01 -7.10946083e-01 8.72507453e-01 -7.35534549e-01
9.60064959e-03 -1.40619203e-01 2.37999603e-01 -1.08967960e+00
-2.67262191e-01 -1.19411421e+00 1.66786641e-01 -3.77346605e-01
-1.53836071e+00 9.30105984e-01 -4.21929061e-01 -7.68974721e-01
-6.04212165e-01 -7.17371702e-01 -1.08654141e+00 7.40025938e-01
-1.19534695e+00 -1.14677596e+00 6.91927678e-04 3.67893726e-01
1.14267695e+00 -3.50243211e-01 1.19844627e+00 3.20926942e-02
-8.96010339e-01 4.28513199e-01 -8.04651435e-03 7.88585603e-01
7.01745212e-01 -1.22964799e+00 3.92717093e-01 1.09274521e-01
-1.29884556e-01 6.33567989e-01 5.15557528e-01 -2.78147191e-01
-1.22370648e+00 -7.60620892e-01 1.10091865e+00 -3.76335561e-01
8.84036243e-01 -6.64857745e-01 -8.03374827e-01 7.04287171e-01
8.02462578e-01 -6.20205045e-01 8.96698534e-01 3.25478464e-01
-2.03675538e-01 2.05840185e-01 -1.15234673e+00 6.33465052e-01
6.04390025e-01 -8.19697618e-01 -9.73822236e-01 3.03051889e-01
1.09001374e+00 -4.08690840e-01 -8.28733683e-01 1.12529151e-01
4.09777910e-01 -8.69348407e-01 6.95654273e-01 -1.02547979e+00
4.54460740e-01 2.29719564e-01 -7.01975450e-03 -1.48475039e+00
2.08575517e-01 -8.14270496e-01 4.34586257e-02 1.62861967e+00
2.89778113e-01 -8.73597980e-01 4.86622214e-01 7.34450459e-01
1.92166958e-02 -7.17199802e-01 -7.93140829e-01 -2.24330395e-01
-4.31667678e-02 -1.92189395e-01 4.43699002e-01 1.18488061e+00
7.80177832e-01 1.37930405e+00 -4.97154772e-01 -8.39626640e-02
-1.18434355e-01 3.50032188e-02 9.52717364e-01 -9.54844594e-01
-1.31045654e-01 -3.89502823e-01 -2.67719567e-01 -1.68933868e+00
6.56839848e-01 -5.19838095e-01 2.55269974e-01 -1.38027465e+00
2.90516615e-02 -2.64655977e-01 1.43635616e-01 4.48155761e-01
-2.88735121e-01 -3.36881220e-01 1.19935330e-02 -2.18110569e-02
-6.44530416e-01 1.07275307e+00 1.05738068e+00 -3.36253941e-01
-3.28786820e-01 8.01724121e-02 -7.11327016e-01 8.20868790e-01
9.89903212e-01 -1.13256589e-01 -4.38165873e-01 -1.23849057e-01
-1.91308707e-01 2.58306801e-01 -1.39130965e-01 -2.44718403e-01
2.28443846e-01 -7.85209015e-02 -2.33015612e-01 -4.16012883e-01
5.52762270e-01 -7.28804290e-01 -8.13765824e-01 2.73317784e-01
-5.93427777e-01 -3.64537816e-03 3.57859023e-02 5.24627984e-01
-5.35634279e-01 -3.85558069e-01 4.66121137e-01 -2.29557976e-01
-9.35967505e-01 -2.10856157e-03 -7.56613016e-01 3.51449221e-01
9.48262811e-01 1.47476971e-01 -3.71809006e-01 -8.24170887e-01
-5.78199327e-01 5.78520238e-01 -2.43088901e-01 7.07156062e-01
5.40854454e-01 -1.06660259e+00 -5.90642631e-01 -1.63189378e-02
-3.35860699e-02 4.53504175e-02 2.32768059e-01 4.95690614e-01
-2.38789573e-01 4.48724747e-01 -2.04917844e-02 -3.80907536e-01
-1.25651586e+00 4.54887487e-02 3.76373172e-01 -5.58230877e-01
-4.27242041e-01 9.51698661e-01 3.37207645e-01 -1.19652116e+00
3.98795038e-01 -1.42767414e-01 -6.63641810e-01 2.97134668e-01
5.48196018e-01 8.13091770e-02 -4.28526461e-01 -8.68187010e-01
-1.70179084e-01 4.78677340e-02 -3.80279541e-01 -4.92868796e-02
1.10842836e+00 -3.67897063e-01 -3.90613168e-01 6.12977147e-01
1.43410254e+00 -8.22509155e-02 -7.58469880e-01 -3.35288078e-01
1.86614603e-01 -2.69048866e-02 -4.37999219e-01 -7.08010197e-01
-7.47203767e-01 1.09623265e+00 6.97203055e-02 8.13771367e-01
6.28393054e-01 -1.59711614e-01 1.32737720e+00 7.07480431e-01
1.82851717e-01 -1.15741849e+00 3.31484884e-01 1.22255957e+00
8.45374584e-01 -1.56191623e+00 -6.45674109e-01 -4.57343847e-01
-1.33734190e+00 1.14177680e+00 9.49734509e-01 2.94991792e-03
5.25088251e-01 5.53636178e-02 5.66606224e-01 -3.04192930e-01
-8.97741616e-01 -1.13001823e-01 -2.47908980e-01 5.64031422e-01
6.17853582e-01 1.34238065e-03 -2.80287981e-01 1.02326918e+00
-2.58166045e-01 -5.63802063e-01 3.52575779e-01 5.81368327e-01
-4.18717384e-01 -1.05104518e+00 8.54483247e-02 2.80549675e-01
-4.50301617e-01 -1.04297340e-01 -8.53322864e-01 7.13851333e-01
-2.90981621e-01 1.44453228e+00 3.14166360e-02 -6.30628288e-01
3.86694282e-01 4.17486280e-01 -3.02090198e-01 -8.81564260e-01
-9.77522850e-01 -2.00320706e-01 6.51527226e-01 -2.30987415e-01
-4.17805791e-01 -9.07622725e-02 -1.54879773e+00 -2.80352354e-01
-5.17193854e-01 7.13364422e-01 4.54696238e-01 1.20944917e+00
2.12455422e-01 3.87123734e-01 1.11607170e+00 -5.81445754e-01
-6.91375136e-01 -1.50127864e+00 -3.82898569e-01 4.20565307e-01
8.95917714e-02 -3.91456306e-01 -4.49567646e-01 -1.09848179e-01] | [12.7887544631958, 7.7094831466674805] |
15d3a513-bbf7-4d9c-9ebe-1cc556737faf | from-big-to-small-adaptive-learning-to | 2203.07375 | null | https://arxiv.org/abs/2203.07375v1 | https://arxiv.org/pdf/2203.07375v1.pdf | From Big to Small: Adaptive Learning to Partial-Set Domains | Domain adaptation targets at knowledge acquisition and dissemination from a labeled source domain to an unlabeled target domain under distribution shift. Still, the common requirement of identical class space shared across domains hinders applications of domain adaptation to partial-set domains. Recent advances show that deep pre-trained models of large scale endow rich knowledge to tackle diverse downstream tasks of small scale. Thus, there is a strong incentive to adapt models from large-scale domains to small-scale domains. This paper introduces Partial Domain Adaptation (PDA), a learning paradigm that relaxes the identical class space assumption to that the source class space subsumes the target class space. First, we present a theoretical analysis of partial domain adaptation, which uncovers the importance of estimating the transferable probability of each class and each instance across domains. Then, we propose Selective Adversarial Network (SAN and SAN++) with a bi-level selection strategy and an adversarial adaptation mechanism. The bi-level selection strategy up-weighs each class and each instance simultaneously for source supervised training, target self-training, and source-target adversarial adaptation through the transferable probability estimated alternately by the model. Experiments on standard partial-set datasets and more challenging tasks with superclasses show that SAN++ outperforms several domain adaptation methods. | ['Mingsheng Long', 'Jianmin Wang', 'Ziyang Zhang', 'Kaichao You', 'Zhangjie Cao'] | 2022-03-14 | null | null | null | null | ['partial-domain-adaptation'] | ['methodology'] | [ 5.28236985e-01 1.91775441e-01 -4.55000699e-01 -4.86734152e-01
-7.40987539e-01 -1.07402360e+00 5.48914135e-01 -1.29477680e-01
-4.10319209e-01 1.31378245e+00 -1.22724980e-01 -1.74663857e-01
-1.11786850e-01 -1.01953411e+00 -1.13773692e+00 -7.14253545e-01
1.58676594e-01 8.36998820e-01 4.91079628e-01 -2.07907602e-01
-3.24747026e-01 5.33103704e-01 -8.93842161e-01 2.27853343e-01
9.69014883e-01 7.84554422e-01 1.48514852e-01 3.30551088e-01
-2.04672500e-01 5.26001215e-01 -8.52212429e-01 -4.81172800e-01
6.34561598e-01 -4.68350619e-01 -7.17572749e-01 -2.06153542e-01
4.60158587e-01 -3.07548523e-01 -4.50255334e-01 1.07830870e+00
6.13141358e-01 2.27100089e-01 9.66310024e-01 -1.54539311e+00
-1.14349866e+00 5.18613875e-01 -4.23252165e-01 3.69327188e-01
-1.22870542e-01 2.11637110e-01 6.45888925e-01 -6.16561234e-01
7.77467787e-01 1.09359550e+00 5.41109264e-01 9.40601349e-01
-1.36172152e+00 -1.07417357e+00 4.21094328e-01 -6.56242818e-02
-1.23367798e+00 -1.47554845e-01 6.88781619e-01 -5.84039927e-01
7.64074028e-01 -3.52441162e-01 1.17865451e-01 1.58964276e+00
-3.88694145e-02 7.09429681e-01 1.05163765e+00 -1.89726546e-01
5.30983448e-01 5.98417878e-01 -2.37198435e-02 1.13448314e-01
2.41625652e-01 4.12048161e-01 -4.05619174e-01 -3.38399291e-01
8.83406758e-01 5.60343489e-02 -1.23860218e-01 -7.82631218e-01
-1.05902636e+00 8.57259989e-01 4.98717904e-01 1.24533795e-01
-3.16847503e-01 -2.95335233e-01 4.72689599e-01 7.03777730e-01
6.02912605e-01 3.89652938e-01 -1.01355314e+00 4.25226808e-01
-6.07605577e-01 3.20417553e-01 6.94988966e-01 1.48306036e+00
8.02627265e-01 7.42157847e-02 -2.80252218e-01 9.59469259e-01
-1.51845202e-01 8.75952005e-01 7.62422323e-01 -6.90354943e-01
6.26181364e-01 5.90067148e-01 8.41723233e-02 -1.96171671e-01
-1.16290234e-01 -6.06053948e-01 -8.96524906e-01 1.75806403e-01
6.54951870e-01 -4.92618799e-01 -9.23655510e-01 2.32072401e+00
5.20767033e-01 3.27739328e-01 3.35568070e-01 5.51811397e-01
3.87598693e-01 5.11228383e-01 4.10693765e-01 7.62894303e-02
9.85779166e-01 -7.84560204e-01 -2.39785209e-01 -5.20145237e-01
3.93856078e-01 -1.69423133e-01 9.64199960e-01 7.51251355e-02
-8.56639743e-01 -6.06529295e-01 -1.02171063e+00 1.57504290e-01
-7.03341186e-01 -1.73441529e-01 2.91830122e-01 5.05932093e-01
-6.87012792e-01 4.75619972e-01 -4.30059731e-01 -3.66715074e-01
9.54468131e-01 5.30363798e-01 -5.69349527e-01 -2.13269830e-01
-1.53749025e+00 7.64312744e-01 5.67078590e-01 -7.45758533e-01
-1.14720547e+00 -1.14111090e+00 -7.88241088e-01 1.82646364e-01
2.28681177e-01 -9.10354435e-01 1.17404044e+00 -1.47177017e+00
-1.57650816e+00 9.21955526e-01 1.54758440e-02 -7.59436548e-01
4.75075930e-01 -1.42014578e-01 -5.44420362e-01 -1.83994044e-02
3.43364418e-01 6.12547994e-01 1.09148693e+00 -1.12886953e+00
-8.10685754e-01 -5.31570137e-01 -6.28531054e-02 1.97891057e-01
-6.52097106e-01 -3.13188672e-01 1.23522021e-01 -8.57203007e-01
-4.28646952e-01 -8.75453711e-01 -8.35127905e-02 2.99527645e-02
-2.20162004e-01 -1.97089314e-01 7.56508231e-01 -4.31121349e-01
1.01797068e+00 -2.26616740e+00 3.53806078e-01 1.50593549e-01
9.00494754e-02 5.04396379e-01 -4.24435079e-01 1.03993677e-01
-4.22626436e-01 -1.74566314e-01 -5.19008815e-01 1.11741349e-01
6.69568628e-02 3.28955114e-01 -7.54308641e-01 2.23364219e-01
5.51875591e-01 9.97281432e-01 -1.05323708e+00 -8.97997245e-02
-1.36035308e-01 1.11258209e-01 -5.91722071e-01 3.66159230e-01
-3.44797224e-01 6.70989931e-01 -6.19343638e-01 5.18698454e-01
8.79544437e-01 -3.62998217e-01 2.06742838e-01 2.42054239e-01
6.36056185e-01 5.61818741e-02 -1.07293582e+00 1.58074677e+00
-3.16312999e-01 1.41367406e-01 2.47035902e-02 -1.35650003e+00
1.05240834e+00 2.06765935e-01 3.92621607e-01 -6.91474259e-01
-2.14017451e-01 3.07612360e-01 -1.79948043e-02 -1.36842215e-02
7.41386116e-02 -6.64421260e-01 -3.55117053e-01 3.40486109e-01
5.42265177e-01 -1.06948867e-01 -1.49787173e-01 2.46831566e-01
1.26271832e+00 3.38247418e-01 4.94957894e-01 -1.70743406e-01
4.89759326e-01 1.25095412e-01 7.05326498e-01 7.25626349e-01
-5.99769831e-01 4.02353972e-01 4.39357966e-01 -1.93002686e-01
-1.20161998e+00 -1.57321393e+00 -1.16094910e-01 1.61227047e+00
9.40380320e-02 5.05493999e-01 -6.20283782e-01 -1.38789189e+00
5.95022678e-01 7.31734097e-01 -8.13780069e-01 -6.68304801e-01
-5.40993631e-01 -5.67216933e-01 5.97041547e-01 7.61161506e-01
4.14939493e-01 -1.09812212e+00 1.85045809e-01 1.71647772e-01
1.32603437e-01 -9.91850734e-01 -5.20001948e-01 5.08419335e-01
-8.27108085e-01 -8.43811750e-01 -1.13122416e+00 -1.01795471e+00
6.34349108e-01 -4.19738665e-02 1.15614438e+00 -7.65850723e-01
1.70450658e-01 3.80110174e-01 -2.03254893e-01 -6.71053588e-01
-5.61267018e-01 4.02947336e-01 4.00537908e-01 -3.53920124e-02
7.86215901e-01 -7.90560544e-01 -3.36405784e-01 4.55711484e-01
-1.03936636e+00 -5.83870053e-01 6.16393268e-01 9.92898703e-01
6.26493394e-01 -9.08159092e-02 1.43003142e+00 -1.24865603e+00
5.21920681e-01 -1.13479614e+00 -4.57542807e-01 3.31138670e-01
-4.27149653e-01 -3.72874625e-02 9.85210955e-01 -9.76036072e-01
-1.30381405e+00 5.45655712e-02 2.31250018e-01 -4.89756793e-01
-3.40540111e-01 2.11427063e-02 -6.29839480e-01 2.80910313e-01
1.20182312e+00 4.04539675e-01 -7.09693413e-03 -2.91499674e-01
4.57329839e-01 6.52249396e-01 5.48470199e-01 -8.16212535e-01
1.14213264e+00 4.98970717e-01 -2.31286481e-01 -3.28801572e-01
-9.23202038e-01 -3.23623389e-01 -1.01137829e+00 4.55393165e-01
5.16311646e-01 -1.26452613e+00 1.21957324e-02 5.64036965e-01
-7.58100212e-01 -8.74167025e-01 -1.04234791e+00 2.00411335e-01
-6.41027272e-01 7.99382553e-02 -2.68321991e-01 -2.42394030e-01
-1.31177351e-01 -6.49882555e-01 8.39045644e-01 1.99449942e-01
-1.69522822e-01 -1.30187368e+00 1.89529926e-01 2.02135280e-01
4.86688375e-01 7.06199259e-02 9.04226243e-01 -1.41365767e+00
-3.02949160e-01 -5.83597645e-02 -1.55873090e-01 7.01196015e-01
3.39344978e-01 -7.81984866e-01 -9.50072765e-01 -4.47035313e-01
-2.89948255e-01 -5.56918025e-01 7.76914477e-01 3.53513092e-01
1.17543757e+00 -8.70415941e-02 -6.08663201e-01 6.47734821e-01
1.20226598e+00 2.26388142e-01 4.26010668e-01 4.87833977e-01
4.48810607e-01 3.45870793e-01 5.69288731e-01 3.23338985e-01
1.82563931e-01 4.17917192e-01 6.72045797e-02 -5.25019728e-02
-2.95186818e-01 -4.71898079e-01 4.82365310e-01 9.67382193e-02
5.91919065e-01 -3.55771214e-01 -7.35940456e-01 7.93142915e-01
-1.54514277e+00 -9.39165056e-01 5.67869902e-01 2.39674640e+00
1.19718337e+00 2.33911470e-01 3.19804400e-01 -4.16565359e-01
9.53284800e-01 -3.47264409e-01 -1.35997009e+00 -1.10271396e-02
-4.76664543e-01 4.15360540e-01 6.84861004e-01 3.49172026e-01
-1.14365375e+00 9.93795037e-01 6.19441509e+00 1.02423072e+00
-9.08574402e-01 2.29546919e-01 4.08580184e-01 -8.85850117e-02
-2.76165456e-01 -3.18653166e-01 -9.95228171e-01 6.59056783e-01
1.00415111e+00 -5.05262315e-01 3.45190525e-01 1.12887514e+00
-4.63204950e-01 4.50323701e-01 -1.33778095e+00 3.72091711e-01
-2.40288734e-01 -9.85879421e-01 3.19929749e-01 9.40308422e-02
1.00313985e+00 2.19101563e-01 3.53552967e-01 8.73183548e-01
9.33223665e-01 -6.85580909e-01 2.94296801e-01 1.26937643e-01
1.16204560e+00 -6.32705390e-01 5.76562285e-01 4.47571456e-01
-7.95142233e-01 -4.01251078e-01 -5.23812473e-01 2.27120236e-01
-7.48034269e-02 3.08815986e-01 -9.66965139e-01 4.65334684e-01
5.25424659e-01 9.17481840e-01 -2.41928890e-01 5.49978375e-01
-2.44113252e-01 6.58242702e-01 -2.70344555e-01 4.10156995e-01
-5.94656877e-02 2.83057261e-02 5.59151530e-01 1.16350985e+00
2.67114013e-01 6.09299690e-02 8.36915746e-02 8.76639843e-01
-4.51631784e-01 -2.66909629e-01 -8.64231408e-01 -8.37549418e-02
8.30719233e-01 7.60021210e-01 -1.32614866e-01 -6.04643345e-01
-4.15086478e-01 1.29018664e+00 6.56988800e-01 6.71429276e-01
-7.87237704e-01 -4.47194129e-01 1.00868189e+00 2.09695518e-01
5.02977788e-01 2.58842319e-01 -3.29215467e-01 -1.15208328e+00
-9.76807028e-02 -8.37133706e-01 7.91156054e-01 -4.29331005e-01
-2.10027504e+00 3.27971578e-01 6.81430250e-02 -1.32694018e+00
-1.53110638e-01 -7.19977796e-01 -4.40014273e-01 1.12296450e+00
-1.93710053e+00 -1.19942689e+00 8.54169857e-03 1.07619822e+00
4.93433744e-01 -7.54154980e-01 8.92254353e-01 3.67510825e-01
-4.26873654e-01 1.14553463e+00 7.37157226e-01 3.05285841e-01
1.29319513e+00 -1.39173162e+00 4.47617501e-01 5.56177676e-01
-4.26309049e-01 6.05488777e-01 2.74639219e-01 -7.06736803e-01
-7.78973997e-01 -1.56209278e+00 5.79059243e-01 -9.64354694e-01
7.47916937e-01 -3.44514430e-01 -1.28543222e+00 1.08388412e+00
-2.16450915e-01 3.40923220e-01 9.54881966e-01 -6.44885004e-02
-8.03733766e-01 -3.21696311e-01 -1.62331343e+00 2.00608820e-01
1.14323223e+00 -4.61208314e-01 -8.06657970e-01 3.98382425e-01
1.00596654e+00 -1.65177122e-01 -1.03706789e+00 3.31907302e-01
2.18998939e-01 -4.42429483e-01 1.27658892e+00 -1.09264481e+00
5.70919812e-01 -8.66478756e-02 -1.89537212e-01 -1.68501425e+00
-5.11293769e-01 -2.98389494e-01 -3.22219074e-01 1.32370806e+00
3.78002554e-01 -9.78077054e-01 7.95696139e-01 5.09167671e-01
-1.49297167e-03 -1.78136840e-01 -1.01734602e+00 -1.25024819e+00
1.01743841e+00 1.77270919e-01 8.02775502e-01 1.34658420e+00
-9.11550224e-02 5.34537911e-01 -9.71397758e-02 3.26505333e-01
6.23705506e-01 3.49504016e-02 8.34270060e-01 -1.36818552e+00
-3.93979818e-01 -3.15752387e-01 -1.49052083e-01 -1.12067997e+00
5.91492891e-01 -1.04343057e+00 -2.63582081e-01 -1.07112432e+00
2.10704699e-01 -6.38080478e-01 -8.01660597e-01 5.75803101e-01
-3.65407318e-01 -1.27442166e-01 -1.42383650e-01 1.74211264e-01
-4.74842519e-01 6.10028386e-01 1.20296001e+00 -2.60414809e-01
-3.36149246e-01 1.72447175e-01 -1.03058684e+00 4.33349252e-01
8.43893647e-01 -5.30980766e-01 -6.91448033e-01 -3.52736533e-01
-2.01851770e-01 -3.54057342e-01 3.10850918e-01 -7.27353036e-01
1.16629906e-01 -3.64279836e-01 6.23090446e-01 3.45539562e-02
1.27293229e-01 -9.93238866e-01 -3.48735005e-01 3.24373513e-01
-5.50511718e-01 -6.36736453e-01 4.53919590e-01 9.94230390e-01
-7.33318776e-02 2.69534439e-02 1.26852715e+00 -1.44699484e-01
-9.39773262e-01 4.95136738e-01 3.08829024e-02 5.30271292e-01
1.36411345e+00 -1.58910155e-01 -3.13491404e-01 -9.64325443e-02
-9.14327443e-01 3.28948617e-01 3.49765390e-01 5.17930329e-01
3.15662831e-01 -1.41633010e+00 -8.51391196e-01 4.18526918e-01
3.46121490e-01 2.60522157e-01 3.71988773e-01 1.05134226e-01
2.08627179e-01 2.07881063e-01 -5.45119286e-01 -4.11031514e-01
-7.39366353e-01 9.88046765e-01 4.70148981e-01 -3.47359866e-01
-2.82976121e-01 1.16968668e+00 8.66116703e-01 -1.15870786e+00
1.29934754e-02 1.57809526e-01 2.84309848e-03 -1.64607733e-01
4.98132825e-01 2.56496400e-01 -1.94954857e-01 -3.85086805e-01
-2.85513014e-01 3.08435231e-01 -2.74990529e-01 1.84189796e-01
1.27252221e+00 -1.86335891e-01 3.05969119e-01 2.16810018e-01
1.02517498e+00 -3.18219811e-01 -1.65680861e+00 -9.77139294e-01
-2.23201707e-01 -2.08551392e-01 -4.88295764e-01 -1.27368975e+00
-7.25648105e-01 8.87322128e-01 5.65081775e-01 -1.60854623e-01
1.20761049e+00 1.44601673e-01 6.99290693e-01 4.11662459e-01
2.92785943e-01 -1.16552675e+00 2.09764197e-01 6.40189290e-01
6.60068393e-01 -1.38655663e+00 -2.70937294e-01 -1.81447208e-01
-9.09497201e-01 5.91882288e-01 1.06674254e+00 -2.41974398e-01
7.94144154e-01 1.62002444e-01 -2.18924835e-01 3.12018484e-01
-5.68413019e-01 1.78517250e-03 1.16684437e-01 1.25801921e+00
2.42854822e-02 1.73259929e-01 3.81724238e-01 1.08187044e+00
1.22320652e-01 2.19065443e-01 1.11252494e-01 7.28218734e-01
-4.83599752e-01 -1.29582286e+00 -3.54145616e-01 3.12364042e-01
-2.89091498e-01 -6.48677051e-02 -4.47338730e-01 8.00583124e-01
3.24935377e-01 4.80569959e-01 4.15672585e-02 -8.65236223e-02
5.26404977e-01 4.54972863e-01 4.27553266e-01 -9.38016117e-01
-4.55683798e-01 -4.53429550e-01 -5.20882368e-01 -1.86194047e-01
-4.29377668e-02 -5.30756891e-01 -1.10851252e+00 -1.81846283e-02
-4.08471450e-02 1.19636543e-01 2.34748363e-01 8.01314354e-01
6.99112415e-01 5.03060043e-01 5.42251408e-01 -4.38709736e-01
-1.22852552e+00 -9.92827058e-01 -9.16638792e-01 7.87473679e-01
4.65540528e-01 -7.39936531e-01 -3.38356435e-01 2.90149063e-01] | [10.36670970916748, 3.1109607219696045] |
0c22cb69-253e-4ddf-981f-88ec1a9ad750 | avist-a-benchmark-for-visual-object-tracking | 2208.06888 | null | https://arxiv.org/abs/2208.06888v1 | https://arxiv.org/pdf/2208.06888v1.pdf | AVisT: A Benchmark for Visual Object Tracking in Adverse Visibility | One of the key factors behind the recent success in visual tracking is the availability of dedicated benchmarks. While being greatly benefiting to the tracking research, existing benchmarks do not pose the same difficulty as before with recent trackers achieving higher performance mainly due to (i) the introduction of more sophisticated transformers-based methods and (ii) the lack of diverse scenarios with adverse visibility such as, severe weather conditions, camouflage and imaging effects. We introduce AVisT, a dedicated benchmark for visual tracking in diverse scenarios with adverse visibility. AVisT comprises 120 challenging sequences with 80k annotated frames, spanning 18 diverse scenarios broadly grouped into five attributes with 42 object categories. The key contribution of AVisT is diverse and challenging scenarios covering severe weather conditions such as, dense fog, heavy rain and sandstorm; obstruction effects including, fire, sun glare and splashing water; adverse imaging effects such as, low-light; target effects including, small targets and distractor objects along with camouflage. We further benchmark 17 popular and recent trackers on AVisT with detailed analysis of their tracking performance across attributes, demonstrating a big room for improvement in performance. We believe that AVisT can greatly benefit the tracking community by complementing the existing benchmarks, in developing new creative tracking solutions in order to continue pushing the boundaries of the state-of-the-art. Our dataset along with the complete tracking performance evaluation is available at: https://github.com/visionml/pytracking | ['Fahad Shahbaz Khan', 'Luc van Gool', 'Salman Khan', 'Hisham Cholakkal', 'Martin Danelljan', 'Akshay Dudhane', 'Christoph Mayer', 'Daniya Najiha', 'Wafa Al Ghallabi', 'Mubashir Noman'] | 2022-08-14 | null | null | null | null | ['visual-tracking', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-1.18113130e-01 -7.67431557e-01 9.44911465e-02 2.54727975e-02
-1.90784752e-01 -1.20967913e+00 6.23262525e-01 -2.72710472e-01
-3.03916872e-01 7.93074131e-01 2.77340472e-01 -6.34849668e-02
4.70533073e-02 -2.15286955e-01 -7.51772523e-01 -7.39638805e-01
-5.08522451e-01 3.18613410e-01 9.22092021e-01 -3.28117311e-01
-1.50445640e-01 6.23066485e-01 -1.86558592e+00 -1.31447852e-01
5.70272684e-01 8.15595508e-01 2.51563728e-01 8.98976803e-01
2.83667475e-01 5.10697901e-01 -8.22601795e-01 -2.79048175e-01
6.64008915e-01 -7.05778599e-02 -4.51123295e-03 2.59533767e-02
1.21759772e+00 -3.31774950e-01 -4.42400843e-01 1.03266621e+00
6.73381090e-01 1.29256845e-01 2.11241439e-01 -1.65796340e+00
-5.11259913e-01 6.27696142e-02 -7.29613900e-01 8.09094608e-01
4.03287798e-01 9.28459585e-01 6.22363925e-01 -6.70077085e-01
7.58863986e-01 1.26277769e+00 1.16219616e+00 4.75416839e-01
-1.08867180e+00 -8.39717567e-01 3.61978143e-01 2.21264809e-01
-1.06283486e+00 -3.79065037e-01 -4.33180220e-02 -6.82720959e-01
8.24140549e-01 5.19927740e-01 9.84625876e-01 1.35203230e+00
2.64505774e-01 4.64959502e-01 1.14563608e+00 2.37180158e-01
6.77759275e-02 -4.08546180e-02 -6.04134472e-03 4.65749949e-01
7.03742146e-01 8.99227619e-01 -3.69971782e-01 -2.01025859e-01
3.20973694e-01 -3.06953322e-02 -5.45788527e-01 -4.31884050e-01
-1.45286822e+00 4.16071534e-01 6.90889537e-01 -6.24131151e-02
-5.52321225e-02 4.42261785e-01 4.97980356e-01 3.58629882e-01
2.88802803e-01 2.03547031e-01 -4.03547198e-01 -2.27121249e-01
-9.67229545e-01 6.29890561e-01 3.87435287e-01 1.06117928e+00
3.16576064e-01 5.48047423e-01 -7.44060338e-01 4.27221239e-01
5.39686322e-01 1.36557293e+00 -1.35272980e-01 -9.15359020e-01
2.84467250e-01 4.35047317e-03 6.54020607e-01 -8.80026162e-01
-4.97409880e-01 -8.24709058e-01 -2.50451744e-01 7.73247600e-01
4.90781784e-01 -3.82255167e-01 -1.25628018e+00 1.74827290e+00
5.50604463e-01 6.67741477e-01 -2.62879401e-01 1.34417617e+00
1.03705955e+00 3.60937357e-01 1.36878237e-01 -9.90919545e-02
1.56799436e+00 -9.31230664e-01 -8.66517544e-01 -3.85489106e-01
1.80308267e-01 -1.12368631e+00 7.73623943e-01 2.91964352e-01
-6.99655831e-01 -5.79397798e-01 -9.18394864e-01 5.63812137e-01
-4.10455644e-01 -1.92486152e-01 5.49800694e-01 8.35222423e-01
-1.18926942e+00 1.59912512e-01 -7.02777445e-01 -6.61981404e-01
4.89950567e-01 1.86499774e-01 -2.64428169e-01 -2.92972952e-01
-1.08987272e+00 1.14009666e+00 -5.08222468e-02 1.91256970e-01
-1.39445758e+00 -1.16976833e+00 -6.27989292e-01 -3.46788973e-01
5.01860857e-01 -6.73721492e-01 1.02033508e+00 -5.12388349e-01
-7.31404364e-01 5.38125038e-01 2.95439822e-04 -5.87243915e-01
8.23467135e-01 -5.29553890e-01 -7.16938317e-01 -2.59980530e-01
2.26082504e-01 6.86342835e-01 8.12783837e-01 -1.32600260e+00
-8.01668584e-01 -1.61043122e-01 2.41310094e-02 1.85100272e-01
2.35672995e-01 3.67812872e-01 -5.78103840e-01 -6.23319805e-01
-5.74047089e-01 -1.23313749e+00 -1.14233810e-02 4.30114597e-01
-2.55815268e-01 2.13068768e-01 1.41057014e+00 -4.11086649e-01
8.60002220e-01 -1.99976230e+00 -1.21144518e-01 -4.30842549e-01
2.86270946e-01 6.53225541e-01 -1.38548076e-01 4.48555738e-01
2.81841785e-01 -3.13027173e-01 1.38909236e-01 -2.37029746e-01
5.84518611e-02 2.34311804e-01 -4.17483032e-01 6.98662996e-01
7.93674812e-02 9.20407951e-01 -1.13662052e+00 -1.31496355e-01
6.15953922e-01 6.05259359e-01 -1.30404472e-01 -2.40391288e-02
-4.52213347e-01 7.86953151e-01 -2.36515746e-01 1.02672660e+00
1.16037786e+00 4.73083407e-02 -6.23227596e-01 -2.62810767e-01
-5.72502613e-01 -3.03426981e-01 -1.35332859e+00 1.25550294e+00
1.81972295e-01 1.02236557e+00 3.14287573e-01 -3.83128561e-02
4.82871711e-01 2.22290188e-01 4.16963577e-01 -7.84643054e-01
1.91401333e-01 6.98747858e-02 1.33074641e-01 -4.39930469e-01
4.57933366e-01 1.13000810e-01 2.27857634e-01 3.35218012e-02
-2.34761596e-01 1.26933679e-01 2.82495052e-01 3.07259500e-01
1.36277544e+00 2.99278855e-01 -2.19305262e-01 -3.32922012e-01
1.56140506e-01 3.67475122e-01 7.86179483e-01 9.40948486e-01
-7.38574862e-01 5.69023132e-01 -1.86615869e-01 -6.24457538e-01
-8.48957062e-01 -1.41625655e+00 -2.24293932e-01 1.00652158e+00
5.00082672e-01 -2.68824071e-01 -1.46184638e-01 -4.39779490e-01
5.09862423e-01 2.61756450e-01 -7.85482883e-01 7.75145888e-02
-4.79001909e-01 -7.97183454e-01 9.09771442e-01 4.82623756e-01
6.43485427e-01 -1.00776923e+00 -9.53142047e-01 1.13430200e-02
-8.57550353e-02 -1.52009726e+00 -4.64753062e-01 -1.16851076e-01
-5.51181674e-01 -1.15224755e+00 -6.81800604e-01 -6.02017455e-02
1.21022664e-01 9.07767653e-01 1.29852033e+00 2.51963049e-01
-6.75009966e-01 6.98185146e-01 -3.20905626e-01 -5.84810555e-01
-8.59791115e-02 -6.16003931e-01 2.91612029e-01 -3.47695857e-01
3.35814655e-02 -1.66019738e-01 -8.95032167e-01 8.21272135e-01
-6.39954746e-01 -2.91285604e-01 3.54709119e-01 3.86007577e-01
3.08382750e-01 -2.73046672e-01 1.17624812e-01 -4.75598544e-01
-7.00097680e-02 -6.54202998e-01 -1.05569839e+00 6.36710599e-02
-2.65547872e-01 -4.51304823e-01 1.31555811e-01 -6.39783680e-01
-1.03471553e+00 -9.62465405e-02 2.87028879e-01 -7.07167447e-01
-1.78303435e-01 -1.65262207e-01 1.78639263e-01 -5.35065889e-01
8.72880042e-01 -1.89167503e-02 -2.40865707e-01 -2.72515804e-01
2.48454571e-01 5.97669333e-02 6.37145519e-01 -3.67405444e-01
1.45857739e+00 8.66888523e-01 7.65645877e-03 -7.38896430e-01
-8.03094089e-01 -4.98501480e-01 1.55154308e-02 -6.67557657e-01
9.41886485e-01 -1.20843530e+00 -6.87908590e-01 5.93005240e-01
-7.14438200e-01 -4.32806730e-01 -5.51292337e-02 5.85800290e-01
-1.20847551e-02 4.21484441e-01 -2.99034029e-01 -8.48706067e-01
-2.43437842e-01 -1.16519487e+00 9.30549324e-01 6.61483467e-01
2.03904405e-01 -8.39496434e-01 3.26940060e-01 4.00564760e-01
9.54357147e-01 7.28169262e-01 -2.09389865e-01 -1.06834151e-01
-9.98108923e-01 1.34317890e-01 -3.00138116e-01 -1.07228821e-02
6.62138611e-02 1.90212101e-01 -9.92890179e-01 -8.00097704e-01
-4.65391874e-01 -2.45674580e-01 9.06090260e-01 5.63189745e-01
3.18345815e-01 3.87089521e-01 -6.27381086e-01 9.07948852e-01
1.48399639e+00 8.99201185e-02 5.29437661e-01 5.24361372e-01
7.79062450e-01 9.21435356e-02 1.12442517e+00 3.38839322e-01
1.65174961e-01 1.14084530e+00 1.00710452e+00 -1.64743140e-01
-7.57862389e-01 2.49130607e-01 5.99792719e-01 -1.15678839e-01
-2.68923461e-01 -3.46603721e-01 -8.32824588e-01 6.49947643e-01
-1.71493006e+00 -1.21517718e+00 -7.37724125e-01 2.36607409e+00
1.31685451e-01 3.26815173e-02 1.74444154e-01 -4.73931104e-01
7.46812522e-01 2.82633334e-01 -7.13777602e-01 2.55370319e-01
-3.81164312e-01 -1.92858145e-01 8.54307294e-01 4.00060803e-01
-1.24004173e+00 9.96390283e-01 6.15556431e+00 4.71005291e-01
-1.16370356e+00 2.67825693e-01 -3.54648978e-01 -4.08307135e-01
1.51606753e-01 6.24684105e-03 -1.31164706e+00 8.30885530e-01
6.25321627e-01 1.97353795e-01 2.53534079e-01 4.37998712e-01
3.09626669e-01 -2.88855404e-01 -4.96234000e-01 7.66716421e-01
-1.85204279e-02 -1.13365638e+00 -4.99922305e-01 1.53824776e-01
7.13439405e-01 1.01520014e+00 1.48439169e-01 3.91444653e-01
6.18336141e-01 -7.37500370e-01 9.34527159e-01 2.43549794e-01
4.93996263e-01 -8.93584639e-02 5.27036965e-01 -2.04444528e-01
-1.59892297e+00 -7.51985535e-02 -4.66728508e-01 -1.33489212e-02
3.78908098e-01 4.26631838e-01 -5.46290874e-01 5.74396491e-01
1.32594454e+00 8.22291553e-01 -8.31747949e-01 1.94239724e+00
-3.01182643e-02 5.18247366e-01 -6.44481659e-01 2.16511369e-01
3.87298316e-01 2.46569052e-01 1.24516690e+00 1.36096835e+00
2.22610921e-01 -2.69713789e-01 4.59428221e-01 4.84924555e-01
4.04498994e-01 -5.08934021e-01 -7.26461887e-01 3.42967838e-01
7.05889642e-01 1.35883760e+00 -8.29593182e-01 -9.45845246e-02
-5.43897271e-01 4.12285686e-01 -1.74135968e-01 4.33619261e-01
-1.53071630e+00 3.26778620e-01 1.18095732e+00 1.90402359e-01
7.03045905e-01 -2.55305231e-01 1.62886500e-01 -1.05781209e+00
1.14480443e-01 -8.57259154e-01 5.58030486e-01 -9.39953446e-01
-1.06466818e+00 6.27009511e-01 2.05615193e-01 -1.62468803e+00
3.88085842e-01 -4.10995901e-01 -6.26978755e-01 7.12886989e-01
-1.63900971e+00 -1.34764254e+00 -9.35351551e-01 7.11082578e-01
4.56742436e-01 -2.27199748e-01 3.12423497e-01 7.25538075e-01
-3.09471488e-01 2.91038811e-01 3.21804173e-02 -2.47007772e-01
1.12253428e+00 -1.12252581e+00 6.72439396e-01 1.16240597e+00
1.53993309e-01 2.83637285e-01 1.22941434e+00 -8.37639272e-01
-1.71480036e+00 -1.19144881e+00 7.61566386e-02 -8.72956932e-01
8.21533024e-01 -5.67924142e-01 -6.87971890e-01 7.80670345e-01
3.02295327e-01 5.76052725e-01 2.14673877e-01 -1.05072204e-02
-4.55009609e-01 -3.49495113e-02 -1.04792440e+00 4.67742532e-01
1.34186208e+00 2.65900761e-01 -3.36791426e-01 4.90278840e-01
5.19089937e-01 -8.61070752e-01 -6.35456383e-01 5.50959945e-01
6.54043972e-01 -1.12489474e+00 1.23714292e+00 -3.56515706e-01
-3.85211706e-01 -1.06974888e+00 -2.28432849e-01 -1.24635410e+00
-4.56607223e-01 -5.77068508e-01 -3.16745251e-01 1.17557561e+00
6.97050840e-02 -7.42278934e-01 6.09966218e-01 1.11341022e-01
-4.69815910e-01 -3.12195718e-01 -9.90657628e-01 -1.22588205e+00
-3.73541653e-01 -2.19737902e-01 3.73104185e-01 6.68791950e-01
-9.87588406e-01 -9.75788608e-02 -8.28417957e-01 5.41979432e-01
1.06009138e+00 3.37558351e-02 1.14607239e+00 -1.39315546e+00
-3.73601109e-01 -1.58039436e-01 -7.02223003e-01 -7.99073756e-01
-4.76957917e-01 -4.14107531e-01 1.43509164e-01 -1.46897066e+00
1.89039797e-01 -7.09453702e-01 -6.52066171e-02 4.03624326e-01
-2.52759606e-01 8.14230263e-01 7.14069545e-01 3.06224644e-01
-9.41408873e-01 3.50973845e-01 1.36168897e+00 -1.65498614e-01
2.46973634e-01 7.18097016e-02 -4.42397624e-01 3.62590343e-01
4.74397361e-01 -6.95775211e-01 -3.23943764e-01 -7.61678457e-01
1.80042088e-01 -2.86989093e-01 8.57201219e-01 -1.63332665e+00
2.48506993e-01 1.83136761e-02 6.56663477e-01 -7.92714655e-01
5.21388471e-01 -1.00045955e+00 6.31227076e-01 8.25501800e-01
5.19529760e-01 3.64033461e-01 8.39910805e-01 7.93823421e-01
1.38657123e-01 3.01323801e-01 8.42400610e-01 7.05937818e-02
-1.14355624e+00 5.12402475e-01 -1.31627917e-01 4.26110864e-01
1.19451463e+00 -4.95177925e-01 -1.08079422e+00 -1.97475821e-01
-6.52491510e-01 6.44528627e-01 6.85207248e-01 9.76113677e-01
1.78925887e-01 -1.21578300e+00 -9.75899458e-01 3.61258024e-03
2.25699574e-01 -5.77290595e-01 3.57453734e-01 1.20842016e+00
-4.49135035e-01 2.68195748e-01 -5.65851331e-01 -1.07117808e+00
-1.67546260e+00 3.71237576e-01 4.44794744e-01 1.82623509e-02
-9.47456360e-01 7.15549648e-01 3.43879938e-01 -8.47682059e-02
3.40074241e-01 -2.77973443e-01 9.07523632e-02 -1.84640944e-01
8.53971779e-01 4.13785636e-01 3.36690992e-02 -8.15103054e-01
-7.27990210e-01 7.50140131e-01 2.33582985e-02 2.10614309e-01
9.31720555e-01 -9.33786407e-02 4.86321360e-01 4.25892174e-02
3.65346670e-01 1.70886710e-01 -1.78669655e+00 -7.65428394e-02
-2.91132778e-01 -8.17542315e-01 -1.18841022e-01 -1.23990178e+00
-1.36084259e+00 4.33350921e-01 1.26562464e+00 1.47770911e-01
8.92977357e-01 1.00954017e-02 5.89142740e-01 -2.29978133e-02
5.85669696e-01 -5.88748515e-01 -1.32938504e-01 6.98953211e-01
7.27532744e-01 -1.35957503e+00 2.18501046e-01 -3.22404146e-01
-6.37043476e-01 5.09935677e-01 8.18577290e-01 -2.20147029e-01
3.74232531e-01 7.65521049e-01 5.56436121e-01 -4.70830113e-01
-7.77922809e-01 -6.39399171e-01 1.52784124e-01 1.01113689e+00
1.40027642e-01 -1.28239423e-01 -5.89501970e-02 -1.16229407e-01
8.83954167e-02 -2.00439885e-01 2.67209113e-01 8.87710154e-01
-4.18895185e-01 -6.90181255e-01 -9.18247044e-01 2.99176782e-01
-3.96778822e-01 1.98859185e-01 -2.99343854e-01 1.10063899e+00
4.87776309e-01 9.70885336e-01 -1.23779833e-01 -1.82641700e-01
6.01701319e-01 -5.55976152e-01 5.82455873e-01 -2.40316138e-01
-1.11275148e+00 -7.38182068e-02 1.90209985e-01 -8.18210363e-01
-5.10181367e-01 -9.52868700e-01 -8.80030155e-01 -4.57736760e-01
-3.58843714e-01 -8.40891004e-02 6.98749483e-01 7.49284029e-01
4.73291755e-01 7.24432290e-01 1.78112388e-02 -1.16168904e+00
-1.75898060e-01 -1.01276112e+00 -1.64001778e-01 3.65441054e-01
6.66822851e-01 -1.47935390e+00 -4.41456646e-01 -1.77921429e-01] | [6.338367938995361, -2.0704896450042725] |
6ee419b6-d68d-428f-9e4e-9513926f6702 | large-batch-optimization-for-dense-visual | 2210.11078 | null | https://arxiv.org/abs/2210.11078v1 | https://arxiv.org/pdf/2210.11078v1.pdf | Large-batch Optimization for Dense Visual Predictions | Training a large-scale deep neural network in a large-scale dataset is challenging and time-consuming. The recent breakthrough of large-batch optimization is a promising way to tackle this challenge. However, although the current advanced algorithms such as LARS and LAMB succeed in classification models, the complicated pipelines of dense visual predictions such as object detection and segmentation still suffer from the heavy performance drop in the large-batch training regime. To address this challenge, we propose a simple yet effective algorithm, named Adaptive Gradient Variance Modulator (AGVM), which can train dense visual predictors with very large batch size, enabling several benefits more appealing than prior arts. Firstly, AGVM can align the gradient variances between different modules in the dense visual predictors, such as backbone, feature pyramid network (FPN), detection, and segmentation heads. We show that training with a large batch size can fail with the gradient variances misaligned among them, which is a phenomenon primarily overlooked in previous work. Secondly, AGVM is a plug-and-play module that generalizes well to many different architectures (e.g., CNNs and Transformers) and different tasks (e.g., object detection, instance segmentation, semantic segmentation, and panoptic segmentation). It is also compatible with different optimizers (e.g., SGD and AdamW). Thirdly, a theoretical analysis of AGVM is provided. Extensive experiments on the COCO and ADE20K datasets demonstrate the superiority of AGVM. For example, it can train Faster R-CNN+ResNet50 in 4 minutes without losing performance. AGVM enables training an object detector with one billion parameters in just 3.5 hours, reducing the training time by 20.9x, whilst achieving 62.2 mAP on COCO. The deliverables are released at https://github.com/Sense-X/AGVM. | ['Ping Luo', 'Yu Liu', 'Liang Chen', 'Zhuofan Zong', 'Guanglu Song', 'Jianming Liang', 'Zeyue Xue'] | 2022-10-20 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [-1.56700656e-01 -2.68507320e-02 5.71765043e-02 -2.75900126e-01
-6.70360863e-01 -4.69438881e-01 3.98734957e-01 -4.85583305e-01
-4.51626033e-01 2.20306382e-01 -3.42557997e-01 -4.33984995e-01
3.17569405e-01 -5.82691669e-01 -1.08950663e+00 -9.14334595e-01
2.14699760e-01 4.33902651e-01 3.81608367e-01 1.10155838e-02
-7.58795664e-02 4.76908058e-01 -1.35625350e+00 1.38748184e-01
6.62037611e-01 1.31493556e+00 6.57265604e-01 7.06426203e-01
-2.67853886e-02 6.15254223e-01 -5.16416967e-01 -5.09423077e-01
4.26560491e-01 -1.12954475e-01 -6.28966630e-01 1.84270784e-01
6.26164019e-01 -2.62811273e-01 -2.80871719e-01 1.05243099e+00
7.31006086e-01 -1.45050585e-01 4.24935341e-01 -1.18103540e+00
-5.65137684e-01 5.13538420e-01 -8.64546776e-01 7.80797601e-02
-5.44247091e-01 5.27887166e-01 1.03939211e+00 -1.06763899e+00
3.53815466e-01 1.21815109e+00 8.26272786e-01 6.44394040e-01
-1.24106061e+00 -7.89944232e-01 3.06018978e-01 1.87714085e-01
-1.35409224e+00 -4.74678397e-01 4.20780420e-01 -5.24146378e-01
9.75830615e-01 3.42334896e-01 6.68178320e-01 1.09116983e+00
-4.51226346e-02 1.16742218e+00 8.73003304e-01 3.85700054e-02
7.26113096e-02 1.37136102e-01 -3.92262861e-02 7.40386009e-01
1.49897039e-01 1.51827028e-02 -3.21467370e-01 2.67537177e-01
6.87316954e-01 -3.72622944e-02 -3.70337926e-02 -7.78146461e-02
-9.91891921e-01 8.90496075e-01 8.63281071e-01 1.39257265e-02
-2.75298625e-01 3.11445922e-01 4.01694328e-01 6.94100559e-02
6.50222480e-01 3.97572160e-01 -8.07582855e-01 2.12167040e-01
-1.00925028e+00 1.87877342e-02 7.44716942e-01 9.51774597e-01
8.36347103e-01 4.24758703e-01 -2.88424283e-01 9.44702923e-01
5.52989900e-01 7.04420865e-01 4.11850542e-01 -7.96104670e-01
3.36581498e-01 4.97401953e-01 -3.04981709e-01 -8.36304367e-01
-6.53839588e-01 -6.92462385e-01 -1.22638559e+00 2.45224759e-01
3.65138769e-01 -3.45068783e-01 -1.20743358e+00 1.48281133e+00
5.03964007e-01 2.79467434e-01 -1.90132141e-01 1.09741640e+00
1.10710526e+00 7.60393381e-01 8.16083774e-02 8.77769962e-02
1.43022144e+00 -1.41787589e+00 -1.37580320e-01 -8.75717878e-01
4.55406904e-01 -7.78988659e-01 1.01907587e+00 2.42208555e-01
-9.30081069e-01 -6.78926468e-01 -8.41752708e-01 -1.90734431e-01
-2.05444381e-01 5.71660221e-01 9.78034317e-01 4.65583712e-01
-1.18926144e+00 5.03062308e-01 -1.05342662e+00 -2.75783777e-01
8.33424568e-01 5.12884438e-01 -3.91472280e-02 2.11129978e-01
-6.96769893e-01 6.21829510e-01 4.72156197e-01 4.26000506e-01
-1.13726425e+00 -8.60632956e-01 -6.06952250e-01 2.44722739e-01
4.53324199e-01 -7.18419313e-01 1.29542100e+00 -1.09516346e+00
-1.70320511e+00 1.03426611e+00 -1.01832084e-01 -6.52528524e-01
5.86914897e-01 -2.58484870e-01 3.91735435e-02 -2.00703859e-01
9.00305994e-03 9.54882562e-01 1.10843861e+00 -9.03610885e-01
-7.78543711e-01 -3.33100289e-01 -2.43064880e-01 2.45334115e-02
-3.07020217e-01 2.24715844e-01 -9.43692267e-01 -3.09352726e-01
7.79089779e-02 -9.55720842e-01 -5.02804816e-01 -1.25628412e-02
-7.79783309e-01 -1.81005403e-01 7.03499019e-01 -6.33280456e-01
9.16518152e-01 -2.28282714e+00 1.17118411e-01 -8.19120333e-02
4.41158772e-01 5.99249780e-01 -1.49850935e-01 -1.33268729e-01
1.95125025e-02 2.26955131e-01 -1.61023989e-01 -7.10033596e-01
8.56031477e-02 1.45391256e-01 -3.40222269e-01 5.97514033e-01
3.47540855e-01 1.24471152e+00 -4.22518224e-01 -4.77362543e-01
2.04298779e-01 5.21212757e-01 -6.02462351e-01 3.28479618e-01
-2.86947280e-01 5.06271064e-01 -3.54995012e-01 8.14153969e-01
7.46907771e-01 -8.00754607e-01 -2.34027300e-02 -3.26982170e-01
-2.61742502e-01 -7.92054832e-03 -1.03234923e+00 1.36831713e+00
-4.10527825e-01 9.62872446e-01 3.31242770e-01 -1.29755557e+00
7.97181845e-01 1.44650312e-02 2.72025347e-01 -6.88466072e-01
2.43109062e-01 2.51424581e-01 -8.15050378e-02 -2.89447933e-01
2.57237077e-01 1.57598689e-01 9.96501371e-02 -6.65116683e-02
3.24466199e-01 -5.27204685e-02 1.63297296e-01 4.46774960e-02
9.16898310e-01 -1.49729326e-02 8.55501592e-02 -8.40482116e-02
2.03407422e-01 2.32716817e-02 7.61030376e-01 9.12129819e-01
-1.05758227e-01 6.15633309e-01 5.92881143e-01 -5.58251321e-01
-1.14956605e+00 -7.31642663e-01 -2.23182634e-01 1.37895644e+00
7.01437294e-02 -4.50155169e-01 -6.51072681e-01 -4.66096640e-01
5.93188740e-02 2.34993890e-01 -3.90549779e-01 4.88710664e-02
-6.10103071e-01 -1.27705526e+00 5.73029578e-01 6.09444916e-01
6.90183818e-01 -1.13957965e+00 -3.86950821e-01 1.57733247e-01
1.75671160e-01 -1.27410877e+00 -1.17413737e-01 3.24304342e-01
-8.08648884e-01 -8.76816452e-01 -6.22605324e-01 -8.71112525e-01
5.82067072e-01 2.31806830e-01 1.23003101e+00 4.44557257e-02
-5.35496354e-01 -1.34814149e-02 -5.32798357e-02 -5.56757927e-01
-1.15756452e-01 2.68582523e-01 -1.70591772e-01 -1.07279956e-01
1.52927086e-01 -4.02001232e-01 -8.11634541e-01 5.15379488e-01
-6.02965057e-01 3.11053604e-01 9.27213490e-01 9.47426975e-01
9.24256086e-01 -3.60394806e-01 2.30497435e-01 -8.50361168e-01
7.43673518e-02 -4.88730162e-01 -1.10021341e+00 2.11030826e-01
-6.44428849e-01 -2.24617258e-01 7.90738404e-01 -4.73489076e-01
-8.93658698e-01 2.21168458e-01 -5.05120218e-01 -5.41936934e-01
-1.14027197e-02 2.67196953e-01 -7.93159679e-02 -2.39384428e-01
5.43602049e-01 2.77686626e-01 -1.39160797e-01 -6.64740503e-01
4.15767670e-01 4.31879044e-01 3.76651078e-01 -1.06257103e-01
8.66964161e-01 4.21328157e-01 -9.87399817e-02 -8.64094436e-01
-1.13980770e+00 -5.10265172e-01 -2.82820255e-01 2.46501714e-02
9.57447946e-01 -1.23697603e+00 -9.00501251e-01 8.30430686e-01
-1.04001987e+00 -7.42429554e-01 -1.71055794e-01 2.78641671e-01
-2.07920417e-01 7.08017722e-02 -7.19327986e-01 -4.89701122e-01
-7.37452507e-01 -1.28228283e+00 1.17578471e+00 5.84581077e-01
3.81796926e-01 -8.83180141e-01 -3.10037285e-01 3.59145850e-01
6.70425475e-01 -2.81217452e-02 3.29746753e-01 -7.62072265e-01
-7.95787811e-01 -6.83129057e-02 -5.93295515e-01 5.91008127e-01
-3.17237765e-01 3.14485699e-01 -1.14356232e+00 -4.73900169e-01
-7.69077465e-02 -5.35734951e-01 1.19536746e+00 6.47077739e-01
1.46140134e+00 -2.56658942e-01 -4.90842342e-01 1.33323383e+00
1.37287009e+00 -1.66890815e-01 4.88428921e-01 3.83141786e-01
1.12335145e+00 1.36013165e-01 4.35366482e-01 3.24097157e-01
3.11743796e-01 6.35861576e-01 6.82157218e-01 -5.20352423e-01
-2.30544969e-01 -2.09360048e-02 4.02484655e-01 7.30183899e-01
4.78080623e-02 -2.85739392e-01 -9.73032236e-01 2.89562762e-01
-1.82434130e+00 -5.94192207e-01 -2.58473128e-01 1.75572777e+00
7.52001941e-01 1.22083485e-01 -7.04210922e-02 -3.69185209e-01
6.25177920e-01 3.23101401e-01 -8.74248326e-01 -1.49599791e-01
-2.41198346e-01 1.44715935e-01 8.56784999e-01 2.20757648e-01
-1.45240438e+00 1.49559605e+00 5.45472956e+00 1.16480827e+00
-1.47230828e+00 2.76433796e-01 9.24240410e-01 -2.55062789e-01
2.74056047e-01 -1.81399211e-02 -1.38957179e+00 4.74023968e-01
8.51165295e-01 2.65992284e-01 4.09796238e-01 1.27329850e+00
2.46026684e-02 1.72427416e-01 -7.33230174e-01 1.14260566e+00
-1.16194367e-01 -1.38788033e+00 -3.54879946e-01 -9.51758400e-03
7.36043036e-01 8.65968645e-01 2.51289636e-01 6.23743534e-01
2.89602071e-01 -1.04468775e+00 7.20832169e-01 1.15002662e-01
7.35216618e-01 -4.08899307e-01 6.34312809e-01 3.75076145e-01
-9.89197731e-01 -1.69975445e-01 -9.08197403e-01 2.28101447e-01
1.85105398e-01 8.76595497e-01 -7.30856657e-01 4.94446494e-02
1.06483972e+00 7.45820820e-01 -7.86734402e-01 1.19340789e+00
-4.63978022e-01 1.02513301e+00 -5.68748593e-01 5.26300035e-02
3.74911934e-01 -2.27472618e-01 5.44568479e-01 1.26589298e+00
2.63468295e-01 -2.83451319e-01 1.05615161e-01 9.00065422e-01
-2.72666574e-01 4.87706028e-02 -1.78851604e-01 5.63293099e-02
1.91366121e-01 1.73972225e+00 -7.77643502e-01 -4.01091486e-01
-4.28462356e-01 1.00605237e+00 5.32000005e-01 3.21579367e-01
-1.19150460e+00 -3.28323729e-02 7.67845154e-01 -2.27210417e-01
7.59242952e-01 -7.55084231e-02 -2.67098665e-01 -1.22323751e+00
8.22449997e-02 -8.85375202e-01 5.90951331e-02 -5.18747985e-01
-1.21984184e+00 5.80566883e-01 -5.33828914e-01 -9.28704798e-01
9.74273309e-02 -9.66617584e-01 -7.12041199e-01 6.88009560e-01
-1.52170575e+00 -1.29966092e+00 -3.34002584e-01 3.95561934e-01
6.35882616e-01 -1.33492053e-01 4.08930749e-01 4.47367281e-01
-1.20140553e+00 6.44116104e-01 2.62835741e-01 2.35170469e-01
5.83191812e-01 -1.31609511e+00 8.57493222e-01 7.77401626e-01
3.07403445e-01 2.35287622e-01 5.45655191e-01 -3.40633959e-01
-1.49030173e+00 -1.37912381e+00 4.59493846e-01 -2.52838105e-01
8.46119881e-01 -5.46835721e-01 -9.46803868e-01 7.90577710e-01
-1.24317482e-02 4.39069360e-01 3.48259300e-01 1.69662282e-01
-1.80123210e-01 -3.24106842e-01 -6.65633857e-01 5.64013422e-01
1.00117302e+00 -2.92945296e-01 7.85614550e-02 7.60532379e-01
8.46810400e-01 -7.88691223e-01 -7.03418911e-01 3.78074557e-01
4.20032144e-01 -1.05471742e+00 1.11997187e+00 -6.04834020e-01
4.08466190e-01 -1.74099997e-01 -7.77957067e-02 -1.19968247e+00
-3.47799659e-01 -6.92060471e-01 -2.75768191e-01 1.31119668e+00
6.42956197e-01 -7.06968367e-01 7.49813259e-01 4.21628952e-01
-3.83303791e-01 -9.40095961e-01 -8.94879818e-01 -6.93766594e-01
-8.65179393e-03 -5.45762539e-01 3.36777538e-01 5.53020477e-01
-7.58216381e-01 5.15290558e-01 -5.50332546e-01 3.07296991e-01
5.47416091e-01 1.45743147e-01 9.09061015e-01 -9.84457076e-01
-5.74613631e-01 -6.50498271e-01 -3.28391582e-01 -1.42707694e+00
7.55191520e-02 -9.02157307e-01 1.70377776e-01 -1.43927443e+00
1.79807425e-01 -5.00404954e-01 -2.00799301e-01 6.12452567e-01
-4.75025803e-01 3.45752388e-01 3.75783294e-01 3.34101528e-01
-8.81102085e-01 5.46753824e-01 1.22280347e+00 -1.84329212e-01
-1.38854936e-01 3.08423281e-01 -5.80487072e-01 9.79353905e-01
1.05337656e+00 -5.61176598e-01 -4.10587676e-02 -7.77089298e-01
3.22518498e-01 -3.30394655e-01 6.43788576e-01 -8.96530926e-01
3.46045464e-01 1.49890617e-01 5.60028434e-01 -4.86716121e-01
2.26402089e-01 -5.21131396e-01 -4.61930735e-03 4.01961237e-01
4.25488055e-02 -8.91362354e-02 3.95964235e-01 3.61025095e-01
-1.80620506e-01 -7.28159249e-02 9.36631978e-01 -1.51021302e-01
-1.04700315e+00 5.12465715e-01 -1.99235156e-02 1.68100238e-01
9.18269336e-01 4.46879081e-02 -4.72756982e-01 5.85042126e-02
-6.51174188e-01 5.20980835e-01 2.08670259e-01 3.07412505e-01
4.05019194e-01 -9.43655014e-01 -7.42125273e-01 1.91790059e-01
-2.52130032e-01 4.83580709e-01 3.28397959e-01 1.17858636e+00
-4.66166496e-01 3.24316889e-01 1.83439076e-01 -9.79067147e-01
-1.13894176e+00 3.74146610e-01 3.95338297e-01 -3.42216462e-01
-8.51284027e-01 1.44980562e+00 5.76816797e-01 -4.84966427e-01
2.61958957e-01 -1.64875269e-01 -1.27387211e-01 1.47074983e-01
5.40262520e-01 1.89106107e-01 -2.60544624e-02 -4.48191315e-01
-3.88264179e-01 6.42402232e-01 -6.69733807e-02 4.35178310e-01
1.45022547e+00 2.44427323e-02 -1.77100301e-01 1.54517040e-01
1.28492677e+00 -3.11491489e-01 -1.76664031e+00 -2.77041703e-01
-1.22709751e-01 -1.94688201e-01 2.19895974e-01 -5.63524067e-01
-1.60109055e+00 9.82057512e-01 6.36357725e-01 1.71766326e-01
1.08574069e+00 3.13629538e-01 7.76961088e-01 4.50036377e-01
5.03058136e-02 -1.04889619e+00 -1.00391008e-01 6.83678329e-01
7.62305260e-01 -1.62288713e+00 4.67901714e-02 -2.10230425e-01
-7.83766806e-01 8.68985176e-01 9.07991886e-01 -1.54647693e-01
5.10239184e-01 2.85988092e-01 6.45770952e-02 -2.65932590e-01
-7.94994295e-01 -3.46949071e-01 3.99700314e-01 1.58866450e-01
2.64465094e-01 1.93083718e-01 9.38052237e-02 5.29280365e-01
-1.91078067e-01 -3.08690667e-01 1.26963392e-01 5.31887472e-01
-3.62023741e-01 -6.30548477e-01 -1.78221345e-01 5.73764086e-01
-7.12391198e-01 -4.06152487e-01 -1.76050544e-01 7.75451958e-01
1.91567183e-01 6.71935320e-01 1.72806501e-01 -3.63991171e-01
2.50989228e-01 -1.93151698e-01 1.01480044e-01 -5.80146790e-01
-7.77048409e-01 2.83004940e-01 -1.86852396e-01 -8.64332259e-01
-1.65224537e-01 -3.47189814e-01 -9.92488861e-01 -2.32019559e-01
-4.96344656e-01 -2.45387197e-01 9.69697297e-01 8.28741252e-01
3.96900326e-01 6.27298534e-01 4.97532815e-01 -1.18448031e+00
-7.53926694e-01 -9.86362040e-01 -4.75739628e-01 5.91172762e-02
2.72714347e-01 -3.30031961e-01 -5.44331729e-01 -4.08812426e-02] | [9.425628662109375, 0.07300645858049393] |
d76a42db-3749-42d5-9f47-e0caffe2718d | novel-object-viewpoint-estimation-through-1 | 2006.03586 | null | https://arxiv.org/abs/2006.03586v1 | https://arxiv.org/pdf/2006.03586v1.pdf | Novel Object Viewpoint Estimation through Reconstruction Alignment | The goal of this paper is to estimate the viewpoint for a novel object. Standard viewpoint estimation approaches generally fail on this task due to their reliance on a 3D model for alignment or large amounts of class-specific training data and their corresponding canonical pose. We overcome those limitations by learning a reconstruct and align approach. Our key insight is that although we do not have an explicit 3D model or a predefined canonical pose, we can still learn to estimate the object's shape in the viewer's frame and then use an image to provide our reference model or canonical pose. In particular, we propose learning two networks: the first maps images to a 3D geometry-aware feature bottleneck and is trained via an image-to-image translation loss; the second learns whether two instances of features are aligned. At test time, our model finds the relative transformation that best aligns the bottleneck features of our test image to a reference image. We evaluate our method on novel object viewpoint estimation by generalizing across different datasets, analyzing the impact of our different modules, and providing a qualitative analysis of the learned features to identify what representations are being learnt for alignment. | ['Mohamed El Banani', 'Jason J. Corso', 'David F. Fouhey'] | 2020-06-05 | novel-object-viewpoint-estimation-through | http://openaccess.thecvf.com/content_CVPR_2020/html/Banani_Novel_Object_Viewpoint_Estimation_Through_Reconstruction_Alignment_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Banani_Novel_Object_Viewpoint_Estimation_Through_Reconstruction_Alignment_CVPR_2020_paper.pdf | cvpr-2020-6 | ['viewpoint-estimation'] | ['computer-vision'] | [ 3.28381747e-01 2.12348744e-01 -1.67335048e-01 -4.99175966e-01
-8.42138886e-01 -8.09766173e-01 7.72254944e-01 -4.02726382e-02
-2.89818525e-01 7.65648782e-02 1.07500270e-01 6.84044361e-02
9.89659950e-02 -4.23602968e-01 -1.09350407e+00 -5.55034697e-01
3.31912428e-01 9.09080744e-01 2.25776762e-01 2.06395052e-02
6.68584228e-01 6.14877880e-01 -1.37931585e+00 1.77481994e-01
2.34914765e-01 1.05856013e+00 3.66461337e-01 7.04225481e-01
2.08940014e-01 3.20863634e-01 -4.27606046e-01 -2.50784457e-01
6.76011443e-01 -4.50987846e-01 -8.93312275e-01 4.37505275e-01
1.01491618e+00 -5.63647568e-01 -6.12769127e-02 8.27458620e-01
2.80434877e-01 4.57093716e-02 8.29167843e-01 -1.12779987e+00
-3.25321287e-01 1.38397112e-01 -4.80458349e-01 9.57101434e-02
5.14348388e-01 2.34117374e-01 1.04555416e+00 -9.98550296e-01
1.16853619e+00 1.06301129e+00 5.68174303e-01 4.17256325e-01
-1.45806015e+00 -7.82649592e-02 3.76502603e-01 2.03365967e-01
-1.35648000e+00 -7.17234194e-01 8.78800333e-01 -7.15106070e-01
8.40498984e-01 1.74429286e-02 7.86686361e-01 8.39090288e-01
2.24441379e-01 4.91266459e-01 8.86838019e-01 -5.17304361e-01
2.25507200e-01 2.72074431e-01 -3.05216074e-01 6.54554784e-01
-3.44461836e-02 6.95025027e-02 -5.90076327e-01 2.11394340e-01
9.91252005e-01 -8.15367624e-02 -3.87685895e-01 -1.19250393e+00
-1.26986670e+00 6.75031483e-01 6.15717113e-01 -6.49597635e-03
-1.43722177e-01 1.46542907e-01 2.55556451e-03 2.43642554e-01
2.77768970e-01 6.92976117e-01 -5.69007397e-01 6.25658184e-02
-6.38556242e-01 3.23707849e-01 7.09550261e-01 1.04670095e+00
1.02658010e+00 -2.54877120e-01 2.11859215e-02 2.79538393e-01
2.27019116e-01 3.10015380e-01 2.09863573e-01 -1.32645869e+00
3.56009841e-01 4.79449034e-01 1.70353606e-01 -9.42544460e-01
-3.88796359e-01 -5.63966393e-01 -7.04032332e-02 2.87702829e-01
6.61049962e-01 2.92631418e-01 -6.78510725e-01 1.87516677e+00
4.47144747e-01 -1.23423494e-01 -1.55274197e-01 1.02660894e+00
2.51620173e-01 3.10211718e-01 -4.16219711e-01 1.35135263e-01
1.12783337e+00 -8.43606770e-01 9.17729810e-02 -5.55474520e-01
6.11586869e-01 -9.05570090e-01 1.10926974e+00 2.86121368e-01
-1.18912923e+00 -4.81323689e-01 -1.02157712e+00 -3.33477110e-01
-7.78908581e-02 1.40470788e-01 1.56028524e-01 3.04712653e-01
-1.01187932e+00 7.85875082e-01 -7.32648909e-01 -4.67656493e-01
3.12968671e-01 2.29361042e-01 -4.43467706e-01 -6.52819425e-02
-5.20014822e-01 1.31602347e+00 6.85978606e-02 -1.10971607e-01
-9.90610600e-01 -6.90422595e-01 -8.65338564e-01 -2.03132210e-03
2.88186431e-01 -1.07898927e+00 1.31672573e+00 -1.40997422e+00
-1.23954523e+00 1.05602574e+00 -1.78802997e-01 -9.15793926e-02
6.86415195e-01 -5.41977063e-02 3.09981257e-01 2.63467968e-01
2.15528324e-01 8.61622095e-01 9.96905386e-01 -1.61108065e+00
-5.03753543e-01 -6.05482638e-01 4.10475016e-01 4.94211197e-01
-1.16818342e-02 -2.67672241e-01 -5.42069137e-01 -2.95733780e-01
6.19148910e-01 -1.03017294e+00 -3.60020101e-02 5.42935729e-01
-3.06522071e-01 1.05691582e-01 5.57458699e-01 -5.06337106e-01
4.96687293e-01 -2.03369546e+00 4.59931374e-01 1.94304362e-01
1.88624471e-01 -2.89467216e-01 -1.95739448e-01 3.34072739e-01
-1.48806959e-01 -7.91064277e-02 1.03748687e-01 -5.57174325e-01
-1.97816491e-01 -5.15329354e-02 -2.28368029e-01 8.01022768e-01
3.76264900e-01 6.59293234e-01 -8.91802549e-01 -2.96093196e-01
3.67138207e-01 6.14411056e-01 -9.54896390e-01 3.17704588e-01
-2.37956554e-01 7.75941372e-01 -2.09374204e-01 3.50822985e-01
5.87466776e-01 -3.13269675e-01 2.29984343e-01 -6.40913188e-01
-1.34367451e-01 4.91420358e-01 -1.14026129e+00 1.86309409e+00
-5.46777546e-01 6.93276107e-01 -1.91034168e-01 -9.57082748e-01
7.52191424e-01 -3.32393087e-02 4.43876922e-01 -5.71915209e-01
2.21930340e-01 1.28925934e-01 8.59015342e-03 -5.15702188e-01
2.24769399e-01 -1.41802147e-01 3.17394495e-01 4.83247817e-01
1.84996143e-01 -3.69346410e-01 -1.50611386e-01 8.70126635e-02
8.01378965e-01 5.27397275e-01 2.72587091e-01 -1.83742911e-01
2.15344399e-01 -1.44635841e-01 3.86794180e-01 5.54383576e-01
5.74091775e-03 1.15010417e+00 6.18681014e-01 -6.08739495e-01
-1.36107385e+00 -1.17728662e+00 -6.92110881e-02 7.57817209e-01
2.89310277e-01 -3.52773547e-01 -5.48448145e-01 -7.09525049e-01
-1.82132497e-02 7.92263567e-01 -6.65294588e-01 -2.02400222e-01
-6.05814755e-01 -4.16537635e-02 -1.93907842e-01 4.51830834e-01
1.64595246e-01 -6.68049514e-01 -9.71010566e-01 -1.34362429e-01
-2.82855541e-01 -1.09348452e+00 -7.38380849e-01 1.46676272e-01
-9.16965902e-01 -1.14019918e+00 -3.10263872e-01 -7.38380790e-01
1.14294302e+00 4.73114580e-01 1.21531224e+00 9.47026070e-03
-1.61539376e-01 6.32876933e-01 -2.05927715e-01 -1.90000638e-01
-2.73768067e-01 7.27173015e-02 -3.48232538e-02 -4.83278563e-04
1.45329144e-02 -6.56949520e-01 -8.31246614e-01 4.54568267e-01
-5.87925732e-01 2.83165604e-01 5.85734248e-01 5.86526036e-01
4.56782699e-01 -4.47345674e-01 1.37265161e-01 -6.43671691e-01
-1.32945195e-01 -1.90796956e-01 -7.24833548e-01 1.58567250e-01
-5.01897454e-01 2.78224409e-01 4.70755488e-01 -3.77007514e-01
-6.68224752e-01 4.95324045e-01 7.52124935e-02 -5.36746442e-01
2.60961838e-02 2.06366211e-01 -2.52668917e-01 -1.23901106e-01
7.34586477e-01 7.39969313e-02 1.00073911e-01 -3.87211472e-01
4.15563703e-01 4.85322252e-02 3.75032663e-01 -6.99101686e-01
1.02783382e+00 6.79520190e-01 4.74225059e-02 -4.23954427e-01
-1.07316363e+00 -3.36116761e-01 -1.18395853e+00 -2.36269057e-01
8.59955192e-01 -9.16284978e-01 -6.71391964e-01 1.90580532e-01
-1.29030097e+00 -2.13938117e-01 -4.02811974e-01 5.52406371e-01
-9.26053882e-01 1.71518236e-01 3.80302556e-02 -4.68060225e-01
1.43773615e-01 -1.42259002e+00 1.44694948e+00 -1.51603684e-01
-2.66597241e-01 -8.75536323e-01 1.93502456e-02 2.66223997e-01
3.30992669e-01 3.32997292e-02 9.97905850e-01 -4.19895858e-01
-1.07459164e+00 -1.77318528e-01 -2.57149100e-01 2.31657326e-01
6.54272512e-02 -5.99662587e-02 -1.05055749e+00 -4.36455697e-01
2.26789922e-01 -3.44527155e-01 6.99203610e-01 4.18920130e-01
1.04709876e+00 -2.98251450e-01 -2.66729802e-01 9.28999543e-01
1.47361887e+00 -6.01573065e-02 3.72201174e-01 4.42652375e-01
6.76979959e-01 7.53244281e-01 4.16969687e-01 1.82074189e-01
5.50195277e-01 1.14193940e+00 6.61677718e-01 1.66831285e-01
-1.01546019e-01 -4.89651322e-01 2.85690904e-01 5.93733072e-01
-3.74046713e-02 -8.55470728e-03 -8.59427989e-01 3.92554909e-01
-1.54017186e+00 -6.86961770e-01 3.77379000e-01 2.60510445e+00
6.82150245e-01 2.65770078e-01 3.45392860e-02 -2.36132815e-01
3.44643891e-01 3.08029409e-02 -7.53019750e-01 -1.30769283e-01
1.75166354e-01 -5.33966161e-02 4.47965890e-01 6.84197307e-01
-9.07222629e-01 7.53791034e-01 6.23336744e+00 2.70329446e-01
-1.40903425e+00 -1.74217880e-01 4.98444617e-01 -1.83172911e-01
-4.10605073e-01 5.20485401e-01 -7.40009546e-01 3.74316750e-03
4.40816879e-01 8.88682995e-03 5.10519385e-01 9.15883183e-01
7.21708965e-03 -1.25917971e-01 -1.75531375e+00 9.11720097e-01
4.66403574e-01 -1.36141431e+00 2.30239779e-01 1.44274101e-01
5.49864054e-01 -4.55995724e-02 6.47925735e-02 -1.85048804e-02
-6.55371919e-02 -9.80409384e-01 1.12749827e+00 5.58272779e-01
8.01598787e-01 -3.10788542e-01 3.29550326e-01 2.88097709e-01
-9.70630884e-01 8.25995430e-02 -2.52952933e-01 -1.65304795e-01
1.72169153e-02 5.51443659e-02 -1.10390234e+00 2.87460923e-01
4.77415591e-01 7.24877477e-01 -8.35297346e-01 9.64204431e-01
-5.86654842e-02 9.85044092e-02 -2.63308644e-01 3.22782636e-01
-9.09104422e-02 -1.27514377e-01 6.65981531e-01 5.71705520e-01
3.33365142e-01 -3.20491880e-01 2.48280719e-01 9.10111129e-01
-1.96161885e-02 -7.66903609e-02 -6.54126942e-01 2.38282591e-01
3.76076669e-01 1.19250596e+00 -7.91003466e-01 -6.15399145e-02
-3.64514232e-01 8.66550028e-01 4.25940484e-01 2.74050325e-01
-6.25984550e-01 1.30775899e-01 4.86309648e-01 5.06514668e-01
5.42652965e-01 -3.20794612e-01 -2.68122971e-01 -1.35280466e+00
1.93914086e-01 -7.86354959e-01 -4.54794280e-02 -1.24485040e+00
-9.49828625e-01 4.14076209e-01 1.42210498e-01 -1.37948978e+00
-4.36067998e-01 -6.88381732e-01 -5.21107554e-01 7.22288847e-01
-1.23334193e+00 -1.26584756e+00 -3.08707297e-01 1.93947196e-01
5.88698626e-01 6.69437200e-02 5.93974948e-01 2.89239604e-02
-2.07155749e-01 5.67179739e-01 -7.60440156e-02 -1.03769213e-01
8.95065546e-01 -1.28260875e+00 5.13282001e-01 6.18480325e-01
3.08029443e-01 6.24590695e-01 9.03966069e-01 -1.67187035e-01
-1.60486233e+00 -6.96389914e-01 8.60927403e-01 -1.21356440e+00
3.87337655e-01 -5.39263487e-01 -6.52806461e-01 1.09787357e+00
-1.08505003e-01 6.07766211e-02 1.09776154e-01 1.51409544e-02
-7.50247419e-01 -2.35668942e-01 -9.07054126e-01 6.27384186e-01
1.11476409e+00 -5.33647180e-01 -3.51854712e-01 2.43179023e-01
5.87895513e-01 -5.65177202e-01 -6.29679203e-01 1.43430576e-01
9.35111284e-01 -1.17271924e+00 1.06256533e+00 -6.72487974e-01
6.15547121e-01 -3.68725598e-01 -3.63424897e-01 -1.30158949e+00
-1.84268162e-01 -2.74081588e-01 4.07448947e-01 9.73320067e-01
2.63433754e-01 -4.46890652e-01 7.62799084e-01 5.27965724e-01
-1.06903635e-01 -9.54789758e-01 -9.21262860e-01 -5.87452650e-01
7.63309747e-02 -3.36388767e-01 4.55193400e-01 7.58038044e-01
-4.10767913e-01 6.03964448e-01 -1.23286009e-01 1.30771548e-01
4.88414556e-01 3.93196493e-01 1.22481322e+00 -1.05455625e+00
-3.22319895e-01 -5.02272367e-01 -7.55305111e-01 -1.33972764e+00
-3.85690704e-02 -9.31838989e-01 6.90558776e-02 -1.20178258e+00
3.95393312e-01 -3.20557624e-01 7.75860026e-02 2.38766700e-01
2.63403028e-01 1.18208326e-01 2.71698296e-01 3.20761979e-01
-5.16772509e-01 4.31880295e-01 1.31406295e+00 4.57181782e-02
-9.91545394e-02 -7.61697516e-02 -7.33718753e-01 8.10799301e-01
5.30438721e-01 -4.79221284e-01 -4.75300223e-01 -7.55586982e-01
5.47732294e-01 -2.01303307e-02 7.43380964e-01 -7.35502601e-01
8.44470561e-02 -2.30851144e-01 7.82165766e-01 -5.26502669e-01
5.57844758e-01 -7.82253683e-01 1.08499624e-01 2.64140397e-01
-5.89452922e-01 1.00334309e-01 -1.15384385e-01 4.39704448e-01
2.88041502e-01 -3.23052377e-01 8.12449634e-01 -1.40402168e-01
-4.62630004e-01 2.72897154e-01 1.22209139e-01 9.08078402e-02
6.98123813e-01 -5.49840808e-01 -1.61701858e-01 -4.66595143e-01
-7.04392552e-01 3.22564691e-02 1.27180624e+00 3.37824523e-01
5.53795695e-01 -1.30735624e+00 -5.49026430e-01 4.67965305e-01
3.39184940e-01 3.19362074e-01 -6.03413843e-02 8.55385244e-01
-5.32682598e-01 1.29065618e-01 -2.74054259e-01 -1.08458233e+00
-1.03622055e+00 4.36746806e-01 6.82190478e-01 1.22480407e-01
-6.17735624e-01 7.74210036e-01 6.70575976e-01 -5.80425918e-01
1.60210043e-01 -3.32315266e-01 1.87643036e-01 -1.33262962e-01
1.66039616e-01 -2.15094164e-02 1.50040999e-01 -9.13450360e-01
-3.77952993e-01 1.00980604e+00 -1.19193442e-01 -3.08488011e-01
1.27062821e+00 -2.74738133e-01 1.03712425e-01 6.39470041e-01
1.50309539e+00 5.01924753e-02 -1.77324128e+00 -3.08496267e-01
-2.21512660e-01 -9.15202975e-01 -1.70826524e-01 -5.66269755e-01
-1.00865448e+00 8.78205895e-01 6.58882380e-01 -2.87280321e-01
7.97046065e-01 3.03865671e-01 3.64253074e-01 4.83277768e-01
3.05878460e-01 -8.49382997e-01 3.83304924e-01 4.53164279e-01
1.14868820e+00 -1.29238927e+00 2.00930923e-01 -2.08678588e-01
-3.80176812e-01 1.24740243e+00 7.44552732e-01 -3.05774301e-01
5.34436703e-01 -1.76457599e-01 1.74357444e-01 -3.04618627e-01
-7.59576678e-01 1.77869834e-02 4.97858256e-01 6.62140012e-01
2.81390399e-01 -3.04331869e-01 3.40984255e-01 -9.75588858e-02
-5.52705824e-01 -4.17582065e-01 5.09660959e-01 7.98362017e-01
-3.89020801e-01 -8.78037393e-01 -7.64245316e-02 3.02269489e-01
2.00929828e-02 1.96329817e-01 -4.60190684e-01 9.04625654e-01
7.68922195e-02 3.21431577e-01 3.71810794e-01 -2.73037881e-01
3.81958693e-01 -5.39726801e-02 1.13918829e+00 -9.23580170e-01
-1.28549308e-01 9.70728993e-02 -2.75267661e-01 -6.82491958e-01
-4.29665327e-01 -1.00389767e+00 -9.26034153e-01 -3.71164419e-02
-3.99430454e-01 -2.25329995e-01 8.85795176e-01 1.03695691e+00
3.40164751e-01 1.50715530e-01 8.41873646e-01 -1.34229565e+00
-6.14795744e-01 -6.53853595e-01 -1.98665366e-01 5.45957148e-01
5.72028399e-01 -8.57178748e-01 -5.59753299e-01 2.33672038e-01] | [8.364216804504395, -2.6861379146575928] |
0a2c63c2-7e1a-404b-866f-636de3228784 | utm-a-unified-multiple-object-tracking-model | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/You_UTM_A_Unified_Multiple_Object_Tracking_Model_With_Identity-Aware_Feature_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/You_UTM_A_Unified_Multiple_Object_Tracking_Model_With_Identity-Aware_Feature_CVPR_2023_paper.pdf | UTM: A Unified Multiple Object Tracking Model With Identity-Aware Feature Enhancement | Recently, Multiple Object Tracking has achieved great success, which consists of object detection, feature embedding, and identity association. Existing methods apply the three-step or two-step paradigm to generate robust trajectories, where identity association is independent of other components. However, the independent identity association results in the identity-aware knowledge contained in the tracklet not be used to boost the detection and embedding modules. To overcome the limitations of existing methods, we introduce a novel Unified Tracking Model (UTM) to bridge those three components for generating a positive feedback loop with mutual benefits. The key insight of UTM is the Identity-Aware Feature Enhancement (IAFE), which is applied to bridge and benefit these three components by utilizing the identity-aware knowledge to boost detection and embedding. Formally, IAFE contains the Identity-Aware Boosting Attention (IABA) and the Identity-Aware Erasing Attention (IAEA), where IABA enhances the consistent regions between the current frame feature and identity-aware knowledge, and IAEA suppresses the distracted regions in the current frame feature. With better detections and embeddings, higher-quality tracklets can also be generated. Extensive experiments of public and private detections on three benchmarks demonstrate the robustness of UTM. | ['Changsheng Xu', 'Bing-Kun Bao', 'Hantao Yao', 'Sisi You'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['multiple-object-tracking'] | ['computer-vision'] | [-1.18360117e-01 -2.61301994e-01 -2.16293707e-01 -1.42044127e-01
-5.45051098e-01 -4.47059721e-01 7.50005066e-01 -1.24240488e-01
-3.37309331e-01 4.60580200e-01 4.50688511e-01 1.47855952e-01
-2.08660886e-01 -6.56140983e-01 -8.58098149e-01 -8.35580289e-01
2.01251522e-01 -1.83382019e-01 5.48508048e-01 -1.47978261e-01
-8.78199469e-04 1.72815576e-01 -1.54892898e+00 1.47809461e-01
8.72876704e-01 1.14818275e+00 1.23248592e-01 1.98213056e-01
-1.52885299e-02 7.34971225e-01 -3.30950052e-01 -2.62278646e-01
3.85121793e-01 -3.50132227e-01 -1.28266051e-01 -1.33159235e-01
3.34956437e-01 -5.52608013e-01 -6.06064379e-01 1.03685987e+00
4.52988774e-01 1.65112674e-01 4.01252717e-01 -1.80861235e+00
-9.90767896e-01 4.96043235e-01 -7.86010027e-01 2.46281505e-01
1.30715400e-01 3.20340604e-01 7.64810443e-01 -1.13871622e+00
3.60802591e-01 1.34876144e+00 6.57791197e-01 5.37909865e-01
-1.03081870e+00 -1.11537349e+00 4.86930162e-01 4.11810786e-01
-1.64460242e+00 -2.91308939e-01 7.25488842e-01 -5.42573392e-01
5.45992613e-01 9.01768580e-02 7.65632212e-01 9.08123374e-01
3.74936089e-02 1.05323005e+00 7.36963928e-01 -2.02990636e-01
-1.58503234e-01 3.55364114e-01 2.68332571e-01 6.65497601e-01
5.78737497e-01 6.06262267e-01 -5.52373469e-01 -3.51942033e-02
8.46287251e-01 3.45789611e-01 -3.64263117e-01 -5.14982939e-01
-1.44164836e+00 6.98869228e-01 7.79766738e-01 1.37150273e-01
-4.83175457e-01 1.36136502e-01 3.29593360e-01 -2.48204432e-02
2.13683963e-01 5.60731217e-02 -1.59480162e-02 9.34970900e-02
-6.08308375e-01 1.58712834e-01 1.14984110e-01 1.13169980e+00
9.56409693e-01 1.09927729e-01 -7.85432756e-01 4.16058451e-01
5.31724751e-01 7.45085299e-01 4.91265506e-01 -6.78526580e-01
2.17531964e-01 9.23226416e-01 2.12350994e-01 -1.11241484e+00
-2.19962969e-01 -6.70266330e-01 -5.80493808e-01 2.07730696e-01
2.48848274e-01 -1.45954132e-01 -7.92028844e-01 2.28011060e+00
6.57139003e-01 7.02017128e-01 -1.08185202e-01 1.01774240e+00
8.82171750e-01 4.34457809e-01 1.94998518e-01 -3.71915959e-02
1.41965508e+00 -1.12628400e+00 -9.40489531e-01 -1.74727231e-01
4.56074327e-01 -5.38020670e-01 7.02434361e-01 -3.16297024e-01
-7.47311413e-01 -9.60063994e-01 -1.20622230e+00 1.44638032e-01
-2.64024585e-01 2.43425116e-01 5.09662271e-01 4.08653438e-01
-7.91482270e-01 2.26095676e-01 -6.61799371e-01 -2.18334630e-01
5.99418998e-01 3.47780257e-01 -3.43974143e-01 -2.30328385e-02
-1.18723416e+00 8.67269695e-01 3.39974165e-01 1.65900007e-01
-7.79198825e-01 -9.11153197e-01 -8.71504962e-01 -2.91378070e-02
4.28472340e-01 -7.63748288e-01 9.11151290e-01 -9.39785182e-01
-1.12671375e+00 3.57232511e-01 -2.28364080e-01 -4.43947852e-01
4.57385302e-01 -4.06514943e-01 -5.09198725e-01 -3.31189483e-02
3.42400342e-01 7.70676672e-01 1.02432406e+00 -1.35591602e+00
-1.02757037e+00 -1.83882192e-01 -8.66264924e-02 1.69315949e-01
-5.54983020e-01 -1.62037775e-01 -7.43749678e-01 -7.94399858e-01
-1.13649787e-02 -8.64804149e-01 8.73768702e-02 2.93236464e-01
-1.90062374e-01 -1.80022761e-01 1.15742171e+00 -5.95898569e-01
1.36522520e+00 -2.41222954e+00 5.39893173e-02 2.24674679e-02
4.44268674e-01 3.34792912e-01 -2.93527663e-01 1.00158304e-01
-6.39475370e-03 -1.48140356e-01 2.94549018e-01 -2.94337481e-01
1.05743445e-01 1.23115569e-01 -4.02948707e-01 4.37305391e-01
6.16222620e-01 1.22516310e+00 -1.15673506e+00 -4.69798207e-01
3.30061644e-01 6.14991605e-01 -4.99390334e-01 4.36593033e-02
-1.52834076e-02 4.31460291e-01 -5.88067234e-01 5.90759277e-01
7.92324305e-01 -2.84732699e-01 -2.68812925e-01 -6.72024369e-01
-3.95167232e-01 -1.78011239e-01 -1.27780139e+00 1.38347793e+00
1.41859530e-02 3.86567146e-01 4.46921997e-02 -5.10285914e-01
7.31887579e-01 1.49078354e-01 5.85857332e-01 -8.44347656e-01
2.81437844e-01 2.81984210e-02 1.40391409e-01 -2.26437688e-01
6.09042645e-01 -1.83019023e-02 1.69263378e-01 2.97497720e-01
-4.69393581e-02 8.82786989e-01 1.23406976e-01 4.40140009e-01
9.40493941e-01 2.99390018e-01 9.01077613e-02 -1.49828106e-01
6.69293582e-01 -1.31319061e-01 1.10895610e+00 6.35079861e-01
-7.69519150e-01 3.80255908e-01 -7.76412264e-02 -1.76322192e-01
-8.41058314e-01 -1.14621687e+00 7.41601214e-02 1.03390908e+00
7.85140097e-01 -3.69599223e-01 -4.20565397e-01 -8.45764637e-01
2.50612587e-01 3.78244489e-01 -9.35301065e-01 -6.66693151e-01
-6.85492992e-01 -5.40559530e-01 4.22574580e-01 8.79714847e-01
6.05216146e-01 -9.88385677e-01 -4.29889500e-01 1.48675248e-01
-2.50596762e-01 -1.00285769e+00 -9.61452842e-01 -3.36958915e-01
-3.89452666e-01 -9.93900001e-01 -6.84128284e-01 -5.42172015e-01
7.32976913e-01 7.47967660e-01 6.07229233e-01 2.06602886e-01
-1.75903350e-01 4.88137692e-01 -4.43375707e-01 -4.88216460e-01
-3.35171781e-02 -3.37469637e-01 3.09052616e-01 5.73411107e-01
4.17298347e-01 -2.88659364e-01 -6.85565710e-01 3.45745295e-01
-7.48029053e-01 1.68873996e-01 8.87919724e-01 8.79612982e-01
5.56719363e-01 -2.18784139e-01 8.88223171e-01 -2.41092116e-01
2.17706099e-01 -5.69355011e-01 -4.40103352e-01 2.35695034e-01
-5.88235915e-01 -5.46638817e-02 3.09269845e-01 -8.09826553e-01
-1.03189552e+00 1.07529640e-01 -1.05844559e-02 -9.09685075e-01
1.88695014e-01 9.96891037e-02 -4.91292179e-01 -1.73453271e-01
2.80433238e-01 4.46992576e-01 2.55008973e-02 -2.24087745e-01
5.36538422e-01 3.57825875e-01 6.31109893e-01 -3.42953593e-01
1.17905307e+00 7.84206212e-01 -2.27288425e-01 -4.48599041e-01
-8.68470550e-01 -6.09889984e-01 -4.50108677e-01 -5.15885234e-01
9.41582799e-01 -1.14601517e+00 -7.49157429e-01 5.20178020e-01
-9.25178111e-01 2.60230839e-01 -5.83286464e-01 5.99994361e-01
-1.62506506e-01 4.10768509e-01 -2.86058664e-01 -8.85014772e-01
-4.70208704e-01 -1.09668112e+00 1.18397892e+00 7.16497838e-01
5.92342317e-02 -6.64734483e-01 -1.34868726e-01 -1.12381922e-02
5.21792889e-01 1.32211268e-01 5.90985715e-01 -6.15304649e-01
-9.07055259e-01 -2.96295106e-01 -5.13741851e-01 1.20404892e-01
3.01305354e-01 -1.53661281e-01 -9.63334382e-01 -5.42264998e-01
-2.37893179e-01 -4.28928062e-02 8.21472108e-01 3.59945923e-01
5.84474266e-01 -1.11926615e-01 -7.16513932e-01 5.38195372e-01
1.29909110e+00 3.05976361e-01 3.96967232e-01 4.17875022e-01
8.79418135e-01 1.51569694e-01 8.63755643e-01 2.63742268e-01
5.51714659e-01 9.29063618e-01 4.57876623e-01 -2.81811863e-01
-3.22868109e-01 -4.94075269e-01 5.98351538e-01 6.88504100e-01
1.70127660e-01 1.53346375e-01 -3.88826072e-01 6.99637651e-01
-2.19085193e+00 -1.24493766e+00 2.00891439e-02 2.30590820e+00
6.20146990e-01 1.34416029e-01 2.92770267e-01 -1.97525024e-02
9.82068956e-01 -1.05116582e-02 -7.93299139e-01 5.05808532e-01
-1.99841157e-01 -3.24509293e-01 2.84197956e-01 2.42680296e-01
-1.20559669e+00 7.88609803e-01 5.28533077e+00 8.64634097e-01
-1.00354767e+00 2.79416323e-01 -3.90619002e-02 -1.50696725e-01
-1.91727161e-01 3.68950143e-02 -1.15013254e+00 6.55927420e-01
4.77263421e-01 -3.80478501e-01 4.84800451e-02 8.80172253e-01
1.28963217e-01 2.68908799e-01 -9.75015700e-01 8.68692219e-01
-8.39933604e-02 -1.35226214e+00 2.69430101e-01 5.21407761e-02
4.83368427e-01 -1.33861735e-01 9.45056751e-02 7.40387976e-01
2.42288783e-01 -3.78106892e-01 1.11387324e+00 6.49116039e-01
5.66713870e-01 -6.81093872e-01 5.84116876e-01 2.56691337e-01
-1.88105273e+00 -3.48357618e-01 -2.42586806e-01 2.05599412e-01
3.80595297e-01 2.09759802e-01 -3.65987271e-01 1.06368279e+00
6.65560424e-01 9.41449940e-01 -6.32061660e-01 1.30429971e+00
2.44967621e-02 2.77209610e-01 -2.24363282e-01 1.77401021e-01
1.82277158e-01 -2.96379775e-01 7.98837662e-01 1.06569552e+00
1.72192976e-01 -3.09518222e-02 4.08573925e-01 1.08257937e+00
1.36133015e-01 -3.19350101e-02 -3.79743695e-01 8.04521665e-02
7.77856290e-01 1.36376810e+00 -4.51015353e-01 -4.87112343e-01
-6.88190937e-01 7.88891733e-01 2.21511379e-01 2.78184950e-01
-1.24136782e+00 -3.18132311e-01 1.00695193e+00 7.33244792e-02
6.10242784e-01 1.08121686e-01 -3.52899209e-02 -1.01834440e+00
9.33802202e-02 -6.88777208e-01 4.52423096e-01 -5.81361294e-01
-1.24653375e+00 4.19772834e-01 -1.38611272e-01 -1.64207554e+00
3.71316016e-01 -2.99318939e-01 -7.00590670e-01 9.12154734e-01
-1.56183481e+00 -1.58128774e+00 -4.87772167e-01 5.80530345e-01
3.83591980e-01 -5.82679175e-03 3.32520545e-01 8.38119566e-01
-9.79457557e-01 8.46808553e-01 -7.42856339e-02 2.46719152e-01
7.41744399e-01 -8.44381154e-01 1.24811649e-01 1.20941389e+00
-4.66283038e-02 8.84878635e-01 3.69456142e-01 -9.46276724e-01
-1.48258758e+00 -1.62314534e+00 4.04719859e-01 -5.90819955e-01
5.36019087e-01 -2.54876345e-01 -1.06210816e+00 7.64608562e-01
-4.30584401e-02 1.53955832e-01 4.49851245e-01 -2.61497557e-01
-5.50998926e-01 -2.04249039e-01 -7.35345304e-01 5.58316886e-01
1.08022952e+00 -3.93072814e-01 -5.67587674e-01 -8.17606673e-02
9.36703384e-01 -2.24533200e-01 -7.02962935e-01 6.66847169e-01
7.01620817e-01 -7.10763991e-01 1.13794684e+00 -4.68560278e-01
-1.50684081e-02 -8.72410059e-01 -3.52277458e-02 -8.89325023e-01
-7.41000891e-01 -4.13143098e-01 -4.94431287e-01 1.61676538e+00
-1.58561568e-03 -5.56158304e-01 5.39511502e-01 3.25988799e-01
-3.24438006e-01 -6.47409737e-01 -8.63412082e-01 -1.08440077e+00
-1.09815754e-01 -1.56788155e-01 8.83181989e-01 9.02701437e-01
-2.56516546e-01 4.47398275e-01 -6.67514443e-01 4.45267767e-01
6.54437661e-01 3.39597762e-01 8.65360796e-01 -1.06520522e+00
-9.60715190e-02 -5.64991474e-01 -4.37549651e-01 -1.19227743e+00
-1.62120387e-01 -9.52211678e-01 1.24567620e-01 -1.49992418e+00
4.99031872e-01 -5.17124653e-01 -6.13628268e-01 5.01801312e-01
-7.03346193e-01 1.27449378e-01 4.82149154e-01 3.49631637e-01
-8.80244613e-01 9.69575286e-01 1.16756582e+00 -8.74996185e-02
-8.75459090e-02 -1.81555614e-01 -9.17563498e-01 5.76579571e-01
2.92625785e-01 -5.02535522e-01 -3.78176302e-01 -2.37691551e-01
-2.55348027e-01 -3.42181087e-01 6.57199502e-01 -1.11133063e+00
4.96160686e-01 9.44368914e-02 5.38098991e-01 -9.75407183e-01
4.19341862e-01 -8.79361212e-01 2.81366140e-01 7.04391837e-01
9.47412848e-02 7.13667944e-02 3.06043625e-01 9.67656434e-01
-2.09169328e-01 2.92912453e-01 8.03713560e-01 2.77935803e-01
-1.15379131e+00 5.85756242e-01 2.11599246e-02 -7.44920000e-02
1.34058928e+00 -3.89714509e-01 -4.75869447e-01 -2.10524604e-01
-3.07390034e-01 6.10547483e-01 4.01788801e-01 8.91126037e-01
6.34381473e-01 -1.97403646e+00 -6.94763958e-01 4.87513810e-01
3.54532003e-01 -3.34564775e-01 3.93225819e-01 1.13131249e+00
2.98173904e-01 2.72237986e-01 -2.21057102e-01 -7.39616871e-01
-1.12187719e+00 9.40012872e-01 2.85749257e-01 -1.22403450e-01
-7.02754080e-01 7.60761321e-01 7.44378924e-01 -4.88415547e-02
2.99198002e-01 1.30516449e-02 -2.64220655e-01 5.86691406e-03
8.60321045e-01 3.63218248e-01 -3.40820730e-01 -9.93887603e-01
-4.88131076e-01 7.20675826e-01 -2.28738785e-01 1.81873277e-01
8.35115194e-01 -3.14594924e-01 2.54290909e-01 1.97448820e-01
7.84392655e-01 6.04332350e-02 -1.58607805e+00 -5.34008205e-01
-8.96359906e-02 -5.99626720e-01 2.98973601e-02 -6.66460216e-01
-1.14799225e+00 6.27423465e-01 7.94336915e-01 -1.09140910e-01
1.18608034e+00 -1.55189916e-01 9.41619396e-01 -1.24162503e-01
4.02677655e-01 -7.03662217e-01 1.98359981e-01 3.18473190e-01
7.48710334e-01 -1.30335486e+00 -2.12909639e-01 -4.73441511e-01
-6.51763082e-01 4.96175021e-01 1.08864188e+00 7.06104785e-02
6.09937608e-01 3.05160880e-01 -4.48736958e-02 -8.69377479e-02
-6.49720371e-01 -5.35685658e-01 5.11669695e-01 6.95997357e-01
-3.70709710e-02 -2.96643347e-01 -1.81226388e-01 1.04112422e+00
4.59132046e-01 -1.05984151e-01 -9.18808281e-02 9.89646792e-01
-4.97737557e-01 -8.98372054e-01 -4.68834579e-01 3.48900199e-01
-8.32613185e-02 3.14580873e-02 -2.40762427e-01 9.36780870e-01
4.74813104e-01 8.02138746e-01 -3.17652598e-02 -7.97279000e-01
5.21391571e-01 -1.16984770e-01 2.72349209e-01 -2.46750876e-01
-5.60633779e-01 1.86363921e-01 -2.93934673e-01 -6.15840316e-01
-3.42190176e-01 -6.46916926e-01 -1.39708316e+00 -1.56798080e-01
-7.50005066e-01 2.71187481e-02 1.97709128e-01 8.85712922e-01
7.06681728e-01 7.69419909e-01 4.53160256e-01 -7.95971334e-01
-5.23429394e-01 -8.55819523e-01 -2.87236065e-01 6.41906321e-01
5.15209198e-01 -1.25075412e+00 -2.08235800e-01 -2.58568162e-03] | [6.31104040145874, -2.1352765560150146] |
d736397b-0cbc-4b5f-83f9-48411da6fe62 | carigans-unpaired-photo-to-caricature | 1811.00222 | null | http://arxiv.org/abs/1811.00222v2 | http://arxiv.org/pdf/1811.00222v2.pdf | CariGANs: Unpaired Photo-to-Caricature Translation | Facial caricature is an art form of drawing faces in an exaggerated way to
convey humor or sarcasm. In this paper, we propose the first Generative
Adversarial Network (GAN) for unpaired photo-to-caricature translation, which
we call "CariGANs". It explicitly models geometric exaggeration and appearance
stylization using two components: CariGeoGAN, which only models the
geometry-to-geometry transformation from face photos to caricatures, and
CariStyGAN, which transfers the style appearance from caricatures to face
photos without any geometry deformation. In this way, a difficult cross-domain
translation problem is decoupled into two easier tasks. The perceptual study
shows that caricatures generated by our CariGANs are closer to the hand-drawn
ones, and at the same time better persevere the identity, compared to
state-of-the-art methods. Moreover, our CariGANs allow users to control the
shape exaggeration degree and change the color/texture style by tuning the
parameters or giving an example caricature. | ['Kaidi Cao', 'Jing Liao', 'Lu Yuan'] | 2018-11-01 | null | null | null | null | ['photo-to-caricature-translation', 'caricature'] | ['computer-vision', 'computer-vision'] | [ 1.17827766e-01 4.79766697e-01 4.61278588e-01 -4.03297544e-01
-2.78756857e-01 -8.20736110e-01 6.28799319e-01 -8.28503609e-01
1.81060195e-01 7.10203707e-01 -1.04346098e-02 6.82720989e-02
4.48974848e-01 -8.77472222e-01 -9.72393572e-01 -6.22017682e-01
7.10960627e-01 3.83681059e-01 -4.00077373e-01 -4.98986870e-01
3.36460327e-03 8.67776692e-01 -1.10663283e+00 2.10749581e-01
9.28974867e-01 8.19171250e-01 -4.56128031e-01 5.44700682e-01
-8.88786390e-02 6.00323975e-01 -8.93569291e-01 -1.19340467e+00
3.01790923e-01 -7.16827512e-01 -3.03327799e-01 2.07060412e-01
8.16324711e-01 -5.32565057e-01 -2.33059466e-01 1.17343307e+00
5.48861444e-01 -3.86879414e-01 7.37597764e-01 -1.42076778e+00
-1.48857129e+00 3.90039444e-01 -9.54997301e-01 -7.23413289e-01
3.57304335e-01 2.95239508e-01 3.45465630e-01 -9.87370670e-01
7.81019211e-01 1.76169252e+00 7.98661828e-01 9.09083664e-01
-1.41336358e+00 -1.01230311e+00 -1.02685034e-01 -2.12750271e-01
-1.51411617e+00 -3.23912591e-01 1.32491410e+00 -3.57026726e-01
-3.65266800e-02 5.89181840e-01 7.73858905e-01 1.33841074e+00
-2.99744634e-03 5.97924352e-01 1.25128198e+00 -2.80210078e-01
4.19774950e-02 8.77802968e-02 -5.82458496e-01 5.86803794e-01
5.90212122e-02 1.88632995e-01 5.55404238e-02 -8.09140950e-02
1.35481536e+00 8.79019722e-02 -3.20915699e-01 -2.16442749e-01
-9.03945088e-01 7.42853284e-01 5.75211823e-01 8.92448798e-02
-2.45767877e-01 3.26515138e-01 6.54053614e-02 2.31769890e-01
3.74631882e-01 6.87868118e-01 -4.04754747e-03 2.88815826e-01
-6.48397088e-01 5.33251345e-01 5.32501757e-01 1.10674334e+00
7.17487097e-01 5.24574637e-01 -2.59846509e-01 9.38741744e-01
-1.60529409e-02 9.75881338e-01 2.36967698e-01 -1.00623596e+00
1.46168560e-01 4.92546290e-01 1.61386847e-01 -1.15804946e+00
-3.25294510e-02 1.15180667e-02 -1.02632463e+00 8.34022284e-01
3.59598100e-01 -3.97082329e-01 -8.70817959e-01 1.89294505e+00
2.32377172e-01 -9.46981981e-02 -4.36222814e-02 1.06737566e+00
9.82470036e-01 4.39390838e-01 1.46638853e-02 1.25197589e-01
1.46348226e+00 -8.87418449e-01 -1.07673776e+00 -1.06810354e-01
1.00904271e-01 -1.17084157e+00 1.51820564e+00 2.88233906e-01
-1.40290987e+00 -6.28594160e-01 -9.57221389e-01 -3.02733481e-01
-2.11911693e-01 4.72961217e-01 3.57237607e-01 6.80947721e-01
-1.03569245e+00 6.64704204e-01 -2.54341364e-01 2.46491507e-02
5.43261409e-01 9.29665267e-02 -4.99384701e-01 2.26903036e-01
-1.15750277e+00 9.39651966e-01 -1.41668841e-01 1.16700262e-01
-5.74064791e-01 -9.74606812e-01 -8.59119415e-01 -2.54089292e-02
1.10354342e-01 -7.57659376e-01 1.13188517e+00 -1.77889752e+00
-2.13817739e+00 1.10329056e+00 1.82822168e-01 -3.22207361e-02
1.04431617e+00 -2.16324612e-01 -4.77720559e-01 -9.53673273e-02
-3.97269160e-01 9.51338947e-01 1.50518107e+00 -1.65064454e+00
1.75890714e-01 -2.88675100e-01 1.50552481e-01 6.62086755e-02
-1.33269146e-01 2.25631390e-02 -5.06193340e-01 -1.38223445e+00
-2.37865537e-01 -1.04168093e+00 2.72078276e-01 5.54842055e-01
-8.40024531e-01 2.13267535e-01 1.27684128e+00 -8.82150888e-01
8.60563219e-01 -2.20779610e+00 3.34658444e-01 2.59281322e-02
2.66305715e-01 3.46331090e-01 -4.25118864e-01 3.33687812e-01
-4.50007051e-01 3.42991352e-01 -1.89741418e-01 -5.83322704e-01
9.35654640e-02 1.69042930e-01 -3.44718963e-01 2.83392757e-01
4.95409340e-01 1.38445902e+00 -6.32494211e-01 -3.70661557e-01
4.85238060e-02 8.18366051e-01 -5.60766518e-01 3.08264852e-01
-4.96540256e-02 6.63417637e-01 -2.51734883e-01 6.18176579e-01
1.28780079e+00 3.52962106e-01 2.48788632e-02 -4.72414613e-01
1.12741813e-01 -4.32555735e-01 -7.60518789e-01 1.39100289e+00
-5.08295596e-01 5.37871003e-01 1.64069995e-01 -2.30009422e-01
1.16719019e+00 1.25981390e-01 -7.22871050e-02 -6.33379996e-01
4.01391029e-01 4.26904522e-02 -2.08123326e-01 -3.12472433e-01
3.61143351e-01 -5.32497585e-01 -1.28809005e-01 2.42153093e-01
-4.64705378e-01 -8.94334793e-01 -3.23428094e-01 -1.65947765e-01
2.80611426e-01 3.28079432e-01 4.81249839e-02 -1.37874082e-01
5.55969417e-01 -4.17581320e-01 3.53406161e-01 7.25062117e-02
2.17648402e-01 1.12252247e+00 6.72509968e-01 -5.44510305e-01
-1.38833356e+00 -1.06669068e+00 2.27688834e-01 4.40621853e-01
1.29430935e-01 1.54786587e-01 -1.52951002e+00 -5.89501023e-01
9.09226835e-02 8.76808822e-01 -1.02312803e+00 -3.59862447e-01
-6.63864970e-01 -1.22857340e-01 7.74952352e-01 5.70183456e-01
6.30264401e-01 -1.23444498e+00 -1.30141914e-01 -2.46243685e-01
9.41808671e-02 -9.01438415e-01 -1.29770458e+00 -8.57452512e-01
-4.48366612e-01 -8.24775100e-01 -1.19747162e+00 -6.16816461e-01
9.89661157e-01 -1.92205697e-01 1.10230350e+00 6.92728311e-02
-2.51417756e-01 -1.52238980e-01 -3.17774147e-01 -6.11210525e-01
-7.99148500e-01 -4.01913792e-01 3.52660529e-02 4.31541413e-01
-2.02624545e-01 -8.49798203e-01 -7.99552619e-01 4.40542072e-01
-1.16024637e+00 3.59102219e-01 5.37600040e-01 6.85937524e-01
4.53719765e-01 -3.14459115e-01 4.08842534e-01 -1.14219773e+00
7.53335178e-01 1.32451370e-01 -3.90154123e-01 1.21945731e-01
-2.44831264e-01 -5.09034656e-02 1.02601099e+00 -9.22901809e-01
-1.06903672e+00 -1.46964386e-01 -2.66648859e-01 -1.08389962e+00
-5.60531691e-02 -2.19182774e-01 -5.78991055e-01 -3.98553193e-01
6.35867596e-01 -1.56199168e-02 3.24684113e-01 -3.82876128e-01
8.15994382e-01 2.91718662e-01 9.05361295e-01 -6.02172554e-01
1.27544999e+00 5.35225809e-01 -8.03141817e-02 -5.92262030e-01
-3.99224430e-01 7.49861777e-01 -5.85067809e-01 -1.33929908e-01
8.56097281e-01 -5.59377670e-01 -8.83049130e-01 8.86820912e-01
-1.42988098e+00 -4.00455296e-01 -6.18835032e-01 -2.41587073e-01
-5.69848835e-01 3.47282469e-01 -6.63852513e-01 -4.87503856e-01
-5.13484895e-01 -9.93024170e-01 1.22403109e+00 2.93329030e-01
-2.72161096e-01 -7.12595522e-01 -9.68984812e-02 2.34475300e-01
5.16597450e-01 8.81514430e-01 9.13984179e-01 1.26542047e-01
-2.72624254e-01 -3.18611592e-01 -3.03218275e-01 5.25724709e-01
2.50102341e-01 3.95821780e-01 -8.55373979e-01 -1.77243516e-01
-2.55087823e-01 -1.14754036e-01 2.31954619e-01 4.25034948e-02
1.29570675e+00 -7.49871373e-01 2.27488413e-01 9.25746858e-01
1.13533545e+00 2.93851882e-01 1.33518839e+00 -2.59426296e-01
9.56449330e-01 5.21733463e-01 2.06533387e-01 2.49605015e-01
4.64118905e-02 9.44449723e-01 5.52820563e-01 -6.97814822e-01
-5.95051169e-01 -7.29404271e-01 3.03665578e-01 3.19424421e-01
-2.15926290e-01 -2.70784330e-02 -1.56765163e-01 5.25700413e-02
-1.25633204e+00 -9.43233967e-01 -3.29792537e-02 1.87065399e+00
9.19532955e-01 -3.36574674e-01 3.89809385e-02 -1.44205987e-01
9.15217161e-01 -3.42241451e-02 -8.04218471e-01 -7.04740882e-01
-2.17033222e-01 3.73198777e-01 2.41560563e-01 5.03756702e-01
-6.87569976e-01 1.20829701e+00 5.97202873e+00 9.65081334e-01
-1.36230981e+00 2.98733450e-02 8.66003036e-01 3.90756354e-02
-6.33454084e-01 -3.37305069e-01 -1.59904763e-01 5.62502682e-01
4.27944139e-02 9.16058272e-02 8.46254528e-01 8.36538672e-01
2.82345980e-01 5.26843131e-01 -8.99743497e-01 1.18308592e+00
2.08114251e-01 -1.24861097e+00 4.77003008e-01 -1.62016287e-01
8.01765919e-01 -1.00230932e+00 5.13624251e-01 1.21891879e-01
2.45639578e-01 -1.26257575e+00 1.17023838e+00 8.15486908e-01
1.59274364e+00 -8.80848229e-01 5.17356694e-01 -1.39822483e-01
-7.72107601e-01 3.14432234e-01 -1.36745304e-01 2.04595476e-01
1.96930721e-01 2.23260537e-01 -2.78294057e-01 3.15218538e-01
3.92868310e-01 6.50157034e-02 -4.85099941e-01 3.42823267e-01
-5.68059087e-01 8.05767477e-02 2.01610401e-02 1.58487976e-01
-7.36993253e-02 -5.64120054e-01 5.38868845e-01 8.93918753e-01
6.08822823e-01 8.30416083e-02 -3.19435298e-01 1.48842454e+00
-4.98044848e-01 7.97951743e-02 -5.97258270e-01 4.72406782e-02
6.13898456e-01 1.25788641e+00 -4.88218963e-02 -2.57132590e-01
1.39838874e-01 1.56133366e+00 8.18638057e-02 4.55665469e-01
-1.12287784e+00 -5.69077075e-01 8.05775404e-01 3.64466757e-01
6.74020723e-02 2.11464688e-01 -3.63287807e-01 -1.00695276e+00
-7.50396326e-02 -1.06456208e+00 -3.49087507e-01 -1.47386229e+00
-1.39731491e+00 9.81899440e-01 -2.48409912e-01 -1.08601964e+00
-4.08357941e-02 -4.36144203e-01 -1.09673822e+00 1.06773007e+00
-1.26334131e+00 -1.71980357e+00 -6.16101563e-01 6.59685969e-01
2.59742469e-01 -3.28387357e-02 7.16713250e-01 1.85230672e-01
-5.39147198e-01 1.21942902e+00 -2.54044592e-01 2.65628487e-01
9.51206446e-01 -9.99396503e-01 6.77734077e-01 6.05773389e-01
-2.13713855e-01 5.39263189e-01 9.31792080e-01 -4.44521785e-01
-1.36226213e+00 -1.19339955e+00 5.17126381e-01 -5.06594837e-01
3.20650667e-01 -3.09159577e-01 -8.30798686e-01 5.66997588e-01
3.92838895e-01 3.48973498e-02 2.54007369e-01 -5.56488991e-01
-6.29142940e-01 -1.59802690e-01 -1.56107163e+00 1.01679182e+00
1.04312396e+00 -3.75532806e-01 -3.44829053e-01 5.62877469e-02
7.71805942e-01 -6.54498518e-01 -7.84582019e-01 2.09373415e-01
6.82299316e-01 -1.01832342e+00 1.00863838e+00 -5.32045782e-01
7.28526056e-01 -4.30788726e-01 1.90741897e-01 -1.59793639e+00
-3.21042001e-01 -1.15544260e+00 1.72762156e-01 1.41241264e+00
1.90557882e-01 -5.99683344e-01 7.18437135e-01 4.86854196e-01
-6.73976392e-02 -4.88944560e-01 -5.06946087e-01 -6.75637245e-01
4.82083827e-01 4.45468277e-02 1.28151131e+00 1.19418764e+00
-4.76450294e-01 1.57923073e-01 -8.57431889e-01 -7.24026784e-02
4.40989226e-01 1.35199144e-01 1.17823744e+00 -8.63514185e-01
-2.07013130e-01 -5.73951900e-01 -1.56797320e-01 -7.40009904e-01
1.64855883e-01 -5.90084016e-01 -3.06189895e-01 -7.99038351e-01
-1.16022535e-01 -2.93574631e-01 5.06199181e-01 4.98343378e-01
-1.84054404e-01 6.54886961e-01 5.64191163e-01 1.69177470e-03
4.05404657e-01 7.63781488e-01 2.09798074e+00 -1.59522250e-01
-1.79777116e-01 -2.31640413e-01 -1.08725810e+00 8.36670041e-01
6.97413027e-01 -8.14182311e-02 -4.31563705e-01 -4.90951687e-01
3.72023918e-02 1.13312021e-01 6.65652990e-01 -6.39110684e-01
-3.74358773e-01 -2.41545796e-01 5.08011580e-01 -1.68009996e-01
5.00284433e-01 -6.74031854e-01 7.26603866e-01 2.99301535e-01
-2.72120982e-01 2.62710571e-01 2.34474987e-01 1.84982270e-01
-9.14361328e-02 3.85947973e-02 1.27215004e+00 2.45534554e-02
-1.44598305e-01 5.08674860e-01 1.92408651e-01 -8.02359805e-02
1.04279661e+00 -1.73357368e-01 -3.77821386e-01 -7.85794139e-01
-7.12795019e-01 -3.30040514e-01 9.94511425e-01 4.82770324e-01
5.67893684e-01 -1.91887844e+00 -9.13047731e-01 5.24166524e-01
-7.29502663e-02 -7.46009573e-02 5.44583142e-01 4.53205556e-01
-8.50348413e-01 -2.12759748e-01 -4.02818024e-01 -1.22777924e-01
-1.22797298e+00 8.91232193e-01 5.25357842e-01 8.42069462e-02
-6.45669401e-01 7.39440560e-01 7.41015553e-01 -5.24319470e-01
-1.05424769e-01 -5.60060218e-02 4.45038863e-02 -2.59533912e-01
4.83652949e-01 1.36608496e-01 -2.90441662e-01 -6.39424384e-01
-4.88124900e-02 9.55697477e-01 1.66607723e-01 3.82368006e-02
9.60050285e-01 9.85630527e-02 -1.83575228e-01 -1.80418596e-01
1.11985838e+00 3.70806336e-01 -1.52435946e+00 1.27629116e-01
-8.57514560e-01 -6.25416517e-01 -2.96612859e-01 -9.98115718e-01
-1.51197624e+00 8.80908906e-01 4.74583179e-01 -4.77655604e-02
1.09154260e+00 -2.92942137e-01 9.23045397e-01 -1.64056376e-01
1.04786053e-01 -7.95366764e-01 3.47177595e-01 1.06356487e-01
1.79921186e+00 -6.82420552e-01 -3.72638226e-01 -5.95358491e-01
-1.07710624e+00 9.87510383e-01 6.16990864e-01 -3.86038363e-01
3.32553506e-01 4.49512959e-01 3.21580589e-01 -8.92147720e-02
-1.21210627e-01 3.45120221e-01 4.23085600e-01 8.21314514e-01
1.81472555e-01 2.45367080e-01 -1.57849833e-01 6.91574633e-01
-7.19862401e-01 -6.00360744e-02 4.09505278e-01 2.07802311e-01
9.93180275e-02 -1.10832369e+00 -7.03458309e-01 -1.75658450e-01
-2.02262044e-01 -7.58500844e-02 -8.74307334e-01 1.06758010e+00
3.74859452e-01 5.17546296e-01 1.11205377e-01 -4.27351654e-01
7.02634513e-01 -7.63273537e-02 6.94095373e-01 -1.43198460e-01
-5.03253579e-01 7.75293261e-02 -1.44037336e-01 -5.04596293e-01
6.34744391e-02 -3.34168673e-01 -9.51496542e-01 -9.29895639e-01
-1.60640031e-02 -2.27552593e-01 5.09150684e-01 5.54873645e-01
4.91109908e-01 5.96683860e-01 7.55607009e-01 -1.08958197e+00
-4.92871225e-01 -1.01661956e+00 -6.81126833e-01 9.03804839e-01
9.20006335e-02 -4.91156965e-01 -1.95867106e-01 2.40192786e-01] | [12.146077156066895, -0.37351420521736145] |
78bb47e2-9bcc-4dfc-b6ff-940b303256ce | what-s-behind-the-couch-directed-ray-distance | 2112.04481 | null | https://arxiv.org/abs/2112.04481v2 | https://arxiv.org/pdf/2112.04481v2.pdf | What's Behind the Couch? Directed Ray Distance Functions (DRDF) for 3D Scene Reconstruction | We present an approach for full 3D scene reconstruction from a single unseen image. We train on dataset of realistic non-watertight scans of scenes. Our approach predicts a distance function, since these have shown promise in handling complex topologies and large spaces. We identify and analyze two key challenges for predicting such image conditioned distance functions that have prevented their success on real 3D scene data. First, we show that predicting a conventional scene distance from an image requires reasoning over a large receptive field. Second, we analytically show that the optimal output of the network trained to predict these distance functions does not obey all the distance function properties. We propose an alternate distance function, the Directed Ray Distance Function (DRDF), that tackles both challenges. We show that a deep network trained to predict DRDFs outperforms all other methods quantitatively and qualitatively on 3D reconstruction from single image on Matterport3D, 3DFront, and ScanNet. | ['David F. Fouhey', 'Justin Johnson', 'Nilesh Kulkarni'] | 2021-12-08 | null | null | null | null | ['3d-scene-reconstruction'] | ['computer-vision'] | [ 3.61749887e-01 3.78912538e-01 4.44942087e-01 -8.83142471e-01
-5.89478731e-01 -5.14215112e-01 8.74927819e-01 -3.49080116e-01
-3.08513135e-01 2.82196313e-01 3.05709571e-01 -5.78086853e-01
-3.39020640e-01 -9.81020868e-01 -1.25772035e+00 -3.52547377e-01
-2.04149023e-01 1.02985251e+00 3.89477819e-01 -8.01845267e-02
5.22790253e-01 1.05157816e+00 -1.44522321e+00 2.24993542e-01
2.73724467e-01 1.18331015e+00 5.26392698e-01 8.19280803e-01
-2.26479366e-01 9.11865711e-01 -6.24597259e-02 3.55416047e-03
7.65662372e-01 -1.44143045e-01 -1.15585279e+00 -3.94729450e-02
9.42647219e-01 -8.17499638e-01 -5.28914273e-01 6.77520096e-01
4.16436344e-01 4.47661988e-02 9.53725636e-01 -7.41144121e-01
-5.80823720e-01 2.58935899e-01 -3.03315133e-01 2.36602142e-01
4.51489955e-01 6.34074807e-02 9.35444653e-01 -1.12696302e+00
9.32021439e-01 1.34562111e+00 6.32032871e-01 5.15352905e-01
-1.27792370e+00 -6.56651855e-02 -1.47724897e-01 -3.14850993e-02
-1.14967346e+00 -1.27614394e-01 7.26573646e-01 -6.05734050e-01
1.53859520e+00 1.23766415e-01 7.22916901e-01 8.75183284e-01
2.35484675e-01 5.39437115e-01 1.25338304e+00 -5.20806134e-01
2.69124836e-01 -3.70209605e-01 -3.78680639e-02 8.27876747e-01
-2.34556556e-01 3.13124746e-01 -3.44006091e-01 1.47961766e-01
1.28495681e+00 -1.95208624e-01 -3.98263693e-01 -7.13269889e-01
-1.37211370e+00 6.03408813e-01 7.29282200e-01 -2.45964695e-02
-2.75787432e-02 3.73293310e-01 -3.55918370e-02 3.16002518e-01
5.81878960e-01 5.18338382e-01 -4.75361973e-01 2.57998377e-01
-6.32976711e-01 2.70427734e-01 7.51402080e-01 1.03073645e+00
9.62613106e-01 -2.01187089e-01 3.52690965e-01 5.17558455e-01
2.39147380e-01 6.03448153e-01 -2.12845877e-01 -1.48173225e+00
2.11563230e-01 2.42510855e-01 1.60315499e-01 -7.34949768e-01
-7.41918325e-01 -3.97938401e-01 -6.38120830e-01 5.88040292e-01
6.42252505e-01 4.23287153e-01 -1.22989047e+00 1.40194798e+00
2.92230070e-01 1.14336580e-01 -1.04002215e-01 1.04772472e+00
8.89055312e-01 6.26683652e-01 -4.33326393e-01 2.97844738e-01
6.10300720e-01 -6.25966728e-01 7.22519457e-02 -2.12138370e-01
7.03227043e-01 -5.32887936e-01 1.09055364e+00 5.07342517e-01
-1.42805529e+00 -4.62477297e-01 -9.89262879e-01 -4.77528065e-01
-2.12751940e-01 -4.01713043e-01 7.24027872e-01 1.87490627e-01
-1.47136557e+00 8.85516167e-01 -6.87844753e-01 -4.98484969e-01
5.48994720e-01 3.65879744e-01 -4.08810735e-01 -4.47353274e-01
-5.50485313e-01 1.05779314e+00 3.05963997e-02 -5.87473996e-02
-1.18910229e+00 -9.29459751e-01 -7.10258961e-01 -1.79591849e-01
-7.84576219e-03 -8.79639804e-01 1.28345001e+00 -4.24400896e-01
-1.22883666e+00 1.54841363e+00 1.36128619e-01 -4.84410167e-01
6.09886110e-01 -2.38398716e-01 5.23095019e-03 3.58128905e-01
-9.90763232e-02 9.46499765e-01 5.09214103e-01 -1.48705196e+00
-1.49833515e-01 -3.35437179e-01 2.98861057e-01 3.03490341e-01
4.74239111e-01 -4.62605238e-01 -1.02247991e-01 2.26910654e-02
6.17767155e-01 -4.67988133e-01 -3.44808191e-01 4.87599701e-01
-6.13670111e-01 1.68950427e-02 4.13681835e-01 -1.71714187e-01
1.31042913e-01 -1.79124725e+00 4.97370847e-02 3.69498223e-01
3.52608651e-01 -2.89922923e-01 -3.68888527e-01 2.95954138e-01
-3.63606252e-02 1.33399054e-01 -2.90763289e-01 -1.59148201e-01
-5.87585457e-02 5.38346827e-01 -4.14164633e-01 8.96319687e-01
1.25265554e-01 1.00352764e+00 -7.10743606e-01 -3.49408627e-01
5.78443766e-01 5.75908720e-01 -9.76980209e-01 2.69755155e-01
-5.42161286e-01 6.38862789e-01 -3.65614563e-01 5.15775800e-01
1.11428010e+00 -3.62497687e-01 -2.43582204e-01 -4.11982864e-01
-3.31140608e-01 4.12416130e-01 -7.84492195e-01 2.13826847e+00
-5.44674456e-01 6.44182265e-01 5.77006228e-02 -1.22420382e+00
9.25909340e-01 -1.94728971e-01 6.58010840e-01 -1.02042723e+00
1.22263737e-01 3.84455532e-01 -1.82434157e-01 -5.62254906e-01
1.14100896e-01 -3.29949111e-01 2.78180778e-01 5.17076433e-01
2.96041816e-01 -9.51883018e-01 -3.29832852e-01 2.22231984e-01
1.47067487e+00 3.98966640e-01 -2.19657421e-01 -6.53537393e-01
2.78305411e-01 7.42316246e-02 -2.39762906e-02 1.05177510e+00
1.33739159e-01 1.02339435e+00 3.35577250e-01 -9.47244942e-01
-1.52023971e+00 -1.53727281e+00 -4.68265086e-01 5.16593874e-01
5.13355553e-01 2.73035411e-02 -4.86445904e-01 -4.54158098e-01
-6.64020777e-02 6.98867142e-01 -6.95696354e-01 2.81867892e-01
-8.25021565e-01 -3.17503929e-01 3.29455525e-01 2.86577970e-01
4.18547183e-01 -7.48363316e-01 -7.11442471e-01 -1.42734766e-01
1.28870785e-01 -1.27099562e+00 2.53691282e-02 5.36776125e-01
-7.95614660e-01 -1.09377146e+00 -4.86023754e-01 -7.67935276e-01
8.25238466e-01 4.36633468e-01 1.56602192e+00 8.76453742e-02
-3.70549530e-01 7.84538209e-01 -1.57731846e-01 -1.02050364e-01
-3.14684927e-01 -1.61726788e-01 -1.48747548e-01 -4.89399344e-01
6.75428510e-02 -1.09142387e+00 -8.52060616e-01 4.49321270e-01
-8.48047554e-01 4.56675678e-01 4.39834982e-01 5.40552974e-01
8.38013172e-01 -2.78193444e-01 -4.47489433e-02 -7.97856808e-01
-2.43583862e-02 -4.25140768e-01 -6.56058967e-01 5.34976609e-02
-2.74541765e-01 3.10093731e-01 5.46845019e-01 -3.91497426e-02
-1.06877685e+00 1.37842730e-01 -5.90680838e-01 -3.18173528e-01
-4.49754953e-01 -1.89655146e-03 1.17335282e-01 -3.85550737e-01
9.83976781e-01 1.23548694e-01 -2.62554199e-01 -5.46158254e-01
3.77823472e-01 1.38561919e-01 6.94207847e-01 -7.57346630e-01
9.61628258e-01 1.04849374e+00 5.98958075e-01 -7.81447470e-01
-1.21135581e+00 -2.91765869e-01 -1.00969720e+00 -2.76195735e-01
9.47982371e-01 -8.05638254e-01 -6.56910479e-01 1.91264927e-01
-1.41676867e+00 -8.59967709e-01 -5.26516676e-01 6.21790946e-01
-1.12287509e+00 8.52096677e-02 -5.29099047e-01 -5.03212988e-01
3.10445298e-02 -8.74023914e-01 1.29753137e+00 -2.07383573e-01
7.17034787e-02 -9.97571766e-01 7.14542195e-02 2.00146094e-01
4.46349382e-01 5.40438235e-01 1.33748257e+00 -1.34797052e-01
-1.13852441e+00 2.40802363e-01 -5.04710078e-01 2.01656923e-01
-2.58585125e-01 -1.55946568e-01 -1.18088293e+00 7.99748302e-02
1.35407507e-01 -7.04514682e-01 9.64411676e-01 6.58343434e-01
1.55180967e+00 1.12654395e-01 -1.21684812e-01 1.22547770e+00
1.69204617e+00 -2.67034262e-01 6.60808980e-01 1.52708128e-01
6.73928738e-01 6.19924664e-01 3.22802544e-01 2.41953477e-01
2.75158882e-01 2.35446617e-01 1.06637895e+00 -1.69116333e-01
-4.66694623e-01 -3.66181523e-01 4.51960899e-02 6.19765580e-01
-3.80289406e-01 -3.80445749e-01 -1.02856505e+00 3.25467497e-01
-1.32128251e+00 -7.53163040e-01 -3.98030132e-01 1.93797064e+00
3.60626578e-01 4.01805252e-01 -3.38140398e-01 -7.26720691e-02
1.00382574e-01 1.43186882e-01 -8.67293835e-01 -3.62024337e-01
-4.82641608e-01 5.01410425e-01 6.48559749e-01 8.66877615e-01
-6.03310287e-01 8.20567071e-01 7.48639297e+00 4.01286572e-01
-9.05654430e-01 4.15507928e-02 6.50811970e-01 -3.82628404e-02
-9.77363110e-01 1.25780061e-01 -5.06734252e-01 -2.20377743e-01
6.41538560e-01 3.71111035e-01 6.12207055e-01 6.68908060e-01
1.13747567e-02 -2.32065544e-01 -1.58035624e+00 1.02629542e+00
3.32231969e-02 -1.65801179e+00 2.08146274e-01 3.00884634e-01
6.43029153e-01 8.44091237e-01 -1.08037777e-01 -5.43428361e-02
4.52966958e-01 -1.40777731e+00 9.75285947e-01 7.30998814e-01
9.65593755e-01 -3.93403202e-01 1.37674212e-01 6.12120688e-01
-6.84007704e-01 1.57362714e-01 -8.04498374e-01 6.87905878e-04
2.90124923e-01 7.67868459e-01 -9.34466362e-01 2.27807686e-01
7.76548862e-01 7.13862419e-01 -3.71743053e-01 9.71824467e-01
9.38315168e-02 2.17266127e-01 -6.34842932e-01 3.10242265e-01
3.99702549e-01 -6.66135326e-02 4.30870384e-01 9.40812886e-01
6.24657810e-01 2.47372270e-01 -1.73327878e-01 1.25322938e+00
-1.62700757e-01 -2.60949671e-01 -1.06740892e+00 3.46285254e-01
6.00366183e-02 8.91579628e-01 -7.93132007e-01 1.44848570e-01
-3.94786745e-01 8.64609718e-01 4.37697738e-01 3.87753755e-01
-6.36138737e-01 1.94637299e-01 5.37513912e-01 6.98594630e-01
2.69794017e-01 -6.19865835e-01 -2.31110513e-01 -1.08593631e+00
-5.13789989e-02 -6.72124512e-03 4.96208183e-02 -1.33201563e+00
-1.39473236e+00 5.44404805e-01 1.19316436e-01 -1.15216506e+00
-8.15327168e-02 -1.23406196e+00 -4.49093312e-01 6.46224737e-01
-1.85982895e+00 -1.00573838e+00 -5.29909074e-01 8.22391331e-01
4.39706445e-01 3.36573273e-01 8.77332211e-01 9.15587172e-02
3.66552681e-01 -2.11892262e-01 -8.31227005e-02 -1.04321897e-01
2.80344039e-01 -1.36898744e+00 7.58910000e-01 3.83439928e-01
1.80089012e-01 1.51699439e-01 6.07492626e-01 -3.65225643e-01
-1.68554175e+00 -7.92112052e-01 4.25462514e-01 -7.55652428e-01
2.77696043e-01 -6.70178831e-01 -6.62327468e-01 8.06466103e-01
5.56924753e-02 5.25510728e-01 2.27509454e-01 -3.48875612e-01
-5.24618864e-01 6.50809854e-02 -1.21645081e+00 2.13124275e-01
1.81198871e+00 -5.84038198e-01 -4.15219277e-01 6.62818730e-01
7.37882316e-01 -6.54622078e-01 -8.02430630e-01 5.81631601e-01
4.50306416e-01 -1.60109079e+00 1.40205896e+00 -4.21586663e-01
7.42531121e-01 -5.03929593e-02 -8.06681693e-01 -1.17273057e+00
-1.38122931e-01 -4.94809747e-01 -6.97370013e-03 2.93127894e-01
3.03897351e-01 -3.49913746e-01 8.67852986e-01 5.56038082e-01
-5.33729732e-01 -8.90722573e-01 -1.04347980e+00 -7.53992796e-01
4.88526344e-01 -8.26315105e-01 4.96470183e-01 6.77767932e-01
-7.48043537e-01 2.08756015e-01 -1.93441510e-01 1.95803687e-01
9.68242466e-01 3.09376776e-01 6.88013315e-01 -1.32135737e+00
-2.72442132e-01 -2.36310646e-01 -4.66647118e-01 -1.80381823e+00
-2.88713723e-02 -1.10502934e+00 1.29686847e-01 -1.76232660e+00
-9.10542607e-02 -7.51965702e-01 2.03543827e-01 -2.79853288e-02
8.88871849e-01 2.01635852e-01 -2.77286023e-01 1.77677155e-01
-4.37389612e-01 4.23012823e-01 1.71564627e+00 -8.55699778e-02
2.06494883e-01 -2.47767046e-01 -1.93807960e-01 9.59721088e-01
5.85537136e-01 -3.25551391e-01 -5.78515291e-01 -1.13488901e+00
7.00199902e-01 1.05937839e-01 7.51926184e-01 -9.92866039e-01
1.58627510e-01 -3.29245180e-01 7.24138677e-01 -1.03271389e+00
7.15187311e-01 -8.36198807e-01 -9.86231267e-02 1.06128685e-01
-3.42497975e-01 -1.85628876e-01 -1.19254351e-01 4.45895523e-01
2.51653194e-01 -9.21748579e-02 8.54761302e-01 -7.22576320e-01
-7.21750200e-01 4.78516340e-01 -1.63825899e-02 1.06168129e-01
6.97562933e-01 -3.80188555e-01 -3.43890429e-01 -3.51903439e-01
-7.90681660e-01 2.20838785e-02 5.89775383e-01 2.28582695e-02
1.11418056e+00 -1.27491915e+00 -7.17710495e-01 4.95945811e-01
-8.46347287e-02 6.23430252e-01 1.65338278e-01 5.21265984e-01
-1.20090961e+00 2.79475123e-01 -2.89438516e-01 -1.11715174e+00
-5.67548931e-01 3.22309405e-01 7.49859512e-01 1.03817128e-01
-1.09303010e+00 1.10634637e+00 5.70784807e-01 -9.37985897e-01
8.86590779e-02 -7.67753422e-01 2.99022228e-01 -6.33441687e-01
1.64905027e-01 -1.16038611e-02 1.51844755e-01 -6.01480186e-01
-2.13029206e-01 1.02757406e+00 2.52254158e-01 -1.31940722e-01
1.97058237e+00 -2.23879516e-01 1.34057431e-02 3.11450362e-01
1.47625935e+00 -3.06274623e-01 -1.65168726e+00 -2.90828675e-01
-3.90245765e-01 -7.46468365e-01 2.61286288e-01 -6.85012579e-01
-9.01814044e-01 1.23809659e+00 3.63243699e-01 3.74247283e-01
8.49073887e-01 5.48078775e-01 8.04518223e-01 7.33777106e-01
6.27577007e-01 -6.31196737e-01 2.10539207e-01 7.27297246e-01
1.03203976e+00 -1.24327672e+00 7.29328692e-02 -3.95276785e-01
1.85771137e-02 1.34300482e+00 5.77180147e-01 -4.61275816e-01
1.09847486e+00 4.17671323e-01 -2.17329070e-01 -7.15576470e-01
-6.29396558e-01 -9.60957482e-02 2.77992636e-01 7.58405387e-01
1.35680497e-01 -3.08326989e-01 5.58461845e-01 -2.13237569e-01
-3.67111951e-01 -3.94817799e-01 6.09553039e-01 7.12797344e-01
-6.05895340e-01 -7.30336607e-01 -1.08063877e-01 3.87674332e-01
-2.47834381e-02 -4.24177013e-02 -3.38234991e-01 8.39354336e-01
2.26949215e-01 3.38302374e-01 4.12140995e-01 -3.34469050e-01
2.99437881e-01 -3.86998266e-01 1.26418006e+00 -6.11239970e-01
8.61314237e-02 -1.96243003e-01 3.29169966e-02 -8.92412186e-01
-7.17492282e-01 -2.88790375e-01 -1.42178059e+00 -3.23310435e-01
7.66184414e-03 -3.33716035e-01 8.44600022e-01 9.52500343e-01
1.03353769e-01 1.10883266e-02 7.18509197e-01 -1.41689718e+00
-3.17319125e-01 -6.02421761e-01 -7.09197998e-01 2.85825431e-01
5.96340120e-01 -6.60299778e-01 -4.68006253e-01 -1.73113108e-01] | [8.586544036865234, -3.4606258869171143] |
812fa37f-e618-4677-af21-b9b2d45ae0c7 | one-ring-to-rule-them-all-a-simple-solution | 2104.05014 | null | https://arxiv.org/abs/2104.05014v1 | https://arxiv.org/pdf/2104.05014v1.pdf | One Ring to Rule Them All: a simple solution to multi-view 3D-Reconstruction of shapes with unknown BRDF via a small Recurrent ResNet | This paper proposes a simple method which solves an open problem of multi-view 3D-Reconstruction for objects with unknown and generic surface materials, imaged by a freely moving camera and a freely moving point light source. The object can have arbitrary (e.g. non-Lambertian), spatially-varying (or everywhere different) surface reflectances (svBRDF). Our solution consists of two smallsized neural networks (dubbed the 'Shape-Net' and 'BRDFNet'), each having about 1,000 neurons, used to parameterize the unknown shape and unknown svBRDF, respectively. Key to our method is a special network design (namely, a ResNet with a global feedback or 'ring' connection), which has a provable guarantee for finding a valid diffeomorphic shape parameterization. Despite the underlying problem is highly non-convex hence impractical to solve by traditional optimization techniques, our method converges reliably to high quality solutions, even without initialization. Extensive experiments demonstrate the superiority of our method, and it naturally enables a wide range of special-effect applications including novel-view-synthesis, relighting, material retouching, and shape exchange without additional coding effort. We encourage the reader to view our demo video for better visualizations. | ['Imari Sato', 'Yinqiang Zheng', 'Richard Hartley', 'Hongdong Li', 'Ziang Cheng'] | 2021-04-11 | null | null | null | null | ['svbrdf-estimation'] | ['computer-vision'] | [ 2.28713840e-01 1.75074726e-01 2.93730706e-01 -1.08798780e-01
-3.87633145e-01 -5.50169706e-01 2.26614237e-01 -6.22180164e-01
3.76228802e-02 6.84053957e-01 9.87802818e-03 -2.18947634e-01
-1.03505338e-02 -7.55802453e-01 -1.01349676e+00 -9.42788720e-01
3.23443413e-01 2.98508108e-01 7.57729635e-02 -2.35742480e-01
1.75883636e-01 8.39924932e-01 -1.48796785e+00 -2.53220916e-01
5.41357279e-01 1.11936522e+00 4.22024131e-01 5.80040097e-01
2.24903360e-01 4.38101709e-01 -1.39993221e-01 -3.15966666e-01
6.27957344e-01 -1.15232937e-01 -7.25300729e-01 2.97705024e-01
5.55271566e-01 -4.88916129e-01 -2.83200592e-01 1.05200958e+00
5.10202348e-01 3.43888134e-01 6.04474247e-01 -8.82746220e-01
-1.25099325e+00 -4.65408117e-02 -9.08743680e-01 -1.50447369e-01
1.67764917e-01 7.32434168e-02 7.20035195e-01 -1.08593380e+00
7.77068257e-01 1.13491106e+00 7.22482145e-01 8.20350528e-01
-1.14406788e+00 -3.64441156e-01 -1.70837790e-02 -1.27368167e-01
-1.36346722e+00 -5.20085156e-01 1.10718393e+00 -3.04610491e-01
6.42056227e-01 4.47782129e-01 7.64152408e-01 8.89897346e-01
3.36386532e-01 2.93791443e-01 9.66021717e-01 -2.92063087e-01
1.69030666e-01 9.33929756e-02 -3.63338649e-01 7.20624506e-01
4.45192643e-02 1.61324799e-01 -6.44967332e-02 -3.18560839e-01
1.57382476e+00 1.80800423e-01 -7.57216632e-01 -6.21404648e-01
-1.36031723e+00 6.04807436e-01 3.81426930e-01 5.85110299e-02
-2.41696343e-01 3.00405890e-01 -1.21631354e-01 4.30176675e-01
6.56706929e-01 2.97457427e-01 -4.00100648e-01 3.91581148e-01
-3.22716892e-01 1.18709825e-01 6.61333263e-01 9.78380561e-01
8.50754976e-01 3.63335520e-01 4.64161992e-01 9.84523833e-01
2.74627805e-01 7.81268477e-01 2.30114654e-01 -1.45540762e+00
1.61862880e-01 2.35724583e-01 4.67495650e-01 -1.12350106e+00
-3.85087401e-01 -3.08907002e-01 -1.25290000e+00 5.70850313e-01
2.33214632e-01 -1.47835329e-01 -6.45888686e-01 1.56006444e+00
6.75826669e-01 2.09406391e-01 -8.41248259e-02 1.18703663e+00
9.27639246e-01 6.45012915e-01 -8.05213749e-01 -2.31263712e-01
1.18887436e+00 -8.48231971e-01 -3.80853891e-01 9.82831325e-03
1.83227211e-01 -7.42960274e-01 1.13714552e+00 3.89785647e-01
-1.48908782e+00 -1.61618769e-01 -9.12007987e-01 -1.67019472e-01
5.07938229e-02 -1.89605325e-01 4.41772312e-01 2.20488548e-01
-1.25540471e+00 7.25319445e-01 -6.10380769e-01 -1.15027308e-01
2.64207453e-01 3.80457133e-01 -5.43365479e-01 -1.14753731e-01
-6.46347880e-01 5.27549982e-01 -1.81082100e-01 3.32599193e-01
-6.38691187e-01 -8.77226532e-01 -6.53781235e-01 -1.39196008e-01
3.38717490e-01 -1.33742774e+00 9.46539402e-01 -9.99549508e-01
-1.93477011e+00 1.06131732e+00 1.95663553e-02 1.51221901e-01
5.00402451e-01 1.05219051e-01 -1.77729934e-01 2.03543380e-01
-2.19115034e-01 3.62320304e-01 1.17542529e+00 -1.82428944e+00
-1.69321418e-01 -4.47211266e-01 1.75563753e-01 4.74163204e-01
-8.47131610e-02 4.87847850e-02 -2.98238516e-01 -8.60065341e-01
3.89688075e-01 -8.97285342e-01 -2.83172071e-01 6.50136173e-01
-4.16381538e-01 2.12546632e-01 9.14649069e-01 -6.83243096e-01
5.49005151e-01 -2.12155604e+00 3.45621198e-01 2.36175321e-02
5.33303976e-01 -1.60323679e-01 -2.48668224e-01 2.38405958e-01
-3.20006758e-01 -3.92444246e-02 -3.67467910e-01 -2.39188328e-01
-4.25028056e-01 1.43539369e-01 -1.63845614e-01 9.17135179e-01
-3.20808351e-01 9.56630528e-01 -7.18902647e-01 1.46632139e-02
2.17550039e-01 8.61424506e-01 -5.95366836e-01 2.67526835e-01
-9.05134901e-02 6.84586346e-01 -4.43170696e-01 5.54165304e-01
8.87441516e-01 -6.48668885e-01 -2.38150120e-01 -3.11443031e-01
-2.63031006e-01 -3.03626955e-01 -1.31676519e+00 1.58574510e+00
-8.08186591e-01 5.49948812e-01 7.19818115e-01 -8.17603469e-01
9.61010158e-01 1.73302874e-01 6.44630969e-01 -3.44523787e-01
1.55041680e-01 1.28086761e-01 -4.65916216e-01 -4.36760426e-01
3.75584602e-01 -4.69148606e-01 4.57772523e-01 5.81268311e-01
-3.34140748e-01 -2.40245596e-01 -5.32263041e-01 -1.74846083e-01
9.32251275e-01 1.77336845e-03 7.15012029e-02 -3.34939986e-01
4.82023031e-01 -6.15115106e-01 4.39940751e-01 3.22252214e-01
3.18144917e-01 1.23669255e+00 9.88664925e-02 -7.14761257e-01
-1.19544530e+00 -1.16565299e+00 -1.70157656e-01 7.06018150e-01
4.66244459e-01 4.12954420e-01 -5.90865731e-01 -2.00611413e-01
-1.49836466e-01 4.28798378e-01 -5.94647706e-01 -1.26514873e-02
-8.27999532e-01 -5.86911857e-01 -1.49329245e-01 3.53783220e-01
6.17080271e-01 -8.98271978e-01 -5.36946595e-01 6.50599599e-02
-1.74519375e-01 -1.06907415e+00 -7.25790143e-01 -2.92084575e-01
-1.08818448e+00 -1.04129791e+00 -1.18169332e+00 -8.27888370e-01
1.01052499e+00 8.47512901e-01 1.15813875e+00 3.13568771e-01
-7.92313088e-03 7.33741522e-01 3.17824036e-02 1.13993876e-01
-4.88190532e-01 -3.70989233e-01 -2.56141573e-02 3.82532388e-01
-6.30313933e-01 -1.12886214e+00 -9.67150390e-01 5.88634372e-01
-1.06517947e+00 3.76076579e-01 1.10727742e-01 7.98808277e-01
7.32823074e-01 -1.28328934e-01 1.86814204e-01 -6.78801179e-01
3.96342069e-01 -2.44597480e-01 -8.15546811e-01 2.50209123e-01
-4.03121263e-01 -2.09675953e-01 8.55245411e-01 -4.98921216e-01
-1.01640332e+00 -6.58050254e-02 -1.12581626e-01 -7.81566560e-01
-7.40508884e-02 -3.94011475e-03 -1.49910048e-01 -7.54366755e-01
5.68915129e-01 3.15765083e-01 1.03113592e-01 -4.67473149e-01
4.58112389e-01 3.32053512e-01 6.67738736e-01 -3.59255701e-01
1.13370931e+00 9.50756371e-01 9.14753824e-02 -1.07847214e+00
-5.16837001e-01 -1.42321810e-01 -4.05817866e-01 -3.63619775e-01
7.12721705e-01 -8.16587389e-01 -1.11720693e+00 8.09388161e-01
-1.37402833e+00 -4.74654496e-01 -5.33692420e-01 1.79797158e-01
-8.59694362e-01 4.96078312e-01 -5.86938620e-01 -5.81801176e-01
-4.20443982e-01 -8.51964653e-01 1.06946063e+00 2.28981510e-01
4.21750009e-01 -1.27206862e+00 -6.15853220e-02 4.43139553e-01
5.85314333e-01 3.79079223e-01 9.84964669e-01 3.14891160e-01
-9.19130206e-01 2.03438222e-01 -3.95194709e-01 3.53743672e-01
3.18418413e-01 -2.86696758e-02 -8.92922878e-01 -4.94642347e-01
4.91566986e-01 -1.53512463e-01 5.42578042e-01 6.82128489e-01
1.36790586e+00 -5.83529174e-01 -2.21928135e-01 1.19173849e+00
1.62432528e+00 2.10031420e-02 6.12966001e-01 1.05313957e-01
1.00927758e+00 4.09255862e-01 -7.61585236e-02 5.31153738e-01
4.40822989e-01 6.58204973e-01 8.82608473e-01 -3.93918693e-01
-2.81406701e-01 1.39923349e-01 1.88653126e-01 1.08064246e+00
-6.42357588e-01 -5.21946847e-01 -5.05964875e-01 2.89136529e-01
-1.40995407e+00 -6.93413317e-01 -7.31154382e-02 2.36607242e+00
4.84265327e-01 -4.39523906e-01 -8.72163326e-02 -2.54133850e-01
8.82425070e-01 2.47830376e-01 -8.95657122e-01 -2.66606778e-01
-1.66346982e-01 -9.74275470e-02 5.53604603e-01 6.02312088e-01
-6.98412061e-01 5.42425454e-01 6.19739056e+00 6.40988648e-01
-1.23321259e+00 7.31563941e-02 6.06268764e-01 3.41626666e-02
-9.01884913e-01 -2.48280570e-01 -4.32862014e-01 2.85496861e-01
3.56577545e-01 -2.12594401e-02 1.08277309e+00 5.15193522e-01
2.61639029e-01 1.65932119e-01 -6.93469226e-01 1.14574707e+00
1.72930509e-01 -1.61504889e+00 8.76667872e-02 5.80620989e-02
9.08804774e-01 2.06290498e-01 1.40146196e-01 -5.58501542e-01
3.30558747e-01 -6.53184950e-01 6.58914745e-01 6.80629969e-01
1.16098773e+00 -7.01706827e-01 9.32344571e-02 3.65844727e-01
-1.05586863e+00 8.59849434e-03 -5.54835379e-01 3.60124618e-01
2.69247234e-01 5.87532282e-01 -1.04289055e-01 5.29105425e-01
7.03351319e-01 8.26736867e-01 -1.26011735e-02 8.70474041e-01
1.30314618e-01 1.10850722e-01 -4.63083237e-01 2.22087115e-01
-1.87395856e-01 -6.03270829e-01 8.95452559e-01 4.18887258e-01
5.33876359e-01 5.58203459e-01 -1.11345403e-01 1.11177552e+00
-2.24586159e-01 7.44565949e-02 -7.02166557e-01 4.66514021e-01
-1.52244931e-04 1.35041761e+00 -8.99442673e-01 6.52872548e-02
-4.88873810e-01 9.60440159e-01 2.67643273e-01 7.28849232e-01
-6.20085716e-01 -1.76194385e-01 5.43173254e-01 4.46181118e-01
3.49512190e-01 -3.81026328e-01 -1.43400937e-01 -1.49446905e+00
1.34631634e-01 -5.85015535e-01 -1.94452584e-01 -1.32046127e+00
-1.55048239e+00 6.23955786e-01 -1.68822765e-01 -1.37526381e+00
1.31011933e-01 -6.71196699e-01 -5.71281552e-01 7.39602923e-01
-1.59522545e+00 -1.12885749e+00 -3.69266391e-01 7.92890310e-01
3.73773426e-01 3.36391828e-03 5.29346943e-01 1.58288047e-01
-2.31921837e-01 1.84572250e-01 4.60854560e-01 -2.09730297e-01
4.56671685e-01 -9.74656582e-01 4.13121819e-01 4.34213132e-01
-3.48613262e-01 3.32486361e-01 4.67058927e-01 -4.77047890e-01
-1.84263468e+00 -1.02530444e+00 2.76353031e-01 -2.78822035e-01
5.58888495e-01 -2.77789265e-01 -1.06340444e+00 7.32102454e-01
1.78361103e-01 3.70449752e-01 2.70802587e-01 -4.09868330e-01
-7.56749464e-03 -2.18933355e-02 -1.28441393e+00 6.71686172e-01
1.37373519e+00 -1.76586449e-01 -2.26791620e-01 5.11351049e-01
8.05117905e-01 -8.18318605e-01 -8.20238233e-01 3.59854162e-01
6.52424991e-01 -1.12419188e+00 1.27002859e+00 -2.92937815e-01
6.30884171e-01 -1.52963042e-01 -2.11211711e-01 -1.37609005e+00
-4.80537087e-01 -8.76072228e-01 -1.48624301e-01 7.75938034e-01
1.04865029e-01 -1.12881756e+00 7.92347789e-01 5.90513945e-01
-3.23453873e-01 -9.97783720e-01 -1.01850188e+00 -8.25348377e-01
9.87452343e-02 -1.74074322e-01 5.94177902e-01 1.08154988e+00
-8.17800760e-01 1.58057719e-01 -5.41731596e-01 4.15507376e-01
7.54683435e-01 3.39002490e-01 5.54472923e-01 -1.34823298e+00
-3.84183139e-01 -2.89393753e-01 -6.88435361e-02 -1.49266970e+00
8.06610659e-02 -9.01181042e-01 1.06741171e-02 -1.33052278e+00
9.57543030e-02 -6.05029166e-01 2.51747578e-01 1.58416927e-01
2.67562717e-01 2.66959101e-01 -9.76616070e-02 2.44175196e-01
-4.88547795e-02 8.28463495e-01 2.08644485e+00 5.06913103e-02
-2.19977006e-01 2.54824638e-01 -6.24869704e-01 8.74472558e-01
5.32085419e-01 -3.76702011e-01 -4.53057379e-01 -7.95069337e-01
5.52201629e-01 5.11341214e-01 8.79013121e-01 -6.47306919e-01
5.53633012e-02 -3.48008931e-01 2.23467246e-01 -3.16410005e-01
5.32990098e-01 -9.83422577e-01 5.19463360e-01 9.71683785e-02
6.10228349e-03 1.11556672e-01 -4.23761792e-02 6.03315175e-01
1.59701422e-01 -2.24339291e-01 1.03039551e+00 -4.45736587e-01
-7.11146295e-02 6.43687725e-01 -4.57702652e-02 1.03246652e-01
8.15098345e-01 -4.78334367e-01 -3.02524865e-01 -6.27341151e-01
-7.76048779e-01 -1.68196857e-01 1.01764524e+00 1.88229188e-01
9.02874589e-01 -1.57660937e+00 -6.26055777e-01 4.65064764e-01
-2.29292974e-01 3.50326598e-01 4.59544927e-01 4.98022646e-01
-6.42384231e-01 -2.58027434e-01 -9.17846859e-02 -6.16505861e-01
-8.67327392e-01 4.55565929e-01 6.90529704e-01 1.15572497e-01
-1.13129151e+00 7.78953195e-01 5.57547450e-01 -6.81727350e-01
-1.57119632e-01 -3.40924829e-01 5.14438190e-02 -4.95575011e-01
2.72541463e-01 5.79764724e-01 -1.94600178e-03 -6.29780829e-01
9.91248786e-02 1.13259351e+00 3.33716631e-01 1.72154512e-02
1.69055367e+00 -3.21662813e-01 -3.06810558e-01 3.96098971e-01
1.30691767e+00 3.41701731e-02 -1.51763439e+00 -7.36542344e-02
-9.65666652e-01 -5.64221084e-01 1.55588478e-01 -3.39351565e-01
-1.74954724e+00 7.23064482e-01 4.20071155e-01 4.03949261e-01
1.18139410e+00 7.22926781e-02 1.09386063e+00 5.14569640e-01
5.86624146e-01 -7.16512203e-01 3.95269096e-02 3.63737613e-01
1.41905570e+00 -1.00906777e+00 7.57283419e-02 -4.85266864e-01
-7.45345056e-02 1.27277350e+00 4.94635195e-01 -3.06479603e-01
8.43596995e-01 3.18595529e-01 -1.25576565e-02 -2.89467394e-01
-5.37720680e-01 4.73443896e-01 2.75717914e-01 5.58545947e-01
-1.06745781e-02 -2.36062273e-01 2.27368698e-01 -5.13068512e-02
-1.60616934e-01 -1.65665343e-01 8.07129622e-01 5.21124303e-01
-2.44344980e-01 -7.06198275e-01 -4.01034474e-01 4.14532989e-01
-1.82921782e-01 1.11496985e-01 1.10742435e-01 7.20841527e-01
-1.33115232e-01 4.33204651e-01 1.20804138e-01 -7.44658262e-02
4.24107820e-01 -5.67634106e-01 6.67594612e-01 -4.45160925e-01
-3.32198471e-01 1.19292013e-01 -3.69606137e-01 -6.03915155e-01
-6.50253177e-01 -5.86982250e-01 -1.02854168e+00 -4.78382587e-01
-4.24389780e-01 -4.02182132e-01 4.69172269e-01 7.05039203e-01
3.83135527e-01 1.53613284e-01 1.11649334e+00 -1.15320134e+00
-2.65592575e-01 -4.26148117e-01 -7.64114559e-01 1.59085527e-01
6.82440698e-01 -6.71805143e-01 -6.04444861e-01 1.08940981e-01] | [9.25064754486084, -3.1404192447662354] |
81fb8616-26e7-41ce-89ef-83c7be38234c | adaptive-online-value-function-approximation | 2204.11842 | null | https://arxiv.org/abs/2204.11842v1 | https://arxiv.org/pdf/2204.11842v1.pdf | Adaptive Online Value Function Approximation with Wavelets | Using function approximation to represent a value function is necessary for continuous and high-dimensional state spaces. Linear function approximation has desirable theoretical guarantees and often requires less compute and samples than neural networks, but most approaches suffer from an exponential growth in the number of functions as the dimensionality of the state space increases. In this work, we introduce the wavelet basis for reinforcement learning. Wavelets can effectively be used as a fixed basis and additionally provide the ability to adaptively refine the basis set as learning progresses, making it feasible to start with a minimal basis set. This adaptive method can either increase the granularity of the approximation at a point in state space, or add in interactions between different dimensions as necessary. We prove that wavelets are both necessary and sufficient if we wish to construct a function approximator that can be adaptively refined without loss of precision. We further demonstrate that a fixed wavelet basis set performs comparably against the high-performing Fourier basis on Mountain Car and Acrobot, and that the adaptive methods provide a convenient approach to addressing an oversized initial basis set, while demonstrating performance comparable to, or greater than, the fixed wavelet basis. | ['George Konidaris', 'Steven James', 'Dean Wookey', 'Michael Mitchley', 'Michael Beukman'] | 2022-04-22 | null | null | null | null | ['acrobot'] | ['playing-games'] | [-9.44711640e-02 -1.87728584e-01 -1.36822999e-01 -6.77156299e-02
-4.49723095e-01 -5.94686985e-01 4.97880161e-01 1.40734334e-02
-4.66121078e-01 1.07275522e+00 -2.76237667e-01 -3.59689713e-01
-2.39722803e-01 -8.73205543e-01 -6.13604605e-01 -8.23718131e-01
-3.89013946e-01 4.24489945e-01 3.92422408e-01 -5.63991487e-01
6.38757050e-02 8.22265327e-01 -1.78067601e+00 -3.76842022e-02
5.43509543e-01 1.03194797e+00 3.30894738e-02 7.02650189e-01
7.35649392e-02 5.87355375e-01 -7.64853299e-01 3.65566522e-01
3.91396523e-01 -4.83151674e-01 -5.68662941e-01 8.67514238e-02
6.90426752e-02 -2.42422462e-01 -2.02969223e-01 9.69463110e-01
4.09142315e-01 6.52173638e-01 4.41549629e-01 -1.20107698e+00
-1.75372139e-01 4.20722723e-01 -4.84833978e-02 3.52484256e-01
1.17949456e-01 1.67690292e-01 6.42879963e-01 -4.87314582e-01
2.70155132e-01 1.09080172e+00 1.06930256e+00 6.43094480e-01
-1.56319594e+00 -6.12313867e-01 1.06818080e-01 9.34343114e-02
-1.28850210e+00 -5.27637005e-01 9.34527338e-01 -2.21518442e-01
9.26280737e-01 2.53301054e-01 1.10139751e+00 6.21865273e-01
1.98292494e-01 3.05861115e-01 1.04255736e+00 -5.57415962e-01
7.92074025e-01 -2.03806952e-01 1.77767381e-01 9.09710407e-01
1.99902669e-01 5.24887979e-01 -3.64057750e-01 -4.16802824e-01
1.28296220e+00 -3.48486565e-02 -4.19701517e-01 -2.95429200e-01
-7.11069345e-01 9.95760918e-01 1.80197775e-01 3.68736804e-01
-7.28345215e-01 6.97519541e-01 2.59316385e-01 7.76172638e-01
2.54702687e-01 6.66161299e-01 -4.29965287e-01 -4.62871909e-01
-8.92376184e-01 5.70545673e-01 7.17105687e-01 5.24501622e-01
6.46260977e-01 7.65816867e-01 3.23087633e-01 5.46576500e-01
-1.83092788e-01 2.04377592e-01 6.36329591e-01 -1.38849950e+00
-1.87827587e-01 2.32292816e-01 4.89687026e-01 -5.99540889e-01
-4.91150379e-01 -4.39398259e-01 -6.46874309e-01 9.07868922e-01
5.65568507e-01 -1.92798644e-01 -8.73746216e-01 1.92104888e+00
5.28545380e-01 2.27197483e-01 1.15695789e-01 4.59826052e-01
-1.02092728e-01 7.35616207e-01 -3.72103691e-01 -5.33544600e-01
1.07067955e+00 -2.95008242e-01 -8.39551985e-01 1.30987927e-01
4.10172760e-01 -4.14891720e-01 1.19377863e+00 6.21824205e-01
-1.45229781e+00 -7.03005672e-01 -1.15349722e+00 5.46934426e-01
-1.76784426e-01 -1.91192433e-01 7.32423067e-01 6.50026560e-01
-1.15997386e+00 9.62976873e-01 -1.38690972e+00 1.38396025e-01
8.44717994e-02 5.42043924e-01 -1.38609573e-01 3.33031476e-01
-1.23335254e+00 1.38239861e+00 5.46715736e-01 -8.30383226e-02
-7.04434633e-01 -8.77480328e-01 -6.31419539e-01 1.84588179e-01
1.33085046e-02 -3.49151015e-01 1.41362906e+00 -7.55815148e-01
-1.73150730e+00 -1.25797451e-01 2.35925522e-02 -8.82889748e-01
2.19582453e-01 9.25529301e-02 -2.10696101e-01 3.47033590e-01
-4.23240155e-01 3.97120178e-01 1.20431495e+00 -9.65380967e-01
-6.78691864e-01 -1.58139780e-01 1.56439304e-01 -7.88039416e-02
-4.62284267e-01 -5.15009642e-01 4.25232887e-01 -4.75163102e-01
1.13522194e-01 -8.30752075e-01 -3.92667443e-01 3.72436158e-02
4.41618502e-01 -3.58224422e-01 1.02963376e+00 -3.34206223e-01
1.37569630e+00 -2.03855038e+00 7.23466352e-02 3.31315309e-01
-5.30035049e-02 1.96113586e-01 -1.49597907e-02 5.28871715e-01
-1.91612750e-01 -1.90831497e-01 -1.12446763e-01 1.29539147e-01
-5.69766574e-02 6.56951368e-01 -3.53838772e-01 6.50231361e-01
2.13288128e-01 3.64581138e-01 -7.22176254e-01 -2.22324476e-01
1.82531938e-01 7.13447690e-01 -8.12299967e-01 -1.02837279e-01
-1.81549281e-01 1.93375260e-01 -4.45625007e-01 1.21401161e-01
1.24538794e-01 -2.17468757e-02 -3.35918479e-02 -1.85281783e-01
-2.04952389e-01 2.30726048e-01 -1.53699195e+00 1.21045196e+00
-6.28707409e-01 4.42788959e-01 4.32765663e-01 -1.60587919e+00
1.24062061e+00 6.02412164e-01 7.88055122e-01 -2.52378941e-01
6.92722350e-02 2.57718176e-01 2.29952380e-01 -8.08754414e-02
2.21522391e-01 -5.05000830e-01 8.75924379e-02 3.74512196e-01
1.24093875e-01 -5.43897867e-01 4.31841403e-01 -5.40204585e-01
1.13542092e+00 7.51790265e-03 2.47137427e-01 -5.74433327e-01
4.79652077e-01 2.26560887e-02 4.91027683e-01 4.00393277e-01
-1.29717598e-02 -9.64771360e-02 3.75372261e-01 -9.38268363e-01
-1.22356749e+00 -8.85161877e-01 -2.15207711e-01 7.42205799e-01
-2.51535088e-01 -2.07999811e-01 -7.24744618e-01 -1.64395556e-01
1.79195598e-01 9.09360766e-01 -6.43917382e-01 -4.34275895e-01
-7.00262368e-01 -4.34561163e-01 5.39710462e-01 6.27785385e-01
5.61266720e-01 -7.62952328e-01 -1.23402536e+00 5.56822419e-01
2.65273482e-01 -5.12573659e-01 -2.78252870e-01 7.52737522e-01
-1.23260009e+00 -8.87151420e-01 -3.54787797e-01 -6.87503517e-01
5.50435841e-01 -2.31993899e-01 8.88032019e-01 -1.94675941e-02
-1.36722431e-01 6.82438552e-01 -1.57792040e-03 -2.59459019e-01
-5.80196559e-01 -3.25913072e-01 4.91317362e-01 -3.95873934e-01
3.82707156e-02 -9.13520753e-01 -3.58231395e-01 1.28610402e-01
-9.47740912e-01 -3.17245901e-01 7.90856779e-02 1.28809202e+00
4.74217653e-01 6.39609993e-01 9.15882826e-01 -1.89695984e-01
1.06059372e+00 -8.34688246e-02 -1.01223087e+00 -8.47046077e-02
-6.53224409e-01 3.94770443e-01 1.02629197e+00 -8.58691335e-01
-5.96569121e-01 2.94423521e-01 -1.92033827e-01 -7.18943834e-01
1.47861943e-01 6.09378636e-01 5.13895094e-01 -3.66695762e-01
9.41191733e-01 3.68915886e-01 5.85328102e-01 -4.83775765e-01
2.53495425e-01 2.68447161e-01 3.75914574e-01 -8.91037881e-01
5.83615422e-01 1.84238330e-01 3.47583562e-01 -9.75572765e-01
-2.45611101e-01 -6.83685243e-02 -4.46752310e-01 -2.19975248e-01
2.83158869e-01 -4.10939783e-01 -1.03933811e+00 -8.30309615e-02
-9.63704884e-01 -4.79176700e-01 -8.94549131e-01 4.55513716e-01
-1.04767406e+00 6.14659935e-02 -6.61106348e-01 -1.15554440e+00
-4.90761027e-02 -8.18187475e-01 4.92204309e-01 -5.21353036e-02
-3.46746802e-01 -1.05142963e+00 6.55744076e-02 -4.93103087e-01
5.93394518e-01 3.55745703e-01 9.40168619e-01 -4.89245057e-01
1.14141025e-01 -4.04297680e-01 7.41585553e-01 4.49247777e-01
2.98174262e-01 -4.76265885e-02 -5.11268675e-01 -5.81075311e-01
4.81360614e-01 -3.24508667e-01 6.07170522e-01 5.58965206e-01
9.51123238e-01 -4.58500296e-01 1.14311703e-01 1.61166549e-01
1.14389372e+00 6.17563188e-01 2.66565591e-01 3.08885425e-01
-3.26577947e-02 3.66727710e-01 3.44904840e-01 6.01381540e-01
-1.91871375e-01 3.62422347e-01 2.27996752e-01 1.39609203e-01
1.86459586e-01 2.06451118e-02 3.72785658e-01 6.80156410e-01
-3.22713733e-01 4.36121613e-01 -7.82861590e-01 3.94515812e-01
-1.67054915e+00 -1.31306124e+00 4.14519191e-01 2.30144811e+00
1.01869988e+00 3.48833740e-01 5.09919405e-01 7.31659889e-01
6.77845538e-01 -2.55713046e-01 -5.86621702e-01 -5.97853482e-01
4.63605136e-01 5.34040272e-01 3.45402837e-01 7.38557279e-01
-7.22159505e-01 4.92258281e-01 7.71667099e+00 7.59961367e-01
-1.20163977e+00 -1.85454309e-01 2.60842294e-01 -8.60947669e-02
-1.56379774e-01 -2.24582717e-01 -7.34342217e-01 5.16263545e-01
1.14560366e+00 -2.96335191e-01 8.57854426e-01 1.01775801e+00
3.23104173e-01 1.11300930e-01 -1.01648164e+00 7.44721532e-01
-5.87411404e-01 -1.42868602e+00 -3.08148682e-01 -1.03006326e-01
6.29979730e-01 -4.11223859e-01 -4.10845224e-03 5.29888630e-01
3.83763820e-01 -8.54361057e-01 4.93614286e-01 3.52567166e-01
5.04042625e-01 -1.24250031e+00 2.64047921e-01 6.80916905e-01
-1.32171762e+00 -3.92382026e-01 -4.87153798e-01 -3.80467325e-01
-9.60666388e-02 2.40088463e-01 -6.42037630e-01 -1.22250505e-01
3.82637233e-01 8.44907090e-02 6.87541366e-02 8.29404294e-01
2.11954340e-01 7.43299723e-01 -8.35172772e-01 -3.17991853e-01
4.00195271e-01 -1.81387976e-01 3.23183268e-01 7.47612417e-01
6.01717234e-01 3.05193782e-01 5.91293871e-01 6.77520394e-01
4.72875774e-01 -4.23550218e-01 -6.08253181e-01 -1.27850905e-01
7.03199744e-01 7.84626484e-01 -6.29336476e-01 -4.09771919e-01
-2.19216824e-01 3.88602287e-01 1.93892330e-01 4.97453004e-01
-7.74413764e-01 -5.11717200e-01 7.53827691e-01 1.19913056e-01
5.45203745e-01 -6.54901206e-01 1.38777375e-01 -7.10930049e-01
-2.97605306e-01 -9.98863816e-01 2.22643062e-01 -2.66983688e-01
-8.94515693e-01 6.49375021e-01 2.28663296e-01 -1.35802436e+00
-8.77520502e-01 -4.38225806e-01 -4.65888381e-01 8.90982389e-01
-1.17363417e+00 -5.82255602e-01 1.87151313e-01 7.66668022e-01
6.01343572e-01 -9.37088206e-02 1.15222490e+00 -5.55773824e-02
-1.28970325e-01 2.35617250e-01 1.95151508e-01 -2.96690613e-01
-9.61115658e-02 -1.20136690e+00 -4.25193273e-03 5.81411481e-01
-3.19783129e-02 6.98763430e-01 1.13341641e+00 -4.54514742e-01
-1.63311386e+00 -6.91981852e-01 2.59974986e-01 -2.11341865e-03
7.91695237e-01 5.50386943e-02 -1.01260018e+00 5.27164519e-01
-1.70341700e-01 3.68211001e-01 4.02947873e-01 9.70553383e-02
8.93912166e-02 -2.99263477e-01 -1.36899555e+00 5.14508724e-01
4.52923238e-01 -4.84106183e-01 -7.36211360e-01 9.96899903e-02
5.13350308e-01 -3.42381090e-01 -1.18512833e+00 2.08854705e-01
6.02930009e-01 -7.56775260e-01 9.92649853e-01 -8.39146852e-01
-3.27537507e-02 -4.47196424e-01 -1.18200324e-01 -1.64731288e+00
-5.79527259e-01 -9.40135300e-01 -7.24810839e-01 6.45703375e-01
9.16060060e-02 -8.46297860e-01 9.09139693e-01 4.40849274e-01
-1.90808535e-01 -1.00236630e+00 -1.39457154e+00 -1.31977832e+00
2.89023161e-01 -3.65879893e-01 5.65157950e-01 6.97443187e-01
4.19729352e-01 1.54462114e-01 1.69699348e-03 1.19959600e-01
4.62777495e-01 6.46531880e-02 3.91599864e-01 -1.24801147e+00
-5.56168079e-01 -6.54695272e-01 -4.79423344e-01 -1.03578258e+00
5.77401519e-02 -4.44673568e-01 3.30515094e-02 -1.22108769e+00
-7.60224462e-01 -5.79177380e-01 -4.34647352e-01 5.59692204e-01
1.65108964e-01 -1.05014155e-02 1.33379893e-02 3.17949206e-02
1.42206073e-01 6.22018576e-01 1.33880687e+00 6.29335418e-02
-5.16069293e-01 3.54876041e-01 -1.41543612e-01 9.08879936e-01
9.00512815e-01 -1.07898168e-01 -8.37945163e-01 8.84657949e-02
6.01385720e-03 5.61191678e-01 2.92392850e-01 -1.20094514e+00
1.63984239e-01 -2.38453150e-01 5.71186483e-01 -3.46383095e-01
7.32267380e-01 -1.12636805e+00 1.39868975e-01 1.06343317e+00
-5.31853139e-01 4.66271162e-01 4.70217347e-01 4.91415977e-01
-1.69261917e-01 -4.76828694e-01 1.05843437e+00 -1.25788957e-01
-4.41310614e-01 4.40358855e-02 -7.08570063e-01 -1.79635957e-01
8.86478245e-01 -3.21316928e-01 7.66515508e-02 -5.21818399e-01
-9.14255619e-01 -1.27634406e-02 2.22571582e-01 -8.01151097e-02
5.76312661e-01 -1.57649386e+00 -3.02724481e-01 4.13471282e-01
-6.85436308e-01 -3.38267356e-01 -2.77458340e-01 5.57650208e-01
-3.62084627e-01 1.88124150e-01 -5.19131660e-01 -4.36771184e-01
-1.09272742e+00 6.31733119e-01 4.39895958e-01 -7.18684122e-02
-8.09693038e-01 5.68993449e-01 -5.37346363e-01 3.53862382e-02
3.04379284e-01 -7.03494489e-01 -9.36821699e-02 -1.12534901e-02
6.76468253e-01 4.48136628e-01 -3.86445709e-02 2.26870645e-04
-2.99177356e-02 3.93769622e-01 2.33922794e-01 -3.99673790e-01
1.57720935e+00 3.51723880e-01 1.57118991e-01 5.62575758e-01
9.05586600e-01 -4.01744783e-01 -1.57800531e+00 1.93306450e-02
-8.52957070e-02 -2.43495658e-01 3.69389534e-01 -2.88836181e-01
-7.94053435e-01 6.84196293e-01 5.91906548e-01 9.51944709e-01
1.26436412e+00 -6.11884177e-01 6.26554251e-01 8.23666334e-01
5.97223580e-01 -1.20902252e+00 -4.49663913e-03 5.88877499e-01
8.82294774e-01 -4.87016350e-01 3.73318493e-02 1.11774668e-01
-2.26836026e-01 1.59362793e+00 3.64331871e-01 -7.48992264e-01
6.52012825e-01 6.52661681e-01 -3.53742212e-01 1.05410174e-01
-9.28187907e-01 -1.67996421e-01 2.03674287e-01 7.52535701e-01
3.22947651e-01 -2.44518936e-01 -4.06845570e-01 3.00914675e-01
-4.23991710e-01 1.01080559e-01 4.23209876e-01 1.21842420e+00
-1.10406518e+00 -1.08635414e+00 -6.31177247e-01 3.77143532e-01
-2.72704810e-01 1.89828679e-01 2.76357740e-01 1.06145680e+00
-1.00408442e-01 6.87776566e-01 2.66778857e-01 -1.17997915e-01
3.46493185e-01 3.78905058e-01 9.16201711e-01 -2.14378759e-01
-4.41302449e-01 2.10082665e-01 1.20505929e-01 -6.14782393e-01
-8.92143101e-02 -5.93155921e-01 -1.50382733e+00 -3.59727681e-01
-4.53070015e-01 5.91456056e-01 6.24267638e-01 7.88650572e-01
4.15400192e-02 6.60234690e-01 6.26129270e-01 -9.73560929e-01
-1.29011655e+00 -8.10366094e-01 -5.12517810e-01 -5.32211214e-02
8.38436425e-01 -9.92842257e-01 -4.31505293e-01 1.81790352e-01] | [4.343340873718262, 2.4022083282470703] |
255c1fe6-d1ea-4d08-8a16-6ed8aec11757 | are-we-really-making-much-progress-in | 2210.12941 | null | https://arxiv.org/abs/2210.12941v3 | https://arxiv.org/pdf/2210.12941v3.pdf | Unsupervised Graph Outlier Detection: Problem Revisit, New Insight, and Superior Method | A large number of studies on Graph Outlier Detection (GOD) have emerged in recent years due to its wide applications, in which Unsupervised Node Outlier Detection (UNOD) on attributed networks is an important area. UNOD focuses on detecting two kinds of typical outliers in graphs: the structural outlier and the contextual outlier. Most existing works conduct experiments based on datasets with injected outliers. However, we find that the most widely-used outlier injection approach has a serious data leakage issue. By only utilizing such data leakage, a simple approach can achieve state-of-the-art performance in detecting outliers. In addition, we observe that existing algorithms have a performance drop with the mitigated data leakage issue. The other major issue is on balanced detection performance between the two types of outliers, which has not been considered by existing studies. In this paper, we analyze the cause of the data leakage issue in depth since the injection approach is a building block to advance UNOD. Moreover, we devise a novel variance-based model to detect structural outliers, which outperforms existing algorithms significantly and is more robust at kinds of injection settings. On top of this, we propose a new framework, Variance based Graph Outlier Detection (VGOD), which combines our variance-based model and attribute reconstruction model to detect outliers in a balanced way. Finally, we conduct extensive experiments to demonstrate the effectiveness and efficiency of VGOD. The results on 5 real-world datasets validate that VGOD achieves not only the best performance in detecting outliers but also a balanced detection performance between structural and contextual outliers. | ['Xuemin Lin', 'Fan Zhang', 'Liping Wang', 'Yihong Huang'] | 2022-10-24 | null | null | null | null | ['graph-outlier-detection'] | ['graphs'] | [-2.78251082e-01 -2.83447742e-01 -1.08656496e-01 9.03536379e-02
-2.53907531e-01 -1.81470126e-01 4.26343739e-01 6.37072384e-01
-5.98522723e-02 3.40684533e-01 8.23924020e-02 -2.20449910e-01
-2.14563429e-01 -1.02530932e+00 -4.54523355e-01 -5.48110723e-01
-2.51549900e-01 1.05835572e-01 6.55095696e-01 -8.54105875e-02
2.95838684e-01 5.05178750e-01 -1.22292316e+00 -4.19537425e-01
1.25263441e+00 6.40503824e-01 -4.20059592e-01 9.23640728e-02
-1.45400941e-01 5.70415437e-01 -7.57411480e-01 -1.83786750e-01
4.33170885e-01 -4.42562014e-01 -3.37815821e-01 1.06355652e-01
5.36556654e-02 -2.36733586e-01 -5.14466941e-01 1.17996621e+00
6.39785469e-01 -1.05502121e-01 4.39812124e-01 -2.05028319e+00
-4.67451990e-01 7.31109619e-01 -9.52786505e-01 3.15138519e-01
2.92493910e-01 8.91811773e-02 8.39443207e-01 -7.14607477e-01
3.61259490e-01 1.10990191e+00 7.12911904e-01 1.62741080e-01
-1.06031561e+00 -8.14946175e-01 3.42463791e-01 1.66495472e-01
-1.54259801e+00 -1.19597934e-01 1.01121783e+00 -2.50978231e-01
6.04769945e-01 4.16828871e-01 4.07465607e-01 8.34258378e-01
1.57258153e-01 5.76405466e-01 9.65907872e-01 -1.10344522e-01
4.78874922e-01 -2.15933457e-01 3.72915536e-01 3.12578678e-01
1.13298309e+00 2.23367661e-02 -3.86389703e-01 -5.20475745e-01
3.00800711e-01 5.76146841e-01 -3.95210087e-01 -3.01062882e-01
-8.73870730e-01 5.91272831e-01 5.83106577e-01 2.63342589e-01
-2.54090369e-01 1.44611791e-01 6.76571846e-01 4.02522296e-01
4.77646708e-01 1.76981300e-01 -9.58380550e-02 -5.07343113e-02
-7.47936487e-01 1.04315661e-01 7.53257573e-01 9.23921168e-01
4.99125034e-01 2.20380157e-01 -2.13953868e-01 5.80706537e-01
5.96887767e-01 3.43289971e-01 2.81371742e-01 -2.26543054e-01
4.31267262e-01 1.17902172e+00 -3.00662458e-01 -1.48422408e+00
-3.82291496e-01 -6.41364694e-01 -1.09860408e+00 -4.65473048e-02
2.04168245e-01 1.54840112e-01 -7.97055304e-01 1.63100469e+00
4.32124972e-01 6.54907703e-01 -4.40315567e-02 6.91749156e-01
7.30630040e-01 3.97948146e-01 -8.68921131e-02 -5.33304155e-01
9.76133525e-01 -7.41104364e-01 -8.72785985e-01 8.99619609e-03
7.93744087e-01 -5.39922416e-01 9.13632810e-01 3.36090565e-01
-5.37608981e-01 -7.13533238e-02 -1.24353933e+00 3.26677620e-01
-3.52151066e-01 -4.05036747e-01 4.77050632e-01 9.52831745e-01
-8.16692650e-01 3.87088567e-01 -8.17367673e-01 -7.30956852e-01
2.82318115e-01 1.51719809e-01 -2.59775043e-01 -1.71177208e-01
-8.65480959e-01 3.53278071e-01 3.88894469e-01 1.76022723e-02
-7.41262197e-01 -5.51586151e-01 -7.26650953e-01 -8.18576571e-03
9.44151878e-01 -3.58378500e-01 6.64947629e-01 -5.59068799e-01
-8.50018859e-01 3.65787297e-01 -1.32306859e-01 -4.10716683e-01
7.50043094e-01 -3.99193435e-04 -8.37661028e-01 -2.11746067e-01
1.87151760e-01 -4.38894153e-01 5.67018569e-01 -1.42084885e+00
-4.14059967e-01 -7.09362209e-01 -3.31678689e-01 -9.85274240e-02
-6.76632643e-01 5.73985726e-02 -4.76661623e-01 -8.23288381e-01
4.46532726e-01 -7.56088793e-01 -3.09278309e-01 -2.79921472e-01
-8.76697481e-01 -3.74810994e-01 1.26599085e+00 -2.51217902e-01
1.94226873e+00 -2.31949449e+00 -3.77914369e-01 8.52658033e-01
7.34110653e-01 2.21811533e-01 5.03843948e-02 7.77229011e-01
-1.24770002e-02 4.84676659e-01 -4.21956897e-01 -3.36113483e-01
1.99311995e-03 1.21837400e-01 -4.21681374e-01 6.53147221e-01
-1.72018096e-01 5.13141572e-01 -9.77360666e-01 -1.49817377e-01
4.27879803e-02 6.49312437e-02 -2.77250290e-01 2.90018260e-01
1.44549996e-01 1.12924337e-01 -4.32256430e-01 9.93395507e-01
1.05826330e+00 -1.15446270e-01 1.69637397e-01 1.02234438e-01
8.59026313e-02 -1.77961349e-01 -1.62371671e+00 8.66698265e-01
2.43405432e-01 1.32794529e-01 -1.61214232e-01 -7.72508979e-01
1.20574653e+00 8.99036378e-02 4.93423253e-01 -5.46713352e-01
-1.51097476e-02 6.51534498e-01 1.10098988e-01 -2.88633704e-01
4.94846910e-01 2.20855206e-01 -1.77010894e-01 4.71517205e-01
-4.92500633e-01 5.52375138e-01 3.76328319e-01 6.92280650e-01
1.74227369e+00 -4.86302376e-01 4.13351715e-01 -2.72122264e-01
5.40723145e-01 -3.47059995e-01 9.60354745e-01 7.98500717e-01
-4.71701175e-01 4.91689652e-01 9.51112032e-01 -1.98741078e-01
-6.36371732e-01 -1.02346134e+00 7.20900372e-02 5.45990169e-01
4.27549034e-01 -8.69582891e-01 -5.09644270e-01 -8.94425571e-01
3.66008878e-01 4.99645978e-01 -4.33176488e-01 -4.07300442e-01
-2.58226097e-01 -1.10423052e+00 6.97207928e-01 3.52536142e-01
6.44116938e-01 -8.94791603e-01 9.87905310e-04 2.02104166e-01
7.62765408e-02 -1.04346907e+00 -3.60938430e-01 -2.05906734e-01
-9.99049366e-01 -1.33071518e+00 -3.81086506e-02 -4.07513827e-01
9.52957869e-01 5.71484923e-01 7.93666959e-01 6.66439235e-01
-9.17717442e-02 1.41741872e-01 -4.88346905e-01 -4.71622199e-01
-2.79133022e-01 6.77956566e-02 3.30980003e-01 1.68653056e-01
4.59123015e-01 -8.08174431e-01 -3.73634428e-01 4.37553912e-01
-1.17379832e+00 -6.63218975e-01 6.33131742e-01 3.37858051e-01
4.26232398e-01 4.23936635e-01 8.75954628e-01 -1.16710091e+00
1.01708889e+00 -1.02016222e+00 -5.81040084e-01 -5.07781282e-02
-1.07726288e+00 -2.07257360e-01 8.52688134e-01 -4.82936613e-02
-5.28981030e-01 -4.16508973e-01 4.86837029e-02 -2.90895462e-01
-7.50164390e-02 6.33636594e-01 -5.52443802e-01 -2.17888672e-02
4.81584251e-01 1.05436087e-01 -4.34352048e-02 -4.05010641e-01
-8.87113959e-02 7.11396933e-01 3.32164139e-01 -3.90194416e-01
1.11519563e+00 5.50416291e-01 2.21100390e-01 -8.64931226e-01
-2.99100250e-01 -6.63180828e-01 -1.49530485e-01 1.54629061e-02
1.76112145e-01 -7.38307357e-01 -5.49323857e-01 8.59670699e-01
-6.99006736e-01 7.85702020e-02 -7.80360773e-03 1.69446692e-01
5.97735085e-02 8.83944988e-01 -5.43618500e-01 -8.85537624e-01
-3.57106745e-01 -1.11859548e+00 5.34745872e-01 1.04484633e-01
1.51587138e-02 -8.70137691e-01 1.25573695e-01 -1.10206045e-01
4.21274900e-01 8.46490920e-01 6.66475117e-01 -1.21500850e+00
-7.49818265e-01 -3.45135629e-01 -3.61557335e-01 -5.82434908e-02
2.40406021e-01 4.09037083e-01 -6.47513032e-01 -7.08656549e-01
-2.06371397e-01 3.71608168e-01 7.43399918e-01 -2.12494154e-02
1.07610226e+00 -1.97967231e-01 -5.41127503e-01 5.90341508e-01
1.70211172e+00 -1.00308195e-01 8.64171147e-01 5.87332785e-01
8.76549661e-01 5.41615225e-02 6.44752622e-01 5.58157563e-01
4.52034563e-01 5.65782964e-01 8.67955625e-01 -4.91881445e-02
2.01168120e-01 -3.31595033e-01 5.30069172e-01 9.23819959e-01
1.94684174e-02 -5.17357588e-01 -1.03014290e+00 6.24838710e-01
-2.05672407e+00 -8.73023689e-01 -7.85322964e-01 2.52681589e+00
2.50960976e-01 4.11704153e-01 3.88079405e-01 5.96994877e-01
8.42250228e-01 2.24629954e-01 -3.73553425e-01 -2.50179559e-01
-1.71982441e-02 -2.61769742e-01 5.14258504e-01 1.05750427e-01
-8.57874870e-01 5.33545434e-01 5.56103897e+00 7.19579041e-01
-8.40800822e-01 -6.03364930e-02 3.50125462e-01 3.34922165e-01
-3.55179071e-01 3.03883016e-01 -6.22696102e-01 9.52064574e-01
8.29361916e-01 -4.25097227e-01 5.74265979e-02 9.00332153e-01
4.36765224e-01 -3.48596275e-02 -7.87126362e-01 8.77819002e-01
7.41567761e-02 -8.02913964e-01 6.24460690e-02 4.32060122e-01
6.75680935e-01 -9.03415829e-02 -3.13086599e-01 2.46382743e-01
2.27051198e-01 -8.43335032e-01 2.47379392e-01 3.64771783e-01
2.61104137e-01 -1.01689255e+00 1.12422943e+00 2.61286706e-01
-1.35941553e+00 -3.28689337e-01 -3.71642798e-01 -1.57018915e-01
-7.11577833e-02 1.24055052e+00 -5.21145165e-01 1.04782307e+00
8.63144577e-01 8.62058818e-01 -8.70112479e-01 1.50558412e+00
-3.52542609e-01 9.34547305e-01 -5.00684857e-01 1.09100193e-01
-4.23150882e-02 -2.79700100e-01 8.14728737e-01 8.70245099e-01
5.15830100e-01 -2.64896512e-01 5.57503819e-01 6.64359152e-01
-3.56335759e-01 4.32124138e-01 -8.66417468e-01 2.17400402e-01
1.04348099e+00 1.25437605e+00 -9.87815797e-01 -1.06694087e-01
-4.28790510e-01 6.41992092e-01 2.02005088e-01 3.13169539e-01
-8.15259337e-01 -6.50844395e-01 5.01820266e-01 3.58258635e-01
-7.25254342e-02 -1.60536766e-01 -4.70934629e-01 -1.11359870e+00
3.86770576e-01 -1.04658449e+00 6.91279590e-01 -1.38084993e-01
-1.47718811e+00 2.83708006e-01 -1.87832668e-01 -1.38053727e+00
3.95419002e-01 -7.79842883e-02 -1.19972301e+00 4.22222406e-01
-1.34699023e+00 -7.46930301e-01 -8.67594182e-01 5.90011716e-01
-1.16263714e-03 4.81566228e-03 2.95041859e-01 4.78667468e-01
-1.18905866e+00 8.30326080e-01 1.50681585e-01 2.28553727e-01
8.80279660e-01 -1.18995988e+00 3.69642884e-01 1.57523000e+00
-1.67404816e-01 6.85719132e-01 6.74044490e-01 -1.15778983e+00
-1.33996677e+00 -1.18739176e+00 5.84279954e-01 -3.21407318e-01
6.45095170e-01 -3.00373554e-01 -1.17970264e+00 7.49023438e-01
-1.71655193e-01 3.29491377e-01 5.20379782e-01 1.02454655e-01
-3.43703866e-01 -4.34346646e-01 -1.39538324e+00 7.15529740e-01
1.21563101e+00 -5.58315068e-02 -2.86105275e-01 4.78225164e-02
7.22119093e-01 -1.54732943e-01 -8.36812854e-01 6.74823701e-01
-8.92997012e-02 -1.29936695e+00 5.74144065e-01 -2.18264326e-01
-3.03780623e-02 -8.92707109e-01 8.24934319e-02 -1.19032025e+00
-1.54544234e-01 -7.69959569e-01 -4.54513460e-01 1.76262045e+00
3.25031281e-02 -1.25214851e+00 5.53410649e-01 3.24941218e-01
-1.26987204e-01 -6.95050418e-01 -7.54941702e-01 -1.12481654e+00
-4.47288781e-01 -3.02425057e-01 1.01467848e+00 1.16844654e+00
2.92235892e-03 -1.52167082e-02 -2.61014909e-01 4.66503650e-01
8.43742371e-01 -2.43776128e-01 1.23329091e+00 -1.51660001e+00
2.34681591e-01 -2.99342066e-01 -9.14406359e-01 -4.25685227e-01
-2.00545698e-01 -8.23158920e-01 -2.41026387e-01 -1.53653896e+00
2.91876584e-01 -4.64927018e-01 -5.82401037e-01 4.73192215e-01
-6.07367456e-01 8.24143291e-02 -2.81834342e-02 4.74416226e-01
-6.79264843e-01 5.65509856e-01 5.51237404e-01 2.69694567e-01
-3.74228716e-01 2.51079183e-02 -8.04442406e-01 6.96611166e-01
8.99333358e-01 -7.04391241e-01 -4.70315844e-01 9.74712428e-03
2.31138483e-01 -3.62693280e-01 2.92928874e-01 -1.20156407e+00
3.45218897e-01 -6.10985309e-02 3.00861392e-02 -6.41836405e-01
-4.59875464e-01 -1.04136050e+00 1.60600066e-01 6.68275774e-01
2.95514107e-01 5.42301834e-01 -9.12449509e-02 1.04217362e+00
-2.58884817e-01 1.67909160e-01 5.25959849e-01 2.62803286e-01
-7.90316880e-01 6.06859982e-01 -1.07188046e-01 1.83208138e-01
1.40590978e+00 -3.12421501e-01 -6.38729453e-01 -3.24697137e-01
-2.31752485e-01 5.04343212e-01 9.16079104e-01 4.30015296e-01
6.61380827e-01 -1.45803106e+00 -6.60086811e-01 2.58980274e-01
5.05186379e-01 -8.64016414e-02 7.62438327e-02 1.07094824e+00
-5.38537621e-01 -1.99883461e-01 9.61526632e-02 -5.10453105e-01
-1.13502610e+00 7.50273407e-01 1.08839355e-01 -4.91844118e-01
-6.24240220e-01 3.11826766e-01 -3.90649915e-01 -4.45557415e-01
2.63214618e-01 -6.20103292e-02 2.07598880e-02 -1.22137964e-01
2.55610853e-01 8.40228677e-01 9.75483507e-02 -4.91509825e-01
-6.28264189e-01 2.74062902e-01 -5.30073568e-02 4.78101045e-01
1.12012720e+00 -2.30255667e-02 -5.98534644e-01 3.87071311e-01
8.39151084e-01 5.57109356e-01 -7.03559339e-01 -2.45957658e-01
4.81362641e-01 -7.30113387e-01 -2.93797076e-01 -2.91242003e-01
-1.22743595e+00 4.18104380e-01 3.66798759e-01 6.77161396e-01
1.31673527e+00 -4.33200151e-01 8.51128221e-01 -1.19723007e-02
3.61689210e-01 -1.01936042e+00 1.74427822e-01 2.90103793e-01
4.07576382e-01 -1.19171906e+00 3.42571259e-01 -8.26994658e-01
-2.64950573e-01 8.96976471e-01 1.03035665e+00 -3.13291460e-01
5.76981902e-01 1.61690578e-01 -8.72205943e-02 -3.00610214e-01
-4.12207752e-01 -1.21315710e-01 -8.25498477e-02 4.91626590e-01
1.08530454e-01 5.16389646e-02 -6.44134045e-01 5.88900745e-01
1.12065837e-01 -3.17494899e-01 1.00673294e+00 9.76222515e-01
-4.69852865e-01 -1.20083499e+00 -4.55119282e-01 6.82465255e-01
-4.04042155e-01 7.55653158e-02 -4.73495722e-01 8.91667068e-01
-1.21650964e-01 1.09052265e+00 -1.15100034e-01 -4.32568729e-01
5.75952590e-01 -2.37823918e-01 -3.65555823e-01 -4.33568388e-01
-5.20476341e-01 -2.66453266e-01 -1.52101845e-01 -6.96155190e-01
-1.00713797e-01 -3.29445839e-01 -1.29173434e+00 -6.54113770e-01
-5.57542324e-01 3.39412600e-01 3.07723671e-01 6.96149170e-01
6.26233935e-01 6.04150593e-01 8.79040897e-01 -6.20926730e-02
-5.33267498e-01 -7.51538336e-01 -9.00389433e-01 6.22686803e-01
1.92026615e-01 -6.11054659e-01 -8.53191435e-01 -8.05890858e-01] | [6.682646751403809, 5.792638778686523] |
64484b42-46e1-46a1-819b-36fc66cf26bb | funnel-activation-for-visual-recognition | 2007.11824 | null | https://arxiv.org/abs/2007.11824v2 | https://arxiv.org/pdf/2007.11824v2.pdf | Funnel Activation for Visual Recognition | We present a conceptually simple but effective funnel activation for image recognition tasks, called Funnel activation (FReLU), that extends ReLU and PReLU to a 2D activation by adding a negligible overhead of spatial condition. The forms of ReLU and PReLU are y = max(x, 0) and y = max(x, px), respectively, while FReLU is in the form of y = max(x,T(x)), where T(x) is the 2D spatial condition. Moreover, the spatial condition achieves a pixel-wise modeling capacity in a simple way, capturing complicated visual layouts with regular convolutions. We conduct experiments on ImageNet, COCO detection, and semantic segmentation tasks, showing great improvements and robustness of FReLU in the visual recognition tasks. Code is available at https://github.com/megvii-model/FunnelAct. | ['Xiangyu Zhang', 'Jian Sun', 'Ningning Ma'] | 2020-07-23 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1342_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123560341.pdf | eccv-2020-8 | ['scene-generation'] | ['computer-vision'] | [-8.56577829e-02 7.32330158e-02 -1.46277785e-01 -5.02000451e-01
-1.47707537e-01 -7.13002980e-01 5.19430876e-01 -5.53506799e-02
-4.85348850e-01 1.99733019e-01 -8.99945199e-02 -7.26139843e-01
1.23663165e-01 -7.30809331e-01 -1.02012992e+00 -3.86079937e-01
-2.12739393e-01 -1.75669730e-01 4.22833532e-01 1.24688752e-01
1.97737098e-01 5.27188301e-01 -1.08844829e+00 3.05058777e-01
7.94267476e-01 1.27849495e+00 4.46279168e-01 5.60548127e-01
-1.51225895e-01 5.94085991e-01 -4.40752596e-01 -4.32271734e-02
2.89620101e-01 6.02410827e-03 -9.25652921e-01 1.84809342e-02
5.77404916e-01 -3.84059340e-01 -5.99924207e-01 9.58873332e-01
3.48663896e-01 1.67298123e-01 5.35936236e-01 -1.19851005e+00
-9.56341743e-01 6.07546926e-01 -9.68714535e-01 2.33222008e-01
9.70946774e-02 2.66865075e-01 6.81137085e-01 -1.00110340e+00
5.01581490e-01 1.29658091e+00 5.71138978e-01 3.44831795e-01
-1.19859040e+00 -6.27771020e-01 6.34529829e-01 1.19870305e-01
-1.38276148e+00 -2.80734211e-01 4.60901380e-01 -2.85074562e-01
1.00091350e+00 2.85615534e-01 4.94225830e-01 8.43971431e-01
-2.97833979e-02 1.19873762e+00 1.36995888e+00 -4.41700697e-01
2.74528891e-01 -1.90822825e-01 5.36389589e-01 7.50069618e-01
-4.26314548e-02 -8.49612281e-02 -2.44811147e-01 4.71650735e-02
1.33878601e+00 1.75784096e-01 -4.37680990e-01 -2.48474717e-01
-1.09633398e+00 4.37030822e-01 9.30285811e-01 3.39664429e-01
-1.60635933e-01 3.85217935e-01 1.99390471e-01 1.63528502e-01
1.82940260e-01 3.72481734e-01 -4.32821661e-01 1.81719586e-01
-8.16160798e-01 4.67087552e-02 2.43210793e-01 1.17946863e+00
7.02752173e-01 1.36056006e-01 -5.50988019e-01 8.89603376e-01
9.58766192e-02 3.55731279e-01 3.94989103e-01 -1.17625737e+00
3.09338719e-01 6.48255527e-01 7.96205774e-02 -6.29278898e-01
-5.74072659e-01 -4.40722048e-01 -7.77955234e-01 2.88544714e-01
6.16541743e-01 1.78475175e-02 -1.41074026e+00 1.80349267e+00
4.36448529e-02 3.19343120e-01 -2.28799596e-01 9.95012105e-01
9.99832988e-01 7.49867618e-01 2.05853477e-01 1.65251359e-01
1.58859086e+00 -1.26487076e+00 -4.22674507e-01 -4.86236006e-01
4.62065250e-01 -4.48747873e-01 1.58291280e+00 2.29597598e-01
-1.26101267e+00 -6.15426302e-01 -9.11311626e-01 -2.41797328e-01
-5.06101787e-01 4.73716319e-01 7.89891601e-01 4.12981927e-01
-1.28318548e+00 3.50935072e-01 -7.98487902e-01 -1.88738480e-01
5.32702267e-01 3.25699955e-01 -4.56594855e-01 -5.62890433e-02
-8.75707984e-01 5.08035958e-01 2.47336015e-01 2.47742489e-01
-9.20154870e-01 -6.87348902e-01 -8.73923004e-01 3.06163847e-01
4.39134181e-01 -4.09043342e-01 1.15785801e+00 -8.30060124e-01
-1.15876591e+00 9.06121492e-01 -1.72601357e-01 -4.97229666e-01
4.90039378e-01 -2.56117702e-01 8.86306986e-02 1.82944432e-01
-6.39026761e-02 1.23975825e+00 7.26441503e-01 -1.13533545e+00
-3.18527669e-02 -4.86819059e-01 2.72599518e-01 2.08888382e-01
-3.02326858e-01 1.69614181e-01 -9.53227997e-01 -7.76183605e-01
2.41651982e-01 -5.68640530e-01 -2.72520095e-01 4.75224525e-01
-6.61744416e-01 -2.85358191e-01 9.03097272e-01 -6.16920829e-01
1.23796999e+00 -2.37417126e+00 -1.06583670e-01 3.51091444e-01
4.78353143e-01 4.43011433e-01 -2.70718157e-01 7.29254261e-02
-3.44559491e-01 2.84444779e-01 -5.36062777e-01 -1.62141711e-01
-1.28514826e-01 3.98589671e-02 -2.52353579e-01 2.24883452e-01
1.15082733e-01 1.23077941e+00 -5.62198520e-01 -2.77042508e-01
1.42417118e-01 2.64201730e-01 -4.27948087e-01 2.41395570e-02
-2.36074582e-01 1.60012946e-01 -4.91350502e-01 7.39654064e-01
9.63380337e-01 -4.47656244e-01 1.79426849e-01 -1.66194692e-01
-3.40274602e-01 1.61708251e-01 -1.14878345e+00 1.65774095e+00
-2.71974266e-01 5.51132262e-01 3.58791798e-01 -8.52695107e-01
8.93558919e-01 1.68295186e-02 3.52271169e-01 -9.80968714e-01
2.99235284e-01 -6.41972944e-02 -2.64680535e-01 -2.12920859e-01
3.10707450e-01 4.91186619e-01 1.05912676e-02 4.83531356e-01
-6.33082613e-02 3.14255089e-01 1.41793802e-01 3.35399032e-01
1.05456984e+00 1.48325026e-01 1.94321632e-01 -3.17697257e-01
1.27214462e-01 -3.11074138e-01 4.43716258e-01 1.01275861e+00
-4.80701149e-01 7.71926880e-01 8.69978368e-01 -1.74130902e-01
-8.16577971e-01 -1.33549726e+00 -3.65493864e-01 8.15802932e-01
4.46084172e-01 -5.30233622e-01 -8.76770020e-01 -7.81090796e-01
7.15276003e-02 4.56461340e-01 -6.84504271e-01 5.52290417e-02
-7.27153361e-01 -5.02667964e-01 6.72968149e-01 1.13379824e+00
9.40406919e-01 -1.09202540e+00 -6.19037092e-01 -2.36884370e-01
1.10867947e-01 -9.98206913e-01 -8.75211954e-01 1.55931816e-01
-9.15101409e-01 -9.39640164e-01 -9.42968249e-01 -7.80035794e-01
1.04993296e+00 4.06062454e-01 8.36305201e-01 1.46276727e-01
-4.17260468e-01 1.14761971e-01 -9.94973183e-02 -6.85287416e-02
4.21284944e-01 -1.80093914e-01 -2.07553923e-01 -2.28411987e-01
-4.99062501e-02 -4.23862755e-01 -8.52679193e-01 5.54272473e-01
-9.04089212e-01 3.22031349e-01 5.08155465e-01 8.55129421e-01
6.65587783e-01 -4.26564723e-01 3.30503881e-01 -7.61274040e-01
4.63312835e-01 -2.75847763e-01 -6.37897611e-01 2.21429840e-01
-3.46447051e-01 -1.69458911e-01 3.45014840e-01 -6.50719464e-01
-8.61293674e-01 1.40221238e-01 -1.13632821e-01 -7.85309494e-01
-3.48469943e-01 3.38494301e-01 -8.89674500e-02 9.82720777e-02
5.74943244e-01 4.82687056e-02 -1.22104958e-01 -6.92714572e-01
6.54959381e-01 4.56305087e-01 6.13949001e-01 -6.94190025e-01
2.49160975e-01 3.79764467e-01 -3.74790549e-01 -6.49470270e-01
-4.05699193e-01 -1.59879074e-01 -5.16066074e-01 -9.72418711e-02
8.38696361e-01 -7.71820545e-01 -6.93105519e-01 6.17954493e-01
-1.05861950e+00 -6.71095669e-01 -1.64748475e-01 3.57203275e-01
-2.35140473e-01 3.05026919e-01 -7.62911439e-01 -6.67606890e-01
-2.47961238e-01 -1.12133181e+00 9.04096365e-01 3.65353405e-01
-1.83715820e-01 -7.37861753e-01 -4.90962148e-01 1.70133412e-01
2.98799843e-01 3.82633284e-02 1.06780338e+00 -2.54068583e-01
-7.62785733e-01 1.83764547e-01 -8.19757104e-01 2.90784478e-01
-1.58348903e-01 -1.22146957e-01 -1.00142193e+00 -2.12015390e-01
-4.09050614e-01 -2.83407807e-01 1.16387844e+00 7.31224418e-01
1.64214957e+00 -3.07352602e-01 -5.10557294e-01 8.12188804e-01
1.26881814e+00 4.01357502e-01 8.44608128e-01 -7.05977762e-03
7.00195312e-01 3.70845616e-01 3.21209997e-01 1.68403149e-01
3.01159739e-01 7.90872633e-01 4.34200227e-01 -6.21632516e-01
-4.05570358e-01 -9.54480767e-02 2.50677377e-01 1.43321350e-01
-2.38546971e-02 -1.25572115e-01 -1.04841065e+00 4.64237124e-01
-1.97341573e+00 -7.12430060e-01 -1.32753253e-01 2.23616648e+00
5.72880328e-01 9.60469544e-02 8.11337754e-02 -1.00137383e-01
7.03308105e-01 1.45701170e-01 -6.20341837e-01 -4.33911085e-01
-1.94742814e-01 2.10987523e-01 5.40190101e-01 4.68915135e-01
-1.10354614e+00 1.04251754e+00 7.08508825e+00 9.07337070e-01
-1.47173262e+00 1.14978626e-01 9.53919470e-01 -2.37428948e-01
-1.89651757e-01 -1.78193048e-01 -5.50592303e-01 5.37329257e-01
3.30729842e-01 4.84201819e-01 3.89575273e-01 8.20968568e-01
9.96081755e-02 -2.23848462e-01 -9.09407556e-01 9.20710444e-01
-2.22509965e-01 -1.29402089e+00 -4.38241400e-02 -1.49486780e-01
5.16117752e-01 9.17655528e-02 3.25080842e-01 1.93148822e-01
1.73836246e-01 -1.15382135e+00 8.34056616e-01 1.90721095e-01
8.96934211e-01 -4.14069295e-01 2.36947671e-01 1.73329622e-01
-1.27829802e+00 -1.94895104e-01 -2.47855186e-01 -5.70539311e-02
3.63650136e-02 5.37010431e-01 -4.87291902e-01 3.60480398e-01
1.06933737e+00 5.16848862e-01 -6.41010165e-01 1.04144645e+00
-2.80755728e-01 5.44678450e-01 -5.56900084e-01 1.96427584e-01
4.51211721e-01 -2.55743802e-01 3.96958143e-01 1.28683341e+00
1.57078445e-01 4.54474017e-02 1.44933954e-01 1.26109755e+00
-3.63502830e-01 -1.41143836e-02 -5.11027515e-01 2.88713127e-01
4.87321556e-01 1.14875650e+00 -1.02438045e+00 -2.15458676e-01
-2.65161812e-01 8.86109889e-01 4.19464529e-01 8.04162681e-01
-9.48635757e-01 -4.50001061e-01 5.99731743e-01 3.06880027e-01
3.83555651e-01 -3.55776906e-01 -7.67338336e-01 -9.65943873e-01
2.67982572e-01 -4.11183953e-01 3.11644942e-01 -9.86852586e-01
-7.17546463e-01 5.96719086e-01 2.40737602e-01 -9.26827252e-01
3.56507838e-01 -1.04277503e+00 -9.57952082e-01 8.93108130e-01
-1.30396068e+00 -1.14520264e+00 -3.18041861e-01 7.35123932e-01
3.13586026e-01 1.62813455e-01 4.53216255e-01 3.11288655e-01
-8.63984823e-01 7.49670744e-01 -9.20999274e-02 2.98154593e-01
5.36063790e-01 -1.19183147e+00 5.78239799e-01 7.63837636e-01
-7.83028454e-03 8.34968388e-01 1.95726961e-01 -5.32536447e-01
-1.16157019e+00 -1.03634441e+00 6.68589413e-01 -2.04074755e-01
2.59862036e-01 -5.26034296e-01 -9.97406006e-01 8.23507667e-01
3.51512939e-01 2.40866125e-01 2.11762756e-01 6.42810091e-02
-6.41770422e-01 -4.95964214e-02 -1.17456424e+00 7.91671991e-01
1.15930617e+00 -4.21081513e-01 -2.33303025e-01 1.77066490e-01
5.29645443e-01 -5.91630459e-01 -5.73302805e-01 4.50978160e-01
5.53948998e-01 -8.42034280e-01 1.08620393e+00 -3.25576901e-01
1.92099810e-01 -4.67495054e-01 1.01157382e-01 -1.09257972e+00
-3.50349307e-01 -4.75218326e-01 -2.17454601e-02 9.76115882e-01
5.61299562e-01 -7.10870385e-01 6.52915418e-01 4.71144736e-01
-3.69199425e-01 -1.09164965e+00 -9.48243976e-01 -7.65929520e-01
1.59177631e-01 -5.61376333e-01 6.19855285e-01 7.73635268e-01
-1.03690594e-01 1.32986829e-01 -1.25632539e-01 -1.66643355e-02
2.80745268e-01 2.28362352e-01 2.80098945e-01 -7.09455490e-01
-3.79525393e-01 -7.20230520e-01 -1.53024957e-01 -1.45480525e+00
-1.90956071e-01 -9.29947495e-01 -3.51950556e-01 -1.72963488e+00
8.72224569e-02 -5.90436280e-01 -4.59708631e-01 1.06104481e+00
-2.64353063e-02 4.58022863e-01 4.42996502e-01 2.82982469e-01
-3.58846843e-01 4.12573904e-01 1.56674635e+00 -2.63863325e-01
-3.33445191e-01 -1.40693516e-01 -6.55773759e-01 7.30204165e-01
8.60390067e-01 -5.94164208e-02 -4.88226324e-01 -8.80305111e-01
-1.19735457e-01 -1.42791435e-01 4.73364383e-01 -6.72106504e-01
1.03605673e-01 6.02471903e-02 6.56257093e-01 -5.79866290e-01
4.88639802e-01 -4.69717085e-01 -1.85485721e-01 3.21194828e-01
-1.87196761e-01 2.21102253e-01 4.33319837e-01 1.14223264e-01
-1.46364659e-01 4.28560264e-02 7.31230795e-01 -1.42775103e-01
-7.52679944e-01 3.62120271e-01 -2.23150298e-01 -2.95731366e-01
1.09003544e+00 -5.30835450e-01 -7.55211174e-01 -2.51031160e-01
-8.48606884e-01 4.98506576e-01 2.85471231e-01 2.57882178e-01
8.09872687e-01 -1.30404353e+00 -3.89587313e-01 3.64438206e-01
6.56725839e-02 1.82346284e-01 3.20493340e-01 1.03673410e+00
-5.93211472e-01 3.80290389e-01 -3.44260216e-01 -6.20185077e-01
-1.13515389e+00 6.05524898e-01 3.67814273e-01 -9.47448239e-02
-7.27357209e-01 8.87910724e-01 8.07499826e-01 -4.96207565e-01
4.15154278e-01 -5.71675479e-01 -1.10399276e-01 -2.16396540e-01
5.76742470e-01 2.82726914e-01 -4.78762947e-02 -1.41041443e-01
-4.98608649e-01 5.36917627e-01 -1.46292225e-01 -9.78314281e-02
8.71473908e-01 6.47024065e-02 -9.75763649e-02 -5.56694977e-02
9.15285707e-01 -3.38108540e-01 -1.58787942e+00 -1.61207214e-01
-2.34690860e-01 -3.28075558e-01 3.57274450e-02 -9.48907197e-01
-1.25180697e+00 1.00873327e+00 6.33873761e-01 6.41095936e-02
1.30233431e+00 2.54875422e-01 4.32081729e-01 2.02139944e-01
1.04621083e-01 -8.44202161e-01 1.64935905e-02 7.27680445e-01
1.15274930e+00 -6.47367299e-01 -2.16536626e-01 -6.07290924e-01
-6.08315587e-01 7.76968777e-01 9.94662702e-01 -1.91784382e-01
7.12111831e-01 5.70433497e-01 6.57932088e-02 -6.38218224e-02
-4.91979569e-01 -1.54805124e-01 2.13500068e-01 5.43164968e-01
4.46874738e-01 9.75442231e-02 -2.12244689e-01 6.71318233e-01
1.41541496e-01 -1.85624391e-01 1.25784606e-01 9.48860884e-01
-3.69656056e-01 -9.45363045e-01 -4.84578848e-01 3.29056770e-01
-2.42941186e-01 -1.10331193e-01 -4.80179697e-01 7.86716461e-01
6.60213083e-02 6.25478745e-01 5.11475027e-01 -2.37373874e-01
3.87709260e-01 -5.24315890e-03 5.53743064e-01 -4.30203795e-01
-5.89865685e-01 3.16795051e-01 -6.12856895e-02 -7.82149553e-01
4.94648479e-02 -3.34276259e-01 -1.63112772e+00 -1.29988119e-01
-1.95931762e-01 -1.12595163e-01 5.74305177e-01 6.22027159e-01
4.89762008e-01 6.64415419e-01 2.75182128e-01 -8.22622180e-01
-1.94617406e-01 -9.78199422e-01 -5.60589314e-01 1.31411225e-01
1.78071633e-01 -5.94256341e-01 1.87897220e-01 -8.52111429e-02] | [9.616273880004883, 0.6383764147758484] |
ab014ce3-8da3-4b1d-8d68-26c69a6daaff | multi-modal-pre-training-for-medical-vision | 2306.06494 | null | https://arxiv.org/abs/2306.06494v1 | https://arxiv.org/pdf/2306.06494v1.pdf | Multi-modal Pre-training for Medical Vision-language Understanding and Generation: An Empirical Study with A New Benchmark | With the availability of large-scale, comprehensive, and general-purpose vision-language (VL) datasets such as MSCOCO, vision-language pre-training (VLP) has become an active area of research and proven to be effective for various VL tasks such as visual-question answering. However, studies on VLP in the medical domain have so far been scanty. To provide a comprehensive perspective on VLP for medical VL tasks, we conduct a thorough experimental analysis to study key factors that may affect the performance of VLP with a unified vision-language Transformer. To allow making sound and quick pre-training decisions, we propose RadioGraphy Captions (RGC), a high-quality, multi-modality radiographic dataset containing 18,434 image-caption pairs collected from an open-access online database MedPix. RGC can be used as a pre-training dataset or a new benchmark for medical report generation and medical image-text retrieval. By utilizing RGC and other available datasets for pre-training, we develop several key insights that can guide future medical VLP research and new strong baselines for various medical VL tasks. | ['Xiao-Ming Wu', 'Lu Fan', 'Ameer Hamza Khan', 'Bo Liu', 'Li Xu'] | 2023-06-10 | null | null | null | null | ['visual-question-answering-1', 'medical-report-generation'] | ['computer-vision', 'medical'] | [ 4.68597770e-01 1.02569133e-01 -4.78062361e-01 -4.06919539e-01
-1.56026626e+00 -2.51162231e-01 5.55791795e-01 3.08969617e-01
-5.25892437e-01 5.96377730e-01 5.71359634e-01 -6.55247211e-01
2.72981077e-01 -4.05247480e-01 -7.51836598e-01 -4.57722902e-01
3.25066358e-01 5.08755624e-01 2.06293315e-01 7.65984803e-02
-5.15671149e-02 1.31118953e-01 -1.35252011e+00 8.69882166e-01
6.19720876e-01 7.55451143e-01 6.47889733e-01 7.96106219e-01
-3.60626765e-02 1.03865397e+00 -3.44228983e-01 -5.88603795e-01
1.37370583e-02 -3.67587864e-01 -9.72503543e-01 -1.13264646e-03
5.76653421e-01 -2.10066006e-01 -4.43278939e-01 7.66557574e-01
9.73709226e-01 -1.62126109e-01 8.34232986e-01 -1.07292104e+00
-1.23139298e+00 4.92351532e-01 -6.39877915e-01 4.70456749e-01
4.70533252e-01 5.83453476e-01 1.07223046e+00 -9.65632021e-01
9.80097353e-01 1.17727685e+00 4.68627721e-01 7.89085150e-01
-8.94545496e-01 -2.71247268e-01 -1.82350218e-01 2.84931421e-01
-1.16675544e+00 -3.93707275e-01 5.62869787e-01 -3.85862261e-01
7.78057218e-01 4.08222973e-01 3.49405795e-01 1.44773698e+00
2.93895781e-01 1.25510120e+00 1.06516242e+00 -4.90886778e-01
-1.98971614e-01 7.10792020e-02 2.11573794e-01 9.62989390e-01
2.16163680e-01 7.96263013e-03 -3.39810342e-01 -2.07456782e-01
6.70207500e-01 -2.58493930e-01 -8.15022349e-01 -3.60445492e-02
-1.52946484e+00 8.73397589e-01 7.36930966e-01 1.39362335e-01
-4.03395742e-01 1.86655566e-01 4.77038741e-01 1.18359059e-01
2.27769554e-01 5.28430462e-01 -1.04253463e-01 9.90680009e-02
-6.68450952e-01 6.87919781e-02 5.94517589e-01 6.72256529e-01
2.12129951e-01 -2.93506891e-01 -1.05080914e+00 1.15903616e+00
3.65251541e-01 8.20399344e-01 5.10393322e-01 -8.25690150e-01
7.55700231e-01 3.30119967e-01 -4.36233461e-01 -7.96929061e-01
-3.52502763e-01 -9.91868973e-02 -8.49627852e-01 -2.63158917e-01
2.86928773e-01 7.84270242e-02 -1.48026991e+00 1.57726634e+00
-3.75634269e-03 -7.64905810e-02 2.18600243e-01 1.08137417e+00
1.76206446e+00 6.85357213e-01 2.78234690e-01 -1.49479672e-01
1.66379118e+00 -1.24201679e+00 -6.50044918e-01 -2.59521872e-01
6.03716075e-01 -9.18844163e-01 1.47432888e+00 -4.14264537e-02
-9.97557521e-01 -5.45336962e-01 -7.24448383e-01 -4.29868639e-01
-6.34553730e-02 2.79811203e-01 7.80478597e-01 1.65799037e-01
-1.15100658e+00 -2.44821519e-01 -8.25109541e-01 -3.76259446e-01
6.10484660e-01 -9.55646038e-02 -4.12297994e-01 -8.61239612e-01
-9.62913930e-01 8.85056436e-01 2.40809470e-01 8.51395726e-02
-1.11095011e+00 -9.48524356e-01 -1.00970554e+00 -3.47310692e-01
4.07575935e-01 -1.18583071e+00 1.28950000e+00 -5.16744494e-01
-8.81706834e-01 1.44506669e+00 1.89054734e-03 -4.76009101e-01
5.67977846e-01 1.02306955e-01 -2.93808550e-01 6.44265115e-01
3.85758817e-01 1.19051313e+00 6.45660818e-01 -1.40470517e+00
-2.56471187e-01 -2.60749198e-02 1.01405177e-02 3.32841605e-01
8.60108286e-02 -7.09004849e-02 -1.25688958e+00 -7.80307770e-01
-2.48098210e-01 -9.82553840e-01 -3.74024451e-01 1.41142651e-01
-7.06620932e-01 -2.75045037e-01 1.91889629e-01 -8.15782487e-01
8.51390600e-01 -1.99678254e+00 -1.29248962e-01 -1.41808510e-01
4.55440193e-01 2.49147296e-01 -4.74463552e-01 1.00208975e-01
-1.08923092e-01 -2.44370680e-02 -3.18211436e-01 -4.86816823e-01
-4.94516134e-01 2.92593986e-01 -8.07737783e-02 1.82280272e-01
2.62270927e-01 1.47068286e+00 -8.93126905e-01 -1.16393900e+00
2.73896009e-01 4.35341597e-01 -5.00746310e-01 3.42110127e-01
-2.67057329e-01 6.72676265e-01 -5.43190241e-01 1.02905023e+00
2.43513808e-01 -8.78530145e-01 -3.58849287e-01 -3.99145961e-01
4.55371648e-01 -9.26973224e-02 -1.87445313e-01 1.91333222e+00
-5.46946049e-01 8.69088948e-01 -1.74677953e-01 -7.12561429e-01
3.66831809e-01 2.64067799e-01 6.77276194e-01 -1.13689327e+00
1.88307136e-01 8.98196921e-03 -2.19216019e-01 -1.00413096e+00
4.10655648e-01 -1.09682083e-01 7.90147297e-03 3.37667882e-01
-2.09090430e-02 -1.20913357e-01 2.97445834e-01 5.30101001e-01
1.32785141e+00 -5.62675357e-01 1.81875929e-01 2.25484535e-01
6.70050144e-01 3.80441129e-01 2.33273491e-01 1.14148736e+00
-2.78436452e-01 1.23662996e+00 7.43137673e-02 -8.52065384e-02
-7.51287520e-01 -1.31445968e+00 -2.98410058e-01 7.17076123e-01
5.92887439e-02 -2.82159716e-01 -3.33931327e-01 -7.64690280e-01
3.39087136e-02 5.94481111e-01 -6.10482752e-01 -1.10545404e-01
-4.99756813e-01 -8.58547628e-01 6.55761063e-01 4.73518819e-01
2.80463159e-01 -1.60580921e+00 -2.52822220e-01 -2.55882945e-02
-6.42086685e-01 -1.66809666e+00 -6.96210146e-01 -2.77746707e-01
-6.18146420e-01 -1.24210668e+00 -1.26924467e+00 -1.02969456e+00
8.60696197e-01 4.81354386e-01 1.66993701e+00 3.52970839e-01
-7.94073761e-01 1.02352500e+00 -4.29647118e-01 -4.77908820e-01
-7.93793261e-01 -1.03076838e-01 -4.47835088e-01 -3.15428644e-01
2.65002064e-02 1.50196910e-01 -8.38230312e-01 -2.29203671e-01
-9.98272300e-01 4.76339877e-01 1.13999820e+00 1.04987049e+00
1.04812324e+00 -7.63026416e-01 3.97654146e-01 -1.10893965e+00
9.19574022e-01 -4.90983218e-01 -2.82468110e-01 7.43863165e-01
-5.10064721e-01 1.38385087e-01 1.82970658e-01 -3.00167292e-01
-9.39690053e-01 -2.09489688e-01 -5.67393839e-01 -5.82198977e-01
1.64614655e-02 8.83382082e-01 3.18348885e-01 -5.97668346e-03
5.72054088e-01 4.80562091e-01 2.52724309e-02 -3.91202152e-01
7.26437390e-01 6.09485030e-01 9.95865345e-01 -5.21386862e-01
5.85335851e-01 4.30719495e-01 -1.99516967e-01 -7.88299441e-01
-1.16327131e+00 -6.99859798e-01 -1.38916776e-01 -1.64748818e-01
1.29292977e+00 -1.09079814e+00 -6.06205881e-01 1.76336631e-01
-1.14032364e+00 -1.75115362e-01 -6.46179542e-02 2.60849297e-01
-3.52723628e-01 5.07129729e-01 -8.36199224e-01 -3.70800287e-01
-9.65013981e-01 -1.62093937e+00 1.30435109e+00 1.47893578e-01
7.83804525e-03 -9.91192281e-01 1.94915026e-01 1.02056658e+00
1.21520653e-01 1.57762989e-01 1.16735828e+00 -2.48235062e-01
-6.78383350e-01 -1.70624182e-02 -5.64053297e-01 4.11293954e-01
6.33530468e-02 -1.12168446e-01 -7.64426768e-01 -4.40458506e-01
-2.67547637e-01 -8.26245248e-01 1.32002485e+00 5.64211369e-01
1.58525801e+00 1.35252029e-01 -5.36046445e-01 7.04825819e-01
1.49362648e+00 -1.71309695e-01 5.15010297e-01 1.51035652e-01
9.92436707e-01 1.87634826e-01 6.22810304e-01 3.78596894e-02
9.29260492e-01 4.93394196e-01 2.38620162e-01 -6.13874793e-01
-7.06885040e-01 -2.35053629e-01 7.26297498e-02 1.10185957e+00
1.48458317e-01 -2.94006616e-01 -1.26334488e+00 7.09396482e-01
-1.55215228e+00 -4.87338036e-01 -2.43077930e-02 1.80535042e+00
1.17727792e+00 -2.61055096e-03 -3.79589736e-01 -4.13313508e-01
3.70003194e-01 2.15660453e-01 -3.37651819e-01 1.97121337e-01
-9.92735103e-02 1.85773030e-01 5.86174250e-01 2.95745909e-01
-1.06042910e+00 7.45091856e-01 6.48229933e+00 8.22399735e-01
-1.13886034e+00 2.23068461e-01 6.87774420e-01 1.12988288e-02
-4.36729223e-01 -4.78950322e-01 -5.88081419e-01 3.72283429e-01
6.43715441e-01 -2.35945117e-02 -2.36193240e-01 6.96312845e-01
9.65432972e-02 -3.21076334e-01 -1.32350266e+00 1.66743875e+00
4.07343954e-01 -1.83959341e+00 5.68780959e-01 -1.37505993e-01
6.98579192e-01 4.89368320e-01 4.04151291e-01 5.75249910e-01
2.90313065e-01 -1.14709818e+00 2.29581609e-01 5.39496303e-01
1.20758212e+00 -1.77006826e-01 9.30839658e-01 5.55151589e-02
-8.41208518e-01 2.94855356e-01 -3.25799912e-01 6.84037149e-01
3.50813657e-01 4.86551583e-01 -1.16082346e+00 8.27322125e-01
5.79068661e-01 6.83842480e-01 -1.00727808e+00 1.23067820e+00
-1.13909833e-01 7.04588115e-01 -6.58510672e-03 2.95520335e-01
3.29982996e-01 2.84035623e-01 2.75057912e-01 1.17584133e+00
-1.02413163e-01 1.47698462e-01 4.25244361e-01 8.10057104e-01
-3.93536478e-01 2.61965841e-01 -5.35577536e-01 -2.43931755e-01
1.41377836e-01 1.17898321e+00 -7.34488368e-01 -3.65393937e-01
-8.94377351e-01 8.38151455e-01 2.66474336e-01 4.27575082e-01
-8.78213942e-01 2.49846160e-01 4.04949307e-01 -9.84282643e-02
-9.77803320e-02 3.72899510e-02 -7.07484111e-02 -1.30220640e+00
-5.33870384e-02 -1.10223567e+00 7.22993910e-01 -1.01069367e+00
-1.66434371e+00 6.87594295e-01 -5.96826561e-02 -1.21690714e+00
-2.69197404e-01 -6.42009199e-01 -3.13105762e-01 5.97105861e-01
-1.93270588e+00 -1.31049633e+00 -5.32261372e-01 8.68120015e-01
7.49722421e-01 -2.42669359e-01 6.74467862e-01 3.35228890e-01
-4.68746781e-01 7.42208004e-01 -1.05830766e-01 3.25222313e-01
9.99307871e-01 -1.01676416e+00 1.85188189e-01 6.33058190e-01
5.07227480e-01 3.51236582e-01 4.49142963e-01 -5.49482107e-01
-1.49283838e+00 -1.20647991e+00 4.29140478e-01 -9.67744410e-01
4.42970365e-01 -9.38959792e-02 -9.20220792e-01 5.56590617e-01
1.69395864e-01 1.20635919e-01 6.97724044e-01 -2.73501128e-01
-4.57043983e-02 1.42760187e-01 -8.44560266e-01 7.00504065e-01
8.80984008e-01 -7.67252803e-01 -8.01736295e-01 7.78924525e-01
1.07629049e+00 -7.30602086e-01 -8.95080507e-01 6.76914334e-01
1.51836425e-01 -4.74621117e-01 1.53030682e+00 -4.10909534e-01
8.15871418e-01 7.62285013e-03 5.30517772e-02 -9.25796568e-01
-1.21953376e-02 -5.92859127e-02 1.86313331e-01 9.41453516e-01
5.67538679e-01 -1.98202327e-01 6.58577740e-01 3.34681928e-01
-2.77125359e-01 -1.06512952e+00 -6.71052754e-01 -3.45976502e-01
9.18799564e-02 -7.95566320e-01 -1.20690607e-01 7.65820622e-01
-5.07076681e-01 5.23057282e-01 -3.31886500e-01 1.01940483e-01
6.05484009e-01 2.51774192e-01 6.53997660e-01 -7.28284001e-01
-3.82472336e-01 -2.89372355e-01 -9.89997983e-02 -1.00700736e+00
1.47836789e-01 -1.49511325e+00 1.68087274e-01 -2.16770959e+00
7.18035400e-01 -3.67255837e-01 -4.68058884e-01 6.60274386e-01
-6.46349072e-01 4.68481809e-01 2.40239039e-01 4.56548333e-01
-9.21148300e-01 5.63156664e-01 1.81246877e+00 -5.99602580e-01
7.98504800e-02 -1.40793145e-01 -6.56572998e-01 5.07487476e-01
1.74796909e-01 -3.68963778e-01 -5.95242202e-01 -6.38503253e-01
-1.38461962e-01 4.86493200e-01 3.96069437e-01 -8.16031516e-01
7.61783868e-02 -3.28970291e-02 2.37922445e-01 -9.71731067e-01
3.32384765e-01 -4.55897957e-01 -3.80558550e-01 4.42076385e-01
-5.09410441e-01 2.73022782e-02 1.26710400e-01 6.26507699e-01
-5.08573174e-01 -4.33701538e-02 6.62540197e-01 -3.74765933e-01
-9.85009491e-01 5.42688549e-01 -1.67252585e-01 5.56710839e-01
9.00184095e-01 1.06909417e-01 -4.93690878e-01 -4.10403013e-01
-5.46424448e-01 7.89513409e-01 2.47975260e-01 7.84597814e-01
1.09548223e+00 -1.08657169e+00 -1.07922089e+00 -7.48053640e-02
8.23694706e-01 1.55303791e-01 4.43797290e-01 8.37467968e-01
-7.21180677e-01 7.23107576e-01 -1.20548522e-02 -1.04009318e+00
-1.41218293e+00 6.53194427e-01 1.84212439e-02 -5.80456436e-01
-1.05128920e+00 9.99552846e-01 4.38830316e-01 -2.44222417e-01
2.60277569e-01 -5.48976719e-01 -2.96306878e-01 -3.01805139e-01
6.16161525e-01 -4.51597095e-01 2.76652500e-02 -5.99777281e-01
-4.61004347e-01 3.82917017e-01 -3.72658819e-01 2.35870462e-02
1.11010504e+00 -1.42838448e-01 8.39241520e-02 3.64737421e-01
1.22509086e+00 -3.43717575e-01 -7.52133608e-01 -5.05415022e-01
-4.71279845e-02 -2.17611209e-01 2.87171230e-02 -8.69704664e-01
-1.34921181e+00 8.09006512e-01 5.80744147e-01 -3.99393141e-01
1.04070389e+00 5.56240857e-01 9.10277307e-01 4.38375860e-01
4.04910266e-01 -7.21747100e-01 5.26663601e-01 3.03135902e-01
1.17297339e+00 -1.86426318e+00 -5.09107858e-02 -4.61156785e-01
-1.07648563e+00 7.96988308e-01 6.33361042e-01 3.38106900e-01
4.30741608e-01 3.12307235e-02 5.13058960e-01 -3.98149371e-01
-8.79006267e-01 -3.82271588e-01 7.72184730e-01 7.24317491e-01
5.55371642e-01 -9.61783808e-03 -2.62924969e-01 3.51834416e-01
-1.55144408e-01 1.14457197e-01 4.12232578e-01 7.11249828e-01
-3.84711474e-02 -9.05242503e-01 -3.15066576e-01 7.50696003e-01
-6.31636202e-01 -4.02400702e-01 2.88222488e-02 7.39387035e-01
-3.09445560e-01 8.21892142e-01 -2.89445430e-01 -4.44843546e-02
2.98654228e-01 -2.00683162e-01 5.43948770e-01 -8.61053884e-01
-2.49481365e-01 -3.18408102e-01 1.78502649e-01 -5.74663997e-01
-3.64458829e-01 -2.70479143e-01 -1.25335217e+00 9.42522734e-02
1.08157001e-01 -2.10504845e-01 3.68695349e-01 9.76313770e-01
1.80764332e-01 8.53627563e-01 2.02085525e-01 -3.30800921e-01
-3.18221986e-01 -8.24109554e-01 -6.36598170e-02 8.59093428e-01
2.51332313e-01 -6.10424101e-01 1.72829647e-02 1.83013290e-01] | [14.938253402709961, -1.6496202945709229] |
c86a11eb-c17f-4385-9879-75627ae16261 | weighted-programming | 2202.07577 | null | https://arxiv.org/abs/2202.07577v2 | https://arxiv.org/pdf/2202.07577v2.pdf | Weighted Programming | We study weighted programming, a programming paradigm for specifying mathematical models. More specifically, the weighted programs we investigate are like usual imperative programs with two additional features: (1) nondeterministic branching and (2) weighting execution traces. Weights can be numbers but also other objects like words from an alphabet, polynomials, formal power series, or cardinal numbers. We argue that weighted programming as a paradigm can be used to specify mathematical models beyond probability distributions (as is done in probabilistic programming). We develop weakest-precondition- and weakest-liberal-precondition-style calculi \`{a} la Dijkstra for reasoning about mathematical models specified by weighted programs. We present several case studies. For instance, we use weighted programming to model the ski rental problem - an optimization problem. We model not only the optimization problem itself, but also the best deterministic online algorithm for solving this problem as weighted programs. By means of weakest-precondition-style reasoning, we can determine the competitive ratio of the online algorithm on source code level. | ['Tobias Winkler', 'Joost-Pieter Katoen', 'Benjamin Lucien Kaminski', 'Adrian Gallus', 'Kevin Batz'] | 2022-02-15 | null | null | null | null | ['probabilistic-programming'] | ['methodology'] | [ 9.71672088e-02 2.98153311e-01 -3.19007009e-01 -3.84882569e-01
-9.14159238e-01 -8.49107623e-01 5.93653023e-01 2.95444280e-01
-5.96573889e-01 3.82487714e-01 -5.91078587e-02 -9.52313960e-01
-4.95417088e-01 -1.26052189e+00 -8.37897301e-01 -5.91241002e-01
-6.27050877e-01 9.79066730e-01 6.01364315e-01 -1.10353865e-01
4.14277762e-01 4.21718508e-01 -1.45683742e+00 2.84385741e-01
7.04443812e-01 4.97577608e-01 -2.63519920e-02 1.00872254e+00
-6.45223618e-01 9.73554432e-01 -6.55838549e-02 -7.05053091e-01
8.08660164e-02 1.89601406e-01 -8.92710984e-01 -7.11757362e-01
3.05897314e-02 -4.91673835e-02 -2.52314568e-01 1.31504154e+00
-7.26044998e-02 -1.01774514e-01 5.56359112e-01 -1.81393611e+00
-2.51286626e-01 1.60722208e+00 -5.47275424e-01 -7.56482929e-02
4.13530707e-01 1.01067409e-01 1.31745124e+00 -6.24439895e-01
3.89074862e-01 1.46885169e+00 5.47274649e-01 4.46791232e-01
-1.70837808e+00 -1.00766569e-01 1.04035020e-01 -9.14125964e-02
-1.16609108e+00 -8.39380771e-02 3.91524792e-01 -5.29787421e-01
8.20612848e-01 7.99735725e-01 3.68180275e-01 3.28456253e-01
1.50113836e-01 9.84225452e-01 1.20508718e+00 -8.45420539e-01
3.49916786e-01 2.45114863e-02 9.37710583e-01 1.04178464e+00
3.58650237e-01 4.17355001e-01 -4.04399097e-01 -1.05464268e+00
1.82070192e-02 -1.78259790e-01 1.31537065e-01 -4.30222362e-01
-1.51330304e+00 9.34357166e-01 -4.42880362e-01 9.05692279e-02
-6.10103607e-02 1.18486357e+00 4.69939888e-01 1.53928995e-01
1.00286767e-01 2.60698825e-01 -7.13676691e-01 -3.42541307e-01
-9.16499376e-01 7.24863648e-01 1.50358570e+00 1.09463072e+00
4.49755490e-01 -1.29029617e-01 -3.92932117e-01 3.56174320e-01
7.45460272e-01 7.15914905e-01 -6.93172961e-02 -1.04923320e+00
4.41828251e-01 1.98232070e-01 2.25973859e-01 -4.62056816e-01
-3.34307283e-01 2.12029800e-01 -2.82126993e-01 4.26410764e-01
9.63088036e-01 -6.99694380e-02 -5.57615876e-01 1.96013188e+00
1.48178220e-01 -6.91699386e-02 -2.60893196e-01 4.16268080e-01
2.26225033e-01 1.14079607e+00 2.21068993e-01 -1.91719204e-01
1.56973565e+00 -6.83944345e-01 -1.81244716e-01 -8.98597389e-03
9.10158694e-01 -3.85232538e-01 9.54610527e-01 7.18378305e-01
-1.48420835e+00 1.14554435e-01 -7.23886669e-01 8.04785416e-02
-2.01244235e-01 -4.12378103e-01 1.07007229e+00 1.02308750e+00
-1.21563911e+00 4.55162734e-01 -8.94670725e-01 2.00199127e-01
-1.43409878e-01 3.25043768e-01 4.23524566e-02 -3.95900309e-02
-1.02754807e+00 7.65926361e-01 5.51686525e-01 -1.56891599e-01
-9.07698095e-01 -9.38369155e-01 -8.87205720e-01 2.48324066e-01
5.99179804e-01 -4.24228132e-01 1.50296879e+00 -5.54543078e-01
-1.42629647e+00 9.31631505e-01 -2.76229590e-01 -5.43346882e-01
7.17350960e-01 4.00770932e-01 4.42898013e-02 -3.22556883e-01
-3.40302557e-01 -6.65357932e-02 5.05049884e-01 -1.19811475e+00
-7.97276199e-01 -3.18148941e-01 7.04423130e-01 -4.79059875e-01
1.53101802e-01 5.90217412e-01 -3.34990561e-01 -3.54065895e-01
7.88133964e-02 -1.11660039e+00 -6.63457215e-01 6.50607198e-02
-5.32247841e-01 -4.30152357e-01 -1.29451558e-01 -2.06565365e-01
1.39021111e+00 -2.17793131e+00 2.56686032e-01 8.40058088e-01
4.92388785e-01 -2.93832690e-01 -1.07085817e-01 4.35863405e-01
1.21691369e-01 4.27278578e-01 -3.35388392e-01 2.82404274e-01
9.75212276e-01 2.74095863e-01 -4.12082613e-01 4.54504520e-01
-1.13416664e-01 8.44500601e-01 -1.08742869e+00 -6.27331197e-01
-2.08636746e-01 -2.37509489e-01 -7.34387159e-01 -2.06865501e-02
-1.00245631e+00 -7.03062236e-01 -4.45245296e-01 4.39053863e-01
8.16401422e-01 -1.07885115e-02 5.14217496e-01 1.93113893e-01
-2.86068112e-01 1.77437603e-01 -1.64704037e+00 1.29844141e+00
-7.34427452e-01 2.80779958e-01 1.91137746e-01 -6.14288747e-01
5.28052330e-01 -5.80815598e-02 3.67000043e-01 -4.63224296e-03
-1.17176399e-01 2.52400368e-01 5.60713410e-02 -3.73188674e-01
6.81715190e-01 -1.45163149e-01 -6.88765824e-01 1.00096428e+00
-2.17031911e-01 -6.31993592e-01 7.39987910e-01 4.76596236e-01
1.14482319e+00 2.88088620e-01 6.78944886e-02 -8.43936265e-01
4.24716473e-01 2.01226652e-01 4.89124179e-01 1.20723748e+00
1.68613628e-01 1.33840084e-01 1.25374162e+00 -4.66568470e-01
-8.77693951e-01 -1.25960147e+00 3.13174836e-02 1.72462988e+00
4.13360400e-03 -7.83651054e-01 -4.95593220e-01 -4.54368025e-01
1.43672034e-01 1.22590792e+00 -4.16240066e-01 1.05204936e-02
-4.86098945e-01 -8.05536866e-01 7.98895717e-01 4.22435343e-01
-2.45995730e-01 -5.65384984e-01 -7.43062079e-01 4.09426302e-01
1.34104103e-01 -5.82001507e-01 -6.84196651e-01 6.75716460e-01
-8.15557897e-01 -9.56915259e-01 -2.53832042e-01 -5.30176580e-01
6.80051923e-01 -1.88276753e-01 1.41241193e+00 1.91177681e-01
5.67557802e-03 4.87429231e-01 6.17232844e-02 -6.30241871e-01
-6.70697570e-01 -3.18537742e-01 -1.42114282e-01 -3.34473968e-01
2.67169058e-01 -2.81183630e-01 1.46527842e-01 2.57503223e-02
-1.01424682e+00 1.05783548e-02 6.36940300e-02 7.82689095e-01
4.75240469e-01 4.49982882e-01 -3.90653342e-01 -1.33551228e+00
7.41032004e-01 -4.27671015e-01 -1.37134063e+00 6.66223109e-01
-4.07487541e-01 8.55431497e-01 5.42961538e-01 -6.87071145e-01
-9.44388449e-01 1.18796512e-01 1.14547469e-01 2.94335306e-01
4.31700826e-01 9.67958152e-01 -1.18900545e-01 -5.59256896e-02
6.71636701e-01 4.35949683e-01 -1.80043772e-01 2.10950702e-01
6.60468519e-01 1.07998982e-01 3.78395021e-01 -1.95449030e+00
9.22909319e-01 2.71992445e-01 3.82123023e-01 -1.54250264e-01
-2.91342378e-01 8.28280672e-02 -9.04369801e-02 -2.29240172e-02
2.71435976e-01 -3.14185619e-01 -1.21397138e+00 2.88476288e-01
-1.47791398e+00 -6.66137695e-01 -3.85731101e-01 2.58134902e-01
-7.93073893e-01 3.42921048e-01 -6.49971604e-01 -1.41588533e+00
-1.23077951e-01 -1.28238940e+00 7.40305126e-01 -2.38019302e-01
-4.05714035e-01 -8.69019151e-01 1.74033538e-01 -1.86904475e-01
4.08648014e-01 1.08987369e-01 1.57004929e+00 -7.74815619e-01
-6.58047199e-01 -3.55176270e-01 -2.53792495e-01 -4.58023569e-05
-9.65647399e-01 8.46167445e-01 -5.66904426e-01 4.23352152e-01
-2.50896484e-01 -8.33716150e-03 3.57241303e-01 2.68995136e-01
1.57900143e+00 -7.16937244e-01 -2.37717837e-01 2.36743510e-01
1.54936469e+00 -6.64174557e-02 5.78897178e-01 1.28936946e-01
2.95548856e-01 5.61495185e-01 4.46178705e-01 6.12138629e-01
6.65085495e-01 4.07572746e-01 4.14343596e-01 8.17773163e-01
6.16555512e-01 -2.01832652e-01 3.85859400e-01 6.84169292e-01
-2.72609770e-01 1.91849157e-01 -1.27247274e+00 5.54298460e-01
-1.88689733e+00 -1.08638537e+00 -5.61492622e-01 2.44323039e+00
1.30008185e+00 4.40412074e-01 2.79019415e-01 -1.76915731e-02
5.51021039e-01 2.51658689e-02 -1.12833837e-02 -9.86990392e-01
3.40631336e-01 5.20195127e-01 8.73316765e-01 8.03835392e-01
-6.14156723e-01 4.63833094e-01 6.44172668e+00 1.16782892e+00
-6.31624639e-01 2.96907693e-01 3.75210404e-01 -7.47542903e-02
-1.13122940e+00 4.36462730e-01 -5.87477148e-01 7.15611100e-01
1.18217778e+00 -7.20884800e-01 9.34174299e-01 1.26636815e+00
-4.31547403e-01 -2.03173250e-01 -1.60389626e+00 6.90775633e-01
-5.55792987e-01 -1.26219523e+00 -4.84419256e-01 2.54431330e-02
4.21900988e-01 -2.31701896e-01 -7.49583021e-02 5.99309385e-01
1.23278463e+00 -8.80477965e-01 1.46301258e+00 7.07314789e-01
5.13764143e-01 -7.70063937e-01 5.01228034e-01 4.65246588e-01
-1.06764674e+00 -2.28561070e-02 -1.87821209e-01 1.61876619e-01
1.25258371e-01 9.26714301e-01 -4.15724069e-02 2.89252371e-01
2.85664052e-01 -3.60822767e-01 1.71532214e-01 1.10915339e+00
-1.79046810e-01 6.63329065e-01 -9.22509491e-01 -5.41384101e-01
1.50492996e-01 -2.82064021e-01 3.57648909e-01 1.40284145e+00
4.46777046e-01 2.13012695e-01 8.43193904e-02 1.11095917e+00
2.76420396e-02 -1.79349691e-01 -3.65857422e-01 -1.10466264e-01
4.91306245e-01 9.64437425e-01 -4.78973716e-01 -4.05721605e-01
-4.99466807e-01 9.91862342e-02 2.04242170e-01 1.13916494e-01
-1.10494983e+00 -4.60905313e-01 3.21056575e-01 -2.91415844e-02
-2.66254306e-01 -5.47647834e-01 -5.75337052e-01 -1.03336895e+00
-1.09022312e-01 -9.41463351e-01 5.87705910e-01 -6.01470351e-01
-1.05113971e+00 2.41679043e-01 4.90051419e-01 -4.61647332e-01
-1.39911398e-01 -9.16275799e-01 -6.15781963e-01 9.60258722e-01
-1.12546098e+00 -8.12913060e-01 2.42898554e-01 2.37382531e-01
2.99904123e-02 2.70627916e-01 8.77605438e-01 -6.62170947e-02
-2.69077212e-01 3.45870793e-01 7.84044787e-02 -3.06542188e-01
8.26594606e-02 -1.60438228e+00 8.64535123e-02 9.11206424e-01
1.49709404e-01 8.27602565e-01 1.01633370e+00 -1.26392663e-01
-2.39437103e+00 -6.99570060e-01 1.18465447e+00 -5.25844455e-01
1.24604332e+00 -3.58990610e-01 -5.76440215e-01 6.60536468e-01
-1.83391511e-01 1.26399070e-01 3.02321583e-01 3.41032982e-01
-8.13064337e-01 -2.52955049e-01 -1.32883489e+00 9.25180078e-01
8.16446662e-01 -5.41685045e-01 -5.37919819e-01 3.99489105e-01
8.18563044e-01 -5.80728531e-01 -7.75081694e-01 2.57434440e-04
9.02769983e-01 -6.96407199e-01 9.17879581e-01 -5.98349392e-01
4.11708623e-01 -3.41904759e-01 -4.23233390e-01 -9.08222377e-01
-2.61248291e-01 -9.33907330e-01 -5.84062159e-01 9.23707485e-01
8.17725182e-01 -8.64705741e-01 6.61230445e-01 1.38305891e+00
1.82704687e-01 -6.17025971e-01 -8.66127610e-01 -8.39076161e-01
3.65720391e-01 -1.29364097e+00 9.83083010e-01 5.45997202e-01
4.68193412e-01 -4.27535951e-01 -1.26241237e-01 2.18302578e-01
9.38229799e-01 3.51380646e-01 5.92806101e-01 -1.12585652e+00
-8.06376219e-01 -9.03042197e-01 -1.62747279e-01 -5.88105202e-01
4.37303632e-01 -1.15621293e+00 3.62053424e-01 -1.13333833e+00
3.48907173e-01 -9.84349966e-01 -2.81136595e-02 6.63160861e-01
7.64272735e-02 -3.96094084e-01 2.02687502e-01 -2.93449968e-01
-5.71656942e-01 -1.37605354e-01 6.43487930e-01 -4.26263154e-01
1.11468092e-01 2.04675794e-01 -6.37316525e-01 6.99752688e-01
6.22157574e-01 -7.96908021e-01 -2.69183815e-01 -4.58277196e-01
1.25736582e+00 5.31196117e-01 2.94091582e-01 -5.16542137e-01
4.88876820e-01 -9.33356702e-01 -8.04512799e-01 -3.65435064e-01
-1.59341738e-01 -9.27231908e-01 3.95333976e-01 7.97040939e-01
-7.59648025e-01 1.26374334e-01 1.06320292e-01 3.59177768e-01
7.83695728e-02 -1.04232192e+00 6.48739219e-01 -1.07000977e-01
-6.31854713e-01 3.17203045e-01 -7.64929593e-01 3.27251405e-01
1.05133045e+00 2.32492596e-01 -3.55356812e-01 -1.04990542e-01
-6.93892658e-01 3.56231123e-01 3.41284037e-01 -3.61667275e-01
3.73756409e-01 -1.09411705e+00 -7.23049700e-01 -2.48559207e-01
3.37644309e-01 -5.58710694e-02 -5.33156283e-02 9.01040614e-01
-9.72337484e-01 1.52530432e-01 2.50031680e-01 -2.18751445e-01
-1.12869883e+00 8.34267497e-01 4.10544351e-02 -5.67449152e-01
3.42792161e-02 8.19352686e-01 -1.92510560e-01 -8.27501059e-01
2.88490862e-01 -8.21523309e-01 6.63072765e-01 -3.12860638e-01
5.79956293e-01 6.60437584e-01 -3.33824784e-01 2.72482187e-01
-5.41882396e-01 1.95322111e-01 2.94511348e-01 -5.03600299e-01
1.32633305e+00 2.70848900e-01 -1.02298319e+00 5.25651932e-01
6.51097953e-01 2.94067174e-01 -5.85207224e-01 -3.96007955e-01
4.31706220e-01 -4.32006091e-01 -2.08437353e-01 -6.57902122e-01
-8.01935196e-01 7.41321445e-01 2.10687071e-02 6.93623126e-01
5.90286613e-01 6.04318865e-02 2.01418191e-01 5.34412563e-01
1.17007256e+00 -8.51683915e-01 -5.36784470e-01 7.11244464e-01
6.18539810e-01 -6.68243289e-01 -9.60010067e-02 -2.83358306e-01
-3.49762410e-01 1.53055716e+00 1.33248106e-01 -1.58243299e-01
8.68286908e-01 1.06941295e+00 -6.28414869e-01 1.90007035e-02
-1.04325604e+00 -4.03983481e-02 -5.51769808e-02 5.18031597e-01
2.63526082e-01 4.35964167e-01 -5.57152271e-01 1.03946114e+00
-9.21451002e-02 2.09479094e-01 7.36620545e-01 1.06913054e+00
-3.16054285e-01 -1.37771213e+00 -6.29736006e-01 6.78794682e-01
-5.82269847e-01 -4.35238987e-01 3.74895245e-01 6.45014226e-01
-2.39666864e-01 6.96104467e-01 -2.19297558e-01 -1.52071193e-01
7.70733729e-02 9.90121514e-02 7.67086387e-01 -6.67779446e-01
-4.51062113e-01 -5.74202001e-01 4.30830777e-01 -4.12965566e-01
5.08662201e-02 -5.57538390e-01 -1.22249603e+00 -9.46270704e-01
-2.54002571e-01 3.44384074e-01 9.31183755e-01 6.87115729e-01
-2.31410354e-01 1.88602477e-01 4.11773473e-01 -5.04936457e-01
-1.29729354e+00 -3.80650163e-01 -8.19778800e-01 -3.69485915e-01
-5.39390594e-02 -1.33702867e-02 -4.75965321e-01 -7.35309646e-02] | [8.469956398010254, 6.597517967224121] |
c59c867d-f0d9-4b9d-9c2f-889573c59c26 | compguesswhat-a-multi-task-evaluation | 2006.02174 | null | https://arxiv.org/abs/2006.02174v1 | https://arxiv.org/pdf/2006.02174v1.pdf | CompGuessWhat?!: A Multi-task Evaluation Framework for Grounded Language Learning | Approaches to Grounded Language Learning typically focus on a single task-based final performance measure that may not depend on desirable properties of the learned hidden representations, such as their ability to predict salient attributes or to generalise to unseen situations. To remedy this, we present GROLLA, an evaluation framework for Grounded Language Learning with Attributes with three sub-tasks: 1) Goal-oriented evaluation; 2) Object attribute prediction evaluation; and 3) Zero-shot evaluation. We also propose a new dataset CompGuessWhat?! as an instance of this framework for evaluating the quality of learned neural representations, in particular concerning attribute grounding. To this end, we extend the original GuessWhat?! dataset by including a semantic layer on top of the perceptual one. Specifically, we enrich the VisualGenome scene graphs associated with the GuessWhat?! images with abstract and situated attributes. By using diagnostic classifiers, we show that current models learn representations that are not expressive enough to encode object attributes (average F1 of 44.27). In addition, they do not learn strategies nor representations that are robust enough to perform well when novel scenes or objects are involved in gameplay (zero-shot best accuracy 50.06%). | ['Stella Frank', 'Alessandro Suglia', 'Ioannis Konstas', 'Emanuele Bastianelli', 'Desmond Elliott', 'Andrea Vanzo', 'Oliver Lemon'] | 2020-06-03 | compguesswhat-a-multi-task-evaluation-1 | https://aclanthology.org/2020.acl-main.682 | https://aclanthology.org/2020.acl-main.682.pdf | acl-2020-6 | ['grounded-language-learning'] | ['natural-language-processing'] | [ 1.69457436e-01 6.14845693e-01 -1.11359507e-02 -6.88836992e-01
-8.75665545e-01 -4.21047449e-01 8.67461145e-01 6.37824535e-01
-3.27280790e-01 6.28700376e-01 2.66849637e-01 5.96850142e-02
-2.35720485e-01 -1.08559752e+00 -8.83262098e-01 -5.87439358e-01
-1.25271857e-01 7.93577552e-01 1.81240097e-01 -4.67384636e-01
1.97263345e-01 2.63773829e-01 -2.24119258e+00 8.08700025e-01
7.14239299e-01 1.24838543e+00 1.27376035e-01 4.69032794e-01
-2.14324728e-01 1.34047711e+00 -5.71854651e-01 -5.90925574e-01
-7.18025938e-02 -5.65179646e-01 -1.14762831e+00 5.27403690e-03
5.52166402e-01 -1.71176121e-01 1.59781530e-01 9.09332633e-01
3.43432337e-01 4.07078505e-01 8.15530777e-01 -1.40641224e+00
-1.04335344e+00 8.94897997e-01 1.54794827e-01 1.21490784e-01
4.89838570e-01 1.69250205e-01 1.33496666e+00 -1.04437757e+00
6.60474420e-01 1.42704642e+00 6.84333444e-01 7.88723052e-01
-1.22376478e+00 -4.00573909e-01 3.50758106e-01 4.42133814e-01
-1.27469301e+00 -7.06427932e-01 8.35758686e-01 -4.24458921e-01
9.71351326e-01 1.70629904e-01 8.65296841e-01 1.28213930e+00
-1.64614409e-01 1.03256238e+00 1.17220759e+00 -4.42364037e-01
5.21003008e-01 2.53356814e-01 1.01599321e-01 6.57626748e-01
2.01272219e-01 2.18680173e-01 -7.47148752e-01 8.85638297e-02
2.32219309e-01 -3.61602694e-01 -1.16797119e-01 -7.07933903e-01
-1.37817776e+00 9.20537114e-01 6.73732996e-01 1.81055754e-01
-4.89515424e-01 2.92094558e-01 3.69743615e-01 3.81183892e-01
4.47865248e-01 9.93116260e-01 -1.15700990e-01 -1.20409310e-01
-6.07951224e-01 3.22258115e-01 5.79512119e-01 1.10753572e+00
9.59475577e-01 3.33902180e-01 -4.09934610e-01 7.01669037e-01
1.25627801e-01 2.26506129e-01 3.02349865e-01 -9.07849789e-01
1.14037499e-01 7.85789073e-01 -4.05808212e-03 -8.60158265e-01
-4.15684313e-01 -4.41262543e-01 -5.96309483e-01 1.84583560e-01
3.28129619e-01 3.14062148e-01 -7.71584690e-01 1.91597211e+00
1.75223529e-01 -7.52802566e-02 3.06670189e-01 9.18934584e-01
1.32809019e+00 5.54519713e-01 6.32897139e-01 2.62457639e-01
1.18390512e+00 -7.24228024e-01 -6.54902637e-01 -4.87841636e-01
9.13527668e-01 -2.05411226e-01 1.74450481e+00 2.85410166e-01
-1.21371198e+00 -7.43315279e-01 -1.21502709e+00 -2.74657965e-01
-8.75422299e-01 -6.46769553e-02 7.03640461e-01 5.21735728e-01
-1.09390759e+00 5.28173923e-01 -2.05257565e-01 -3.50089341e-01
5.40822327e-01 1.28370523e-01 -3.26062828e-01 -2.55104117e-02
-1.38932049e+00 1.13929963e+00 7.81012654e-01 -2.26528659e-01
-1.36197340e+00 -5.33832312e-01 -1.14996743e+00 2.49568164e-01
5.04130244e-01 -8.49145412e-01 1.04261410e+00 -1.15455317e+00
-1.16736424e+00 1.34427702e+00 1.89681321e-01 -4.70052570e-01
1.97094053e-01 -2.20887765e-01 -3.83521408e-01 1.37405530e-01
1.94440156e-01 8.21556747e-01 7.10105121e-01 -1.51242638e+00
-4.63024586e-01 -4.06078398e-01 6.44990683e-01 3.27518106e-01
-2.98131704e-01 -3.19794863e-01 2.62200981e-01 -5.42241871e-01
-6.55378699e-02 -7.12125599e-01 -2.85916943e-02 -8.32419172e-02
-3.71713489e-01 -3.62645090e-01 4.26607877e-01 -1.48389846e-01
7.93437123e-01 -2.30591083e+00 1.59291670e-01 -1.00499220e-01
2.63144374e-01 1.59958843e-02 -3.86134654e-01 3.75049293e-01
-6.91727251e-02 2.90281415e-01 -6.53519258e-02 -3.80617887e-01
1.93108782e-01 2.50481158e-01 -3.71298403e-01 2.06380621e-01
3.26987833e-01 1.11788166e+00 -1.24949920e+00 -4.47839290e-01
2.35554412e-01 2.02095374e-01 -6.80299282e-01 3.29714447e-01
-2.94171751e-01 1.74860388e-01 -2.98091143e-01 6.73461139e-01
2.96498779e-02 -1.40313849e-01 -1.75089940e-01 -9.93415266e-02
1.54384777e-01 4.70936269e-01 -1.05193949e+00 1.80715227e+00
-6.20925248e-01 5.70703804e-01 -3.19437057e-01 -9.53005373e-01
1.17974281e+00 1.72502726e-01 1.80256277e-01 -8.05667937e-01
-1.53390234e-02 5.98812327e-02 -1.06337048e-01 -4.19359714e-01
7.58695781e-01 -3.56676221e-01 -3.23926777e-01 2.99689204e-01
4.96066093e-01 -5.12895823e-01 3.40496562e-02 3.07898998e-01
8.06536615e-01 3.74676883e-01 5.16238451e-01 -6.66089594e-01
3.57673585e-01 2.45259404e-01 1.59025714e-01 9.82446730e-01
-9.95574296e-02 4.93203819e-01 4.95710403e-01 -7.42625952e-01
-1.02343357e+00 -1.10893667e+00 1.49447590e-01 1.67316735e+00
1.90269828e-01 -6.26915574e-01 -7.14743733e-01 -5.62553465e-01
-2.36970037e-01 1.23662341e+00 -9.98372436e-01 -7.50646055e-01
-8.75087753e-02 -2.03573793e-01 5.15149891e-01 6.91343009e-01
4.63031173e-01 -1.54793072e+00 -9.90644336e-01 8.39421824e-02
-2.62467980e-01 -9.87839341e-01 2.74592787e-01 4.65458244e-01
-6.01243675e-01 -1.05670369e+00 -1.98002189e-01 -5.92118859e-01
4.05161440e-01 -5.05106226e-02 1.82799947e+00 1.67522162e-01
1.66400760e-01 7.47753263e-01 -6.90594733e-01 -8.15675318e-01
-4.10038561e-01 9.20295864e-02 -1.23770274e-01 6.17450383e-03
5.05184472e-01 -3.41800749e-01 -2.95970261e-01 1.91941217e-01
-8.13881099e-01 1.47634968e-01 6.02846980e-01 8.83590698e-01
6.83437109e-01 -3.75499159e-01 5.96623719e-01 -1.03006351e+00
8.04657221e-01 -5.99214733e-01 -2.07907215e-01 4.46201652e-01
-5.25858223e-01 1.53873831e-01 6.13373280e-01 -3.89418572e-01
-8.70485306e-01 -2.66637895e-02 -1.80494010e-01 -2.81385690e-01
-4.94910687e-01 5.84888518e-01 -4.18387711e-01 2.64662266e-01
1.21619534e+00 2.28020266e-01 -2.65175343e-01 -1.63591757e-01
6.72283351e-01 3.72684956e-01 5.64578533e-01 -1.01988554e+00
5.33604681e-01 3.40840131e-01 -5.94732212e-03 -8.15957665e-01
-1.35131025e+00 -3.24365526e-01 -6.65149271e-01 -3.67073208e-01
8.56128871e-01 -8.06437492e-01 -7.73312688e-01 1.23543233e-01
-9.12771583e-01 -6.24502957e-01 -1.05885589e+00 3.24824691e-01
-1.24391103e+00 -2.17691272e-01 -3.21115971e-01 -8.24257314e-01
-1.31781593e-01 -9.87062156e-01 1.21148455e+00 -2.25864336e-01
-4.58032578e-01 -8.56997907e-01 -2.33400613e-01 2.31031269e-01
4.28535342e-01 5.80585480e-01 1.23510218e+00 -9.54210758e-01
-4.07977849e-01 5.83462082e-02 2.24661734e-02 2.36366719e-01
-2.63253331e-01 -3.42543811e-01 -1.32570314e+00 -8.74691382e-02
-1.52831115e-02 -1.07243490e+00 8.61797094e-01 1.52726039e-01
1.41190255e+00 -4.65219408e-01 9.97761488e-02 5.97605109e-01
1.21566343e+00 1.54275537e-01 8.51038456e-01 5.66190422e-01
6.19818747e-01 8.41415584e-01 7.72010088e-01 3.79891634e-01
4.63831097e-01 7.13937342e-01 7.98023522e-01 5.22413142e-02
-2.76596278e-01 -6.59960985e-01 1.97062105e-01 2.29046255e-01
-2.83528179e-01 -1.97067842e-01 -1.22176743e+00 7.17819154e-01
-1.73076081e+00 -1.14410174e+00 4.28643040e-02 2.06897879e+00
5.89593530e-01 9.69088301e-02 1.29503027e-01 3.41850489e-01
4.05489892e-01 1.98494375e-01 -4.48225856e-01 -5.31343579e-01
-3.74061197e-01 1.89142793e-01 -3.04776043e-01 2.22715691e-01
-9.68334377e-01 1.19332385e+00 6.30839825e+00 5.88157296e-01
-6.03992164e-01 1.85772210e-01 6.13901258e-01 1.04843192e-01
-6.91967309e-01 -1.48506099e-02 -4.94337201e-01 -1.79445632e-02
9.69204545e-01 6.25680154e-03 3.16301227e-01 1.14457798e+00
-4.62089837e-01 7.45327026e-02 -1.57540679e+00 9.17958677e-01
2.01052874e-01 -1.44319332e+00 5.02771020e-01 -1.14268705e-01
3.56773049e-01 -4.09701198e-01 1.55891567e-01 8.24350476e-01
4.46084440e-01 -1.37599456e+00 1.17003846e+00 7.02181816e-01
7.33944893e-01 -7.97993183e-01 6.31374717e-01 3.64989758e-01
-9.18200970e-01 -2.53575239e-02 -4.44731981e-01 -2.61487484e-01
-2.81272292e-01 1.26018211e-01 -8.54363680e-01 2.05421627e-01
8.40111732e-01 7.07602441e-01 -8.16248953e-01 7.43571639e-01
-5.78130543e-01 4.00246680e-01 1.05517022e-01 -3.37395638e-01
4.64940041e-01 3.17007333e-01 5.22574842e-01 1.05357134e+00
2.89365262e-01 2.64937252e-01 7.26047605e-02 1.06811225e+00
-7.35125737e-03 1.76770136e-01 -9.68094826e-01 1.71940923e-02
5.74651659e-01 1.02407372e+00 -5.52369833e-01 -3.83012772e-01
-1.77929625e-01 6.42348289e-01 6.36565447e-01 2.96822488e-01
-5.06453395e-01 -1.89337477e-01 6.25542998e-01 3.72526288e-01
5.65969944e-02 1.01726927e-01 -1.59175903e-01 -1.07478392e+00
-2.71180183e-01 -1.00080311e+00 5.29632807e-01 -1.22573245e+00
-1.26859498e+00 7.94982016e-01 1.65739417e-01 -1.04197669e+00
-4.81018990e-01 -6.74673438e-01 -2.46094763e-01 5.32486975e-01
-1.34931970e+00 -1.50684547e+00 -6.72642648e-01 9.22475219e-01
5.25935471e-01 -3.43497157e-01 1.21128237e+00 -2.54619777e-01
9.28758830e-02 3.99566799e-01 -5.75341225e-01 8.18356350e-02
4.40694034e-01 -1.49897051e+00 3.68584067e-01 6.58985972e-01
4.87463087e-01 5.25636196e-01 7.70717859e-01 -3.36152792e-01
-1.06703091e+00 -9.22307551e-01 8.97663713e-01 -7.74691045e-01
4.21602875e-01 -6.80255353e-01 -1.03084970e+00 7.98086345e-01
6.67043477e-02 2.59744793e-01 8.21016192e-01 6.55477941e-01
-6.22750461e-01 3.60628441e-02 -1.21775436e+00 6.09373093e-01
1.39584601e+00 -6.81177795e-01 -8.92701268e-01 2.98832834e-01
6.81285322e-01 -2.79425561e-01 -7.91727304e-01 4.26827163e-01
2.28371337e-01 -1.27333570e+00 1.13171661e+00 -1.11273766e+00
5.92776299e-01 -8.54206085e-02 -6.86293542e-01 -1.45268393e+00
-6.74380302e-01 -2.25747153e-01 -1.88670948e-01 1.19220519e+00
3.79745930e-01 -3.15432101e-01 6.59216702e-01 4.71267819e-01
-5.94161868e-01 -8.05791080e-01 -8.69954646e-01 -7.20999122e-01
5.19746095e-02 -7.13768482e-01 9.36682463e-01 1.12863338e+00
-7.30884597e-02 6.32628977e-01 -2.02135190e-01 -1.26549110e-01
5.32166541e-01 1.17147051e-01 7.11832523e-01 -1.45997691e+00
-4.58625406e-02 -3.97883683e-01 -6.98113441e-01 -5.44815361e-01
4.42255169e-01 -9.72267509e-01 2.90005859e-02 -1.68118370e+00
8.35445747e-02 -6.32437646e-01 -3.66054028e-01 8.33036363e-01
-4.78443429e-02 1.06030636e-01 3.59085679e-01 1.70496300e-01
-8.70575726e-01 7.01545238e-01 1.12782133e+00 -3.47129047e-01
1.08822718e-01 -2.69548804e-01 -9.02917147e-01 7.99196303e-01
8.19652259e-01 -3.76843154e-01 -7.32003868e-01 -3.88480067e-01
5.63819587e-01 -1.53880060e-01 8.04617226e-01 -1.20255148e+00
3.59398685e-02 -1.73722118e-01 4.64350760e-01 -1.39435574e-01
6.16346955e-01 -7.54174650e-01 -5.28320894e-02 1.19817257e-01
-8.14159036e-01 1.39522746e-01 2.12699443e-01 3.53397787e-01
-3.91315550e-01 -4.16292518e-01 7.00592518e-01 -3.53392094e-01
-1.39995670e+00 1.37609631e-01 -1.78588122e-01 3.52249026e-01
1.04805052e+00 -5.05013824e-01 -5.06205082e-01 -4.86889362e-01
-8.92295778e-01 -1.00538388e-01 5.41793704e-01 5.80036640e-01
9.42284465e-01 -1.53614199e+00 -7.67921209e-01 7.05604926e-02
8.13445449e-01 -1.11397095e-01 1.25733256e-01 2.82838076e-01
-1.28952086e-01 1.17158055e-01 -5.94603837e-01 -6.41539574e-01
-8.04252565e-01 9.52871680e-01 4.20913070e-01 4.15017270e-02
-4.63200778e-01 9.85885441e-01 2.31033280e-01 -3.73621285e-01
1.28841832e-01 -1.77782357e-01 -4.20753837e-01 4.19287115e-01
3.30190122e-01 7.86560774e-02 2.22722128e-01 -8.34038496e-01
-3.79994214e-01 2.97983915e-01 2.81089634e-01 -3.94928828e-02
1.22496045e+00 3.08934208e-02 1.67180568e-01 8.82604301e-01
8.35213959e-01 -3.01601261e-01 -1.07237554e+00 -2.10291192e-01
1.40344262e-01 -2.34441936e-01 -6.32612184e-02 -8.93694341e-01
-8.31900001e-01 1.08369696e+00 4.29324180e-01 2.23544806e-01
1.06839657e+00 4.04820889e-01 8.03534910e-02 7.49636889e-01
5.92984438e-01 -1.00766599e+00 4.82666254e-01 5.04092276e-01
1.23796964e+00 -1.25827336e+00 -1.88683078e-01 -9.37824994e-02
-1.00659394e+00 9.35890615e-01 6.96017027e-01 8.55221897e-02
2.94641107e-01 -7.14402422e-02 -5.92465885e-02 -5.46774924e-01
-8.68264139e-01 -5.31236768e-01 6.03993058e-01 1.08863425e+00
5.02532005e-01 1.20057561e-01 2.96711564e-01 6.62845492e-01
-5.75442791e-01 -3.85456711e-01 3.85661036e-01 6.73741758e-01
-5.49652696e-01 -5.25682688e-01 7.46869529e-03 3.48497510e-01
-1.27015278e-01 -1.77787632e-01 -6.82903051e-01 1.00867486e+00
2.63223886e-01 9.35205817e-01 8.94880891e-02 -4.77064371e-01
5.28938770e-01 1.88436374e-01 4.04263616e-01 -9.52886164e-01
-6.91956997e-01 -7.10170627e-01 4.00455803e-01 -7.62435138e-01
-6.11138999e-01 -5.00957191e-01 -1.09838283e+00 -1.56862363e-01
2.80402582e-02 2.50331163e-01 3.28378350e-01 9.32016373e-01
2.23723650e-01 4.55866456e-01 3.16767581e-02 -6.91318572e-01
-3.92060906e-01 -8.15209448e-01 -4.56822872e-01 9.18926120e-01
1.93479314e-01 -9.22811210e-01 -1.90307185e-01 -7.35652223e-02] | [10.360855102539062, 2.2295026779174805] |
906901da-2322-4e02-8007-5c9f86d373f5 | isointense-infant-brain-segmentation-with-a | 1710.05956 | null | http://arxiv.org/abs/1710.05956v4 | http://arxiv.org/pdf/1710.05956v4.pdf | Isointense Infant Brain Segmentation with a Hyper-dense Connected Convolutional Neural Network | Neonatal brain segmentation in magnetic resonance (MR) is a challenging
problem due to poor image quality and low contrast between white and gray
matter regions. Most existing approaches for this problem are based on
multi-atlas label fusion strategies, which are time-consuming and sensitive to
registration errors. As alternative to these methods, we propose a
hyper-densely connected 3D convolutional neural network that employs MR-T1 and
T2 images as input, which are processed independently in two separated paths.
An important difference with previous densely connected networks is the use of
direct connections between layers from the same and different paths. Adopting
such dense connectivity helps the learning process by including deep
supervision and improving gradient flow. We evaluated our approach on data from
the MICCAI Grand Challenge on 6-month infant Brain MRI Segmentation (iSEG),
obtaining very competitive results. Among 21 teams, our approach ranked first
or second in most metrics, translating into a state-of-the-art performance. | ['Christian Desrosiers', 'Jing Yuan', 'Jose Dolz', 'Ismail Ben Ayed'] | 2017-10-16 | null | null | null | null | ['infant-brain-mri-segmentation'] | ['medical'] | [ 1.37233198e-01 2.83214778e-01 1.00963645e-01 -6.15296960e-01
-6.55063450e-01 -3.22922051e-01 1.92241475e-01 3.04810643e-01
-7.83850431e-01 4.88116324e-01 1.96672887e-01 -1.87247247e-01
-6.93828240e-02 -4.58678991e-01 -7.93766916e-01 -5.05792260e-01
-3.03572059e-01 7.80293882e-01 6.29673421e-01 -7.88324401e-02
3.61518674e-02 5.56055069e-01 -1.00042272e+00 3.63092601e-01
1.09844995e+00 7.76730716e-01 1.29578561e-01 5.31128585e-01
-2.96372354e-01 7.68786848e-01 -1.55795261e-01 -4.36187238e-01
2.14644402e-01 -4.28243190e-01 -1.32021058e+00 -3.02856803e-01
5.12060821e-01 -2.24577218e-01 -1.03779197e-01 1.09253633e+00
7.78159678e-01 -3.47416885e-02 4.39979762e-01 -7.61018097e-01
-3.19447786e-01 8.96406114e-01 -6.83508337e-01 8.10401380e-01
-2.83319280e-02 -1.47620618e-01 2.65850574e-01 -6.69748902e-01
5.52082837e-01 9.11357105e-01 9.00378525e-01 6.87448800e-01
-1.29257202e+00 -6.53609455e-01 9.32041556e-02 2.29241759e-01
-9.88206863e-01 -3.69938344e-01 3.98688406e-01 -6.71611965e-01
9.91455853e-01 -2.69037392e-02 6.86834276e-01 7.34396994e-01
1.87256709e-01 6.66618764e-01 1.37351584e+00 -1.69064358e-01
9.37995985e-02 -3.85746449e-01 2.57487744e-01 7.49475479e-01
1.29237413e-01 -2.20400110e-01 -2.73007564e-02 1.08514823e-01
8.13631892e-01 5.25049530e-02 -4.52747226e-01 -5.68369150e-01
-1.33996093e+00 7.15487063e-01 7.65762269e-01 8.01158309e-01
-5.98543525e-01 -5.09059355e-02 2.67863721e-01 2.96629705e-02
6.62806332e-01 2.89420426e-01 -3.73283386e-01 1.44965695e-02
-1.26603520e+00 -5.04043996e-02 5.88536501e-01 4.74617243e-01
4.56656158e-01 -2.40079030e-01 -1.17722481e-01 8.58730257e-01
2.26872206e-01 -4.20250446e-02 7.13734448e-01 -9.09492314e-01
4.12300587e-01 3.54464978e-01 -4.92175072e-01 -6.45658493e-01
-1.09045720e+00 -4.81641471e-01 -8.82580400e-01 3.85322481e-01
6.43681109e-01 -3.80115777e-01 -1.42216527e+00 1.65245068e+00
3.99911106e-01 2.44124621e-01 -2.01525822e-01 1.04480624e+00
1.22221708e+00 -1.06201628e-02 3.88683081e-02 -1.53469533e-01
1.05597520e+00 -1.31481254e+00 -5.92241645e-01 -5.58290929e-02
7.34168947e-01 -4.91148651e-01 4.49563384e-01 2.92898536e-01
-1.53386855e+00 -1.99340165e-01 -8.96699429e-01 1.02075063e-01
-2.26682395e-01 -4.74525124e-01 5.18588841e-01 5.76147556e-01
-1.65207255e+00 9.32032704e-01 -1.27556157e+00 -7.65424222e-02
9.27615225e-01 7.44456708e-01 -7.02972412e-01 -1.68135062e-01
-9.31319892e-01 1.31237757e+00 3.36218923e-01 -1.29302973e-02
-8.21960270e-01 -1.05957174e+00 -6.76043034e-01 -1.83970183e-01
2.65622765e-01 -3.13332319e-01 1.13899755e+00 -7.07879782e-01
-1.29805863e+00 1.12236667e+00 1.97624251e-01 -6.28641963e-01
8.01072598e-01 -1.60637870e-01 -2.49076396e-01 6.80341899e-01
1.65402070e-01 1.05593657e+00 5.07425666e-01 -1.02819979e+00
-3.48426580e-01 -6.07329905e-01 -1.17239252e-01 8.41100216e-02
1.99030116e-01 4.11141098e-01 -2.47638032e-01 -5.80630362e-01
5.13182342e-01 -6.87008739e-01 -4.86693680e-01 -3.19214463e-01
-2.34009534e-01 -3.80317494e-02 5.70674360e-01 -1.06746733e+00
7.07134604e-01 -1.57148480e+00 1.43874750e-01 1.76660597e-01
7.85968244e-01 2.50477046e-01 -5.39527424e-02 -4.05130833e-01
-5.71461022e-01 6.48394134e-03 -5.69718421e-01 -2.82199413e-01
-4.98977542e-01 1.84909832e-02 3.79792899e-01 6.99203312e-01
1.83329865e-01 9.85950291e-01 -1.02729869e+00 -5.62187493e-01
1.52486414e-01 7.12960482e-01 -4.61210638e-01 1.29635110e-01
3.29234511e-01 1.04745793e+00 -8.63142237e-02 2.43812695e-01
6.12002730e-01 -2.26984948e-01 -8.29139724e-03 -2.14151457e-01
-1.26685619e-01 2.93326795e-01 -7.80474424e-01 2.13925004e+00
-2.54056960e-01 2.78557360e-01 3.47022802e-01 -1.42847252e+00
5.65085411e-01 6.32423103e-01 1.05505669e+00 -7.90915668e-01
3.89369071e-01 3.87432486e-01 4.37021792e-01 -5.19957304e-01
-1.52305841e-01 -1.92172885e-01 5.99197686e-01 5.08110642e-01
4.37102735e-01 -1.81918159e-01 2.29690194e-01 1.55267715e-01
1.23380589e+00 1.60593301e-01 -2.06400022e-01 -6.37854099e-01
3.47030044e-01 -3.19507688e-01 6.44678354e-01 5.64104199e-01
-3.81850958e-01 1.24525023e+00 5.33564150e-01 -5.57264626e-01
-8.96302700e-01 -7.98566759e-01 -1.99219108e-01 6.08219564e-01
-1.61249638e-01 1.36715651e-01 -1.41812098e+00 -1.07497513e+00
-5.25235772e-01 3.29458803e-01 -8.90187681e-01 1.92795455e-01
-1.06418025e+00 -9.29983497e-01 5.21638513e-01 5.88702559e-01
3.98358852e-01 -1.06307101e+00 -9.28152502e-01 5.12539089e-01
-2.30171040e-01 -1.16058314e+00 -3.45330477e-01 4.92981672e-01
-1.10378301e+00 -1.17628682e+00 -1.44640517e+00 -8.32277536e-01
7.83435285e-01 3.63585874e-02 1.18136394e+00 2.05641642e-01
-3.43869507e-01 6.88966960e-02 -3.42584342e-01 -2.34077722e-01
-3.62009227e-01 2.94695616e-01 -1.50551230e-01 -1.43171743e-01
-2.29763854e-02 -7.03060687e-01 -7.69173384e-01 1.41497210e-01
-9.21763599e-01 1.98863879e-01 4.11530972e-01 6.76343083e-01
4.94842023e-01 -3.70654911e-01 6.01326942e-01 -8.96350205e-01
3.28262568e-01 -6.74669862e-01 -5.15983820e-01 3.02655786e-01
-6.29692376e-01 -1.56885497e-02 3.01427484e-01 -2.66749263e-01
-7.30223179e-01 6.92624301e-02 -5.29879570e-01 -3.39643806e-01
-4.80857730e-01 1.14521638e-01 2.72415817e-01 -4.41252291e-01
4.41118628e-01 -2.54600316e-01 2.96621859e-01 -5.82216740e-01
2.22001851e-01 1.60565853e-01 6.80376172e-01 -3.58676314e-01
1.97878182e-01 5.64712107e-01 6.25404939e-02 -2.54253030e-01
-7.66276419e-01 -3.66873354e-01 -1.20075846e+00 -2.60664582e-01
1.40705609e+00 -5.58287323e-01 -2.98170924e-01 5.36435604e-01
-1.19366610e+00 -4.94842052e-01 -8.76769423e-02 8.12936068e-01
-3.41916949e-01 3.40523869e-01 -8.92393947e-01 -4.82860059e-02
-5.96659124e-01 -1.87870777e+00 6.41004801e-01 3.08060050e-01
3.71650886e-03 -1.06560874e+00 1.84588268e-01 5.66068828e-01
8.19727659e-01 4.98673588e-01 9.43343043e-01 -9.83625412e-01
-7.22687244e-02 1.37855187e-01 -3.57356519e-01 3.66105318e-01
7.96436891e-02 -4.97597992e-01 -8.79060745e-01 -4.23366308e-01
1.77136630e-01 -2.70050347e-01 9.24150288e-01 7.34684110e-01
1.23343873e+00 2.37096950e-01 -2.39491865e-01 7.32628882e-01
1.14126921e+00 2.90068030e-01 4.68430758e-01 2.78353035e-01
1.03399694e+00 9.25935745e-01 5.07755540e-02 -2.02148080e-01
6.01954043e-01 5.08331954e-01 5.88573456e-01 -6.68279827e-01
-3.69815081e-01 4.87083584e-01 -2.19912738e-01 9.90319312e-01
-3.16456556e-01 3.52442503e-01 -1.37080824e+00 8.40715706e-01
-1.61847234e+00 -5.72844803e-01 -3.59787226e-01 2.02670169e+00
9.30255771e-01 5.82192801e-02 3.23599130e-01 6.36294335e-02
7.63420820e-01 -1.17527962e-01 -4.68916953e-01 -3.45064551e-01
-1.09552562e-01 8.20211649e-01 6.11283004e-01 4.17843372e-01
-1.13794398e+00 6.58578873e-01 6.65730190e+00 4.80437458e-01
-1.29124737e+00 7.76236296e-01 9.25161600e-01 -1.83052540e-01
-6.91701323e-02 -3.19372952e-01 -2.83922881e-01 4.79113072e-01
8.85713935e-01 5.11981547e-01 3.79626274e-01 2.05961168e-01
-1.47821769e-01 -2.09525585e-01 -1.06419849e+00 7.74933636e-01
6.14178441e-02 -1.29288983e+00 -4.38530594e-01 -1.94168881e-01
9.40846562e-01 7.75672853e-01 -1.75827622e-01 -7.06209540e-02
3.48555654e-01 -1.41084182e+00 5.74542403e-01 4.49714810e-01
6.57138288e-01 -7.87913084e-01 7.62281418e-01 1.42143279e-01
-7.00585723e-01 3.73333186e-01 -2.27288455e-02 3.74512106e-01
1.76555127e-01 5.46622574e-01 -5.60943246e-01 5.66019058e-01
1.22153091e+00 3.41277003e-01 -3.99819732e-01 1.38259494e+00
-1.59056708e-01 6.47063434e-01 -3.16951841e-01 5.73180437e-01
4.45415616e-01 -2.54311234e-01 4.02498454e-01 1.34889770e+00
1.77531857e-02 2.57156879e-01 8.44480991e-02 8.25947404e-01
-2.26369545e-01 2.22432837e-01 -4.37243283e-01 5.48622727e-01
-2.05605716e-01 1.49435568e+00 -1.29236376e+00 -4.72172797e-01
-6.50525749e-01 6.99445009e-01 3.67409289e-01 1.30302772e-01
-7.09669113e-01 -2.40251496e-01 1.58844978e-01 2.01374605e-01
2.71047205e-01 -4.22143191e-02 -3.88183266e-01 -9.35804248e-01
-1.48163572e-01 -8.44637036e-01 3.31609130e-01 -3.97799790e-01
-9.83206213e-01 1.07272184e+00 1.21820763e-01 -6.55947745e-01
3.45484763e-02 -3.11577797e-01 -6.00381911e-01 9.67213273e-01
-1.74077547e+00 -8.14026058e-01 -1.68110937e-01 5.69185376e-01
2.09819868e-01 -8.45311582e-02 6.59365296e-01 8.24943960e-01
-4.84626621e-01 4.99309331e-01 -2.16079980e-01 2.06792235e-01
4.33226764e-01 -1.32539451e+00 5.05466938e-01 1.01202047e+00
-3.89951915e-02 3.58589083e-01 2.03191221e-01 -6.40167892e-01
-6.64734066e-01 -9.34422910e-01 5.12585878e-01 -2.09923059e-01
5.98858833e-01 -7.43707716e-02 -1.25037336e+00 5.70093453e-01
5.30577302e-01 3.62348735e-01 5.10811269e-01 -3.40121001e-01
-7.20841736e-02 2.36113518e-01 -1.48802233e+00 2.44586989e-01
9.67789114e-01 -5.83464392e-02 -6.84585154e-01 3.47527802e-01
7.97694564e-01 -8.76803517e-01 -1.18441510e+00 8.08282673e-01
2.35618547e-01 -1.01404178e+00 8.28463078e-01 -5.01404762e-01
5.20202816e-01 -1.23032577e-01 3.92737031e-01 -1.45814621e+00
-1.75838590e-01 -4.39875960e-01 2.06219569e-01 8.43800008e-01
4.44842458e-01 -6.63541973e-01 6.82381570e-01 6.76407099e-01
-4.56387520e-01 -1.03558600e+00 -9.48233068e-01 -3.56545359e-01
5.99268615e-01 -2.84855843e-01 6.61796272e-01 1.19113636e+00
-3.71072859e-01 6.97779804e-02 -9.41586588e-03 -6.18464388e-02
7.21906900e-01 -2.71844834e-01 -4.02566530e-02 -1.27569675e+00
3.30806002e-02 -8.49322438e-01 -3.54582280e-01 -5.90826988e-01
5.91972433e-02 -1.35229743e+00 5.61247543e-02 -1.77434778e+00
3.11140150e-01 -6.20021045e-01 -5.74752629e-01 6.72628999e-01
-1.71545684e-01 6.11311793e-01 4.56822515e-02 -7.74275288e-02
-5.04010260e-01 1.57354288e-02 1.43237054e+00 -4.64898832e-02
-5.36327548e-02 -1.26335338e-01 -4.83463556e-01 8.43169689e-01
8.22135448e-01 -5.56607187e-01 -1.64170370e-01 -1.02859783e+00
-2.27340221e-01 1.33589908e-01 2.08501950e-01 -1.07341206e+00
2.22140878e-01 4.28606004e-01 4.87450391e-01 -5.87880492e-01
-7.23119359e-03 -7.58493304e-01 -1.03367426e-01 4.88640487e-01
-2.60359377e-01 3.89312357e-01 1.35317504e-01 -2.47876078e-01
-2.52150834e-01 -2.71163285e-01 1.17703331e+00 -4.08269107e-01
-2.17853233e-01 7.08441019e-01 -2.93631554e-01 2.21982509e-01
9.07450199e-01 -9.44493487e-02 -6.19662516e-02 6.57462850e-02
-1.01100373e+00 3.70431632e-01 2.89994061e-01 4.72587585e-01
5.78141570e-01 -9.56901968e-01 -8.16111088e-01 8.65216777e-02
-4.35330063e-01 4.22628790e-01 2.62258679e-01 1.54895401e+00
-8.24762702e-01 2.12418482e-01 -5.55465341e-01 -9.01090026e-01
-1.08682489e+00 2.64498621e-01 7.19770312e-01 -4.13757771e-01
-1.00194538e+00 1.08161938e+00 1.27608135e-01 -6.52349591e-01
3.40464771e-01 -5.70636749e-01 -6.66405976e-01 -2.80926898e-02
5.36438406e-01 3.25611949e-01 5.48230886e-01 -9.07885969e-01
-4.48117495e-01 7.40864158e-01 -3.15309614e-01 -1.70779943e-01
1.61979198e+00 3.64395343e-02 -4.46691990e-01 4.88637462e-02
1.34301841e+00 -5.26576579e-01 -1.04619098e+00 -3.01028728e-01
3.41053843e-01 -1.25406697e-01 3.65500063e-01 -8.60760570e-01
-1.79195821e+00 1.14431071e+00 1.04024005e+00 1.29740015e-01
8.98350716e-01 -2.29883734e-02 9.62455332e-01 -1.99870825e-01
3.30513626e-01 -8.67090464e-01 -2.07203969e-01 5.04633784e-01
7.90999651e-01 -1.48664963e+00 -3.17692071e-01 -1.39238656e-01
-5.34236252e-01 1.08173382e+00 6.23476446e-01 -1.21627592e-01
6.13317668e-01 5.35920024e-01 4.41481918e-01 -4.93628114e-01
-2.15051591e-01 -1.42628342e-01 5.07426143e-01 5.67737401e-01
7.86754251e-01 -1.46224266e-02 -4.30702329e-01 3.12157273e-01
-2.88369553e-03 -1.70011193e-01 2.80468971e-01 9.02048647e-01
-1.99809462e-01 -1.06837726e+00 -1.51365519e-01 7.02436149e-01
-1.14299095e+00 -1.77372649e-01 1.04620390e-01 4.22844440e-01
2.23005980e-01 8.64436746e-01 -1.91608630e-02 3.77055965e-02
3.55009168e-01 -6.73519215e-03 7.56542861e-01 -6.62867367e-01
-1.06635690e+00 6.77875653e-02 -1.83979526e-01 -8.34219575e-01
-6.06414974e-01 -6.93373561e-01 -1.82891560e+00 5.09311333e-02
-1.41052842e-01 1.80968553e-01 9.75126147e-01 1.18590736e+00
8.26161504e-02 8.41020226e-01 5.55330873e-01 -9.71973836e-01
-6.25993088e-02 -1.07192862e+00 -2.87847310e-01 4.69324678e-01
4.06031817e-01 -6.51625693e-01 8.50643739e-02 -2.64380574e-01] | [14.230156898498535, -2.3860318660736084] |
03b3d525-25cb-4940-af0d-7a2dfb0f9475 | st-fl-style-transfer-preprocessing-in | 2203.13680 | null | https://arxiv.org/abs/2203.13680v1 | https://arxiv.org/pdf/2203.13680v1.pdf | ST-FL: Style Transfer Preprocessing in Federated Learning for COVID-19 Segmentation | Chest Computational Tomography (CT) scans present low cost, speed and objectivity for COVID-19 diagnosis and deep learning methods have shown great promise in assisting the analysis and interpretation of these images. Most hospitals or countries can train their own models using in-house data, however empirical evidence shows that those models perform poorly when tested on new unseen cases, surfacing the need for coordinated global collaboration. Due to privacy regulations, medical data sharing between hospitals and nations is extremely difficult. We propose a GAN-augmented federated learning model, dubbed ST-FL (Style Transfer Federated Learning), for COVID-19 image segmentation. Federated learning (FL) permits a centralised model to be learned in a secure manner from heterogeneous datasets located in disparate private data silos. We demonstrate that the widely varying data quality on FL client nodes leads to a sub-optimal centralised FL model for COVID-19 chest CT image segmentation. ST-FL is a novel FL framework that is robust in the face of highly variable data quality at client nodes. The robustness is achieved by a denoising CycleGAN model at each client of the federation that maps arbitrary quality images into the same target quality, counteracting the severe data variability evident in real-world FL use-cases. Each client is provided with the target style, which is the same for all clients, and trains their own denoiser. Our qualitative and quantitative results suggest that this FL model performs comparably to, and in some cases better than, a model that has centralised access to all the training data. | ['Rob Otter', 'Sean Moran', 'Fran Silavong', 'Varun Babbar', 'Antonios Georgiadis'] | 2022-03-25 | null | null | null | null | ['covid-19-image-segmentation', 'covid-19-detection'] | ['computer-vision', 'medical'] | [-2.66436078e-02 3.51489484e-01 -7.53936842e-02 -3.94875139e-01
-1.23860729e+00 -6.58787072e-01 1.91816211e-01 -2.98502684e-01
-4.12604988e-01 6.56996012e-01 2.94499636e-01 -4.47924793e-01
-2.25655690e-01 -7.82276690e-01 -5.32637656e-01 -9.20418561e-01
1.58598963e-02 8.56572092e-01 -2.27339402e-01 3.51889208e-02
-6.97101593e-01 5.48970401e-01 -5.84997594e-01 6.88116133e-01
3.53079885e-01 1.08651245e+00 -7.00417310e-02 9.23352540e-01
1.50180280e-01 1.23843181e+00 -6.84070766e-01 -7.59632707e-01
8.38836372e-01 -5.23962140e-01 -8.98359597e-01 -2.30878238e-02
1.99993819e-01 -5.55148423e-01 -1.86782300e-01 5.59936702e-01
6.90911114e-01 -5.35107076e-01 3.00945729e-01 -1.28341246e+00
-2.27498636e-01 4.51247007e-01 -2.16188833e-01 1.22830413e-01
6.92783892e-02 6.07941568e-01 5.63149631e-01 -2.31526688e-01
9.75370467e-01 6.97007060e-01 9.21481669e-01 7.09514260e-01
-1.19833422e+00 -6.94381118e-01 -3.48013163e-01 -2.99074948e-01
-9.98648345e-01 -1.82488561e-01 2.43601710e-01 -2.78279603e-01
6.92291439e-01 6.23305082e-01 6.15758717e-01 1.28472221e+00
6.77347541e-01 3.43342394e-01 1.29337263e+00 -7.87975788e-02
2.18843147e-01 2.32878864e-01 -5.81695735e-01 4.66987342e-01
1.28553271e-01 -1.70463230e-03 -3.77296627e-01 -5.69214165e-01
7.43743002e-01 2.57139683e-01 -4.78510529e-01 -3.25211108e-01
-1.13996065e+00 9.48140502e-01 7.10492730e-01 4.65463012e-01
-6.35946393e-01 2.41990000e-01 7.07888186e-01 7.63755441e-01
3.53441149e-01 1.25672519e-01 -6.25071943e-01 -2.18432806e-02
-1.01977146e+00 -2.02964187e-01 7.70742536e-01 7.64898360e-01
3.99186999e-01 -8.69169831e-02 -6.97638541e-02 3.54522705e-01
7.62343556e-02 4.79400575e-01 5.46013594e-01 -1.18771684e+00
4.06727076e-01 4.80786443e-01 -2.96626478e-01 -6.47905707e-01
-3.54180515e-01 -4.83740062e-01 -1.30103087e+00 2.24500537e-01
1.82795629e-01 -4.37138736e-01 -9.11444247e-01 1.69037771e+00
5.19869745e-01 8.09237584e-02 2.50568211e-01 7.59040058e-01
5.89020252e-01 2.22083390e-01 -9.57507640e-02 -2.95285899e-02
1.33087158e+00 -6.85447872e-01 -5.58926046e-01 1.53824806e-01
6.45116150e-01 -5.78852952e-01 5.37160277e-01 6.15686476e-01
-1.11615789e+00 -1.20906301e-01 -6.73723698e-01 2.93008775e-01
-2.07254022e-01 -5.54207683e-01 3.98107439e-01 1.11033213e+00
-1.47350752e+00 2.89380133e-01 -1.20066679e+00 -2.55933255e-01
9.36599672e-01 7.04551816e-01 -7.39889264e-01 -4.07792538e-01
-8.76959741e-01 8.79418910e-01 2.26112846e-02 -9.98395495e-03
-1.08574009e+00 -8.12409580e-01 -5.02767563e-01 -1.58747837e-01
1.41847923e-01 -1.22976732e+00 1.15459716e+00 -1.35460973e+00
-1.09121931e+00 9.58001435e-01 5.21746516e-01 -7.25519598e-01
1.13614821e+00 3.21262300e-01 -5.94450533e-01 4.09784466e-01
1.00089073e-01 5.63008368e-01 1.02960873e+00 -1.24356437e+00
-5.58486044e-01 -4.33243811e-01 -3.02409142e-01 -5.17296605e-02
-2.39263847e-01 3.60761024e-02 -2.99013406e-01 -6.23345435e-01
-3.00448000e-01 -8.74550879e-01 -4.76291299e-01 2.47979969e-01
-5.42829968e-02 3.77192110e-01 1.12134659e+00 -6.99241400e-01
7.24995911e-01 -2.18891811e+00 -2.02354461e-01 6.61219358e-01
6.76823080e-01 1.04148649e-01 3.68676335e-02 1.35608882e-01
-2.57771853e-02 2.34976351e-01 -2.87699282e-01 -1.92597210e-01
-3.72347802e-01 8.52866948e-01 2.96396643e-01 7.15921521e-01
1.95815433e-02 1.00302100e+00 -7.29711115e-01 -7.20349431e-01
4.07555401e-02 5.85408509e-01 -6.12460077e-01 4.20504034e-01
8.56492817e-02 9.84836161e-01 -5.02505124e-01 5.98931134e-01
6.66564405e-01 -6.47756636e-01 5.40355146e-01 -3.96163203e-02
3.15263033e-01 -3.40433449e-01 -9.25834894e-01 1.81699121e+00
-6.59648061e-01 2.00893477e-01 7.47698545e-01 -7.27645040e-01
5.35665929e-01 9.60254371e-01 9.14057851e-01 -7.37923145e-01
3.53739798e-01 4.41332906e-01 -8.34964868e-03 -4.41837877e-01
6.00019507e-02 -4.44102198e-01 -1.28838390e-01 8.08659673e-01
1.60489693e-01 -1.60026714e-01 -6.11864090e-01 4.48156983e-01
1.68826771e+00 -5.13744593e-01 -4.88565788e-02 4.94701080e-02
3.75915438e-01 3.06777470e-02 4.62735862e-01 7.81525910e-01
-3.14318985e-01 8.44884753e-01 2.62131363e-01 -4.53875184e-01
-1.19964063e+00 -1.03175521e+00 -1.74001008e-01 6.93152666e-01
-3.48991871e-01 -1.07713379e-01 -7.30512381e-01 -9.70679462e-01
-7.63751641e-02 2.07963750e-01 -7.68794358e-01 3.23589556e-02
-6.63495004e-01 -4.24423665e-01 7.89758146e-01 5.47759950e-01
5.75437307e-01 -1.04092324e+00 -1.14479423e+00 4.11061138e-01
-1.94752589e-01 -9.87987041e-01 -3.60039204e-01 3.83303910e-01
-7.31945992e-01 -1.10376036e+00 -8.90431106e-01 -2.97640890e-01
5.42927086e-01 -1.25491813e-01 1.31346166e+00 2.89695293e-01
-4.90840256e-01 7.41463721e-01 -3.73206645e-01 -4.63405222e-01
-9.66856539e-01 2.33260080e-01 -4.29715663e-01 6.90800846e-02
-1.11395933e-01 -2.52562553e-01 -9.41864610e-01 4.57982123e-01
-1.28658378e+00 -2.61143595e-01 4.81899232e-01 1.17751300e+00
6.03624582e-01 -2.01705649e-01 4.92391169e-01 -1.35183954e+00
3.42706472e-01 -9.01834786e-01 -2.74454772e-01 2.11142316e-01
-7.24266231e-01 -2.35648558e-01 7.28276551e-01 1.34119600e-01
-1.04041243e+00 2.37229336e-02 -2.84300059e-01 -6.96189344e-01
-1.78006127e-01 2.74172723e-01 1.36867151e-01 -3.38601232e-01
7.08405793e-01 -1.95538163e-01 5.08255661e-01 -3.19320083e-01
2.42880285e-01 8.33207667e-01 6.31056249e-01 -3.64383221e-01
6.47327960e-01 8.88165414e-01 -1.19379617e-01 -1.04328156e-01
-5.33343852e-01 -2.83260435e-01 -3.19723487e-01 -2.42142394e-01
1.05946648e+00 -1.02532017e+00 -3.71068418e-01 3.45608771e-01
-7.65928626e-01 -4.09701377e-01 -7.71191061e-01 2.04438120e-01
-6.17340088e-01 -1.28593072e-01 -6.79158390e-01 -1.73307359e-01
-8.74217689e-01 -1.23662627e+00 9.24875855e-01 -1.60983577e-01
-2.09290326e-01 -9.48779762e-01 2.09879637e-01 6.97441041e-01
9.33089495e-01 8.39100182e-01 7.63698161e-01 -7.55642891e-01
-6.13836527e-01 -2.43600383e-01 -7.08224578e-03 5.51436782e-01
3.25238287e-01 -2.52349913e-01 -9.61036921e-01 -6.92333400e-01
4.05358732e-01 -4.86901700e-01 3.15524310e-01 2.41206214e-01
1.16298378e+00 -5.31856120e-01 -7.23642781e-02 1.01094341e+00
1.65130579e+00 -9.28829312e-02 4.39380288e-01 3.93985838e-01
4.98145700e-01 2.67701268e-01 2.15650946e-01 4.39758688e-01
2.34796762e-01 5.87042011e-02 7.80516207e-01 -7.36296773e-01
-4.73563634e-02 1.62712112e-01 -9.59124863e-02 6.27335668e-01
1.12783229e-02 -2.53269225e-01 -1.02896750e+00 5.30442953e-01
-1.64429426e+00 -7.43957639e-01 -1.13889773e-03 1.84694695e+00
6.01031721e-01 -8.37955847e-02 2.34721247e-02 -1.18568048e-01
3.77258360e-01 -7.02545941e-02 -6.62829399e-01 -6.71713531e-01
-9.75389183e-02 6.87788308e-01 8.15977395e-01 -3.20778966e-01
-6.62763059e-01 2.40894169e-01 6.34417105e+00 4.86219138e-01
-1.34076607e+00 9.08148825e-01 1.11674452e+00 -2.82901883e-01
-2.32354924e-01 -4.53418612e-01 1.84810922e-01 4.08628404e-01
1.26988065e+00 -1.12776190e-01 2.47945353e-01 8.32932115e-01
9.84034836e-02 1.19004548e-01 -9.55532730e-01 9.39630926e-01
-1.80325508e-01 -1.76756060e+00 -2.31028005e-01 4.66479182e-01
1.01059830e+00 6.81700706e-01 2.30146781e-01 -1.07395284e-01
8.22484612e-01 -1.27996755e+00 4.82962877e-01 3.65767002e-01
1.47831285e+00 -9.44227517e-01 1.19592297e+00 2.80916274e-01
-7.62012005e-01 -2.51986206e-01 5.28831184e-02 7.48204648e-01
-5.59096644e-03 1.90010071e-01 -1.00327337e+00 1.04428530e+00
1.08502924e+00 1.92647800e-01 -3.86752695e-01 5.45646012e-01
3.18796098e-01 6.53363943e-01 -4.33921695e-01 8.23241591e-01
4.95194405e-01 1.14985563e-01 1.29488856e-01 1.08039737e+00
1.79619655e-01 1.37260720e-01 -1.07479684e-01 3.76118541e-01
-3.48380446e-01 3.91883925e-02 -6.86171353e-01 3.93807888e-01
3.81915271e-02 1.37137520e+00 -6.72648013e-01 -3.48120809e-01
-5.10219336e-01 8.87014508e-01 -7.21681491e-02 -6.10130616e-02
-7.53663361e-01 3.88416737e-01 5.18906951e-01 2.15092704e-01
4.68820959e-01 3.47958744e-01 -2.56588608e-01 -9.52000082e-01
-2.03575790e-01 -1.53469133e+00 1.07886064e+00 -5.03513455e-01
-1.67029488e+00 1.12185836e+00 -3.37626159e-01 -1.16808748e+00
-5.06318808e-01 -2.95363486e-01 -5.35033584e-01 6.54602647e-01
-1.35366642e+00 -1.50412071e+00 -4.54004765e-01 1.42495084e+00
-6.60288753e-03 -2.60906607e-01 1.10365593e+00 2.88427174e-01
-1.73480585e-01 8.16363513e-01 3.69888067e-01 3.23887259e-01
6.41671538e-01 -1.06068146e+00 -9.40763801e-02 5.73728621e-01
-1.41119197e-01 2.32573241e-01 2.89625227e-01 -4.79218781e-01
-1.47069931e+00 -1.42509115e+00 2.22261205e-01 -4.49376822e-01
4.15244877e-01 -1.71358898e-01 -6.99035764e-01 8.95183027e-01
4.71700490e-01 5.09060323e-01 1.02116299e+00 -4.69244450e-01
-1.20434664e-01 -3.28910559e-01 -2.01062369e+00 2.35346444e-02
6.26263738e-01 -4.87475663e-01 -1.06042795e-01 5.37164032e-01
6.12010002e-01 -4.64963704e-01 -1.26849747e+00 1.73915043e-01
3.44320327e-01 -1.13141894e+00 5.67337394e-01 -4.70854223e-01
1.02926835e-01 -2.23723203e-02 -1.91072047e-01 -1.18802178e+00
-8.88197422e-02 -6.97033346e-01 2.70302564e-01 8.77072692e-01
2.17545345e-01 -1.03741491e+00 1.13488209e+00 9.08285320e-01
1.06708765e-01 -6.88979089e-01 -1.47679985e+00 -5.40917456e-01
2.18813136e-01 -5.19751489e-01 9.90272880e-01 1.16212034e+00
-6.70791149e-01 -3.55537713e-01 -5.71893267e-02 7.12240040e-02
6.69214845e-01 -1.28990829e-01 7.87916899e-01 -8.60750914e-01
-4.87493485e-01 1.74844079e-02 -5.20155728e-01 -6.72379509e-02
-2.19253823e-01 -8.41723621e-01 -4.25603241e-01 -1.32287121e+00
1.35256916e-01 -8.73835802e-01 -4.15761799e-01 7.55442083e-01
4.46577966e-01 5.51614881e-01 3.40507954e-01 3.48199874e-01
-2.15920866e-01 -6.29977509e-02 1.30624938e+00 -1.81188881e-01
3.10161352e-01 4.01153527e-02 -7.22555816e-01 3.85574639e-01
6.82034194e-01 -7.87323713e-01 -2.76746064e-01 -5.96947312e-01
1.47403628e-01 5.00749052e-01 6.03868425e-01 -9.15184796e-01
3.38658005e-01 8.63713101e-02 5.02281368e-01 -2.96245307e-01
3.02233417e-02 -1.66708422e+00 1.00471878e+00 7.40601242e-01
1.16747603e-01 2.63104945e-01 -1.16253138e-01 4.76384729e-01
-2.26882264e-01 1.46666273e-01 9.69197392e-01 -6.26125097e-01
-6.40122741e-02 4.75789398e-01 -2.70359755e-01 6.77268282e-02
1.23847115e+00 -2.98902929e-01 -1.60957307e-01 -5.05249381e-01
-8.66494954e-01 3.29694957e-01 7.60489166e-01 -1.21421985e-01
3.77132714e-01 -8.99984956e-01 -1.08400071e+00 3.71734053e-01
-1.48566797e-01 4.23353523e-01 5.07486939e-01 1.04752481e+00
-7.03135490e-01 -4.05521169e-02 -1.82637960e-01 -8.39930952e-01
-1.12936032e+00 4.73027289e-01 7.74714351e-01 -7.42959976e-01
-8.75227988e-01 7.62512326e-01 -3.25537361e-02 -5.49134552e-01
-3.65142114e-02 -2.98883766e-02 6.83310449e-01 -1.61683097e-01
2.49055594e-01 8.01237896e-02 7.42547929e-01 -6.34649694e-01
-3.38458627e-01 -3.67137715e-02 -1.73048526e-01 1.73964107e-03
1.74910164e+00 3.01838163e-02 -8.17568600e-02 -1.06810510e-01
1.34495342e+00 -1.84283122e-01 -1.04795337e+00 -6.74342215e-02
-3.76545697e-01 -5.14804423e-01 1.08676486e-01 -1.09602427e+00
-1.91209114e+00 2.96073794e-01 9.03151751e-01 5.87627350e-04
1.52155352e+00 4.48495336e-02 8.95680726e-01 -1.86226040e-01
6.09291673e-01 -7.96894491e-01 -1.24591686e-01 -1.59765929e-01
4.97571200e-01 -1.05487382e+00 -2.48535350e-02 1.26676932e-01
-7.81301141e-01 9.01463866e-01 3.39042157e-01 -9.01206136e-02
5.65473914e-01 7.51497209e-01 6.12640381e-01 -3.86174113e-01
-8.63884866e-01 6.46952271e-01 -4.15636957e-01 6.87356949e-01
-8.64029229e-02 2.33892441e-01 2.92050391e-01 2.25789547e-01
-2.64135569e-01 1.25619397e-01 4.57496703e-01 1.28894806e+00
4.67683434e-01 -1.26218390e+00 -7.61589050e-01 7.69113362e-01
-9.00254130e-01 1.31662577e-01 -1.24522008e-01 8.52200210e-01
5.41831374e-01 7.80695558e-01 1.03548393e-01 -9.97068360e-03
1.85320511e-01 -8.75769109e-02 2.24953488e-01 -4.61968154e-01
-1.59101689e+00 1.39670134e-01 -1.68171063e-01 -7.79416025e-01
-4.76226240e-01 -6.12978101e-01 -1.08170545e+00 -5.21441638e-01
-2.41884544e-01 1.03124000e-01 8.22530448e-01 5.39125085e-01
4.23635960e-01 5.57677329e-01 1.07162821e+00 -3.94041806e-01
-6.34472609e-01 -3.79019260e-01 -5.18472970e-01 4.77855414e-01
5.03714502e-01 6.96574599e-02 -2.28307560e-01 -1.25533883e-02] | [6.126415252685547, 6.523619174957275] |
11d10114-74d9-4976-8c36-c0aef6d308ea | proceedings-10th-international-workshop-on | 2202.02144 | null | https://arxiv.org/abs/2202.02144v1 | https://arxiv.org/pdf/2202.02144v1.pdf | Proceedings 10th International Workshop on Theorem Proving Components for Educational Software | This EPTCS volume contains the proceedings of the ThEdu'21 workshop, promoted on 11 July 2021, as a satellite event of CADE-28. Due to the COVID-19 pandemic, CADE-28 and all its co-located events happened as virtual events. ThEdu'21 was a vibrant workshop, with an invited talk by Gilles Dowek (ENS Paris-Saclay), eleven contributions, and one demonstration. After the workshop an open call for papers was issued and attracted 10 submissions, 7 of which have been accepted by the reviewers, and collected in the present post-proceedings volume. The ThEdu series pursues the smooth transition from an intuitive way of doing mathematics at secondary school to a more formal approach to the subject in STEM education, while favouring software support for this transition by exploiting the power of theorem-proving technologies. The volume editors hope that this collection of papers will further promote the development of theorem-proving based software, and that it will collaborate on improving mutual understanding between computer scientists, mathematicians and stakeholders in education. | ['Pedro Quaresma', 'Walther Neuper', 'João Marcos'] | 2022-02-02 | null | null | null | null | ['automated-theorem-proving', 'automated-theorem-proving'] | ['miscellaneous', 'reasoning'] | [-2.57140517e-01 5.96158385e-01 3.02260578e-01 -7.35247135e-02
-4.42905724e-01 -7.67554581e-01 5.79268336e-01 5.05542815e-01
-1.88316867e-01 9.01781499e-01 -1.46760151e-01 -8.78096044e-01
-3.08833301e-01 -8.75497401e-01 -1.03187466e+00 -8.22637901e-02
-3.79750371e-01 2.56033003e-01 3.87628347e-01 -3.81883860e-01
5.43838382e-01 5.01534998e-01 -1.87376118e+00 5.77129126e-02
9.58695769e-01 2.75202930e-01 1.85483947e-01 6.91317976e-01
-3.21052745e-02 6.69896603e-01 -4.73797798e-01 -6.10841095e-01
1.02355592e-01 -5.35291195e-01 -1.12903345e+00 -4.66360211e-01
3.33742291e-01 2.71769106e-01 -3.03511117e-02 1.06070697e+00
1.98403094e-02 1.72988907e-01 -2.09871195e-02 -1.53057146e+00
-8.37827846e-02 8.21625590e-01 -3.07498664e-01 1.24429449e-01
6.74218535e-01 -2.74201483e-01 9.01614189e-01 -5.98085999e-01
7.58264422e-01 1.02593970e+00 6.86878204e-01 1.22057498e-01
-8.54461551e-01 -5.71581662e-01 -1.37920246e-01 5.62220693e-01
-1.08469605e+00 -1.32823214e-01 6.16198659e-01 -6.63691700e-01
6.34618104e-01 6.13948762e-01 1.04870784e+00 3.87350917e-01
2.30533063e-01 9.21854019e-01 1.03371978e+00 -8.40392590e-01
4.10210818e-01 6.31815314e-01 8.62359181e-02 5.62830627e-01
3.40068907e-01 -3.53478715e-02 -1.24524504e-01 2.10125074e-02
7.13766158e-01 -4.79800910e-01 -1.90257058e-01 -5.17605722e-01
-1.39349270e+00 6.10479951e-01 -1.02355644e-01 7.19490409e-01
-5.14939055e-02 2.02818923e-02 3.06803972e-01 7.60014355e-01
7.95445964e-03 4.71576184e-01 -4.90693033e-01 -6.66098237e-01
-9.15610373e-01 8.55837345e-01 1.37387967e+00 7.35589981e-01
3.18311304e-01 -3.89688015e-01 6.20866179e-01 2.54113436e-01
2.96837032e-01 -5.46337385e-03 -1.50678903e-01 -1.22121370e+00
1.85387284e-01 5.48656166e-01 2.83053275e-02 -9.44260716e-01
-1.19914643e-01 -1.67839020e-01 -4.75248188e-01 3.09893757e-01
6.08644962e-01 -2.96268135e-01 -4.51264232e-02 1.30521560e+00
3.84156734e-01 1.41594276e-01 -2.15946630e-01 6.85747266e-01
6.64429247e-01 8.18031907e-01 -2.95068055e-01 -4.24382001e-01
9.45736527e-01 -5.31409264e-01 -4.90455300e-01 6.90003157e-01
9.81370091e-01 -1.29316163e+00 4.02127862e-01 1.19662082e+00
-1.51735234e+00 -2.81509876e-01 -1.12222350e+00 3.21255475e-01
-4.58759040e-01 -4.26144183e-01 5.87598503e-01 6.75503314e-01
-1.28119338e+00 6.52594864e-01 -7.31142759e-01 -3.18752259e-01
1.00539170e-01 3.44083548e-01 -1.85886845e-02 -1.20046839e-01
-1.17158401e+00 9.57229376e-01 2.48800963e-01 -1.86688825e-01
-1.82700619e-01 -1.20188773e+00 -5.90389848e-01 -1.27773238e-02
4.38953340e-01 -2.63628006e-01 1.11166847e+00 -3.47652197e-01
-1.44215119e+00 1.02548027e+00 3.93450886e-01 -4.03964281e-01
8.39975357e-01 2.58367836e-01 -1.99201718e-01 -1.14651665e-01
5.88955916e-02 2.80766845e-01 -1.91702887e-01 -7.68234611e-01
-7.92987764e-01 -2.09653363e-01 4.83832747e-01 -1.05692938e-01
5.45475855e-02 3.67650628e-01 1.19445972e-01 -3.35234821e-01
9.83081460e-02 -7.24989593e-01 6.63852273e-03 -2.59224504e-01
1.47747159e-01 -7.38601387e-01 8.41467857e-01 -1.15645982e-01
9.98578906e-01 -1.90106595e+00 6.74506649e-02 2.17269033e-01
2.81837791e-01 3.10721874e-01 4.60949451e-01 1.05803502e+00
-2.90443808e-01 4.78208922e-02 7.33847246e-02 3.12226295e-01
6.20296061e-01 -8.95547718e-02 -3.35850626e-01 4.75663811e-01
-1.82677895e-01 3.75735253e-01 -1.12563384e+00 -4.63314533e-01
6.42296076e-01 2.19619572e-01 -3.55113089e-01 -2.09655628e-01
-2.80501723e-01 -6.44355342e-02 -4.08315301e-01 3.44933756e-03
8.49987626e-01 4.82082143e-02 2.19539016e-01 6.70214832e-01
-9.65278089e-01 5.04576266e-01 -1.72278476e+00 1.55969095e+00
-3.56088400e-01 8.03735852e-01 4.83855605e-01 -1.33589184e+00
8.59660506e-01 6.68739378e-01 4.67143565e-01 -1.24653265e-01
2.20546454e-01 3.82262409e-01 1.82303786e-01 -1.83151633e-01
4.03486907e-01 -3.07553560e-02 1.72784582e-01 6.39706314e-01
-3.85894626e-02 -7.98201025e-01 8.00623059e-01 4.91221398e-01
9.32322562e-01 6.55645013e-01 8.33429247e-02 -7.63994634e-01
8.88409734e-01 -7.73224533e-02 2.05729008e-01 4.04925823e-01
-2.76591301e-01 1.07530197e-02 6.87444627e-01 -6.89404845e-01
-9.28851426e-01 -7.52898455e-01 -6.19671345e-01 6.22500896e-01
-2.58465260e-01 -9.67535973e-01 -6.99583828e-01 -4.29069966e-01
-4.34493452e-01 8.26519012e-01 -3.59753132e-01 3.23563635e-01
-4.77401406e-01 9.49254632e-02 2.61082441e-01 -5.35269231e-02
4.94287729e-01 -9.98946488e-01 -7.20654905e-01 1.67104378e-01
-1.73930693e-02 -8.24662209e-01 1.62748739e-01 2.90750206e-01
-6.22108519e-01 -1.05316913e+00 -6.82319403e-01 -1.26876736e+00
3.30408335e-01 -1.17904156e-01 8.23311627e-01 2.95931995e-01
-3.64499509e-01 4.94823366e-01 -1.65010273e-01 -8.94767165e-01
-7.21191645e-01 -8.01378340e-02 -1.37795821e-01 -6.97984874e-01
3.19902569e-01 -6.53436601e-01 -6.39274120e-02 -7.48733506e-02
-8.29169691e-01 3.27058792e-01 -8.24098811e-02 5.70507407e-01
-1.05111357e-02 1.14770971e-01 4.78198051e-01 -6.54313743e-01
2.92191356e-01 2.06897929e-01 -1.05100334e+00 3.39860409e-01
-6.59659147e-01 -2.92330503e-01 7.56498635e-01 1.55567542e-01
-6.12376988e-01 -4.05501008e-01 -4.89154086e-02 2.26639241e-01
-3.37833583e-01 4.03027475e-01 -1.37998924e-01 -3.47243637e-01
4.73866373e-01 1.69648111e-01 -1.95597783e-01 -1.09437235e-01
-6.63348138e-02 4.75959897e-01 1.98606864e-01 -1.04959393e+00
9.35056269e-01 -3.33471268e-01 4.51218188e-01 -1.14266002e+00
-3.34638447e-01 -3.51507872e-01 -3.64618659e-01 -4.65018034e-01
3.25279415e-01 -6.17201149e-01 -9.89103615e-01 4.61123526e-01
-1.28875315e+00 -2.51893640e-01 -4.33836728e-01 7.51466393e-01
-2.30232999e-01 3.01716536e-01 -4.65809375e-01 -7.85249352e-01
3.06512207e-01 -9.88600194e-01 3.41933936e-01 3.35970670e-01
-2.43701354e-01 -1.21330249e+00 3.64710987e-01 4.26392853e-01
9.65483338e-02 1.75268337e-01 6.03243709e-01 -5.48779011e-01
-6.90878808e-01 -3.73796225e-01 6.46090508e-03 2.88293123e-01
-5.42796493e-01 5.95826566e-01 -7.54212916e-01 -5.71130626e-02
-1.72265783e-01 -3.28843623e-01 1.23431779e-01 1.47615716e-01
8.65630388e-01 -7.27193132e-02 -1.40416592e-01 -1.41078487e-01
1.49190068e+00 7.00601116e-02 7.34205425e-01 5.70067167e-01
-4.85917367e-02 5.59380293e-01 2.18742266e-01 1.80232525e-01
5.65493703e-01 4.74789977e-01 2.02319562e-01 4.36353534e-01
1.19866811e-01 8.02315995e-02 3.04729789e-01 1.45843279e+00
-4.16165859e-01 4.55764979e-01 -1.13655853e+00 7.10550725e-01
-1.58591413e+00 -9.46551740e-01 -1.02370489e+00 2.27408099e+00
8.98002028e-01 4.48013037e-01 2.77653843e-01 4.46568400e-01
6.33697927e-01 -9.09014717e-02 5.05492926e-01 -1.08520806e+00
2.24897474e-01 7.30397701e-01 -1.85050517e-01 7.98748851e-01
-4.72911954e-01 5.22284329e-01 6.02673912e+00 7.90522456e-01
-7.86876440e-01 -2.90528834e-01 1.33715540e-01 4.77065235e-01
-6.19356751e-01 3.44281912e-01 -4.59168315e-01 2.63973922e-01
1.12072861e+00 -9.37375546e-01 4.87497538e-01 4.98486161e-01
6.23134449e-02 -1.84846967e-01 -9.31675851e-01 3.61570448e-01
-7.73483589e-02 -1.57204211e+00 -5.40941656e-01 3.54866087e-01
1.01361656e+00 -5.87837175e-02 -2.89593548e-01 5.11350393e-01
3.98521543e-01 -8.41274023e-01 5.68406463e-01 3.73114318e-01
1.22831531e-01 -1.10600495e+00 5.97003818e-01 5.98770440e-01
-1.02960598e+00 2.76392639e-01 -2.83190748e-03 -9.27077949e-01
-5.53637385e-01 4.43638414e-01 -6.72456801e-01 9.80515599e-01
3.29515278e-01 3.47442061e-01 -7.52189569e-03 1.53542411e+00
-1.78526610e-01 6.97357178e-01 -3.72826844e-01 -7.92802691e-01
1.51759312e-01 -3.50229770e-01 8.12116623e-01 1.14727235e+00
-2.18827669e-02 3.47583652e-01 4.44517359e-02 1.14887381e+00
2.68465966e-01 1.59748152e-01 -2.74885714e-01 1.18143633e-02
2.38042504e-01 1.34059775e+00 -8.31114054e-01 -1.75549611e-01
-2.85238236e-01 2.29145974e-01 -3.27694029e-01 3.71804722e-02
-8.13547611e-01 -9.60388422e-01 1.35420978e-01 4.39228415e-01
2.66491234e-01 -3.23690444e-01 -5.27569175e-01 -8.12817395e-01
4.39481959e-02 -8.53349090e-01 -1.75406277e-01 -7.20543265e-01
-5.81622839e-01 1.88063323e-01 1.93709567e-01 -6.18683815e-01
-1.72556460e-01 -5.66122890e-01 -8.34190845e-01 7.99112558e-01
-1.05353034e+00 -7.86074281e-01 2.75342017e-01 4.79933321e-02
6.25527948e-02 9.51441228e-02 1.08270907e+00 1.81767121e-01
-1.88779742e-01 3.99745882e-01 -1.47329466e-02 -1.33388355e-01
3.43212157e-01 -1.45091653e+00 3.45425569e-02 7.88430154e-01
1.48301288e-01 5.66464782e-01 1.04525197e+00 -1.13538086e-01
-1.48527384e+00 -4.86318260e-01 1.89344668e+00 -2.73372114e-01
1.24100220e+00 -1.95741057e-01 -6.20234728e-01 5.98726988e-01
4.24804002e-01 -4.09512788e-01 2.97601163e-01 9.74071771e-02
4.58711460e-02 -1.55669767e-02 -8.40208352e-01 2.75364578e-01
5.71482062e-01 -1.34394184e-01 -1.11574852e+00 5.33185244e-01
5.20153999e-01 -7.99089253e-01 -1.03255129e+00 4.64204811e-02
3.28293383e-01 -9.51586187e-01 4.88466591e-01 -1.32080331e-01
7.96163201e-01 -3.49776596e-01 7.66412988e-02 -8.90273690e-01
1.61457762e-01 -1.17949712e+00 3.98705184e-01 1.27310240e+00
2.52806097e-01 -5.48351824e-01 7.80389845e-01 3.24830145e-01
-3.89459491e-01 -6.66436613e-01 -9.85171378e-01 -6.24940336e-01
7.17766821e-01 -7.38748193e-01 3.73170257e-01 1.13402116e+00
9.22434926e-01 9.84103009e-02 3.77943963e-01 -4.56026882e-01
7.11298823e-01 5.86692058e-03 9.44133997e-01 -1.48017442e+00
-1.62087500e-01 -9.15877819e-01 -9.07951117e-01 -7.15632141e-01
-1.84041500e-01 -1.04837453e+00 -2.11062774e-01 -1.67500496e+00
-3.29910547e-01 -1.19554877e-01 1.10612608e-01 3.74551862e-01
3.19408804e-01 -7.32099116e-02 1.54145509e-01 -3.15086901e-01
-7.60230660e-01 1.34516031e-01 1.52969825e+00 2.99591810e-01
1.16522741e-02 3.44488829e-01 -4.73781556e-01 6.54850483e-01
4.58462566e-01 -8.40494111e-02 -5.26450686e-02 1.51351541e-01
6.21920109e-01 2.49240607e-01 6.58808291e-01 -1.30371177e+00
3.58787119e-01 -2.83815563e-01 2.69029643e-02 -5.71772695e-01
-8.66669938e-02 -8.13980043e-01 9.19885188e-02 6.93151891e-01
-3.74514222e-01 -6.99511096e-02 5.18717229e-01 -1.67820156e-01
-2.90719450e-01 -3.39306056e-01 3.27461243e-01 -2.25604206e-01
-2.86264449e-01 -1.83575049e-01 -4.51731026e-01 3.59694332e-01
1.43432891e+00 -3.34510922e-01 -1.24177903e-01 -1.11956358e-01
-6.81208909e-01 4.09947217e-01 3.23555112e-01 2.19314843e-02
1.34630546e-01 -8.17468345e-01 -9.88221884e-01 1.73207864e-01
-6.15960777e-01 -2.93985438e-02 1.51721478e-01 9.80085731e-01
-8.08456838e-01 7.53560126e-01 -2.67366976e-01 -4.57908809e-01
-1.43024659e+00 3.98080051e-02 1.67099282e-01 -2.49382943e-01
-7.97590971e-01 8.96073878e-01 -4.12205040e-01 -6.90040767e-01
5.11189282e-01 -2.62648255e-01 1.26535490e-01 -2.82935232e-01
6.66407526e-01 5.39017081e-01 1.80173740e-01 7.70886093e-02
-3.43525797e-01 1.18093848e-01 2.18368977e-01 -1.48404688e-01
1.81738460e+00 1.13558158e-01 -5.20136952e-01 7.84428298e-01
8.74098063e-01 1.70735851e-01 -6.29476488e-01 2.73265660e-01
6.81826994e-02 2.82741226e-02 1.15653217e-01 -9.13028300e-01
-4.53888923e-01 7.72498548e-01 -7.79494792e-02 7.49280751e-01
7.44824708e-01 -3.93389136e-01 5.41232884e-01 1.69775978e-01
6.10929668e-01 -9.40473020e-01 -5.03054738e-01 7.69476354e-01
8.39497089e-01 -6.10720158e-01 2.64804691e-01 -3.66640240e-01
-3.06055956e-02 1.75305617e+00 4.29140985e-01 -2.51228571e-01
6.41498864e-01 4.37641442e-01 -4.04284656e-01 -1.31264731e-01
-9.07504022e-01 3.04884017e-01 1.60944283e-01 6.56258523e-01
8.58139396e-01 9.67237651e-02 -8.11575413e-01 3.89650881e-01
-8.27307165e-01 7.68435061e-01 1.00542748e+00 1.00121057e+00
-4.27665174e-01 -1.58154190e+00 -5.87994695e-01 -2.22582459e-01
-4.47215438e-01 -7.55906850e-02 -5.48977137e-01 1.44339502e+00
3.15407187e-01 6.65406227e-01 3.04018054e-02 -5.52190393e-02
2.56332248e-01 9.11817923e-02 1.16451466e+00 -4.15605038e-01
-9.82759356e-01 -2.16859087e-01 1.97669178e-01 1.58523738e-01
-2.84466356e-01 -8.39392304e-01 -1.35020161e+00 -1.34193373e+00
-3.66041839e-01 1.14074242e+00 9.01471972e-01 9.21640515e-01
-1.57406181e-01 6.74775362e-01 4.54489052e-01 -3.54027331e-01
-5.47308385e-01 -6.79743946e-01 -6.97846174e-01 -5.81999660e-01
-1.24021448e-01 -1.84687331e-01 -3.52367729e-01 -9.56205204e-02] | [8.88636589050293, 6.86614990234375] |
00ef12c2-1262-4b42-a243-ee6cc6cdd835 | co-attention-propagation-network-for-zero | 2304.03910 | null | https://arxiv.org/abs/2304.03910v1 | https://arxiv.org/pdf/2304.03910v1.pdf | Co-attention Propagation Network for Zero-Shot Video Object Segmentation | Zero-shot video object segmentation (ZS-VOS) aims to segment foreground objects in a video sequence without prior knowledge of these objects. However, existing ZS-VOS methods often struggle to distinguish between foreground and background or to keep track of the foreground in complex scenarios. The common practice of introducing motion information, such as optical flow, can lead to overreliance on optical flow estimation. To address these challenges, we propose an encoder-decoder-based hierarchical co-attention propagation network (HCPN) capable of tracking and segmenting objects. Specifically, our model is built upon multiple collaborative evolutions of the parallel co-attention module (PCM) and the cross co-attention module (CCM). PCM captures common foreground regions among adjacent appearance and motion features, while CCM further exploits and fuses cross-modal motion features returned by PCM. Our method is progressively trained to achieve hierarchical spatio-temporal feature propagation across the entire video. Experimental results demonstrate that our HCPN outperforms all previous methods on public benchmarks, showcasing its effectiveness for ZS-VOS. | ['Heng-Tao Shen', 'Xingguo Huang', 'Dan Huang', 'Fumin Shen', 'Yazhou Yao', 'Gensheng Pei'] | 2023-04-08 | null | null | null | null | ['video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 3.71599972e-01 -1.72647163e-01 -1.85996398e-01 7.85173941e-03
-4.20812100e-01 -3.40078145e-01 4.29484636e-01 -2.51217633e-01
-2.69803196e-01 4.03681517e-01 2.45643303e-01 -8.34472626e-02
3.42712134e-01 -5.57279646e-01 -8.82740796e-01 -6.95172846e-01
-1.01849616e-01 1.75459255e-02 7.95366526e-01 2.48069763e-01
2.17470273e-01 3.18741560e-01 -1.35991108e+00 3.51560652e-01
7.21401572e-01 1.01959777e+00 4.58802760e-01 1.05472505e+00
-1.49652317e-01 1.47317600e+00 -4.66817826e-01 -3.07887495e-01
2.78767616e-01 -5.24641097e-01 -1.03771555e+00 4.45691556e-01
6.19004965e-01 -7.41631687e-01 -5.55196166e-01 9.07215059e-01
8.08438882e-02 1.48733675e-01 4.08333033e-01 -1.52184343e+00
-6.24538720e-01 4.88619685e-01 -7.40858436e-01 7.52360642e-01
6.36235625e-02 5.17209768e-01 1.04065907e+00 -7.35834837e-01
7.72924960e-01 1.26811659e+00 4.75996226e-01 5.45195878e-01
-1.08707869e+00 -4.25724000e-01 6.61648154e-01 4.38459516e-01
-1.24709344e+00 -4.84657556e-01 6.93669558e-01 -7.20307469e-01
6.70095801e-01 7.91739300e-02 8.46301675e-01 8.37439597e-01
2.31687129e-01 1.46043956e+00 3.67692441e-01 -3.76722030e-02
1.52345337e-02 -2.45367348e-01 5.96835390e-02 7.36461282e-01
-2.52985749e-02 -1.75482005e-01 -7.05593646e-01 1.53059781e-01
1.01595259e+00 8.59220549e-02 -6.25274360e-01 -4.27490562e-01
-1.28659606e+00 4.84741926e-01 2.68656343e-01 1.95557788e-01
-4.74910557e-01 3.80238950e-01 3.30540985e-01 -8.79926682e-02
3.89384896e-01 3.43254134e-02 -4.54727679e-01 -2.49458760e-01
-1.34498227e+00 8.24509263e-02 5.07862628e-01 1.05274343e+00
7.57139683e-01 1.86363980e-02 -6.25073135e-01 4.45385337e-01
3.68366241e-01 2.41720662e-01 2.42204800e-01 -1.50453079e+00
2.02744842e-01 5.64288437e-01 1.43649265e-01 -8.51555884e-01
6.53689057e-02 -2.42997929e-01 -6.49288893e-01 1.54176336e-02
3.82742077e-01 -1.58232942e-01 -1.01289392e+00 1.73649216e+00
5.41212797e-01 8.04897845e-01 -1.56807646e-01 1.08241427e+00
6.99763060e-01 9.06728983e-01 2.39703745e-01 -1.24986954e-01
9.81155932e-01 -1.32326496e+00 -6.94390893e-01 -2.49209270e-01
3.33031684e-01 -4.77333009e-01 5.35499811e-01 -2.91420687e-02
-1.39911675e+00 -6.88532531e-01 -8.90089512e-01 -2.20994174e-01
3.33951488e-02 -2.02348039e-01 2.86174238e-01 3.03502262e-01
-1.03993213e+00 6.02635682e-01 -1.08191764e+00 -4.40133028e-02
8.34974229e-01 3.64788860e-01 1.71529036e-02 -9.06338990e-02
-9.74658668e-01 3.89846087e-01 3.81098121e-01 3.36087584e-01
-1.33688867e+00 -9.69832122e-01 -9.23376441e-01 1.87926963e-01
5.43458104e-01 -9.73177433e-01 1.18309689e+00 -1.41884792e+00
-1.49640512e+00 5.27856410e-01 -4.93442506e-01 -4.64906514e-01
6.53307021e-01 -4.49378908e-01 -1.25149593e-01 5.00005603e-01
1.94900572e-01 9.64644551e-01 1.09518218e+00 -1.19768977e+00
-1.10577226e+00 1.60280541e-01 1.13449275e-01 1.52284428e-01
-1.16444342e-01 1.81226626e-01 -1.04108715e+00 -5.09774804e-01
-1.78630412e-01 -6.46654069e-01 -1.97358236e-01 1.72643483e-01
-5.22780657e-01 -1.49466582e-02 1.17187977e+00 -7.33632743e-01
1.39552355e+00 -2.07530308e+00 4.34690356e-01 -2.20400393e-01
4.48643625e-01 5.55324018e-01 -7.80312568e-02 -3.43346335e-02
2.28425667e-01 -8.37953016e-02 -3.42129529e-01 -3.45705599e-01
-1.79881677e-01 3.01193565e-01 -1.12733558e-01 5.46820402e-01
6.22296512e-01 1.07586658e+00 -1.20081472e+00 -9.27681506e-01
4.13198590e-01 6.25800788e-01 -7.64172494e-01 3.75267535e-01
-4.10665363e-01 6.96376443e-01 -2.90551633e-01 7.20103383e-01
6.21213078e-01 -6.96526825e-01 -2.51440275e-02 -1.35460407e-01
-1.78115770e-01 7.21333548e-02 -9.69995499e-01 1.64434206e+00
1.89762202e-03 8.53871822e-01 1.47253588e-01 -6.99857175e-01
1.70395166e-01 2.39403814e-01 8.95279646e-01 -3.70023012e-01
1.89911097e-01 -1.28371745e-01 -4.94015850e-02 -6.09495103e-01
4.14511561e-01 3.26467097e-01 3.88598889e-01 1.47633523e-01
3.50409836e-01 5.58878362e-01 4.07749802e-01 4.87885088e-01
1.06547105e+00 4.62855220e-01 -9.41116884e-02 -9.53402519e-02
7.21822381e-01 -2.04473987e-01 1.04125571e+00 7.07179368e-01
-6.13413632e-01 7.48016775e-01 6.37118459e-01 -2.65497953e-01
-9.08718646e-01 -1.20552552e+00 2.97814399e-01 1.13514912e+00
5.20932853e-01 -3.82655025e-01 -8.86485040e-01 -7.99991488e-01
-1.82202652e-01 4.45163131e-01 -6.86317861e-01 -5.35314269e-02
-7.50768602e-01 -4.83417183e-01 2.05011562e-01 6.58985376e-01
5.21468580e-01 -9.93831694e-01 -1.04082596e+00 4.33764011e-01
-5.62814713e-01 -1.40088344e+00 -9.57283020e-01 -2.59562522e-01
-5.95051944e-01 -1.18286085e+00 -1.06643438e+00 -8.00566912e-01
4.68802989e-01 5.00993311e-01 1.14696503e+00 2.16034010e-01
-3.26974928e-01 5.70569813e-01 -2.84461349e-01 -6.30957559e-02
-3.42863262e-01 -3.19883041e-02 -5.24146318e-01 6.47374034e-01
1.94332510e-01 -5.28417051e-01 -9.30058956e-01 2.14156896e-01
-9.04896915e-01 3.30097079e-01 5.50344169e-01 5.50409615e-01
4.97209609e-01 -3.34067971e-01 2.13908061e-01 -8.13154221e-01
-1.55368552e-01 -6.80245638e-01 -6.18355215e-01 2.84024268e-01
-1.10118069e-01 -2.19695076e-01 1.97109222e-01 -4.39816266e-01
-1.04909992e+00 2.17686042e-01 1.24199830e-01 -9.35561478e-01
-6.77544773e-02 -1.37674101e-02 -4.01191533e-01 3.83389406e-02
-1.02329634e-01 4.39234614e-01 -4.10311371e-01 -2.47143775e-01
4.38951313e-01 3.42454851e-01 8.77265275e-01 -2.14988396e-01
6.47983253e-01 7.18832076e-01 -2.14408293e-01 -8.49983096e-01
-9.60987091e-01 -7.19059050e-01 -9.85656381e-01 -6.15166426e-01
1.38042343e+00 -9.90522385e-01 -6.74401879e-01 5.57864428e-01
-1.37706423e+00 -5.33442795e-01 -1.96390316e-01 1.84370682e-01
-5.92767477e-01 4.58599925e-01 -9.15055394e-01 -8.47441792e-01
-1.55897304e-01 -1.21264672e+00 1.24837387e+00 3.70431364e-01
-3.57968025e-02 -1.09217322e+00 -1.17858164e-01 3.80272657e-01
1.65273011e-01 3.71500164e-01 5.55523694e-01 -1.86024278e-01
-1.25347674e+00 2.26486027e-01 -4.32189763e-01 3.17990690e-01
1.73466384e-01 2.90324360e-01 -1.00135291e+00 -1.56519696e-01
-3.02522659e-01 1.40679270e-01 1.02414608e+00 6.42488241e-01
1.16725397e+00 -2.71234661e-01 -3.48982483e-01 7.81932473e-01
1.25186670e+00 2.98282951e-01 6.60956800e-01 1.70046985e-01
1.32476676e+00 3.65960747e-01 3.69968921e-01 3.19194049e-01
5.38433492e-01 5.12371600e-01 5.11634588e-01 -2.09613487e-01
-3.63410234e-01 -5.40939718e-02 6.32185757e-01 7.58266509e-01
-1.15164667e-01 -3.64549637e-01 -7.06041455e-01 9.03048575e-01
-2.02976775e+00 -1.18362856e+00 -1.97290525e-01 1.89579999e+00
6.78646505e-01 1.46978885e-01 2.09212422e-01 -9.61571187e-02
9.08895016e-01 3.31146806e-01 -6.30870163e-01 1.05095468e-01
-1.43036246e-01 -2.02902541e-01 5.23620784e-01 5.44651031e-01
-1.48453355e+00 1.13748944e+00 5.90153599e+00 4.86006498e-01
-1.07341766e+00 2.46531680e-01 6.55293405e-01 -3.32049966e-01
7.23212725e-03 -1.67961549e-02 -6.64597809e-01 7.77201533e-01
7.59907842e-01 3.02772177e-03 2.37804025e-01 4.91477221e-01
2.16767669e-01 -1.40633017e-01 -1.18795836e+00 7.38736868e-01
3.39535670e-03 -1.50311756e+00 7.95136485e-03 -1.23669907e-01
9.64427829e-01 1.47143528e-01 -7.91648328e-02 7.54300728e-02
4.18253362e-01 -6.48175120e-01 1.01939356e+00 5.90463340e-01
3.26988965e-01 -6.30564034e-01 5.97238839e-01 1.49172306e-01
-1.50614762e+00 -1.08291112e-01 -1.06651053e-01 2.68898249e-01
6.21298134e-01 3.43402922e-01 -4.60543066e-01 6.28412485e-01
7.01615751e-01 1.10203862e+00 -4.18506354e-01 1.25975180e+00
-2.81773624e-03 7.31175482e-01 -1.12195853e-02 4.67148513e-01
5.51189780e-01 -1.25990838e-01 6.94883347e-01 1.46743691e+00
-4.11269367e-02 -4.36725887e-03 3.11559409e-01 1.11529648e+00
3.43107060e-02 -1.38147622e-01 -8.50248486e-02 -6.87731132e-02
8.48098695e-02 1.09029698e+00 -9.81163025e-01 -7.36148000e-01
-7.34653056e-01 1.31103110e+00 1.75022736e-01 5.71943641e-01
-1.20396817e+00 -1.28361538e-01 9.21059549e-01 1.66626677e-01
8.60260844e-01 -1.69687673e-01 5.61468862e-02 -1.23934996e+00
-2.12836012e-01 -4.53626066e-01 3.81479055e-01 -7.22542226e-01
-9.77788448e-01 3.95075351e-01 -2.41020605e-01 -1.10136962e+00
8.44791383e-02 -2.63516665e-01 -6.94138587e-01 6.92936718e-01
-1.76194930e+00 -1.23245168e+00 -3.09039712e-01 6.22704983e-01
9.28122222e-01 2.68242508e-01 7.10836425e-02 6.87230229e-01
-8.83720815e-01 2.74576008e-01 -2.35705853e-01 4.98362869e-01
5.02096236e-01 -1.18283856e+00 4.77466017e-01 1.29158199e+00
1.40028372e-01 2.12189704e-01 4.98704851e-01 -8.41953814e-01
-1.29088700e+00 -1.68197179e+00 6.78298950e-01 -4.67094094e-01
6.07103825e-01 -2.72599012e-01 -1.14801705e+00 8.79386485e-01
2.84046888e-01 4.43016678e-01 3.74546975e-01 -5.30710161e-01
-1.48351416e-01 1.50910541e-01 -4.72154081e-01 6.35461152e-01
1.13379061e+00 -3.76399934e-01 -3.13098341e-01 1.24522477e-01
9.09727454e-01 -4.14254040e-01 -6.53663337e-01 2.33786613e-01
3.78827900e-01 -1.01392043e+00 1.04073739e+00 -6.76868916e-01
7.03564942e-01 -6.06601179e-01 -1.74683947e-02 -8.34022641e-01
-4.26364690e-01 -9.02734995e-01 -6.72281921e-01 1.28571045e+00
3.52221727e-03 -2.37848237e-02 8.62375021e-01 4.58181649e-01
-3.69443059e-01 -6.24221742e-01 -7.34869421e-01 -6.03163242e-01
-1.70774177e-01 -4.66888398e-01 3.44974577e-01 8.68397534e-01
-4.90763605e-01 2.49499798e-01 -5.70763469e-01 5.36639512e-01
6.81206524e-01 2.29209423e-01 6.45232916e-01 -1.02573562e+00
-3.23712111e-01 -6.77598834e-01 -4.67302471e-01 -1.47748494e+00
2.17993990e-01 -6.89719319e-01 2.90034652e-01 -1.54569340e+00
3.13123912e-01 2.37124301e-02 -4.29152220e-01 2.23122776e-01
-5.81123412e-01 2.72004485e-01 5.11177123e-01 2.23633900e-01
-1.36730194e+00 5.08982718e-01 1.50306499e+00 -2.40756109e-01
-1.94440112e-01 -1.57752067e-01 -2.85563022e-01 8.78112674e-01
3.81668091e-01 -4.55141515e-01 -3.76426667e-01 -5.55025816e-01
-3.40660453e-01 1.48806468e-01 5.48308611e-01 -9.95116770e-01
4.88568097e-01 -2.16688156e-01 5.23845911e-01 -7.38969564e-01
1.63064390e-01 -6.85017049e-01 -3.83785693e-03 5.00661314e-01
-1.30893379e-01 -2.11181298e-01 7.02657178e-02 8.59054327e-01
-1.59442306e-01 1.12736195e-01 8.29697490e-01 -9.72838402e-02
-1.03440523e+00 6.77745283e-01 -7.04311490e-01 2.69884229e-01
1.03523302e+00 -3.16098660e-01 -5.37649319e-02 -2.05144480e-01
-6.45421207e-01 6.19743049e-01 3.79237384e-01 5.44982970e-01
5.58470309e-01 -1.07822323e+00 -5.85653067e-01 1.60153076e-01
-7.43708983e-02 2.58758903e-01 4.67531174e-01 1.16808724e+00
-6.76353037e-01 2.46853650e-01 -1.71085328e-01 -9.66984332e-01
-1.18212521e+00 7.81579614e-01 4.48262602e-01 2.95709725e-02
-9.22370017e-01 1.09117723e+00 7.06698716e-01 3.84097546e-01
3.51530969e-01 -4.47589129e-01 -1.50398910e-01 -1.12987220e-01
7.36404836e-01 5.03473818e-01 -5.16785264e-01 -9.53268290e-01
-2.85794616e-01 5.18162608e-01 -1.29139759e-02 7.80638754e-02
1.09424829e+00 -4.86795008e-01 -3.12157832e-02 5.70307612e-01
1.22665024e+00 -2.73784459e-01 -2.24928308e+00 -3.00437987e-01
1.06032029e-01 -6.35175765e-01 1.19620979e-01 -3.05204630e-01
-1.54270065e+00 9.39989328e-01 3.11444908e-01 7.63943493e-02
1.22046041e+00 -3.02615240e-02 1.18520451e+00 -1.62407607e-01
3.26751284e-02 -8.58443379e-01 1.33024648e-01 4.16348904e-01
2.45189354e-01 -1.18455875e+00 -2.95623839e-01 -5.04460096e-01
-7.01034069e-01 9.72065449e-01 7.01979578e-01 -1.41704902e-01
5.67142904e-01 3.13240677e-01 -1.13151157e-02 -2.28224206e-03
-9.56441522e-01 -3.52444112e-01 4.25903440e-01 4.71975088e-01
1.83906242e-01 -3.87762576e-01 2.82037556e-01 4.20543402e-01
4.99958724e-01 1.02588177e-01 4.50196505e-01 1.04713333e+00
-4.84710932e-01 -7.01753676e-01 -2.82539576e-01 2.04439834e-01
-7.12396622e-01 -4.01786119e-02 -3.25742155e-01 5.00445664e-01
4.07084197e-01 7.76389599e-01 4.17439580e-01 -2.29613304e-01
-9.07494500e-02 5.59012964e-03 4.48153079e-01 -3.06101859e-01
-5.56659043e-01 4.44703847e-01 -3.31671596e-01 -9.02133286e-01
-7.64063537e-01 -8.53586435e-01 -1.16278505e+00 -1.18114628e-01
-1.46785051e-01 -4.09876555e-02 9.71103683e-02 1.01035321e+00
3.71070713e-01 1.10209763e+00 4.10327673e-01 -1.16315019e+00
2.68456429e-01 -5.21300316e-01 -3.52888167e-01 5.10243714e-01
8.23400974e-01 -6.03793621e-01 -1.66895688e-01 7.73287296e-01] | [9.197833061218262, -0.1326318234205246] |
a49dfac7-7f44-4a83-8222-ab2b58fa2af3 | towards-accurate-facial-landmark-detection-1 | 2208.10808 | null | https://arxiv.org/abs/2208.10808v1 | https://arxiv.org/pdf/2208.10808v1.pdf | Towards Accurate Facial Landmark Detection via Cascaded Transformers | Accurate facial landmarks are essential prerequisites for many tasks related to human faces. In this paper, an accurate facial landmark detector is proposed based on cascaded transformers. We formulate facial landmark detection as a coordinate regression task such that the model can be trained end-to-end. With self-attention in transformers, our model can inherently exploit the structured relationships between landmarks, which would benefit landmark detection under challenging conditions such as large pose and occlusion. During cascaded refinement, our model is able to extract the most relevant image features around the target landmark for coordinate prediction, based on deformable attention mechanism, thus bringing more accurate alignment. In addition, we propose a novel decoder that refines image features and landmark positions simultaneously. With few parameter increasing, the detection performance improves further. Our model achieves new state-of-the-art performance on several standard facial landmark detection benchmarks, and shows good generalization ability in cross-dataset evaluation. | ['Jae-Joon Han', 'Seungju Han', 'Seon-Min Rhee', 'Zidong Guo', 'Hui Li'] | 2022-08-23 | towards-accurate-facial-landmark-detection | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Towards_Accurate_Facial_Landmark_Detection_via_Cascaded_Transformers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Towards_Accurate_Facial_Landmark_Detection_via_Cascaded_Transformers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['face-alignment', 'facial-landmark-detection'] | ['computer-vision', 'computer-vision'] | [-1.34830967e-01 1.91843718e-01 -2.95723110e-01 -7.82592773e-01
-9.61697996e-01 -1.17526151e-01 3.89289081e-01 -1.07679449e-01
-3.38072985e-01 -2.24836683e-03 2.57001668e-01 3.04455400e-01
2.12937906e-01 -5.31280935e-01 -7.05006301e-01 -5.32791853e-01
6.06251433e-02 4.35301483e-01 1.07182100e-01 -1.73457205e-01
1.86340764e-01 6.89504147e-01 -1.33726001e+00 -2.83381164e-01
4.37005341e-01 1.29589140e+00 -1.80287287e-02 1.08975261e-01
2.15911135e-01 1.79659411e-01 -1.07235953e-01 -5.74492991e-01
3.12304139e-01 -8.46538285e-04 -4.72740054e-01 6.92856237e-02
1.02716339e+00 -5.15911877e-01 -2.03135505e-01 1.10835326e+00
6.99146509e-01 -1.74789980e-01 4.84864831e-01 -9.76539254e-01
-4.72005010e-01 2.23071381e-01 -1.10887754e+00 -6.33609369e-02
3.35070223e-01 -9.19745937e-02 1.10788143e+00 -1.47778904e+00
4.67722118e-01 1.41427386e+00 6.99006200e-01 6.99897885e-01
-1.19861197e+00 -1.13986421e+00 5.07773280e-01 2.53608197e-01
-1.90855777e+00 -9.90175605e-01 9.33281898e-01 -4.67355102e-02
6.81566298e-01 -3.96448839e-03 6.57189846e-01 6.78834736e-01
5.17770983e-02 6.77430689e-01 4.05991673e-01 -1.91583201e-01
-1.70069054e-01 -2.70575285e-01 -5.56107819e-01 1.39905417e+00
5.92715712e-03 -1.00834385e-01 -6.80812657e-01 7.91395083e-02
1.01016486e+00 1.63102672e-01 -6.95142746e-02 -4.32509243e-01
-7.83884525e-01 5.93887508e-01 7.86077023e-01 5.71644977e-02
-3.26479584e-01 2.91912436e-01 1.04278952e-01 -2.35604346e-01
7.67711639e-01 1.64114788e-01 -3.81297767e-01 2.62377113e-01
-9.46438551e-01 -1.07835770e-01 1.51240313e-03 9.77741778e-01
8.52494121e-01 -1.81862220e-01 -3.35694432e-01 7.47340620e-01
9.01236773e-01 6.57404482e-01 1.62411883e-01 -8.12648594e-01
3.70281458e-01 6.89543724e-01 -1.13984033e-01 -1.23697972e+00
-6.45434678e-01 -4.65882093e-01 -7.65543222e-01 5.50003015e-02
1.54194981e-01 3.81733887e-02 -1.06213439e+00 1.93171632e+00
6.77525222e-01 4.88957494e-01 -4.00218427e-01 9.42853868e-01
8.26994479e-01 4.02333677e-01 1.60060093e-01 -2.30260253e-01
1.51156795e+00 -9.54363823e-01 -5.30102193e-01 -3.41463834e-01
4.19960141e-01 -8.31636906e-01 7.36459494e-01 -6.04534894e-03
-1.01235640e+00 -6.39856339e-01 -6.96304083e-01 -4.59720463e-01
2.26605117e-01 4.06191647e-01 5.39102852e-01 4.52368170e-01
-1.21264660e+00 2.62146890e-01 -9.97095942e-01 -3.12943459e-01
8.77075434e-01 7.70619512e-01 -5.75653791e-01 1.15664616e-01
-6.17393374e-01 6.43522263e-01 -3.20369840e-01 4.81267184e-01
-9.02527034e-01 -7.34912217e-01 -1.16978419e+00 1.39172196e-01
3.40717673e-01 -6.62805021e-01 1.28728056e+00 -5.97456813e-01
-1.58869553e+00 1.15584338e+00 -9.33351338e-01 -1.44433841e-01
4.53162402e-01 -3.69730264e-01 -1.74512595e-01 3.06961060e-01
3.70639980e-01 1.01119685e+00 1.10491884e+00 -1.05370641e+00
-6.57252133e-01 -7.48357296e-01 -3.29687178e-01 1.92912906e-01
-5.18282950e-01 1.88958108e-01 -1.04053986e+00 -4.00532722e-01
3.98522973e-01 -8.05113494e-01 -3.45781177e-01 4.62350398e-01
-3.66747558e-01 -5.63965678e-01 7.56099582e-01 -4.11495358e-01
9.74248886e-01 -2.10982609e+00 1.39739394e-01 4.77693528e-01
5.79448819e-01 1.25389248e-01 -3.04284722e-01 -1.13377742e-01
1.25539914e-01 1.12162992e-01 2.36977473e-01 -8.65852475e-01
-1.95280582e-01 -2.67754346e-01 -1.60946354e-01 7.54947126e-01
4.28831875e-01 1.14628482e+00 -6.62810802e-01 -7.41951942e-01
1.49522379e-01 7.28747010e-01 -6.13883913e-01 1.01357833e-01
-5.46952151e-02 5.89251876e-01 -7.64717400e-01 9.71179307e-01
7.97965646e-01 -2.97534496e-01 -1.04724005e-01 -5.39511919e-01
-1.61785409e-02 9.84204933e-03 -7.89573133e-01 1.96613336e+00
-6.86071217e-01 5.06675661e-01 -3.26491818e-02 -5.40299416e-01
1.09419930e+00 1.64933652e-01 5.36667287e-01 -9.02963459e-01
1.75725698e-01 -4.16010804e-02 -2.69480854e-01 -1.62093669e-01
1.32649943e-01 1.17743693e-01 1.12111524e-01 1.85607746e-01
9.78966355e-02 2.67210513e-01 -2.02175081e-01 -3.79993916e-02
5.65289259e-01 1.40588909e-01 2.69942671e-01 -9.87994373e-02
6.98377371e-01 -6.13320351e-01 7.74841964e-01 -1.76221691e-02
-1.58883557e-01 6.70428991e-01 3.98571253e-01 -6.84543133e-01
-7.70274222e-01 -7.70538330e-01 -3.35318267e-01 1.26183724e+00
3.77594531e-01 -7.62655318e-01 -7.11337328e-01 -7.36498654e-01
1.28088761e-02 -6.59851506e-02 -8.67304087e-01 -1.92023769e-01
-6.50640845e-01 -5.43435752e-01 4.35955584e-01 4.89221334e-01
5.91886640e-01 -6.28467560e-01 -3.77952665e-01 1.05534784e-01
7.68421516e-02 -1.19584870e+00 -1.00008106e+00 -4.41115022e-01
-7.13397861e-01 -1.09845424e+00 -7.67665684e-01 -1.11630487e+00
1.13958240e+00 4.41771805e-01 7.81692445e-01 3.48251432e-01
-2.99886733e-01 2.09076941e-01 -1.56662129e-02 -3.03242654e-01
3.37205827e-01 6.48349822e-02 7.11836442e-02 5.73552012e-01
4.39787686e-01 -3.14476103e-01 -9.77774560e-01 6.50712669e-01
-2.12732911e-01 -4.70771901e-02 7.25991905e-01 7.86429524e-01
9.54119980e-01 -2.98982501e-01 2.99737662e-01 -5.67209899e-01
5.01237772e-02 -3.16699483e-02 -7.94826746e-01 2.52542675e-01
-5.31805575e-01 7.19588995e-02 4.78706270e-01 -1.00756161e-01
-9.51834142e-01 5.23827672e-01 -3.91268313e-01 -5.72411358e-01
1.32738709e-01 8.32537413e-02 -5.38820982e-01 -7.19901919e-01
2.76229948e-01 2.26437122e-01 7.13331923e-02 -5.51998675e-01
2.40241066e-01 2.61278361e-01 3.50720495e-01 -4.41008449e-01
9.30134475e-01 6.94746196e-01 2.94736087e-01 -6.84199035e-01
-8.18837404e-01 -3.16443563e-01 -8.51662874e-01 -2.09459513e-01
6.58533990e-01 -1.15731752e+00 -1.06235838e+00 3.36654335e-01
-1.22068906e+00 -6.30434752e-02 3.35076511e-01 1.67380437e-01
-3.09543759e-01 2.18197539e-01 -4.02909547e-01 -6.46263301e-01
-5.71493506e-01 -1.36452031e+00 1.88410985e+00 4.96331602e-01
-5.58608118e-03 -6.92717373e-01 -1.49118200e-01 1.89189360e-01
2.47856379e-01 2.16649473e-01 4.73473608e-01 -2.33522594e-01
-7.90325165e-01 -4.33874607e-01 -3.32491964e-01 -1.25938416e-01
1.12960920e-01 1.71464384e-01 -9.75229502e-01 -3.45588654e-01
-4.49244142e-01 -1.61252961e-01 8.46406519e-01 2.22475305e-01
1.29909277e+00 -2.55084038e-01 -6.60760641e-01 1.09027886e+00
1.06017160e+00 -8.52868631e-02 3.51808101e-01 -8.90088733e-03
9.85532045e-01 4.37798470e-01 6.53722048e-01 3.87664139e-01
7.07081676e-01 1.04352331e+00 5.36493957e-01 -4.02529210e-01
-5.10742925e-02 -6.22227490e-01 1.50489748e-01 2.48533204e-01
-1.59264430e-01 3.55508775e-01 -7.99830139e-01 2.94678241e-01
-1.64635134e+00 -6.35067582e-01 2.53690958e-01 2.07793236e+00
6.70167625e-01 -1.11309879e-01 1.95401132e-01 -2.61678249e-01
6.53263748e-01 2.43764848e-01 -6.25307620e-01 2.10092187e-01
2.36367211e-01 3.41643423e-01 2.41467014e-01 6.44957900e-01
-1.22213614e+00 1.55260479e+00 5.95749998e+00 8.13213170e-01
-1.36766481e+00 9.67919901e-02 9.63497877e-01 -7.12064058e-02
3.30059491e-02 -3.35772455e-01 -1.25508881e+00 2.81221449e-01
2.78643817e-01 -1.15406506e-01 -3.19542177e-02 1.00246632e+00
1.87186047e-01 1.36632353e-01 -1.12559533e+00 1.23301911e+00
3.14184487e-01 -1.26225352e+00 7.05367997e-02 1.30032346e-01
5.96894026e-01 -2.23067358e-01 4.09013480e-01 -3.14611122e-02
-1.90532640e-01 -1.15516007e+00 6.90622807e-01 3.74220163e-01
1.14784455e+00 -9.31776702e-01 5.82232475e-01 -7.73997083e-02
-1.53656805e+00 1.39424428e-01 -4.68835562e-01 2.89331049e-01
7.52252042e-02 2.05017269e-01 -8.88948560e-01 1.65428489e-01
4.88780946e-01 9.82103705e-01 -8.10574472e-01 1.12211823e+00
-4.95917976e-01 1.67551726e-01 -4.94250387e-01 3.51004899e-02
2.15963763e-03 -7.54510006e-03 1.41947806e-01 9.57279563e-01
3.55539531e-01 6.01797514e-02 1.12700172e-01 4.69183624e-01
-2.42755949e-01 4.23758924e-01 -4.18202370e-01 6.06624901e-01
5.94136357e-01 1.51008677e+00 -7.25017309e-01 2.87543051e-02
-3.56316239e-01 9.69880104e-01 5.51796556e-01 1.61030442e-01
-8.54645371e-01 -6.83964193e-02 9.27611649e-01 2.18573794e-01
2.84414083e-01 -1.62836567e-01 -2.05235809e-01 -9.38186705e-01
2.34095603e-01 -5.41810751e-01 2.25567371e-01 -3.67712796e-01
-7.93704212e-01 7.86757529e-01 -4.03153032e-01 -1.18059897e+00
-2.18729954e-02 -4.23546851e-01 -6.09961390e-01 6.78636670e-01
-1.69332576e+00 -1.59161699e+00 -5.47148168e-01 6.63188040e-01
3.79582316e-01 -1.07163273e-01 7.99892664e-01 3.82243752e-01
-9.17859137e-01 1.21952879e+00 -4.32720035e-01 5.45373917e-01
9.15884435e-01 -8.61342311e-01 6.99902952e-01 8.14045191e-01
5.07820308e-01 7.71428645e-01 2.08332688e-01 -4.68440503e-01
-1.36610186e+00 -1.42254138e+00 9.65177476e-01 -3.69630486e-01
2.91048676e-01 -5.86332202e-01 -6.52805448e-01 6.83275938e-01
-3.44266683e-01 5.57381928e-01 6.40537977e-01 2.45575026e-01
-6.99292779e-01 -6.04713857e-01 -1.05214405e+00 6.78655803e-01
1.50234807e+00 -5.28976560e-01 -5.81348240e-02 5.20021558e-01
4.16189015e-01 -7.55334675e-01 -6.74809992e-01 4.29874808e-01
8.71692181e-01 -7.80007660e-01 1.01662505e+00 -4.56680000e-01
1.31763190e-01 -3.39954346e-01 2.69194245e-01 -9.84314144e-01
-4.79149550e-01 -9.16114211e-01 -1.16195820e-01 1.11844599e+00
4.43297118e-01 -4.44617242e-01 1.05465817e+00 5.78470588e-01
2.46862485e-03 -1.11226940e+00 -1.23414314e+00 -3.48249942e-01
-4.11707371e-01 -1.88695550e-01 8.46225202e-01 4.27613437e-01
-1.50402635e-01 4.51530427e-01 -4.46700066e-01 4.82280910e-01
7.32086182e-01 2.14748144e-01 9.56108868e-01 -1.33440685e+00
3.04471314e-01 -6.59548938e-01 -8.92464221e-01 -1.62408578e+00
4.45366681e-01 -7.97245920e-01 -8.73057637e-03 -1.14607012e+00
1.26906097e-01 -4.63063449e-01 -1.30582273e-01 7.78448522e-01
-3.74925226e-01 8.05299163e-01 1.69001445e-01 9.18165892e-02
-9.41140592e-01 8.47186565e-01 1.34409368e+00 4.13237046e-03
-1.64410204e-01 5.62149808e-02 -8.02921534e-01 7.90864348e-01
6.29384756e-01 -3.58473033e-01 -5.07292114e-02 -6.55782700e-01
-3.51119936e-02 -2.07886681e-01 3.51675093e-01 -8.24903429e-01
4.90283579e-01 5.58175072e-02 8.03182423e-01 -4.12509561e-01
6.09179139e-01 -8.40590358e-01 -3.76921922e-01 3.45533937e-01
-1.84932888e-01 8.49451497e-02 1.91419721e-01 5.15402555e-01
-1.56812578e-01 3.72919261e-01 9.06802893e-01 2.96382934e-01
-7.94988155e-01 1.07239270e+00 5.00648558e-01 -2.82973766e-01
1.23703837e+00 -3.92521732e-02 1.51554674e-01 -2.56568491e-01
-6.28968835e-01 4.27588314e-01 4.76860404e-01 5.77017009e-01
8.26306343e-01 -1.48946536e+00 -7.76334047e-01 6.61552250e-01
2.01500311e-01 2.65050709e-01 2.04276249e-01 9.40967262e-01
-4.48976517e-01 4.02076036e-01 -1.00439675e-01 -8.29094052e-01
-1.51396370e+00 3.57222944e-01 5.26566386e-01 2.34130144e-01
-6.40266478e-01 1.25443971e+00 4.20073152e-01 4.65257419e-03
4.15137380e-01 -2.49607533e-01 -3.11184853e-01 1.13915958e-01
8.93005073e-01 -1.04326241e-01 1.00862809e-01 -1.11554205e+00
-6.23409748e-01 1.46260405e+00 -2.92513877e-01 2.02787310e-01
1.20492625e+00 -1.52971402e-01 -1.08554177e-01 -2.33167306e-01
1.37385356e+00 2.64599472e-01 -1.50587261e+00 -4.08642948e-01
-1.09616876e-01 -7.49755383e-01 1.44585967e-01 -2.80439973e-01
-1.47462165e+00 8.77089083e-01 6.08398378e-01 -5.02000928e-01
1.11374581e+00 1.13493055e-01 7.37512171e-01 1.63540974e-01
5.59911489e-01 -6.26421094e-01 1.78120226e-01 2.31792763e-01
9.75441158e-01 -1.46257365e+00 -4.29948345e-02 -7.48169780e-01
-3.99390161e-01 1.17810225e+00 8.34397733e-01 -1.05049662e-01
7.39961863e-01 5.93169816e-02 1.21808656e-01 -2.40081057e-01
-5.34714162e-01 -5.11109114e-01 5.97899258e-01 5.19568145e-01
4.46134299e-01 -1.12185068e-01 6.91730976e-02 4.15975034e-01
-1.92089513e-01 -2.44392887e-01 -3.01108241e-01 3.02577615e-01
-4.25577432e-01 -9.27871644e-01 -2.34727338e-01 2.18212396e-01
-3.78923208e-01 -4.47406732e-02 -2.76161283e-01 6.77468777e-01
4.49895673e-02 6.82189286e-01 2.03650892e-01 -4.21678603e-01
4.04992074e-01 -2.38906533e-01 5.63753247e-01 -7.07862139e-01
-8.47881883e-02 3.87781888e-01 -3.85147423e-01 -9.55100477e-01
-1.31981567e-01 -6.90245867e-01 -1.25703561e+00 -1.61080852e-01
-2.33039960e-01 6.21665968e-03 4.76824433e-01 8.22743654e-01
7.68765807e-01 1.09156869e-01 8.62682402e-01 -1.16600740e+00
-3.08144957e-01 -9.34684932e-01 -1.72687858e-01 1.19719185e-01
5.13297856e-01 -9.45745289e-01 1.08957365e-01 -1.73004001e-01] | [13.49160385131836, 0.41838449239730835] |
0eb4724f-4d84-43b7-af81-2e375d9767bb | pixel-level-face-image-quality-assessment-for | 2110.11001 | null | https://arxiv.org/abs/2110.11001v3 | https://arxiv.org/pdf/2110.11001v3.pdf | Pixel-Level Face Image Quality Assessment for Explainable Face Recognition | An essential factor to achieve high performance in face recognition systems is the quality of its samples. Since these systems are involved in daily life there is a strong need of making face recognition processes understandable for humans. In this work, we introduce the concept of pixel-level face image quality that determines the utility of pixels in a face image for recognition. We propose a training-free approach to assess the pixel-level qualities of a face image given an arbitrary face recognition network. To achieve this, a model-specific quality value of the input image is estimated and used to build a sample-specific quality regression model. Based on this model, quality-based gradients are back-propagated and converted into pixel-level quality estimates. In the experiments, we qualitatively and quantitatively investigated the meaningfulness of our proposed pixel-level qualities based on real and artificial disturbances and by comparing the explanation maps on faces incompliant with the ICAO standards. In all scenarios, the results demonstrate that the proposed solution produces meaningful pixel-level qualities enhancing the interpretability of the complete face image quality. The code is publicly available | ['Arjan Kuijper', 'Kiran Raja', 'Florian Kirchbuchner', 'Naser Damer', 'Marco Huber', 'Philipp Terhörst'] | 2021-10-21 | null | null | null | null | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 4.36348349e-01 -6.32519424e-02 1.16659537e-01 -8.10295582e-01
-1.63162515e-01 -4.76636291e-02 3.50475729e-01 -3.68007302e-01
-1.26543999e-01 6.70460999e-01 -8.36543068e-02 -5.64820617e-02
-3.88583839e-01 -8.03988338e-01 -6.15932286e-01 -7.32047915e-01
1.10594463e-02 -2.09700331e-01 -4.43591446e-01 -5.99790551e-02
4.29222703e-01 9.86626267e-01 -1.93696809e+00 3.70353103e-01
9.72682953e-01 1.40794122e+00 8.95846356e-03 6.32777929e-01
-2.94571323e-03 5.21402597e-01 -7.48107553e-01 -2.65537083e-01
3.88331145e-01 -6.18200600e-01 -3.96790624e-01 5.59920490e-01
4.96631801e-01 -3.04926157e-01 4.33473922e-02 1.34862757e+00
3.18707526e-01 -1.47569522e-01 7.93332040e-01 -1.36926723e+00
-7.47254789e-01 1.29016966e-01 -3.56773436e-01 1.19602874e-01
3.27396959e-01 5.51738441e-01 5.61691880e-01 -9.16581511e-01
3.83935481e-01 1.62000847e+00 3.10967147e-01 5.20069957e-01
-1.25897050e+00 -7.62929797e-01 -1.22536749e-01 4.33325410e-01
-1.28119493e+00 -7.50036061e-01 1.09329867e+00 -2.97479242e-01
5.49483895e-01 2.54447162e-01 6.78507209e-01 6.23812020e-01
3.00815701e-01 1.36992618e-01 1.45776856e+00 -6.60467088e-01
3.16229850e-01 4.17392999e-01 5.90447057e-03 7.82408535e-01
2.91172504e-01 3.19927335e-01 -6.79469883e-01 2.53950089e-01
8.45885515e-01 -2.52921462e-01 -3.61627221e-01 -2.63611190e-02
-5.43403327e-01 4.14263546e-01 5.53504407e-01 4.34269130e-01
-5.75201333e-01 9.97670963e-02 -2.05214873e-01 3.93758297e-01
2.76116431e-01 1.00297831e-01 -4.38268073e-02 7.74787441e-02
-9.56213474e-01 -4.08367068e-01 6.65703773e-01 4.50231016e-01
8.39910865e-01 4.17523265e-01 -1.60509959e-01 7.52956152e-01
6.11561477e-01 8.89864087e-01 2.66356528e-01 -1.13838780e+00
7.12577179e-02 8.22887659e-01 1.07830033e-01 -1.46626520e+00
-1.44783527e-01 -5.08688390e-01 -8.38960171e-01 9.94695187e-01
3.96716505e-01 -3.14571969e-02 -9.68464136e-01 1.58866608e+00
2.52315640e-01 9.29076448e-02 2.02607721e-01 9.55098212e-01
4.73502636e-01 6.88471615e-01 8.29234719e-03 -5.69320977e-01
1.35207534e+00 -5.34998059e-01 -1.01858449e+00 -3.12526897e-02
-2.64474481e-01 -6.72715425e-01 1.20807159e+00 6.51145160e-01
-9.80859637e-01 -9.27103996e-01 -1.29305589e+00 4.98462260e-01
-1.07359640e-01 4.43428248e-01 2.12508634e-01 1.19566500e+00
-1.27185142e+00 7.68497229e-01 -5.77595413e-01 -3.42796773e-01
4.31533247e-01 3.31198364e-01 -4.45357114e-01 2.13639177e-02
-8.58075559e-01 9.87366676e-01 1.85436875e-01 6.12491369e-01
-9.40168440e-01 -4.63405877e-01 -4.63902056e-01 1.90150440e-01
-9.07321051e-02 -1.13966623e-02 6.65564120e-01 -1.57732522e+00
-1.57828009e+00 6.17393970e-01 -3.22200418e-01 -1.28027573e-01
3.89594018e-01 9.15263742e-02 -8.01919639e-01 3.77117217e-01
-3.66141319e-01 5.50729513e-01 1.21092236e+00 -1.62379408e+00
-4.08670604e-01 -5.02809644e-01 -7.93929622e-02 -2.68407376e-03
-5.19340575e-01 -9.51348711e-03 -2.84059495e-01 -2.38592014e-01
1.71127424e-01 -3.97976071e-01 2.50879645e-01 6.27551556e-01
7.20537640e-03 1.23257622e-01 7.81118155e-01 -8.49700570e-01
9.07904029e-01 -2.37715530e+00 -3.03202152e-01 5.41483104e-01
-1.39208362e-01 2.30135575e-01 -3.02196175e-01 -1.03512079e-01
-5.16561791e-02 5.42161353e-02 -2.71648556e-01 8.37300122e-02
-1.45455480e-01 2.27622837e-02 1.71520457e-01 5.25904357e-01
7.14789510e-01 4.09624815e-01 -5.26600361e-01 -4.30128723e-01
3.06150109e-01 1.01006377e+00 -4.66977149e-01 3.12189788e-01
-4.08838168e-02 4.08482194e-01 -3.17797005e-01 7.13240564e-01
1.00494385e+00 9.09569338e-02 1.23293586e-01 -5.24295390e-01
1.11039743e-01 -4.12519217e-01 -1.41027975e+00 1.01600778e+00
-6.17660999e-01 7.52280951e-01 2.66080499e-01 -7.30379760e-01
1.34596801e+00 4.03434277e-01 2.04655886e-01 -8.68620694e-01
2.21313655e-01 2.40786821e-01 8.10597464e-02 -5.40993512e-01
6.08344749e-02 -1.89044803e-01 7.59495497e-01 2.67372191e-01
8.32043663e-02 -1.67285070e-01 -1.10656600e-02 -4.09832627e-01
6.44705653e-01 9.47495773e-02 2.40637183e-01 -5.00042617e-01
1.05911303e+00 -4.95414764e-01 5.12323856e-01 3.28620493e-01
-4.14366364e-01 3.79614323e-01 4.76313382e-01 -3.34792614e-01
-9.11214769e-01 -9.25474584e-01 -3.66390944e-01 3.58040214e-01
1.75276875e-01 3.11201304e-01 -1.24262619e+00 -2.76395887e-01
-8.60735849e-02 5.74904025e-01 -6.38357043e-01 -2.57646680e-01
-2.03942522e-01 -5.60866833e-01 3.47447902e-01 1.05249777e-01
8.85652840e-01 -1.36974764e+00 -5.78210771e-01 -3.34524252e-02
1.15204185e-01 -8.83220434e-01 1.67878382e-02 -1.34145051e-01
-8.89498770e-01 -9.94333804e-01 -4.52635825e-01 -5.79451740e-01
1.12096560e+00 5.32710217e-02 8.71247709e-01 4.01982576e-01
-4.71770078e-01 3.19750786e-01 -8.52707475e-02 -1.22816920e-01
-5.12320101e-01 -6.15296364e-01 1.11505553e-01 6.72139287e-01
2.18007758e-01 -4.36217695e-01 -6.60626352e-01 4.13273335e-01
-8.60734761e-01 -3.26389372e-01 7.71216333e-01 7.06368327e-01
3.15749884e-01 4.23624992e-01 6.84608996e-01 -3.93626243e-01
8.09969306e-01 9.89812892e-03 -8.80369842e-01 4.50441837e-01
-9.55718100e-01 2.55515933e-01 6.31322801e-01 -2.37700552e-01
-1.42452550e+00 3.64433229e-02 -1.70593083e-01 -1.05584927e-01
-2.90956169e-01 1.56947404e-01 -6.30052745e-01 -4.38880682e-01
8.54111552e-01 8.87730420e-02 2.62247771e-01 -2.64799654e-01
2.15848237e-01 8.65453780e-01 5.34185290e-01 -4.33891743e-01
8.09745610e-01 3.28088254e-01 1.31047219e-01 -1.07081306e+00
-1.90073550e-01 6.56761825e-02 -4.69642937e-01 -9.40688908e-01
6.91282272e-01 -5.75266838e-01 -1.00860453e+00 6.32035613e-01
-1.03291178e+00 -1.86846107e-02 -1.45131350e-01 3.12218159e-01
-2.75698394e-01 2.94620365e-01 -2.46917248e-01 -1.30947340e+00
-3.05401027e-01 -1.30437887e+00 8.23603451e-01 6.00036740e-01
2.72855043e-01 -8.46653163e-01 -4.33124244e-01 3.43273908e-01
5.36182702e-01 1.70263246e-01 8.16266894e-01 6.81118667e-02
-5.99600554e-01 -2.31482424e-02 -3.91651124e-01 8.47263277e-01
4.39162821e-01 3.49884659e-01 -1.38752878e+00 -8.94580707e-02
3.84162217e-01 -9.37161148e-02 4.27441448e-01 4.31429744e-01
1.03860152e+00 -5.14380217e-01 2.42360324e-01 5.33234417e-01
1.68569458e+00 4.06498134e-01 9.24760938e-01 -1.36243384e-02
1.99871317e-01 8.89972508e-01 5.91018379e-01 4.26783979e-01
-4.38445896e-01 5.72875142e-01 5.16196251e-01 -2.06070051e-01
-4.43232715e-01 -2.25159954e-02 5.54266214e-01 3.34392428e-01
-2.04093128e-01 -2.10733548e-01 -5.20584762e-01 1.22258149e-01
-1.15224588e+00 -1.01939559e+00 2.04051957e-02 2.07412434e+00
6.81269050e-01 5.02204113e-02 -3.27309698e-01 5.79231083e-01
8.82159948e-01 -9.82874408e-02 -5.09965479e-01 -4.35509652e-01
-1.28454596e-01 1.92282438e-01 7.20583573e-02 7.41444588e-01
-5.76662242e-01 5.69621980e-01 6.75875807e+00 5.92666924e-01
-1.38125908e+00 -1.89224899e-01 1.02320313e+00 2.31521234e-01
-2.95041054e-01 -3.47860634e-01 -4.62637484e-01 4.52142358e-01
1.00148058e+00 -8.28873739e-02 6.62898898e-01 7.45639145e-01
7.00356364e-01 -3.57640654e-01 -9.67192769e-01 1.17754376e+00
2.06278563e-01 -8.74263287e-01 1.19463734e-01 3.90202664e-02
6.01768553e-01 -8.17229748e-01 3.61970037e-01 -4.28004950e-01
-3.01698983e-01 -1.27059376e+00 7.74630845e-01 8.34102631e-01
8.04021060e-01 -6.97077572e-01 7.25770712e-01 7.10938731e-03
-9.22186255e-01 -2.24701256e-01 -4.13810641e-01 -1.10287173e-02
-2.27515072e-01 8.35850894e-01 -6.68731391e-01 3.00727665e-01
4.90411818e-01 2.85628110e-01 -7.40733266e-01 8.12100530e-01
-3.47351283e-01 6.20362759e-01 -1.97938532e-01 -6.11386774e-03
-2.60841042e-01 -3.92657101e-01 2.28694633e-01 9.20224130e-01
4.97140437e-01 7.94312283e-02 -5.35652339e-01 1.25393176e+00
-1.55374452e-01 6.27536252e-02 -2.90159345e-01 -3.78722362e-02
4.37137932e-01 1.36450458e+00 -7.67324030e-01 -1.51340574e-01
-2.09274843e-01 9.15482581e-01 -1.37062177e-01 4.42872316e-01
-7.42339194e-01 -1.30086035e-01 5.75876176e-01 1.43729029e-02
-1.42653376e-01 6.76535293e-02 -6.03184819e-01 -7.18512237e-01
3.56021345e-01 -9.66727555e-01 -1.53118834e-01 -9.05280352e-01
-1.03484845e+00 8.98031771e-01 -2.48218283e-01 -1.07217741e+00
-1.65741388e-02 -9.02995884e-01 -5.47263980e-01 1.19025064e+00
-1.50944138e+00 -7.58516192e-01 -8.36856067e-01 4.94684964e-01
3.40430140e-01 -3.04420769e-01 5.51080823e-01 3.54945958e-01
-5.62561929e-01 6.14802718e-01 -1.34479105e-01 -6.86533079e-02
3.63628089e-01 -8.48194122e-01 7.90598616e-02 1.20511794e+00
-9.90735949e-04 4.48168933e-01 8.86384428e-01 -4.44200128e-01
-1.42643940e+00 -8.45636964e-01 4.91423666e-01 -5.64864315e-02
5.39675578e-02 -1.26606673e-02 -9.22637045e-01 3.12608178e-03
2.21073061e-01 4.39203642e-02 2.29386240e-01 -4.00875062e-01
-1.27029210e-01 -7.69732058e-01 -1.76870465e+00 3.40622783e-01
7.66393244e-01 -5.38535297e-01 -4.12089080e-01 -4.37138304e-02
8.59657899e-02 3.72546434e-01 -7.65471876e-01 3.00121993e-01
7.04913676e-01 -1.23206282e+00 7.66817093e-01 -2.10861359e-02
2.08982930e-01 -5.98373771e-01 -5.75980619e-02 -1.18417799e+00
-2.18478739e-01 -1.58493474e-01 2.93657154e-01 1.37025857e+00
5.40951610e-01 -6.17082953e-01 7.42400587e-01 7.34718144e-01
2.55895495e-01 -4.39744204e-01 -8.77043545e-01 -7.30920136e-01
-4.11413759e-01 -2.64546514e-01 7.09991634e-01 4.90769714e-01
-1.92056671e-01 -1.45289838e-01 -2.21841201e-01 5.20369172e-01
8.67577314e-01 -2.62573659e-01 4.55206484e-01 -1.16369188e+00
-7.27611929e-02 -3.70466173e-01 -7.46054649e-01 -5.06955683e-01
-1.82383973e-02 -4.23299938e-01 1.36511862e-01 -1.52140033e+00
1.24577314e-01 -2.60889322e-01 -3.15283328e-01 3.31863791e-01
-7.91540593e-02 4.58185017e-01 1.68820515e-01 1.00444570e-01
1.20052807e-01 5.19016802e-01 1.26502097e+00 -2.50451148e-01
-1.21020824e-01 -2.90341049e-01 -5.12279153e-01 4.71740574e-01
7.10836947e-01 -1.69282049e-01 -5.10018766e-01 -2.14071527e-01
-1.42845601e-01 2.87233666e-02 3.61372113e-01 -1.61268556e+00
-1.17386088e-01 -3.35412741e-01 7.75603294e-01 -1.24404710e-02
3.99703026e-01 -1.30033994e+00 4.38966990e-01 7.03167260e-01
-2.71216393e-01 -2.44448230e-01 5.77662103e-02 2.40093425e-01
-3.81190717e-01 -3.47859204e-01 1.25295055e+00 1.75623670e-01
-7.78612018e-01 -2.82791965e-02 -9.96558070e-02 -5.27937233e-01
1.06075037e+00 -5.54461718e-01 -1.79258958e-01 -4.60018665e-01
-5.16635656e-01 -3.55720073e-01 4.80820566e-01 2.33760640e-01
9.13806200e-01 -1.33788836e+00 -6.65842295e-01 7.59059072e-01
-5.85022494e-02 -7.95679867e-01 2.33173773e-01 4.54046071e-01
-8.66691232e-01 1.43093854e-01 -8.31216991e-01 -6.17657840e-01
-1.31995130e+00 4.14728254e-01 7.02202678e-01 4.35214490e-01
-6.83633611e-02 5.44612229e-01 -4.92591746e-02 -3.28350887e-02
2.95901835e-01 -4.03663874e-01 -3.20274949e-01 -2.79679865e-01
8.43177497e-01 3.17693949e-01 1.12336934e-01 -8.12854946e-01
-3.54164958e-01 7.68203676e-01 6.37541234e-01 -3.83619457e-01
1.10216379e+00 -9.56983119e-02 -1.61947548e-01 1.52350619e-01
1.02504981e+00 -1.44476295e-01 -1.53127933e+00 3.41193229e-01
-9.70261395e-02 -7.87273586e-01 1.64842099e-01 -1.12270367e+00
-1.50245249e+00 9.39956307e-01 1.30846643e+00 -2.49495322e-04
1.73756123e+00 -5.03697991e-01 -3.76690291e-02 2.35990971e-01
3.66133541e-01 -1.03390479e+00 1.25823990e-01 -1.53571427e-01
1.15522921e+00 -1.10039604e+00 -6.73061460e-02 -5.84157407e-01
-3.85054648e-01 1.30603468e+00 5.71457744e-01 1.62919983e-01
5.67163646e-01 2.38434702e-01 3.25787365e-01 1.97020099e-02
-4.53357369e-01 9.48641598e-02 4.09414619e-01 8.30406904e-01
3.50492388e-01 -4.37304936e-02 -2.35036135e-01 2.63621867e-01
1.25280842e-02 2.43433192e-01 3.52750808e-01 5.54192364e-01
-7.16613352e-01 -9.30182397e-01 -8.57482016e-01 2.92347968e-01
-1.43177152e-01 2.32473552e-01 -2.67223835e-01 4.16842788e-01
3.24298471e-01 1.44623268e+00 -7.38239214e-02 -3.89794350e-01
3.85012329e-01 9.93856117e-02 5.85798264e-01 -6.21534660e-02
-3.19339037e-01 -2.06287995e-01 -7.03016743e-02 -5.84568918e-01
-5.73764741e-01 -3.17334861e-01 -1.11438954e+00 -4.34874475e-01
-2.37836465e-01 1.36441618e-01 1.08193171e+00 7.05091238e-01
2.51176715e-01 5.54556191e-01 8.99289072e-01 -6.38076782e-01
-4.35032934e-01 -1.03145981e+00 -6.96695328e-01 6.09907806e-01
2.05181345e-01 -6.80423677e-01 -6.89122140e-01 4.21562791e-01] | [13.073864936828613, 0.7844659090042114] |
acca680b-db1b-481e-9bfa-34345587a27d | no-attention-is-needed-grouped-spatial | 2206.10810 | null | https://arxiv.org/abs/2206.10810v2 | https://arxiv.org/pdf/2206.10810v2.pdf | A Simple Baseline for Video Restoration with Grouped Spatial-temporal Shift | Video restoration, which aims to restore clear frames from degraded videos, has numerous important applications. The key to video restoration depends on utilizing inter-frame information. However, existing deep learning methods often rely on complicated network architectures, such as optical flow estimation, deformable convolution, and cross-frame self-attention layers, resulting in high computational costs. In this study, we propose a simple yet effective framework for video restoration. Our approach is based on grouped spatial-temporal shift, which is a lightweight and straightforward technique that can implicitly capture inter-frame correspondences for multi-frame aggregation. By introducing grouped spatial shift, we attain expansive effective receptive fields. Combined with basic 2D convolution, this simple framework can effectively aggregate inter-frame information. Extensive experiments demonstrate that our framework outperforms the previous state-of-the-art method, while using less than a quarter of its computational cost, on both video deblurring and video denoising tasks. These results indicate the potential for our approach to significantly reduce computational overhead while maintaining high-quality results. Code is avaliable at https://github.com/dasongli1/Shift-Net. | ['Xiaogang Wang', 'Simon See', 'Ka Chun Cheung', 'Hongsheng Li', 'Hongwei Qin', 'Yi Zhang', 'Xiaoyu Shi', 'Dasong Li'] | 2022-06-22 | a-simple-baseline-for-video-restoration-with | http://openaccess.thecvf.com//content/CVPR2023/html/Li_A_Simple_Baseline_for_Video_Restoration_With_Grouped_Spatial-Temporal_Shift_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_A_Simple_Baseline_for_Video_Restoration_With_Grouped_Spatial-Temporal_Shift_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-denoising', 'video-restoration'] | ['computer-vision', 'computer-vision'] | [ 4.11420502e-02 -5.37921906e-01 -7.61081427e-02 -1.95316344e-01
-7.07394719e-01 -3.21388304e-01 3.83728862e-01 -4.02222782e-01
-3.06354225e-01 7.49407291e-01 5.05963027e-01 -2.29727641e-01
5.73002920e-02 -5.53381145e-01 -8.64417255e-01 -7.06242263e-01
7.75692686e-02 -5.53729773e-01 1.91272721e-01 -1.73103526e-01
2.62335777e-01 5.23894966e-01 -1.45937705e+00 3.67452979e-01
1.00769281e+00 9.72609937e-01 4.12243843e-01 7.06603467e-01
1.61815852e-01 1.06187618e+00 -3.34379077e-01 -2.37679765e-01
2.91939348e-01 -3.18871170e-01 -6.48381114e-01 1.89086109e-01
8.02231073e-01 -1.12546825e+00 -9.74308074e-01 1.08824241e+00
5.16443968e-01 3.31575871e-01 1.04246333e-01 -8.66306961e-01
-9.71317530e-01 1.95526257e-01 -4.11135137e-01 6.62881792e-01
2.85099596e-01 3.54676247e-01 7.17354119e-01 -9.13721621e-01
3.45962197e-01 1.24673653e+00 6.77145183e-01 5.87907910e-01
-1.12297809e+00 -6.77111745e-01 1.72411457e-01 6.38887942e-01
-1.11577404e+00 -9.71509933e-01 6.66201890e-01 -2.95859456e-01
9.04934824e-01 1.86963454e-01 5.62298894e-01 8.00391197e-01
2.15775847e-01 7.95983613e-01 7.51238108e-01 -1.34244263e-01
-9.90878344e-02 -6.95537269e-01 -3.30545902e-02 5.51440716e-01
1.56639054e-01 1.15714446e-01 -5.61746657e-01 1.99497700e-01
1.22022259e+00 3.87092620e-01 -9.72291172e-01 -8.06454644e-02
-1.07885337e+00 4.46013808e-01 5.45902073e-01 2.02307358e-01
-3.66873533e-01 4.58416313e-01 4.35089231e-01 1.90772906e-01
7.69228697e-01 7.90267531e-03 -2.95256525e-01 -2.18634501e-01
-1.00233507e+00 1.51742443e-01 3.78407508e-01 5.89249671e-01
8.09937358e-01 3.26792032e-01 -1.80108383e-01 8.10754776e-01
1.88369945e-01 2.76635319e-01 3.55331033e-01 -1.57149816e+00
3.50687951e-01 1.66018873e-01 2.70850003e-01 -1.12809181e+00
-8.10599923e-02 -2.07545802e-01 -1.14497626e+00 2.76755840e-01
2.09706649e-01 -2.68510450e-02 -9.94532287e-01 1.66905117e+00
1.62008956e-01 7.26487041e-01 -1.25395790e-01 1.19797230e+00
9.07409906e-01 6.89352155e-01 -1.05735838e-01 -2.81285197e-01
9.66323912e-01 -1.18159866e+00 -8.89879763e-01 -6.99659437e-02
1.96012735e-01 -8.54779303e-01 8.45470309e-01 1.45861462e-01
-1.45674837e+00 -6.09173000e-01 -8.69496584e-01 -5.25438428e-01
2.76086599e-01 1.42196370e-02 6.56839550e-01 2.93168336e-01
-1.52236795e+00 8.82269442e-01 -1.08257806e+00 -1.48453012e-01
7.43932545e-01 3.56842548e-01 -4.65816259e-01 -5.45877337e-01
-9.26695347e-01 5.32620072e-01 7.46666715e-02 3.89479488e-01
-9.58730340e-01 -8.12310517e-01 -1.03323328e+00 2.47190639e-01
2.86464423e-01 -9.54587638e-01 1.08179688e+00 -9.78083551e-01
-1.59374583e+00 3.16796690e-01 -4.85840976e-01 -3.43536586e-01
4.12716031e-01 -5.27457833e-01 -2.37999588e-01 4.87249523e-01
-2.73147039e-02 6.29724741e-01 9.24759746e-01 -1.17752743e+00
-4.56801057e-01 -1.10295303e-01 3.28485757e-01 3.05201739e-01
-4.55683023e-01 9.81981456e-02 -7.16287076e-01 -1.02450597e+00
8.91885310e-02 -6.64550245e-01 -1.12303659e-01 3.63801062e-01
-1.58354610e-01 6.14254400e-02 9.15262878e-01 -1.09145665e+00
1.24399829e+00 -2.16746807e+00 1.90773472e-01 -4.24009740e-01
5.01508594e-01 5.86026430e-01 -1.77962765e-01 1.16008483e-01
-6.16671257e-02 1.02786414e-01 -4.00279850e-01 -4.57467139e-01
-2.89570004e-01 9.95032266e-02 -3.88082147e-01 6.93578601e-01
1.06748708e-01 9.17185426e-01 -8.18251431e-01 -1.21186830e-01
6.27261519e-01 1.04183769e+00 -7.35291898e-01 2.71219462e-01
1.11764237e-01 6.26962781e-01 -1.58027411e-01 5.62391818e-01
9.13434625e-01 -3.43706369e-01 1.46723464e-02 -6.05723560e-01
-8.91540870e-02 2.65467703e-01 -9.13244843e-01 1.80778742e+00
-3.61746699e-01 1.01407492e+00 4.19050068e-01 -1.10485852e+00
4.20196086e-01 2.90986419e-01 6.27516210e-01 -6.33908391e-01
3.78033891e-02 1.23913929e-01 -2.86471099e-01 -5.67278862e-01
5.68143427e-01 1.76961482e-01 5.73259830e-01 2.96677291e-01
-5.58540151e-02 3.47628266e-01 3.42141032e-01 1.23966239e-01
1.14152229e+00 3.56930435e-01 -2.22094338e-02 -1.37092352e-01
6.11192346e-01 -5.41949093e-01 8.12919736e-01 4.62756157e-01
-4.02817279e-01 8.77610505e-01 1.31193742e-01 -7.65380859e-01
-1.02632141e+00 -9.23303604e-01 7.20540285e-02 7.49709785e-01
5.40786624e-01 -4.25840795e-01 -7.69039154e-01 -2.12768316e-01
-2.56860435e-01 1.72238916e-01 -3.41499925e-01 -8.15401897e-02
-9.23873842e-01 -6.78621233e-01 3.02934647e-01 4.52375829e-01
7.93341339e-01 -8.72378528e-01 -3.18067044e-01 2.93607861e-01
-6.29356086e-01 -1.18590736e+00 -8.85197937e-01 -5.13318896e-01
-1.10469472e+00 -1.04149187e+00 -9.88446295e-01 -8.53333175e-01
6.12749994e-01 1.01647091e+00 1.12167203e+00 4.28082734e-01
-1.72298893e-01 2.80856222e-01 -2.36709729e-01 5.24817109e-01
-7.02743009e-02 -3.12612474e-01 -4.57031429e-02 1.03237942e-01
7.48740807e-02 -9.06524658e-01 -1.14116168e+00 2.31326550e-01
-1.19426477e+00 1.90158129e-01 4.41827744e-01 9.41679239e-01
4.08823788e-01 -1.08050026e-01 3.74900311e-01 -2.68844724e-01
5.56131184e-01 -3.09915811e-01 -3.83368582e-01 6.87075034e-02
-2.89278239e-01 -1.58418566e-01 5.85272133e-01 -3.01061809e-01
-1.11444867e+00 -2.40189329e-01 -2.97279079e-02 -7.97915459e-01
-1.51490971e-01 3.21172774e-01 -1.13449708e-01 -3.62186313e-01
3.38418156e-01 3.68184686e-01 1.32468134e-01 -6.37780249e-01
2.80442297e-01 4.93007123e-01 7.13907778e-01 -3.65271181e-01
6.01513565e-01 7.07640171e-01 -2.42854163e-01 -7.41375327e-01
-6.75042450e-01 -3.18009406e-01 -4.80705947e-01 -2.79871464e-01
7.46420622e-01 -1.29889750e+00 -8.16950202e-01 9.01257932e-01
-1.29496753e+00 -4.69453126e-01 9.70474705e-02 5.26711643e-01
-4.41319913e-01 9.00762677e-01 -1.15376115e+00 -3.10850024e-01
-4.85379457e-01 -1.32341933e+00 9.16624367e-01 2.70808578e-01
3.22742969e-01 -9.67179835e-01 -2.49986723e-01 6.81931674e-01
7.19792008e-01 5.65316202e-03 4.40321416e-01 2.45553643e-01
-1.16438031e+00 1.47459790e-01 -5.52594781e-01 6.26476586e-01
3.37182879e-01 -2.36184984e-01 -7.26066768e-01 -6.26391649e-01
1.95480082e-02 -6.59865215e-02 1.27965558e+00 7.50842392e-01
1.52262902e+00 -4.07418519e-01 -2.57778987e-02 1.10991025e+00
1.28803396e+00 1.01675941e-02 1.01548779e+00 4.15738642e-01
1.05635226e+00 1.38079062e-01 2.25795314e-01 3.80064160e-01
6.56138539e-01 6.32774711e-01 5.48222542e-01 -3.12812120e-01
-5.78690827e-01 2.63311028e-01 5.36855638e-01 8.65332961e-01
-3.85981143e-01 -2.97560245e-01 -6.20439589e-01 5.94545364e-01
-1.97790086e+00 -1.21253896e+00 1.00441962e-01 2.02936149e+00
9.26504016e-01 -3.05087268e-01 -2.60445893e-01 -3.37383933e-02
9.37418580e-01 4.27135676e-01 -5.58261156e-01 4.53692116e-02
-2.54766911e-01 2.78755520e-02 3.34534734e-01 8.22086871e-01
-1.23575616e+00 8.42428446e-01 5.98114634e+00 7.28891969e-01
-1.09935844e+00 9.96758193e-02 7.85609126e-01 -2.23503053e-01
-2.80535817e-01 -1.19684063e-01 -4.54727590e-01 6.44758463e-01
6.30811214e-01 -1.01264022e-01 7.94336319e-01 2.19499052e-01
5.90630591e-01 -1.26681691e-02 -7.36833274e-01 1.19638598e+00
3.06993186e-01 -1.70978022e+00 1.28145427e-01 9.44606364e-02
8.06860030e-01 7.78418332e-02 -8.98037553e-02 -1.73108056e-01
9.50633362e-02 -8.68370295e-01 4.68772858e-01 6.37393773e-01
7.55148172e-01 -6.39259756e-01 5.74248970e-01 -5.80917969e-02
-1.16926372e+00 -5.48039526e-02 -3.63142550e-01 -2.11684644e-01
4.75220084e-01 7.57335544e-01 1.00941680e-01 5.23050547e-01
1.11775458e+00 1.38846850e+00 -1.32488310e-01 1.31418538e+00
-1.83428884e-01 4.80276644e-01 -3.54096740e-02 6.76900387e-01
-4.66533974e-02 -2.72851586e-01 5.95911026e-01 1.07727671e+00
5.53589344e-01 2.48793423e-01 1.80971816e-01 5.92955947e-01
-3.50018889e-01 -2.78092563e-01 -2.97431737e-01 2.79426724e-01
3.25496644e-01 1.09961414e+00 -3.26128900e-01 -4.60526317e-01
-6.39990985e-01 1.22062027e+00 2.23636672e-01 6.64440393e-01
-8.88174951e-01 -3.49191606e-01 1.16463852e+00 1.06309578e-01
5.23406923e-01 -5.05379319e-01 -1.45403489e-01 -1.64138591e+00
3.06981087e-01 -7.97199011e-01 5.03138714e-02 -9.21676934e-01
-1.11385143e+00 6.48989499e-01 -4.39156532e-01 -1.40810990e+00
4.85649444e-02 -4.10217643e-01 -4.63303268e-01 8.64753187e-01
-1.97905087e+00 -1.08181369e+00 -7.60841250e-01 9.08917844e-01
7.02484906e-01 8.13089684e-02 4.38690305e-01 7.80959249e-01
-7.64618337e-01 3.76188010e-01 2.41924167e-01 1.63251474e-01
9.19581831e-01 -7.55534470e-01 4.56922889e-01 1.40667963e+00
-2.42236018e-01 5.41280210e-01 4.69488651e-01 -5.70863783e-01
-1.31952357e+00 -1.13865542e+00 6.25728965e-01 1.53556531e-02
5.50858498e-01 6.26145303e-02 -1.17564964e+00 6.24585986e-01
4.55784261e-01 5.02897441e-01 5.10575533e-01 -4.29030329e-01
-3.21132183e-01 -3.45075101e-01 -9.16323900e-01 6.97272480e-01
1.13915396e+00 -5.05629361e-01 -2.25054190e-01 2.66733855e-01
7.57440805e-01 -5.08854687e-01 -8.72686028e-01 4.08499241e-01
3.93388212e-01 -1.23194325e+00 1.27037990e+00 -2.59939641e-01
8.04152250e-01 -5.83934247e-01 -1.72752872e-01 -1.18900168e+00
-6.19104266e-01 -8.96802843e-01 -5.77384830e-01 1.08777535e+00
-1.74311429e-01 -6.23074412e-01 5.57672381e-01 7.06686199e-01
-4.25708503e-01 -6.08668029e-01 -8.89818013e-01 -6.11204505e-01
-1.45729050e-01 -2.14912742e-01 3.26708913e-01 9.41045761e-01
-3.92535210e-01 -5.01752011e-02 -7.61132836e-01 2.54475266e-01
8.74605715e-01 -7.58254975e-02 5.49329042e-01 -7.99568474e-01
-2.25940585e-01 -4.11373198e-01 -4.52760190e-01 -1.55574048e+00
2.39682004e-01 -5.83010197e-01 -3.55467200e-02 -1.77978861e+00
2.24063918e-01 -7.88445175e-02 -3.84640515e-01 4.57848549e-01
-4.34580237e-01 6.27994239e-01 2.99024314e-01 5.05809486e-01
-5.90913296e-01 7.83538878e-01 1.51548052e+00 -6.38193712e-02
-1.19796067e-01 -3.59230787e-01 -7.48599827e-01 6.25521779e-01
8.61238897e-01 -7.70595297e-02 -2.50371903e-01 -1.11741996e+00
-3.53903532e-01 1.08504012e-01 6.98679209e-01 -9.20259118e-01
3.88689399e-01 -1.06233440e-01 4.66361612e-01 -2.73306936e-01
5.56874394e-01 -4.98489410e-01 1.86753441e-02 3.23326230e-01
2.67490167e-02 9.32227820e-02 2.93884128e-01 5.88020802e-01
-5.08258045e-01 2.27234468e-01 7.92239308e-01 -1.73918635e-01
-9.32481289e-01 5.73246598e-01 -3.68269563e-01 -1.46244526e-01
8.19971859e-01 -2.71653950e-01 -6.07133329e-01 -5.82715392e-01
-5.60055614e-01 -1.49961538e-03 6.15279973e-01 4.10080880e-01
7.91402459e-01 -1.32495725e+00 -8.17222297e-01 2.12957010e-01
-5.56462824e-01 8.55195429e-03 7.63463557e-01 9.42240119e-01
-8.14698815e-01 1.94969535e-01 -3.77245128e-01 -6.24380291e-01
-1.27432656e+00 3.68818969e-01 4.88080919e-01 1.16112873e-01
-1.02182543e+00 7.26645887e-01 4.00097966e-01 1.30329609e-01
1.19722649e-01 -2.20121637e-01 -7.37894550e-02 -2.90330589e-01
8.60783339e-01 6.37471557e-01 -9.20960009e-02 -6.71660483e-01
-2.19041437e-01 6.55880392e-01 -1.45012751e-01 2.48243287e-01
1.62758505e+00 -6.53817713e-01 -4.41846550e-01 -2.46474490e-01
1.19775164e+00 -1.91368893e-01 -1.97419441e+00 -2.61626750e-01
-5.53456783e-01 -1.03262854e+00 5.11252820e-01 -4.33470100e-01
-1.66577542e+00 8.26246142e-01 6.01363838e-01 -4.95380312e-02
1.61064029e+00 -2.97708392e-01 1.11681736e+00 1.09877966e-01
1.74693003e-01 -6.17738128e-01 1.26546785e-01 3.69721055e-01
9.43126976e-01 -1.33691359e+00 1.25665203e-01 -5.37325621e-01
-1.60801545e-01 1.17905819e+00 6.26028657e-01 -3.24603021e-01
5.86694956e-01 1.52101919e-01 3.69010828e-02 2.55329341e-01
-6.38313770e-01 -3.62248309e-02 2.09746987e-01 4.54541802e-01
5.57611942e-01 -2.61326075e-01 -1.69552565e-01 1.69974208e-01
3.78634453e-01 2.73149490e-01 8.02506864e-01 8.06817472e-01
-4.03605998e-01 -9.94783819e-01 -2.74547726e-01 2.26149246e-01
-6.84195399e-01 -4.74655390e-01 3.21531802e-01 3.78542244e-01
-1.76485494e-01 1.09052122e+00 1.54467747e-01 -2.58532524e-01
5.20087630e-02 -5.40520191e-01 4.58366483e-01 -9.12707001e-02
-1.63756311e-01 2.41786778e-01 -1.51934698e-01 -9.74322140e-01
-8.78917933e-01 -5.06676257e-01 -9.06879663e-01 -7.89729416e-01
2.95513845e-03 -3.03729445e-01 2.21481502e-01 8.65276039e-01
8.66976202e-01 6.38509154e-01 5.12054145e-01 -1.41564071e+00
-1.76653311e-01 -7.90364683e-01 -2.39110067e-01 4.25334781e-01
8.88842225e-01 -5.33779144e-01 -4.10943836e-01 4.49948907e-01] | [11.137659072875977, -1.9736356735229492] |
74ed9c8c-f630-4fd9-b87c-7d3c0daa22d4 | english-to-hindi-multi-modal-neural-machine | null | null | https://aclanthology.org/D19-5205 | https://aclanthology.org/D19-5205.pdf | English to Hindi Multi-modal Neural Machine Translation and Hindi Image Captioning | With the widespread use of Machine Trans-lation (MT) techniques, attempt to minimizecommunication gap among people from di-verse linguistic backgrounds. We have par-ticipated in Workshop on Asian Transla-tion 2019 (WAT2019) multi-modal translationtask. There are three types of submissiontrack namely, multi-modal translation, Hindi-only image captioning and text-only transla-tion for English to Hindi translation. The mainchallenge is to provide a precise MT output.The multi-modal concept incorporates textualand visual features in the translation task. Inthis work, multi-modal translation track re-lies on pre-trained convolutional neural net-works (CNN) with Visual Geometry Grouphaving 19 layered (VGG19) to extract imagefeatures and attention-based Neural MachineTranslation (NMT) system for translation.The merge-model of recurrent neural network(RNN) and CNN is used for the Hindi-onlyimage captioning. The text-only translationtrack is based on the transformer model of theNMT system. The official results evaluated atWAT2019 translation task, which shows thatour multi-modal NMT system achieved Bilin-gual Evaluation Understudy (BLEU) score20.37, Rank-based Intuitive Bilingual Eval-uation Score (RIBES) 0.642838, Adequacy-Fluency Metrics (AMFM) score 0.668260 forchallenge test data and BLEU score 40.55,RIBES 0.760080, AMFM score 0.770860 forevaluation test data in English to Hindi multi-modal translation respectively. | ['Sivaji yopadhyay', 'B', 'Rohit Pratap Singh', 'Sahinur Rahman Laskar', 'Partha Pakray'] | 2019-11-01 | null | null | null | ws-2019-11 | ['hindi-image-captioning'] | ['computer-vision'] | [-1.94359809e-01 -7.47337043e-02 -2.30163485e-01 -2.04158977e-01
-1.45953596e+00 -7.86329985e-01 9.67122436e-01 -4.06322658e-01
-5.86079121e-01 1.11936712e+00 5.29941022e-01 -8.60030353e-01
5.51518619e-01 -5.99776745e-01 -9.57534850e-01 -3.52009982e-01
6.67054117e-01 8.90883029e-01 -4.82381582e-01 -5.83433032e-01
-1.77388683e-01 2.55845249e-01 -6.17203832e-01 7.92226553e-01
1.10274649e+00 6.56865954e-01 1.19962081e-01 9.12489295e-01
8.17728713e-02 4.61041749e-01 -5.05782485e-01 -9.50797021e-01
1.80612549e-01 -8.34679008e-01 -1.06197941e+00 -9.72640961e-02
5.63591003e-01 -5.24450094e-02 -3.86502624e-01 9.29591537e-01
7.27722585e-01 -2.34517250e-02 8.99770439e-01 -8.99699032e-01
-1.22215915e+00 5.72798789e-01 -7.55795896e-01 3.83305341e-01
3.39361489e-01 4.98009920e-02 6.58515334e-01 -1.44319582e+00
8.60446095e-01 1.27250159e+00 6.29504740e-01 4.98313367e-01
-1.12393951e+00 -6.83014691e-01 -3.97070587e-01 3.05186331e-01
-1.06010079e+00 -3.17160606e-01 2.06730038e-01 -3.44246686e-01
1.35488558e+00 4.24234182e-01 3.75164628e-01 1.25105822e+00
8.00780535e-01 8.19579899e-01 1.64100850e+00 -5.19535005e-01
-6.45178080e-01 4.09113556e-01 -2.56952673e-01 7.00117052e-01
-3.16481829e-01 2.49403328e-01 -4.99317080e-01 6.56628489e-01
6.67410135e-01 -2.85792708e-01 7.16330856e-02 4.87376451e-01
-1.81156194e+00 6.38184607e-01 5.82078576e-01 5.53357244e-01
-4.43390518e-01 -7.03155845e-02 6.50346279e-01 1.11529732e+00
3.24158132e-01 2.75077611e-01 -2.97005504e-01 -1.43389612e-01
-9.20266271e-01 1.75749391e-01 3.52952778e-01 1.06751883e+00
6.69472754e-01 2.63778150e-01 -7.01873243e-01 1.06934929e+00
1.83313876e-01 1.20995641e+00 8.79673421e-01 -4.61320758e-01
1.14524770e+00 3.10351938e-01 -2.56496668e-02 -7.52602875e-01
-2.52091587e-01 -4.24533814e-01 -1.19683826e+00 3.24824937e-02
2.37999335e-01 -4.41114843e-01 -1.18713391e+00 1.48556864e+00
-1.30425051e-01 -6.39216006e-01 3.11221153e-01 9.97146308e-01
1.14629030e+00 1.04200876e+00 8.42593610e-02 -9.11321938e-02
1.41508901e+00 -1.45673692e+00 -5.94486356e-01 7.37189725e-02
8.92341256e-01 -1.63162911e+00 1.24943781e+00 1.01492025e-01
-1.27473009e+00 -9.20170426e-01 -1.02591980e+00 -2.66877681e-01
-5.09601176e-01 5.04327416e-01 1.59730464e-02 4.21870351e-01
-1.49808896e+00 2.12158516e-01 -4.86626625e-01 -8.92913222e-01
6.32004216e-02 4.77238536e-01 -7.70649254e-01 -1.18441753e-01
-1.19417679e+00 1.55896783e+00 2.57654369e-01 2.39018306e-01
-1.05511296e+00 -2.53769666e-01 -6.20008409e-01 -5.25024056e-01
-3.85139495e-01 -6.92090571e-01 1.13901663e+00 -1.54024518e+00
-1.51248991e+00 1.36618280e+00 -2.82587716e-03 -4.37082857e-01
8.81833315e-01 -3.10590072e-03 -6.89023614e-01 1.54772680e-03
4.19466674e-01 1.01458609e+00 3.91526550e-01 -9.60823536e-01
-5.68312228e-01 -2.98968524e-01 -1.92219555e-01 8.44318032e-01
-4.93190484e-04 5.11841059e-01 -2.77226776e-01 -8.93493116e-01
-7.02394471e-02 -1.11102593e+00 2.80687958e-01 -4.17667836e-01
-7.43533731e-01 -3.63112651e-02 8.07190359e-01 -1.29656410e+00
7.62391388e-01 -1.66083586e+00 3.17911595e-01 -2.76620299e-01
-1.27691060e-01 3.61387938e-01 -5.06522954e-01 5.49343348e-01
-3.29387575e-01 4.73601818e-02 9.26127136e-02 -2.69206315e-01
4.77317069e-03 2.43785642e-02 -1.35427475e-01 5.56926847e-01
9.62784514e-02 1.53196239e+00 -5.79762757e-01 -6.46962285e-01
3.16182375e-01 3.31458539e-01 -2.94100549e-02 4.56612349e-01
1.61951184e-01 5.69819808e-01 -7.02098086e-02 6.94726884e-01
5.46131313e-01 2.26940185e-01 -2.73141026e-01 -1.95039168e-01
-2.24047020e-01 5.77808060e-02 -4.30267066e-01 1.79658270e+00
-1.05675197e+00 9.07459974e-01 -1.84785545e-01 -8.66368890e-01
7.83942759e-01 6.87108397e-01 8.14627334e-02 -1.12002957e+00
4.23336238e-01 6.26930356e-01 3.01357079e-02 -5.72617233e-01
4.41832304e-01 -3.09668362e-01 -2.39213973e-01 5.19682229e-01
2.58562028e-01 1.13947153e-01 9.10457075e-02 -1.20658047e-01
3.80618840e-01 4.54896152e-01 -6.18500821e-02 -4.78001714e-01
5.66896915e-01 3.32712978e-01 -1.00186899e-01 5.69792449e-01
-2.10107982e-01 8.57132256e-01 -6.91918284e-02 -5.69969833e-01
-1.35199678e+00 -1.11465526e+00 1.39830813e-01 9.63476181e-01
-2.21046999e-01 8.11990872e-02 -8.73603761e-01 -7.74689853e-01
-8.25304747e-01 6.57616794e-01 -6.31124675e-01 -5.66854775e-02
-8.42788398e-01 -8.32049191e-01 6.56563163e-01 4.91522372e-01
1.00398803e+00 -1.21140599e+00 -1.94077939e-01 1.15647592e-01
-8.75678062e-01 -9.85878348e-01 -8.67743134e-01 -6.55222759e-02
-5.87291539e-01 -2.69688100e-01 -1.46242666e+00 -1.31624281e+00
5.02684057e-01 -1.13683334e-02 1.10928464e+00 -4.02894646e-01
9.18277726e-03 -1.23030066e-01 -2.03563005e-01 -1.42581537e-01
-5.25561869e-01 3.30779076e-01 -1.73227057e-01 -2.23510280e-01
6.65727437e-01 -2.88924485e-01 -3.66836101e-01 3.10487717e-01
-1.86295629e-01 6.04654014e-01 9.00758505e-01 1.15918517e+00
5.85474312e-01 -7.96553314e-01 2.03898862e-01 -3.39158565e-01
6.85066879e-01 -3.10828656e-01 -1.61865339e-01 6.10040426e-01
-3.85184705e-01 -2.73933589e-01 8.31184924e-01 -4.80155259e-01
-8.26854885e-01 -3.37530494e-01 -2.23234415e-01 -2.29515448e-01
-1.38696834e-01 3.25909346e-01 -1.19266324e-01 2.01851755e-01
8.12504590e-01 5.53886175e-01 -4.61588651e-01 -1.08882673e-01
5.68912446e-01 1.09545791e+00 6.78411782e-01 -4.48843956e-01
7.49107778e-01 -1.48737282e-01 -2.99182117e-01 -4.79862452e-01
-4.29257214e-01 8.10541213e-02 -7.19574809e-01 -2.16770619e-01
1.31453550e+00 -1.04414725e+00 -4.05695200e-01 4.79963750e-01
-1.36958182e+00 -2.77384907e-01 1.83322802e-01 8.22880566e-01
-7.22107530e-01 -1.03948899e-01 -8.70478809e-01 -3.12390327e-01
-9.32054698e-01 -1.32430053e+00 8.90508652e-01 2.47496068e-02
-2.10415751e-01 -8.66533816e-01 1.14138626e-01 7.56507218e-01
3.70545626e-01 2.50453979e-01 1.01287961e+00 -6.51627839e-01
-3.18344742e-01 -1.58179119e-01 -6.85990751e-01 1.31199032e-01
-1.09430976e-01 -5.40839791e-01 -8.37854564e-01 -2.74266005e-01
-2.78531611e-01 -5.07784724e-01 5.74650168e-01 3.39438260e-01
3.39612722e-01 -3.40939879e-01 4.84958887e-02 6.40194178e-01
1.46566129e+00 2.07481891e-01 7.74890006e-01 4.15537268e-01
8.85817051e-01 4.34639692e-01 4.22173917e-01 -3.91401827e-01
6.62792981e-01 9.51078415e-01 -1.37039542e-01 -7.09987640e-01
-7.31501818e-01 -3.88901860e-01 7.71367252e-01 1.19381726e+00
-3.52306217e-01 -4.90833282e-01 -9.41947103e-01 6.21948183e-01
-1.69651878e+00 -7.90577650e-01 -3.97027194e-01 2.06887674e+00
8.92944336e-01 -1.57280743e-01 3.60876471e-01 -3.53970319e-01
8.54870617e-01 -3.04731190e-01 -8.42150599e-02 -1.12765813e+00
-3.49748522e-01 4.24843207e-02 5.26329398e-01 8.32198024e-01
-9.11168516e-01 1.37634277e+00 5.09704065e+00 7.49324322e-01
-1.38916647e+00 6.92598701e-01 1.00131547e+00 2.17675120e-01
-6.70670494e-02 -3.86822492e-01 -6.73357964e-01 3.19616526e-01
1.42040932e+00 -1.10607386e-01 7.25954950e-01 1.91270903e-01
2.77366221e-01 2.46159092e-01 -7.74594903e-01 1.17801058e+00
4.69013929e-01 -1.08045638e+00 3.45183372e-01 -3.54941301e-02
8.83786976e-01 5.14608920e-01 4.41391945e-01 5.70074916e-01
3.40794563e-01 -1.37007535e+00 7.74413645e-01 2.89165914e-01
1.41226649e+00 -1.04933012e+00 1.05721653e+00 5.06138988e-02
-9.80916858e-01 3.71054798e-01 -4.02593136e-01 2.46753022e-01
5.80289438e-02 -1.95246842e-02 -9.54971254e-01 7.64175296e-01
5.61862886e-01 5.38007379e-01 -3.24688643e-01 5.44930518e-01
-2.31237933e-01 3.72507066e-01 7.63612986e-02 -9.04344246e-02
7.51764059e-01 -2.51512706e-01 3.29144239e-01 1.56015217e+00
5.97271204e-01 -3.65798593e-01 2.53174037e-01 5.50303698e-01
-3.78671050e-01 6.20087206e-01 -6.10140681e-01 -1.60756886e-01
-7.19772279e-02 1.18550408e+00 -6.93759680e-01 -4.78820980e-01
-4.77262914e-01 1.60204673e+00 1.56256765e-01 6.05556190e-01
-1.06417322e+00 -7.03840792e-01 5.05472608e-02 -9.92378891e-02
-5.72405010e-02 -1.56673953e-01 -5.13892114e-01 -1.43479550e+00
1.57737185e-03 -1.20316982e+00 1.60004869e-01 -9.77758110e-01
-9.77705896e-01 1.32524312e+00 -1.48113325e-01 -1.11182976e+00
-3.09819341e-01 -5.83225906e-01 -4.32937831e-01 1.56195760e+00
-1.32910013e+00 -2.05711412e+00 1.68490082e-01 7.04562664e-01
1.06135762e+00 -5.09367406e-01 9.52390671e-01 4.67388153e-01
-4.63205576e-01 1.06730223e+00 1.69751912e-01 3.78043503e-01
9.81883287e-01 -1.15180969e+00 5.23784816e-01 9.46465671e-01
1.80347368e-01 5.03008068e-01 6.19459331e-01 -7.09506452e-01
-1.29810143e+00 -1.28052747e+00 1.79593289e+00 -5.08880675e-01
5.07424295e-01 -4.70392793e-01 -2.63035506e-01 7.11285353e-01
9.71137583e-01 -4.12960231e-01 3.10758382e-01 -1.66242719e-01
-1.54975295e-01 3.46987061e-02 -1.06961763e+00 8.07032526e-01
6.68148100e-01 -6.10212982e-01 -5.43501198e-01 7.92862177e-01
5.19596398e-01 -5.74425220e-01 -9.27652240e-01 5.49529120e-02
5.51306009e-01 -5.09455919e-01 8.62889886e-01 -6.74879491e-01
8.28193843e-01 -1.71816900e-01 -3.29523563e-01 -1.37136579e+00
-2.67937124e-01 -7.68486619e-01 3.48374188e-01 8.93287003e-01
1.00086761e+00 -3.32470179e-01 4.03282702e-01 -7.64360204e-02
-3.61174107e-01 -6.62803292e-01 -9.27249610e-01 -6.45515382e-01
7.71393180e-01 5.78645058e-02 3.72713983e-01 9.58795846e-01
-1.91407874e-01 9.11374927e-01 -8.41417551e-01 -1.81538031e-01
1.17552876e-01 5.80075942e-02 5.76939464e-01 -4.08471853e-01
-1.66089460e-01 -3.43237489e-01 -1.54875234e-01 -7.71401525e-01
7.55157229e-03 -1.26868117e+00 -7.90714473e-02 -1.84471059e+00
3.83628637e-01 2.74523020e-01 -1.90255880e-01 6.21104956e-01
7.16846660e-02 8.98017824e-01 1.82154611e-01 1.53239906e-01
-2.03523204e-01 3.74723315e-01 1.74544394e+00 -3.62509370e-01
9.96591970e-02 -2.15409398e-01 -3.04899812e-01 2.92074550e-02
9.04677093e-01 -3.08094978e-01 -4.14324611e-01 -1.09306037e+00
1.28606737e-01 2.44792685e-01 1.10053040e-01 -6.14640892e-01
-6.07053339e-02 -9.66165662e-02 9.82365310e-01 -8.06285739e-01
2.15133771e-01 -4.61581618e-01 1.63925136e-03 6.72841132e-01
-3.12476397e-01 1.01388919e+00 1.26113385e-01 -1.69785529e-01
-2.68473059e-01 2.26822734e-01 9.15486097e-01 -4.92986679e-01
-2.80556023e-01 2.99957752e-01 -4.43062812e-01 7.83873051e-02
8.11083555e-01 -2.88882196e-01 -1.21086381e-01 -4.64817464e-01
-9.36668277e-01 -1.77396033e-02 2.66791850e-01 6.78274632e-01
5.04376531e-01 -1.73828173e+00 -1.46849155e+00 9.25073773e-02
1.30362317e-01 -9.73901927e-01 6.67442903e-02 1.20596266e+00
-8.95306408e-01 7.70146549e-01 -6.88773513e-01 -3.90057564e-01
-1.15632629e+00 2.39132836e-01 5.38462818e-01 -3.39717627e-01
-4.79766846e-01 9.73155975e-01 9.52550545e-02 -9.31193173e-01
-1.39573753e-01 -4.93301265e-02 -1.10848725e-01 -1.88626900e-01
4.73375648e-01 2.77683914e-01 3.27124327e-01 -1.35586524e+00
-2.59973258e-01 6.66831136e-01 -3.05612206e-01 -6.81798339e-01
1.00534546e+00 -2.85865247e-01 -1.53405994e-01 3.48116070e-01
1.40128028e+00 -4.57409501e-01 -3.83476406e-01 3.66872624e-02
-4.48040426e-01 -2.28014916e-01 -1.23064302e-01 -1.55054128e+00
-8.12771201e-01 1.35400343e+00 1.05683494e+00 -6.21040821e-01
8.67176533e-01 -3.35407227e-01 1.07695198e+00 2.99402595e-01
1.69362232e-01 -1.22968936e+00 -7.36259297e-02 7.28266895e-01
1.22174442e+00 -1.41685557e+00 -5.19414365e-01 2.01408282e-01
-8.38066101e-01 1.15907621e+00 6.96766555e-01 -4.64913771e-02
1.39816627e-01 -1.23362340e-01 5.91847420e-01 1.54447347e-01
-4.85423774e-01 1.34278566e-01 5.74828029e-01 8.69242013e-01
9.14050639e-01 4.37619865e-01 -5.63222110e-01 2.38617346e-01
-7.13274717e-01 -3.16524148e-01 1.60710454e-01 4.83816922e-01
-1.37608901e-01 -1.01941741e+00 -4.22916651e-01 8.81509632e-02
-6.37267113e-01 -3.87274712e-01 -5.32127321e-01 7.48883843e-01
1.99754432e-01 1.14440739e+00 -2.65959114e-01 -5.82098246e-01
2.44080558e-01 -2.08296403e-01 6.78103387e-01 -4.15977180e-01
-1.13431680e+00 2.77125627e-01 2.95257688e-01 -2.97788471e-01
-1.04837447e-01 -3.43692124e-01 -8.56761754e-01 -5.87099373e-01
-1.73924983e-01 1.67430058e-01 8.60602379e-01 1.10881639e+00
9.28408802e-02 2.82877594e-01 4.29366350e-01 -3.79709870e-01
-1.34578362e-01 -1.25797832e+00 7.62138963e-02 2.48768061e-01
2.23522335e-01 8.96031260e-02 3.79136622e-01 1.44928873e-01] | [11.4982271194458, 1.5505329370498657] |
612d6254-a3d1-49b0-a7b1-e6db9066389e | viewset-diffusion-0-image-conditioned-3d | 2306.07881 | null | https://arxiv.org/abs/2306.07881v1 | https://arxiv.org/pdf/2306.07881v1.pdf | Viewset Diffusion: (0-)Image-Conditioned 3D Generative Models from 2D Data | We present Viewset Diffusion: a framework for training image-conditioned 3D generative models from 2D data. Image-conditioned 3D generative models allow us to address the inherent ambiguity in single-view 3D reconstruction. Given one image of an object, there is often more than one possible 3D volume that matches the input image, because a single image never captures all sides of an object. Deterministic models are inherently limited to producing one possible reconstruction and therefore make mistakes in ambiguous settings. Modelling distributions of 3D shapes is challenging because 3D ground truth data is often not available. We propose to solve the issue of data availability by training a diffusion model which jointly denoises a multi-view image set.We constrain the output of Viewset Diffusion models to a single 3D volume per image set, guaranteeing consistent geometry. Training is done through reconstruction losses on renderings, allowing training with only three images per object. Our design of architecture and training scheme allows our model to perform 3D generation and generative, ambiguity-aware single-view reconstruction in a feed-forward manner. Project page: szymanowiczs.github.io/viewset-diffusion. | ['Andrea Vedaldi', 'Christian Rupprecht', 'Stanislaw Szymanowicz'] | 2023-06-13 | null | null | null | null | ['single-view-3d-reconstruction', '3d-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 2.61173189e-01 4.04325664e-01 2.10218892e-01 -3.17049831e-01
-8.84242594e-01 -8.02532911e-01 7.62733161e-01 -6.85189605e-01
4.70436215e-02 3.22363764e-01 1.85416371e-01 -1.92482859e-01
1.97455108e-01 -9.36128020e-01 -9.64835405e-01 -6.26961768e-01
3.64082307e-01 1.11382568e+00 9.70951319e-02 1.23630494e-01
-3.90077829e-02 7.01702237e-01 -1.56193185e+00 5.01062632e-01
4.44978088e-01 8.37336123e-01 6.58206820e-01 1.12827528e+00
-2.06350088e-01 3.64406526e-01 -4.22133863e-01 -2.23916039e-01
6.21711016e-01 -4.93467808e-01 -4.39027995e-01 7.69680858e-01
6.85783803e-01 -7.14993238e-01 3.01660877e-03 6.71622753e-01
4.31899548e-01 -1.49759084e-01 8.70390952e-01 -1.15323246e+00
-7.41519570e-01 8.41399580e-02 -5.82218349e-01 -1.04150862e-01
3.09782058e-01 1.31377965e-01 7.37285495e-01 -1.22587240e+00
9.95135307e-01 1.21754026e+00 3.34738910e-01 7.57062554e-01
-1.48078060e+00 -1.20924726e-01 2.56409168e-01 -5.80038130e-01
-1.24724913e+00 -7.02558517e-01 8.01785111e-01 -6.60757840e-01
8.03906798e-01 4.59041119e-01 8.92876863e-01 1.10156250e+00
3.40821385e-01 6.06476188e-01 1.16772294e+00 -2.51121014e-01
2.03954190e-01 6.49212599e-02 -6.96422100e-01 6.32901490e-01
-4.96927984e-02 1.81440651e-01 -6.10071778e-01 -1.03462577e-01
1.61929250e+00 1.48459613e-01 -2.27794304e-01 -8.56888115e-01
-1.23031008e+00 6.75458074e-01 2.21267372e-01 -1.29507646e-01
-4.42829698e-01 2.36235961e-01 -2.45226115e-01 2.76560843e-01
7.09247053e-01 2.29848281e-01 -2.16590017e-01 1.79952636e-01
-8.65712881e-01 5.33561170e-01 5.65757394e-01 1.16078234e+00
7.66581595e-01 1.97342664e-01 2.97858715e-01 5.94185770e-01
5.24509370e-01 8.13637435e-01 3.61769162e-02 -1.40147030e+00
2.99093425e-01 3.38329345e-01 1.57286629e-01 -6.57855213e-01
-7.80081302e-02 -4.47710574e-01 -8.50717068e-01 8.55242252e-01
2.68264621e-01 2.69019827e-02 -1.21069503e+00 1.56060600e+00
6.78923547e-01 -2.81625152e-01 -1.84706122e-01 1.09527946e+00
6.61193669e-01 5.95742941e-01 -4.95073587e-01 -1.09103493e-01
9.78982270e-01 -5.60223162e-01 -2.43187025e-01 -3.63582641e-01
7.29402229e-02 -1.12403011e+00 1.03867090e+00 4.51816291e-01
-1.58125556e+00 -2.74858564e-01 -8.04261923e-01 -1.79967657e-01
6.19753264e-02 -1.62220851e-01 3.69442642e-01 4.66510892e-01
-1.19462800e+00 2.78191656e-01 -8.79887998e-01 2.23545525e-02
4.42083746e-01 1.66055992e-01 -3.91010731e-01 -2.00706348e-01
-3.89804900e-01 8.08646977e-01 -8.47006366e-02 -8.49461928e-02
-1.16050577e+00 -7.92177856e-01 -6.92934453e-01 -2.08957404e-01
2.96136647e-01 -1.34747112e+00 1.23545539e+00 -8.52569640e-01
-1.32758093e+00 1.26565194e+00 -1.55692637e-01 -3.64613831e-02
8.50561082e-01 5.75745143e-02 1.24207512e-01 1.11096375e-01
1.65867478e-01 7.50635087e-01 1.17116463e+00 -1.98807514e+00
-1.51502430e-01 -5.49153984e-01 -9.62996390e-03 5.60532212e-01
4.57120806e-01 -3.92071098e-01 -4.94392723e-01 -4.95729774e-01
6.92305624e-01 -7.25853503e-01 -3.86176020e-01 4.01051909e-01
-4.67691988e-01 4.29733157e-01 6.18287444e-01 -5.32905400e-01
5.20252645e-01 -1.93552148e+00 3.46517414e-01 8.74370039e-02
4.84153330e-01 -3.10369343e-01 -4.93068015e-03 3.06874096e-01
7.24465847e-02 1.54648766e-01 -2.84536421e-01 -1.01534832e+00
-3.30621526e-02 5.42320132e-01 -4.87760514e-01 4.05087888e-01
2.91350223e-02 7.88460970e-01 -5.96621752e-01 -4.49711531e-01
4.75504667e-01 7.02510893e-01 -8.43710184e-01 4.32457268e-01
-5.66588223e-01 8.14938307e-01 -3.28269809e-01 5.94263971e-01
9.28311229e-01 -4.93114680e-01 1.15947209e-01 -1.28709316e-01
-8.20157528e-02 9.96082276e-02 -1.31524992e+00 1.99791789e+00
-7.86592484e-01 1.21603444e-01 2.55322188e-01 -4.41538692e-01
7.02954113e-01 2.05111265e-01 5.13245344e-01 -4.60554063e-01
-1.21912912e-01 1.51616484e-01 -3.91041100e-01 -9.98658389e-02
3.77809733e-01 -5.44442892e-01 1.12854339e-01 8.17834675e-01
1.02391906e-01 -9.02998924e-01 -3.41795564e-01 4.01047796e-01
6.18039668e-01 5.09309232e-01 -7.01313913e-02 -3.90946828e-02
-1.55917883e-01 -2.31613800e-01 2.48737186e-01 6.55777276e-01
5.11419892e-01 1.51697457e+00 2.46929511e-01 -5.04177570e-01
-1.53347492e+00 -1.73052585e+00 -1.07270397e-01 4.78939503e-01
1.32710516e-01 -2.21558198e-01 -5.07279217e-01 -4.89557296e-01
-1.38099015e-01 6.83974981e-01 -6.04237497e-01 3.48110616e-01
-5.04419446e-01 -4.53302115e-01 -1.52210474e-01 2.03459889e-01
5.06249852e-02 -6.18427336e-01 -8.70878458e-01 1.29421398e-01
-1.17456421e-01 -1.00592256e+00 -6.29565775e-01 1.20399818e-01
-1.11180556e+00 -9.90241468e-01 -8.07859540e-01 -4.75820810e-01
1.09856665e+00 3.98538649e-01 1.55149496e+00 2.15934273e-02
-2.26820469e-01 6.71801329e-01 1.41990528e-01 -1.96777150e-01
-8.45416665e-01 -3.48737299e-01 -3.02310158e-02 -1.20847806e-01
-4.58400816e-01 -1.04644012e+00 -8.47468793e-01 3.76211822e-01
-9.69751298e-01 4.91709590e-01 4.73343045e-01 7.44360566e-01
1.18754268e+00 -2.99429744e-01 9.89547893e-02 -8.20299149e-01
1.40808418e-01 -2.07457960e-01 -7.76044190e-01 3.59555520e-02
-4.79880154e-01 1.98961105e-02 3.97589654e-01 -3.70010197e-01
-1.11510921e+00 1.77170515e-01 -9.44844484e-02 -9.82866585e-01
-1.72724724e-01 4.77001667e-02 -2.93419003e-01 1.30878121e-01
6.82396114e-01 3.59264910e-01 3.24478984e-01 -5.80193043e-01
6.56365931e-01 1.18711226e-01 4.35556918e-01 -5.30709028e-01
6.26499951e-01 8.10094118e-01 4.61329408e-02 -8.11521232e-01
-6.03685319e-01 -1.06169090e-01 -6.74588084e-01 -4.20707732e-01
8.29618037e-01 -9.48492587e-01 -3.22594374e-01 2.93163240e-01
-1.23666799e+00 -3.84420782e-01 -7.03942001e-01 2.31157646e-01
-1.04993355e+00 1.03158668e-01 -2.08973199e-01 -7.84167111e-01
-1.93332121e-01 -1.24392486e+00 1.50046384e+00 -2.75243819e-01
-2.39862829e-01 -9.34916973e-01 -5.62764853e-02 3.00560743e-01
2.59177327e-01 3.18008363e-01 7.67034769e-01 4.38140705e-02
-1.21343458e+00 -2.24924297e-03 1.30193323e-01 3.27873766e-01
5.15245944e-02 7.71854892e-02 -1.08050895e+00 -1.16806835e-01
4.24108118e-01 -2.55815893e-01 5.15290797e-01 7.61565089e-01
9.45666730e-01 -2.97418833e-01 -2.95643419e-01 7.34463215e-01
1.53692162e+00 -9.31188241e-02 4.83816922e-01 -6.31218255e-02
8.08235526e-01 4.43798214e-01 3.32702659e-02 4.84539300e-01
4.59927142e-01 6.58356071e-01 7.47127354e-01 -1.09508656e-01
-4.53742296e-01 -6.10646307e-01 -7.59623721e-02 5.85253954e-01
-1.66099742e-01 -5.19946218e-01 -6.83617949e-01 4.78806406e-01
-1.40385842e+00 -8.99796844e-01 -1.62557378e-01 2.47019839e+00
4.25440878e-01 1.57220408e-01 1.86010431e-02 -3.08233410e-01
2.63342977e-01 1.89892799e-01 -5.78817606e-01 -3.97786461e-02
-1.94290906e-01 -5.94239533e-02 3.50864917e-01 9.03695166e-01
-6.64555728e-01 5.93845010e-01 6.50318480e+00 4.65252191e-01
-1.04681039e+00 1.54588982e-01 8.31820726e-01 -5.36059439e-01
-1.20775461e+00 1.05155110e-01 -6.21676028e-01 2.80069709e-01
3.63201737e-01 2.43581664e-02 4.96643543e-01 6.36345983e-01
2.39424586e-01 -2.84116596e-01 -1.10694540e+00 1.26707160e+00
1.45879686e-01 -1.66087961e+00 3.07837307e-01 5.04955351e-01
9.39602256e-01 1.51475668e-01 2.62964100e-01 -4.27167386e-01
5.62774301e-01 -1.02509534e+00 1.20819092e+00 7.19402075e-01
9.94783580e-01 -5.81504226e-01 -2.76034214e-02 7.76521206e-01
-7.08231628e-01 3.52024168e-01 -2.66222388e-01 2.46352956e-01
6.75473750e-01 7.66538680e-01 -8.01164269e-01 4.79145527e-01
5.81676900e-01 3.28143209e-01 -1.31325433e-02 5.66894770e-01
-1.19965576e-01 5.83303943e-02 -6.17512047e-01 5.41402102e-01
-7.15102106e-02 -3.47877473e-01 7.64913201e-01 5.21463275e-01
6.17362380e-01 2.70139247e-01 1.02187522e-01 1.38956761e+00
6.32837266e-02 -4.02469546e-01 -1.02745438e+00 3.46982479e-01
2.91997403e-01 9.85130489e-01 -7.97352910e-01 -1.48647040e-01
-2.22838461e-01 1.21477675e+00 1.68074086e-01 2.87720621e-01
-5.00921488e-01 7.16216147e-01 5.49532592e-01 6.86606646e-01
4.18387383e-01 -5.22825301e-01 -5.93777061e-01 -1.30176163e+00
1.24114618e-01 -5.96682787e-01 1.02389557e-02 -1.19279933e+00
-1.29331946e+00 7.32374847e-01 1.63941041e-01 -1.43692958e+00
-6.34627223e-01 -3.83458197e-01 -2.94057220e-01 1.15933967e+00
-1.02934206e+00 -1.31973517e+00 -1.20240748e-01 3.92797917e-01
7.05485642e-01 1.74546063e-01 9.48730111e-01 2.47016922e-02
2.42923886e-01 1.02345899e-01 -8.52096602e-02 -5.23115516e-01
4.85780597e-01 -1.28646815e+00 7.16200650e-01 6.19328737e-01
3.55021834e-01 4.93009359e-01 7.18237638e-01 -6.51410520e-01
-1.57761538e+00 -7.04052389e-01 6.32699847e-01 -1.00599062e+00
7.52994698e-03 -6.15342081e-01 -6.61758363e-01 7.65694201e-01
1.07227445e-01 1.05255067e-01 4.60971892e-01 -3.22653353e-01
-1.06301695e-01 2.66676158e-01 -1.18357444e+00 5.94814777e-01
1.37989414e+00 -5.90025723e-01 -1.24923415e-01 2.91811347e-01
5.64959407e-01 -9.64359820e-01 -8.50507021e-01 5.00749797e-02
5.90157151e-01 -1.34785271e+00 1.25594127e+00 -1.62656948e-01
6.24500751e-01 -3.80162060e-01 -3.89794648e-01 -1.43045938e+00
-3.25728767e-02 -6.25254214e-01 -1.89791948e-01 7.22113311e-01
4.33682472e-01 -5.32425523e-01 9.39925969e-01 7.01740623e-01
-2.77226716e-01 -6.88090980e-01 -9.39637780e-01 -6.35435402e-01
1.68707252e-01 -5.91983795e-01 6.11811042e-01 6.45124733e-01
-7.82655239e-01 3.41804206e-01 -3.66769463e-01 3.34427506e-01
9.44007933e-01 5.42230070e-01 1.07626653e+00 -9.39462066e-01
-6.45809650e-01 -2.82934129e-01 -9.06677395e-02 -1.48308802e+00
-3.23704898e-01 -7.42260814e-01 -6.46889284e-02 -1.79563057e+00
4.27649952e-02 -6.53389096e-01 7.38425970e-01 3.28807645e-02
3.15762699e-01 3.47793013e-01 1.80456385e-01 4.14286941e-01
-1.41533902e-02 4.75713968e-01 1.84195960e+00 2.06805751e-01
-1.66944042e-02 7.55710229e-02 -5.80066681e-01 7.40881920e-01
4.09991920e-01 -2.47379720e-01 -8.51948559e-01 -8.91304374e-01
2.70710319e-01 4.34411645e-01 7.47563183e-01 -5.56228459e-01
1.39782801e-02 -2.66917706e-01 8.99028301e-01 -9.82503533e-01
8.64457846e-01 -9.40223038e-01 8.55444551e-01 -2.25570966e-02
2.73822229e-02 1.31763637e-01 -6.50868118e-02 5.40401518e-01
2.07950413e-01 -3.11175082e-02 7.38628209e-01 -8.17863166e-01
-3.36347312e-01 5.50343096e-01 -1.19430766e-01 -1.87735818e-02
7.44975269e-01 -5.85827529e-01 1.27635032e-01 -5.71286976e-01
-1.15942931e+00 -8.15035626e-02 1.14720249e+00 2.60740906e-01
9.61423874e-01 -1.33927095e+00 -7.05646694e-01 7.11908281e-01
-1.05680004e-01 7.62138963e-01 5.56991816e-01 2.82053858e-01
-5.62832236e-01 -1.14305494e-02 -3.82738896e-02 -9.67108369e-01
-1.10892272e+00 4.01632696e-01 5.66558182e-01 -4.14068997e-02
-9.58238542e-01 8.73655558e-01 7.50895321e-01 -6.58086956e-01
-3.00877750e-01 -2.88030773e-01 5.21512806e-01 -2.91549921e-01
3.91988009e-01 -5.15790395e-02 7.17127174e-02 -6.00318491e-01
-4.82383370e-02 6.45870328e-01 9.89907533e-02 -7.46927142e-01
1.42612195e+00 -3.38296115e-01 1.74261227e-01 3.83930176e-01
1.09408867e+00 7.43721128e-02 -1.78114569e+00 -6.74617058e-03
-9.66138661e-01 -8.32985044e-01 1.68305665e-01 -8.02584350e-01
-1.00276327e+00 8.77690613e-01 3.72595996e-01 1.49624661e-01
8.20427358e-01 2.68469185e-01 5.10203183e-01 -2.44539291e-01
6.10047936e-01 -5.72636425e-01 2.02443451e-01 2.52099544e-01
1.27991188e+00 -1.04250109e+00 1.40217870e-01 -4.34769392e-01
-6.18365467e-01 9.72347319e-01 5.87138653e-01 -1.49881318e-01
6.15128636e-01 5.45955598e-01 1.40480027e-01 -5.34213305e-01
-7.83802807e-01 1.80998206e-01 3.52653861e-01 7.55045712e-01
1.14146315e-01 1.43226804e-02 3.25722277e-01 7.34668747e-02
-3.12781751e-01 -2.40186110e-01 4.99168754e-01 9.81986344e-01
-2.23867044e-01 -1.14783907e+00 -5.51217616e-01 3.64315093e-01
-1.12329409e-01 6.81545679e-03 -2.42322311e-01 6.39957070e-01
-9.07147024e-03 3.86301816e-01 3.10623556e-01 -6.60902634e-02
3.54023486e-01 -1.39625281e-01 1.02838051e+00 -7.98133790e-01
3.63216549e-02 6.50060475e-01 -1.02519713e-01 -4.72710997e-01
-3.07448059e-01 -6.80425882e-01 -1.00256813e+00 -3.46200794e-01
-1.16243593e-01 -3.06555331e-01 7.04567492e-01 6.14467680e-01
5.15965879e-01 3.12013328e-01 7.70366549e-01 -1.23482454e+00
-4.26665604e-01 -3.89758736e-01 -6.16511345e-01 9.53440294e-02
4.86199617e-01 -6.22090280e-01 -3.69727999e-01 4.09414709e-01] | [9.072075843811035, -3.141407012939453] |
4fb79f97-bfd3-4864-af4f-b51e96f6d120 | seggpt-segmenting-everything-in-context | 2304.03284 | null | https://arxiv.org/abs/2304.03284v1 | https://arxiv.org/pdf/2304.03284v1.pdf | SegGPT: Segmenting Everything In Context | We present SegGPT, a generalist model for segmenting everything in context. We unify various segmentation tasks into a generalist in-context learning framework that accommodates different kinds of segmentation data by transforming them into the same format of images. The training of SegGPT is formulated as an in-context coloring problem with random color mapping for each data sample. The objective is to accomplish diverse tasks according to the context, rather than relying on specific colors. After training, SegGPT can perform arbitrary segmentation tasks in images or videos via in-context inference, such as object instance, stuff, part, contour, and text. SegGPT is evaluated on a broad range of tasks, including few-shot semantic segmentation, video object segmentation, semantic segmentation, and panoptic segmentation. Our results show strong capabilities in segmenting in-domain and out-of-domain targets, either qualitatively or quantitatively. | ['Tiejun Huang', 'Chunhua Shen', 'Wen Wang', 'Yue Cao', 'Xiaosong Zhang', 'Xinlong Wang'] | 2023-04-06 | null | null | null | null | ['panoptic-segmentation', 'personalized-segmentation', 'video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 7.10462511e-01 9.47797392e-03 -4.27678764e-01 -6.01505578e-01
-7.88292944e-01 -9.00734603e-01 4.20849204e-01 -3.67864631e-02
-2.19974563e-01 3.57815981e-01 -1.91596538e-01 -3.50279272e-01
1.52185097e-01 -7.45196700e-01 -7.13805795e-01 -6.69270933e-01
3.06218237e-01 6.41956151e-01 4.98965025e-01 3.77612934e-02
4.91192758e-01 3.64519805e-01 -1.27483582e+00 6.70818865e-01
7.44499266e-01 8.06006074e-01 4.69978005e-01 7.63048410e-01
-7.29969084e-01 7.43203104e-01 -7.81955004e-01 -2.90330470e-01
3.13435137e-01 -6.68934166e-01 -1.16294789e+00 7.48809814e-01
6.61709011e-01 9.85286087e-02 2.28393346e-01 1.21639538e+00
-1.18378505e-01 2.22761869e-01 7.75000870e-01 -1.27625549e+00
-7.54676163e-01 7.77604282e-01 -7.05213547e-01 1.30023673e-01
4.00632501e-01 1.80275455e-01 1.20301521e+00 -8.05765808e-01
6.99191988e-01 1.41048396e+00 3.57138157e-01 5.90790749e-01
-1.56398046e+00 -3.81919324e-01 7.03376830e-01 -1.86066777e-01
-1.00281096e+00 1.18619792e-01 7.52424479e-01 -5.90188563e-01
4.47666854e-01 4.40986544e-01 9.33642805e-01 8.67707312e-01
-1.82807036e-02 1.26128244e+00 1.31795537e+00 -4.09219593e-01
5.00922441e-01 -2.28836805e-01 1.59810930e-01 5.67016244e-01
1.29985005e-01 -1.94180310e-01 -3.26452434e-01 2.46993184e-01
8.97286713e-01 3.45496275e-02 7.94983953e-02 -2.42724836e-01
-1.26540446e+00 8.25504124e-01 3.66460949e-01 1.37010053e-01
1.54275283e-01 2.35527158e-01 2.38749459e-01 1.00459732e-01
6.94566607e-01 5.72708547e-01 -7.78122842e-01 4.37775761e-01
-1.14902961e+00 8.18747431e-02 7.49933779e-01 1.25487638e+00
9.86973941e-01 6.00193031e-02 -4.14226115e-01 7.38697648e-01
8.25291723e-02 5.25172472e-01 1.71730518e-01 -1.29249692e+00
3.60276669e-01 4.28996176e-01 -1.58248946e-01 -5.32075226e-01
-3.39939892e-01 1.08981378e-01 -3.41271996e-01 3.20234895e-01
5.79909682e-01 -3.78306657e-01 -1.58046734e+00 1.66293931e+00
4.42804039e-01 4.71997321e-01 -6.87192660e-03 9.02211368e-01
9.76131618e-01 9.94075179e-01 5.81230998e-01 -7.84736276e-02
1.39198828e+00 -1.27617884e+00 -4.12412345e-01 -5.68935394e-01
4.37488139e-01 -8.20320785e-01 1.23405313e+00 5.82324028e-01
-1.05068982e+00 -7.89516568e-01 -6.96660459e-01 -3.46297801e-01
-7.28599906e-01 -1.23359725e-01 6.72969878e-01 6.09987140e-01
-1.17905188e+00 4.52233851e-01 -4.62231666e-01 -4.42293733e-01
5.45617640e-01 1.12366676e-01 1.14929804e-03 -2.82134652e-01
-7.18863845e-01 4.15606469e-01 9.85234559e-01 -1.77898690e-01
-9.67677295e-01 -5.63253582e-01 -9.24905658e-01 -1.54192716e-01
7.15764880e-01 -6.63843036e-01 1.20831442e+00 -1.58945715e+00
-1.24931920e+00 1.33918154e+00 1.90327436e-01 -1.62006944e-01
2.58111864e-01 -2.12937802e-01 -3.31634969e-01 2.74699479e-01
2.94287026e-01 1.41474867e+00 1.10813510e+00 -1.54438663e+00
-1.10828304e+00 -1.52480036e-01 3.91867518e-01 5.46719432e-01
5.24275124e-01 5.18346950e-02 -1.04735279e+00 -9.90565479e-01
3.05081457e-01 -9.36422706e-01 -2.86751807e-01 -9.45171043e-02
-8.57650518e-01 -1.33779943e-01 1.34998906e+00 -5.77571690e-01
7.79897392e-01 -2.11482096e+00 2.58036822e-01 -6.30739704e-02
6.23721369e-02 -9.91326645e-02 -3.03489417e-01 -6.60498301e-03
-9.50983539e-02 3.51620287e-01 -6.00502074e-01 -1.29807428e-01
-5.58452830e-02 6.41496003e-01 -1.12350360e-01 6.75599650e-02
3.23195904e-01 1.06097221e+00 -1.04535127e+00 -9.14486766e-01
3.53297830e-01 -9.08020288e-02 -4.37139660e-01 1.84759408e-01
-1.01606357e+00 6.36979640e-01 -5.51817179e-01 9.35247779e-01
5.16774476e-01 -3.01656008e-01 2.13943914e-01 4.62032184e-02
3.52755538e-03 -1.68448269e-01 -1.07647228e+00 2.01722121e+00
-1.14897214e-01 7.44430363e-01 1.09433383e-01 -1.24426520e+00
5.08310676e-01 -9.96068418e-02 4.63992864e-01 -6.54380381e-01
-5.47804609e-02 -1.01422228e-01 -1.84800059e-01 -6.21758580e-01
6.58008933e-01 -4.25216369e-02 -5.34346700e-01 5.66094995e-01
1.67371735e-01 -7.23131597e-01 6.07556045e-01 3.80483717e-01
2.81677425e-01 6.42971098e-01 1.59786507e-01 -5.01036465e-01
1.59309104e-01 4.79427695e-01 4.01184112e-01 7.98451185e-01
-7.60691687e-02 8.36009860e-01 7.08142042e-01 -9.76625681e-02
-8.51536512e-01 -1.14511931e+00 3.94455940e-02 1.84791422e+00
5.86974740e-01 3.49418074e-02 -1.14525664e+00 -9.25790131e-01
-1.72633216e-01 6.78966582e-01 -7.09604561e-01 4.05795544e-01
-5.29336929e-01 -7.79799938e-01 1.47425413e-01 5.95108628e-01
5.48606753e-01 -1.43299794e+00 -4.94250238e-01 -1.04956195e-01
-1.03894919e-01 -1.29667020e+00 -7.44995475e-01 3.77214402e-01
-1.08544099e+00 -1.32151163e+00 -5.70882380e-01 -1.26257622e+00
6.82900548e-01 3.72495681e-01 1.46015155e+00 9.53259021e-02
-3.38233709e-01 7.24147320e-01 -4.87871200e-01 -1.91955626e-01
-2.50143796e-01 -1.20139591e-01 -7.93479204e-01 -1.00435562e-01
2.32690051e-01 1.74310114e-02 -5.32054305e-01 2.76426733e-01
-1.16730702e+00 4.48659837e-01 5.04150927e-01 5.34988403e-01
1.11202705e+00 -5.56832813e-02 2.51207590e-01 -1.53695726e+00
2.21005782e-01 -4.44854558e-01 -5.37932456e-01 6.35272026e-01
-2.16432720e-01 -2.48386756e-01 3.14055979e-01 -4.17472810e-01
-1.05033314e+00 1.76797196e-01 2.87425607e-01 -1.77358344e-01
-5.38273633e-01 4.17513132e-01 -2.94309318e-01 1.79069459e-01
3.33712965e-01 5.92617318e-02 -4.75477695e-01 -2.96099931e-01
9.11315560e-01 1.57550275e-01 6.69518888e-01 -9.99823809e-01
3.56395096e-01 2.21908689e-01 -1.26974538e-01 -8.89587760e-01
-1.20497787e+00 -5.16364813e-01 -8.83631527e-01 -1.85491875e-01
1.53862906e+00 -8.55098128e-01 -3.04928850e-02 4.68582988e-01
-7.92553306e-01 -1.01468086e+00 -5.86557508e-01 -1.04318351e-01
-5.90589225e-01 3.72318059e-01 -6.70540512e-01 -4.28405255e-01
1.76687874e-02 -1.20938861e+00 1.17515039e+00 5.52293420e-01
9.66159999e-02 -1.18938112e+00 -4.41417247e-01 3.93350691e-01
-2.69414008e-01 4.43506300e-01 1.05176508e+00 -6.98448598e-01
-8.11361909e-01 2.73731977e-01 -5.03530443e-01 2.25786105e-01
5.80511875e-02 2.82851547e-01 -8.84495080e-01 -3.82070802e-02
-2.07601696e-01 -3.80889237e-01 9.28854525e-01 7.07430601e-01
1.62614179e+00 -5.81129342e-02 -3.99146289e-01 7.64764607e-01
1.54989541e+00 5.58437288e-01 5.53467393e-01 -1.06238686e-01
9.59757566e-01 7.47402906e-01 9.22390521e-01 -7.73135498e-02
2.58324258e-02 3.38173270e-01 2.88539141e-01 -4.62303072e-01
-4.32208359e-01 -1.33161740e-02 1.38293058e-01 2.86495507e-01
1.73244730e-01 -5.55673897e-01 -8.34788144e-01 5.75049102e-01
-1.79265273e+00 -8.22598100e-01 -1.90412357e-01 1.75351250e+00
8.12289178e-01 1.89413577e-01 2.68635988e-01 -3.11406851e-01
1.04056728e+00 9.37961265e-02 -7.22463310e-01 -4.38506335e-01
-1.22636110e-01 2.47189149e-01 2.88928062e-01 3.68306428e-01
-1.59472036e+00 1.47523773e+00 7.25473452e+00 9.91021276e-01
-1.15007484e+00 -2.51751617e-02 1.33247972e+00 2.01407805e-01
-2.40327403e-01 7.89887831e-02 -4.66807783e-01 5.93138635e-01
2.45238811e-01 3.80990565e-01 3.34197193e-01 8.19904685e-01
-4.88188714e-02 -5.35825849e-01 -1.18488383e+00 7.62568891e-01
9.79447961e-02 -1.17265892e+00 1.98058143e-01 -4.24874216e-01
1.21775687e+00 -1.31215945e-01 1.80930227e-01 3.14718872e-01
6.18728876e-01 -8.95999849e-01 9.36190605e-01 2.71555543e-01
1.03964841e+00 -3.61956120e-01 9.71879438e-02 -7.39816576e-02
-1.15914738e+00 5.35506345e-02 -2.25826688e-02 3.24678063e-01
2.36263089e-02 3.07865381e-01 -6.27702892e-01 3.85387063e-01
6.54260516e-01 8.61987054e-01 -7.03105748e-01 1.02542520e+00
-4.09299433e-01 7.19723463e-01 -1.28801703e-01 2.21829206e-01
5.07789373e-01 -6.40434384e-01 1.07511133e-01 1.76993048e+00
-5.58239222e-02 3.76490563e-01 9.71920073e-01 9.86264884e-01
-9.79485288e-02 -6.86799139e-02 -1.84539750e-01 8.83588195e-03
2.40799457e-01 1.23190176e+00 -1.68593895e+00 -7.55363703e-01
-4.05998886e-01 1.05405784e+00 -2.28587821e-01 9.42754626e-01
-7.67721176e-01 -2.47854114e-01 3.18748713e-01 -1.80994377e-01
4.70110029e-01 -1.02188848e-01 -7.16568828e-01 -1.01246011e+00
-5.36225736e-01 -7.07604110e-01 7.45548546e-01 -1.20719051e+00
-1.21564412e+00 2.74916351e-01 4.54150528e-01 -1.05430305e+00
1.03356011e-01 -7.78215706e-01 -7.86090970e-01 5.65448463e-01
-1.34132159e+00 -1.35634696e+00 -3.06810677e-01 7.60853589e-01
1.15832138e+00 1.34355113e-01 3.60235929e-01 -1.15941860e-01
-6.48664594e-01 -1.16878733e-01 -8.50607529e-02 4.41063255e-01
5.04598141e-01 -1.64748931e+00 4.44849819e-01 1.09922421e+00
3.76937836e-01 1.99434653e-01 4.97670114e-01 -7.88315713e-01
-8.22339296e-01 -1.26353025e+00 2.53750920e-01 -4.87271130e-01
4.82556254e-01 -5.00514448e-01 -6.07060850e-01 8.23828995e-01
2.94768065e-01 1.16102457e-01 5.71022153e-01 7.90546089e-02
-3.80745679e-01 2.04294026e-01 -1.11958003e+00 6.83669627e-01
1.09753466e+00 -4.99279648e-01 -7.18999207e-01 4.69696373e-01
1.06904984e+00 -6.65332973e-01 -4.63274956e-01 -6.44349912e-03
2.09848940e-01 -6.75425529e-01 1.21648347e+00 -7.86774516e-01
6.97847784e-01 -3.75366122e-01 -3.21731269e-01 -9.70388830e-01
-1.16698295e-01 -3.26677740e-01 2.29988590e-01 1.04156625e+00
4.83084798e-01 -2.13549823e-01 9.03259277e-01 6.12117648e-01
-4.69262362e-01 -4.32712197e-01 -3.56592655e-01 -5.48791289e-01
1.08314238e-01 -5.72207868e-01 3.74122828e-01 1.04199302e+00
-3.88043314e-01 4.17434365e-01 4.71462831e-02 -2.83199325e-02
5.27755678e-01 6.63803279e-01 5.37206531e-01 -1.18071902e+00
-2.77018964e-01 -5.90682507e-01 -2.58927960e-02 -1.18367505e+00
7.24999085e-02 -9.38936770e-01 4.36331630e-01 -1.79593456e+00
3.20132762e-01 -7.05185771e-01 -2.74532348e-01 6.38145745e-01
-4.08030063e-01 5.41648507e-01 4.59761411e-01 2.10677530e-03
-1.17432201e+00 -2.47393489e-01 1.65661919e+00 -5.11353374e-01
-1.93293661e-01 1.74268021e-03 -7.89645433e-01 7.45938540e-01
8.25923145e-01 -4.38169450e-01 -7.31235027e-01 -4.68409538e-01
-7.70901963e-02 1.88863307e-01 3.26027334e-01 -8.54359925e-01
-1.29271537e-01 -8.95880103e-01 5.95348120e-01 -7.32722104e-01
2.26324394e-01 -7.18638182e-01 -4.61304234e-03 4.23575826e-02
-4.54073310e-01 -2.82247573e-01 2.16257244e-01 6.00228727e-01
-1.38491333e-01 -4.27573472e-01 8.36219847e-01 -4.78421152e-01
-1.58218813e+00 2.11048335e-01 -3.35905999e-01 6.06332839e-01
1.09960294e+00 -5.89182556e-01 -2.87918150e-01 -4.68667969e-02
-1.27645481e+00 2.85903186e-01 5.32622933e-01 2.80529439e-01
4.72608954e-01 -8.70158195e-01 -3.95992577e-01 3.67646441e-02
7.66490176e-02 2.87836820e-01 3.23642761e-01 4.52724099e-01
-6.02890670e-01 1.86503649e-01 -9.45642963e-02 -1.00530791e+00
-1.10162044e+00 8.28977346e-01 2.09288195e-01 2.67652124e-02
-5.42444766e-01 9.55998957e-01 9.67105031e-01 -2.85048485e-01
1.60397574e-01 -8.41884136e-01 -1.49745673e-01 1.16655491e-01
2.73611069e-01 -5.34572126e-03 -3.99345934e-01 -3.53781670e-01
4.74749953e-02 9.35575902e-01 4.61956188e-02 -5.34910746e-02
7.68317163e-01 -3.07581812e-01 -4.60130833e-02 4.17190731e-01
9.66316462e-01 -3.29041064e-01 -1.74647701e+00 -8.33798200e-02
2.10695580e-01 -3.37058991e-01 -1.62169874e-01 -1.16914511e+00
-1.34052825e+00 6.46316051e-01 2.89569527e-01 4.88508046e-01
1.35410631e+00 2.07222357e-01 5.16776562e-01 3.20132196e-01
3.01258206e-01 -1.37557602e+00 4.67306316e-01 5.18272817e-01
4.99716491e-01 -1.35832810e+00 2.05437705e-01 -6.95316672e-01
-1.04738295e+00 1.05023623e+00 6.80317640e-01 -1.56897560e-01
5.64216971e-01 7.66974092e-02 1.68930396e-01 -3.09984446e-01
-2.44444102e-01 -6.12812996e-01 5.25805652e-01 7.75123537e-01
3.51227671e-01 2.61561751e-01 1.03787482e-02 1.30231604e-01
2.14810878e-01 -3.04429382e-01 4.34007138e-01 9.92310882e-01
-6.51689947e-01 -1.03625286e+00 -4.84696329e-01 4.22103703e-01
-5.00019610e-01 -1.22794971e-01 -5.30642509e-01 1.05719948e+00
4.95437115e-01 9.09298778e-01 4.09872144e-01 -1.11066528e-01
-9.90430340e-02 2.02332780e-01 4.45148766e-01 -1.08535135e+00
-5.79437792e-01 4.39516574e-01 2.96452884e-02 -4.13460523e-01
-9.26997662e-01 -7.51100004e-01 -1.44495094e+00 1.41857982e-01
-1.41794845e-01 7.13163102e-03 5.24619043e-01 1.03227031e+00
-2.70011146e-02 6.25743270e-01 2.93911815e-01 -8.87842953e-01
5.55041790e-01 -5.05268693e-01 -6.70963764e-01 6.93206549e-01
2.17442006e-01 -2.80836016e-01 -4.11885940e-02 6.77420259e-01] | [9.606740951538086, 0.5053882598876953] |
c22b4d2a-7016-4a79-b8f9-46801b1f77d9 | current-state-of-community-driven | 2212.14177 | null | https://arxiv.org/abs/2212.14177v2 | https://arxiv.org/pdf/2212.14177v2.pdf | Current State of Community-Driven Radiological AI Deployment in Medical Imaging | Artificial Intelligence (AI) has become commonplace to solve routine everyday tasks. Because of the exponential growth in medical imaging data volume and complexity, the workload on radiologists is steadily increasing. We project that the gap between the number of imaging exams and the number of expert radiologist readers required to cover this increase will continue to expand, consequently introducing a demand for AI-based tools that improve the efficiency with which radiologists can comfortably interpret these exams. AI has been shown to improve efficiency in medical-image generation, processing, and interpretation, and a variety of such AI models have been developed across research labs worldwide. However, very few of these, if any, find their way into routine clinical use, a discrepancy that reflects the divide between AI research and successful AI translation. To address the barrier to clinical deployment, we have formed MONAI Consortium, an open-source community which is building standards for AI deployment in healthcare institutions, and developing tools and infrastructure to facilitate their implementation. This report represents several years of weekly discussions and hands-on problem solving experience by groups of industry experts and clinicians in the MONAI Consortium. We identify barriers between AI-model development in research labs and subsequent clinical deployment and propose solutions. Our report provides guidance on processes which take an imaging AI model from development to clinical implementation in a healthcare institution. We discuss various AI integration points in a clinical Radiology workflow. We also present a taxonomy of Radiology AI use-cases. Through this report, we intend to educate the stakeholders in healthcare and AI (AI researchers, radiologists, imaging informaticists, and regulators) about cross-disciplinary challenges and possible solutions. | ['Rahul Choudhury', 'M. Jorge Cardoso', 'Hyeonhoon Lee', 'Haris Shuaib', 'Risto Haukioja', 'Matthew Lungren', 'David Bericat', 'Tessa Cook', 'Richard D. White', 'Eric Kerfoot', 'Sebastien Ourselin', 'Gigon Bae', 'Ming Melvin Qin', 'Andreas Michael Bucher', 'Tobias Penzkofer', 'Khaled Younis', 'Rickmer Braren', 'Felix Nensa', 'Jens Kleesiek', 'Christopher P Bridge', 'Sidney Bryson', 'Brad Genereaux', 'Laurence Jackson', 'Ralf Floca', 'Carolina Ramirez', 'Barbaros Selnur Erdal', 'Vikash Gupta'] | 2022-12-29 | null | null | null | null | ['medical-image-generation'] | ['medical'] | [ 2.24188864e-01 5.18797398e-01 -6.56460300e-02 -3.40701967e-01
-1.03250504e+00 -6.54708862e-01 -1.14214756e-01 3.68308336e-01
-3.53077739e-01 2.83641398e-01 4.51687992e-01 -8.74694705e-01
-3.13043654e-01 -3.72554392e-01 -5.11647284e-01 -3.72856349e-01
4.84879948e-02 9.71997976e-01 -1.10627443e-01 2.88525522e-01
-7.40878237e-03 7.66806304e-01 -8.12455535e-01 5.97042680e-01
8.19718659e-01 7.32082129e-01 9.23558399e-02 9.03993547e-01
1.79875419e-02 1.38029897e+00 -4.98696864e-01 -8.67256522e-02
4.61843163e-01 -5.58538318e-01 -9.18349504e-01 3.38921040e-01
5.25140092e-02 -3.70819718e-01 -2.67319828e-01 5.01146495e-01
7.60005653e-01 -4.35110599e-01 4.79300678e-01 -9.58713233e-01
-5.26456833e-01 7.39777207e-01 -4.05318797e-01 6.23843431e-01
2.59071976e-01 6.78089082e-01 4.18398440e-01 -3.87020051e-01
8.27967525e-01 4.87732381e-01 6.63331807e-01 4.69893396e-01
-7.69431114e-01 -5.40839434e-01 -1.28428981e-01 3.08609664e-01
-1.06471622e+00 -8.42786729e-02 2.73576438e-01 -6.31056786e-01
9.62200403e-01 5.63359618e-01 1.20548534e+00 5.49208820e-01
4.74369287e-01 8.87860239e-01 9.50372994e-01 -5.58876455e-01
2.76714534e-01 2.14305684e-01 2.03035757e-01 6.90118909e-01
4.77249652e-01 -3.96966606e-01 -2.22975627e-01 -3.61079067e-01
9.96460080e-01 -1.02870397e-01 -2.71033853e-01 -5.27718365e-02
-1.49193954e+00 6.09488606e-01 4.24092352e-01 6.25388265e-01
-8.57691109e-01 1.44438356e-01 3.78647000e-01 7.14886114e-02
-1.33991435e-01 9.31752264e-01 1.74730107e-01 -1.86726198e-01
-6.86713934e-01 8.11990947e-02 8.38956237e-01 6.34205878e-01
-1.69971392e-01 -1.27168864e-01 3.30006406e-02 6.57421529e-01
3.19826305e-01 4.14194882e-01 5.96123517e-01 -1.28936386e+00
-2.40454778e-01 6.09469295e-01 2.74447501e-02 -8.84552300e-01
-5.93290925e-01 -4.89500195e-01 -5.74589550e-01 4.18332666e-02
2.70457000e-01 -1.59593031e-01 -1.10920978e+00 9.14087772e-01
1.20322712e-01 -1.24144308e-01 -1.11280344e-01 1.22732687e+00
8.01548362e-01 2.90425658e-01 4.56908405e-01 -2.53075808e-01
1.73570740e+00 -9.40765202e-01 -7.27789640e-01 -4.29295331e-01
8.48139822e-01 -7.72182047e-01 9.80366230e-01 5.85354745e-01
-1.58276260e+00 -1.95421934e-01 -9.65583265e-01 1.95402354e-01
6.51335493e-02 -1.74189836e-01 7.41954625e-01 6.93523049e-01
-8.72364044e-01 -3.79221924e-02 -1.37711978e+00 -5.46785772e-01
6.08703136e-01 4.51274395e-01 -1.64149061e-01 -3.76102298e-01
-5.71464062e-01 1.45532167e+00 -5.51734976e-02 7.47934207e-02
-4.97559518e-01 -1.18637061e+00 -3.71262878e-01 -2.55269319e-01
3.41771483e-01 -1.27223027e+00 1.69423807e+00 -8.28653038e-01
-9.81203020e-01 1.09570885e+00 1.73795044e-01 -5.55856645e-01
4.53662962e-01 -3.93136926e-02 -5.73785484e-01 3.60797256e-01
2.64654845e-01 7.03330994e-01 1.21368859e-02 -1.09280658e+00
-7.46139467e-01 -2.54671246e-01 -2.82902092e-01 4.29360747e-01
-8.69779587e-02 1.08446762e-01 -3.44815671e-01 -3.13484639e-01
1.15648851e-01 -1.22025061e+00 -8.80188704e-01 2.78030843e-01
1.04462594e-01 1.25617221e-01 4.01204824e-01 -4.20015544e-01
1.23964059e+00 -2.05825973e+00 -3.32976848e-01 4.00239319e-01
5.72640538e-01 2.94198990e-01 3.06193411e-01 1.53766438e-01
-1.40576616e-01 1.79346606e-01 1.05330618e-02 6.47086322e-01
-3.75508070e-01 2.79287875e-01 1.18989408e-01 2.31079653e-01
1.01175547e-01 1.20215726e+00 -1.01619732e+00 -7.07086205e-01
3.94060731e-01 3.31050456e-01 -4.24120307e-01 1.56545222e-01
7.83093497e-02 7.36420810e-01 -7.50872910e-01 7.74337888e-01
3.52692157e-01 -9.97345448e-01 5.23442686e-01 -1.64804146e-01
-1.62595540e-01 1.47903308e-01 -8.79426956e-01 1.51591694e+00
-4.60038662e-01 3.52312416e-01 9.46502313e-02 -6.78655088e-01
4.51943666e-01 7.10732162e-01 1.07487237e+00 -3.75629187e-01
2.10723296e-01 5.18438160e-01 8.21060598e-01 -9.40789342e-01
-1.90201133e-01 -2.65223563e-01 2.91682869e-01 8.45191419e-01
-5.53396344e-01 -6.49790287e-01 2.46069223e-01 1.55632183e-01
1.58414483e+00 -2.74527401e-01 4.19568747e-01 -6.61137402e-02
3.08378249e-01 9.10946369e-01 2.86278814e-01 9.04827595e-01
-4.40593511e-01 6.10734344e-01 1.21171586e-01 -9.41931248e-01
-1.20208395e+00 -9.13594961e-01 -2.52073318e-01 4.87862438e-01
-1.88728333e-01 -1.16272099e-01 -6.33625448e-01 -5.49978316e-01
-1.96867883e-01 6.50208175e-01 -3.26772273e-01 6.79530427e-02
-7.97877073e-01 -6.92981362e-01 4.18483853e-01 6.19701207e-01
5.79147004e-02 -1.30441022e+00 -1.56492591e+00 6.17278934e-01
-3.01273823e-01 -1.23064816e+00 -1.14417985e-01 -1.59587860e-01
-8.46132696e-01 -1.01007700e+00 -7.76236653e-01 -6.99240506e-01
8.27724278e-01 2.73168474e-01 1.08119655e+00 1.92534328e-01
-1.18110144e+00 9.55717981e-01 -2.39679992e-01 -1.18167579e+00
-8.16222012e-01 -2.69385219e-01 -4.61287171e-01 -7.09285080e-01
5.34186423e-01 -2.08315849e-01 -1.10953152e+00 1.94587275e-01
-1.04457581e+00 4.31283534e-01 1.18724024e+00 7.28929758e-01
6.88654482e-01 -4.92110014e-01 4.27978069e-01 -1.18032026e+00
6.65147424e-01 -6.76934600e-01 -3.19885416e-03 4.15987492e-01
-7.19748914e-01 -2.25433856e-01 7.84745731e-04 -2.76149690e-01
-8.67736518e-01 1.73455670e-01 3.03607397e-02 -1.68460444e-01
-1.89928055e-01 8.90630186e-01 9.01182532e-01 -1.23801403e-01
1.05392885e+00 -2.36461788e-01 4.47685689e-01 1.93031847e-01
1.01895206e-01 7.85247147e-01 5.12663305e-01 -3.84151340e-01
2.67060667e-01 5.04577637e-01 -1.19561233e-01 -5.73879302e-01
-7.25822508e-01 -4.74659175e-01 1.42093487e-02 -5.52478135e-01
9.22300339e-01 -6.83563590e-01 -3.80702168e-01 -2.12934330e-01
-9.61877763e-01 -1.07447460e-01 -7.58517981e-01 8.61857891e-01
-3.98227930e-01 1.29931718e-01 -7.28372276e-01 -4.94732529e-01
-7.59546280e-01 -1.77005827e+00 6.69157028e-01 3.36580455e-01
-8.49613965e-01 -7.23089516e-01 -3.25182229e-02 7.70639300e-01
6.58161342e-01 4.46249634e-01 1.01610315e+00 -4.79747862e-01
-7.61936486e-01 -4.13602352e-01 -2.93505222e-01 -6.05886877e-02
3.76163945e-02 -1.11046575e-01 -5.01797020e-01 1.80444524e-01
2.47578904e-01 -1.15480177e-01 -4.04049940e-02 8.92266870e-01
1.17693830e+00 3.29083167e-02 -4.64085609e-01 1.79707766e-01
1.09774768e+00 7.80037761e-01 5.93983352e-01 5.08200765e-01
3.14506650e-01 3.94339055e-01 5.89835346e-01 2.91022539e-01
4.22496498e-01 1.26525164e-01 3.42340507e-02 -6.17961526e-01
-1.79700956e-01 3.79525632e-01 -3.08625191e-01 1.06900132e+00
-4.20861006e-01 4.29282710e-02 -1.62326193e+00 6.19594514e-01
-1.80156958e+00 -4.58824873e-01 -1.91825405e-01 1.70634067e+00
7.22449362e-01 1.13805786e-01 3.19871418e-02 -2.62060612e-01
1.74488053e-01 -6.03138983e-01 -6.38821423e-01 -5.10648310e-01
3.62350196e-01 3.94593120e-01 5.97759306e-01 1.52940348e-01
-6.42677307e-01 5.14034629e-01 7.39853048e+00 1.78899691e-01
-1.26862407e+00 1.58279300e-01 8.98352861e-01 -3.56532000e-02
-2.74566978e-01 -3.13245237e-01 -1.74238291e-02 -4.49258760e-02
1.03798306e+00 -6.55123472e-01 1.11106388e-01 8.72375309e-01
2.81073481e-01 -2.40295738e-01 -1.18100679e+00 8.52670908e-01
-5.33221848e-03 -1.70273817e+00 -1.82641134e-01 1.25374854e-01
5.05077422e-01 9.51935202e-02 1.56442001e-02 -1.90645441e-01
4.45504516e-01 -1.18427312e+00 1.79369062e-01 5.39075673e-01
6.15480602e-01 -1.39400586e-01 8.98489773e-01 -2.34680083e-02
-6.64789677e-01 -1.68667302e-01 3.85650550e-03 1.82316288e-01
2.85463512e-01 4.03880358e-01 -1.52499127e+00 4.44421589e-01
7.51858830e-01 1.13443367e-01 -2.36770302e-01 1.21418178e+00
2.04724550e-01 5.42271852e-01 -2.22005635e-01 2.41006851e-01
3.44727546e-01 5.96951619e-02 3.98728758e-01 1.36221623e+00
1.73997372e-01 7.53394127e-01 2.25736246e-01 5.40469229e-01
4.15794343e-01 2.22374544e-01 -5.53554595e-01 -3.76099795e-01
3.60301226e-01 1.46694839e+00 -1.03658712e+00 -4.94338036e-01
-6.22884691e-01 4.08966631e-01 -3.14398587e-01 4.79808375e-02
-8.83431315e-01 9.41484049e-02 2.88176417e-01 4.67047006e-01
-3.79385859e-01 3.05602942e-02 -7.56872416e-01 -5.54090858e-01
-1.21139206e-01 -1.54353726e+00 6.56705379e-01 -9.18568611e-01
-1.17625797e+00 5.69298029e-01 -5.54860123e-02 -1.25100398e+00
-5.74388325e-01 -5.89689791e-01 -1.98474750e-01 5.62499881e-01
-1.08006251e+00 -1.20739257e+00 -5.76352239e-01 8.10580328e-02
4.84825581e-01 -1.66947283e-02 9.86992061e-01 2.84180939e-01
-1.13094158e-01 2.76498377e-01 -3.83975983e-01 6.99236840e-02
6.11178100e-01 -9.55249786e-01 -4.53660041e-02 3.15430939e-01
-2.71360934e-01 8.37014198e-01 4.76849228e-01 -4.94325846e-01
-1.75691998e+00 -5.20386338e-01 4.44297314e-01 -6.01343155e-01
6.64085925e-01 4.83810276e-01 -6.30994141e-01 9.63663161e-01
2.77555674e-01 -5.11030108e-02 1.15007114e+00 -3.93000156e-01
1.86731040e-01 8.31478909e-02 -1.17380512e+00 7.88375258e-01
6.87259555e-01 -9.72027406e-02 -5.68997145e-01 6.90403759e-01
4.43317503e-01 -7.86517441e-01 -1.26358700e+00 5.64355969e-01
7.75261164e-01 -5.34340501e-01 8.50481749e-01 -4.79528606e-01
6.12231970e-01 -3.05765808e-01 6.09681487e-01 -9.20645237e-01
-4.12202865e-01 -3.12580019e-01 2.69731700e-01 1.23880610e-01
5.74701607e-01 -5.89740217e-01 7.80706465e-01 1.31244934e+00
-4.87795234e-01 -9.97280717e-01 -5.46233594e-01 -1.51177319e-02
-8.49933997e-02 -4.10919785e-01 1.86039612e-01 1.02323258e+00
3.89744639e-01 -1.05536180e-02 4.05317962e-01 4.37885746e-02
3.04824799e-01 -1.02056868e-01 5.13234556e-01 -9.06931281e-01
-4.57307011e-01 -5.74138999e-01 -7.06117511e-01 -3.68835032e-01
-8.45025122e-01 -8.39609027e-01 -2.71331131e-01 -2.16888046e+00
4.20696795e-01 -5.56378126e-01 -2.18001813e-01 5.60670853e-01
5.38683198e-02 1.80098325e-01 1.96640804e-01 5.09223998e-01
-3.59755814e-01 -7.75838792e-01 1.59315252e+00 8.63504317e-03
-2.27211118e-01 -3.13266426e-01 -1.15976417e+00 7.39831626e-01
5.75320959e-01 -4.02781218e-01 -3.79089296e-01 -5.45490682e-01
2.27179810e-01 1.91125900e-01 6.89641014e-02 -1.35550737e+00
6.55538082e-01 -2.36666173e-01 4.84549850e-01 -3.87825370e-01
-6.11984506e-02 -1.07675552e+00 7.68179536e-01 1.04858458e+00
-3.03704590e-01 3.63581330e-02 2.99735904e-01 -9.02252346e-02
-2.00395241e-01 -1.71994984e-01 7.95620024e-01 -6.59165859e-01
-3.28221977e-01 1.31195020e-02 -7.96102643e-01 -4.23171511e-03
1.48293149e+00 -3.25958401e-01 -2.91057885e-01 -4.52384084e-01
-1.06253886e+00 4.58417803e-01 1.70661733e-01 2.42880043e-02
4.85897362e-01 -7.55131364e-01 -9.60983574e-01 -1.17923781e-01
5.16332984e-02 3.93463403e-01 5.53223968e-01 1.37753391e+00
-1.35917974e+00 4.97069478e-01 -2.91405201e-01 -7.29270697e-01
-1.06793606e+00 3.90822023e-01 4.53035116e-01 -5.96212924e-01
-8.06756079e-01 5.65369308e-01 1.12553209e-01 -9.50623825e-02
8.73952135e-02 -1.49375990e-01 2.39261016e-01 -4.29285556e-01
6.58542573e-01 1.59251839e-01 1.78220838e-01 -1.26097370e-02
-3.72913182e-01 3.66085052e-01 -5.05606651e-01 -2.89100349e-01
1.35081804e+00 1.22148015e-01 7.30615482e-02 3.74327540e-01
4.60598707e-01 -1.97215244e-01 -3.75353724e-01 1.33978263e-01
9.05069858e-02 -9.72171426e-02 9.20215994e-02 -1.19556439e+00
-9.11289752e-01 6.51065171e-01 6.80501103e-01 2.04383388e-01
1.12298381e+00 2.92328835e-01 7.30652750e-01 2.23822162e-01
1.97241381e-01 -1.05603802e+00 9.67434347e-02 -1.97193533e-01
9.77193475e-01 -1.05470657e+00 3.98515880e-01 -5.91444194e-01
-1.20833516e+00 1.17947578e+00 5.54484665e-01 3.89458612e-02
5.60262620e-01 7.94729710e-01 7.18765199e-01 -5.54180801e-01
-5.16062140e-01 1.72184020e-01 2.03179941e-02 4.65671122e-01
9.55693662e-01 2.65387863e-01 -7.09375560e-01 3.29342484e-01
-6.32311031e-02 7.08061278e-01 5.87554812e-01 1.39018226e+00
-2.71336824e-01 -1.12864363e+00 -6.05496883e-01 7.04536200e-01
-7.26140141e-01 3.21440771e-02 -2.43669719e-01 8.03203106e-01
3.97327635e-03 7.90130734e-01 3.50254849e-02 1.11947007e-01
5.01562238e-01 -3.13621052e-02 4.92654979e-01 -6.51802003e-01
-9.64242220e-01 3.32160473e-01 2.55108058e-01 -4.61512715e-01
-2.53150135e-01 -6.44416451e-01 -1.64963090e+00 -6.89899549e-02
-9.88718867e-03 2.22727582e-02 1.00119209e+00 5.72536051e-01
5.48543394e-01 9.36245561e-01 8.00085142e-02 -8.48447829e-02
-3.19543481e-01 -6.25346959e-01 -1.85947090e-01 2.16936603e-01
-2.35091448e-01 -8.98380652e-02 4.48994711e-02 2.72629946e-01] | [14.849830627441406, -2.3291335105895996] |
d9de030b-f161-4ebb-8fd2-ed4ea8e6b3f0 | context-aware-visual-tracking-with-joint-meta | 2204.01513 | null | https://arxiv.org/abs/2204.01513v1 | https://arxiv.org/pdf/2204.01513v1.pdf | Context-aware Visual Tracking with Joint Meta-updating | Visual object tracking acts as a pivotal component in various emerging video applications. Despite the numerous developments in visual tracking, existing deep trackers are still likely to fail when tracking against objects with dramatic variation. These deep trackers usually do not perform online update or update single sub-branch of the tracking model, for which they cannot adapt to the appearance variation of objects. Efficient updating methods are therefore crucial for tracking while previous meta-updater optimizes trackers directly over parameter space, which is prone to over-fit even collapse on longer sequences. To address these issues, we propose a context-aware tracking model to optimize the tracker over the representation space, which jointly meta-update both branches by exploiting information along the whole sequence, such that it can avoid the over-fitting problem. First, we note that the embedded features of the localization branch and the box-estimation branch, focusing on the local and global information of the target, are effective complements to each other. Based on this insight, we devise a context-aggregation module to fuse information in historical frames, followed by a context-aware module to learn affinity vectors for both branches of the tracker. Besides, we develop a dedicated meta-learning scheme, on account of fast and stable updating with limited training samples. The proposed tracking method achieves an EAO score of 0.514 on VOT2018 with the speed of 40FPS, demonstrating its capability of improving the accuracy and robustness of the underlying tracker with little speed drop. | ['Yongsheng Liang', 'Fanyang Meng', 'Xin Li', 'Qiuhong Shen'] | 2022-04-04 | null | null | null | null | ['visual-object-tracking'] | ['computer-vision'] | [-3.75265002e-01 -5.91727614e-01 -1.57691628e-01 -1.58650223e-02
-3.49905461e-01 -5.73645115e-01 4.13996011e-01 4.35416996e-02
-5.30851007e-01 5.65236032e-01 -1.15399376e-01 7.90559202e-02
3.99467461e-02 -3.46530110e-01 -7.54808962e-01 -8.81726682e-01
1.20269237e-02 1.00011081e-01 8.38238835e-01 1.36254663e-02
-1.03081793e-01 4.75355774e-01 -1.58123612e+00 -3.33329290e-01
8.86653423e-01 1.17444718e+00 4.11474764e-01 6.69051230e-01
2.38589644e-02 4.86338288e-01 -6.89985275e-01 -4.80222881e-01
1.72404066e-01 -1.43984035e-02 5.84303476e-02 5.58542386e-02
7.01570809e-01 -4.42366600e-01 -6.15779936e-01 1.16483665e+00
5.55924952e-01 5.86958192e-02 2.63189584e-01 -1.15580225e+00
-2.71206707e-01 2.81277925e-01 -6.54783309e-01 4.92731333e-01
-4.36246134e-02 3.65328431e-01 7.99829125e-01 -7.92828500e-01
4.54787999e-01 9.90339458e-01 8.50521684e-01 5.78084946e-01
-8.60781074e-01 -7.93472648e-01 5.43921888e-01 3.02808285e-01
-1.20690858e+00 -4.02200192e-01 6.56200171e-01 -4.24111634e-01
3.29526067e-01 9.83709171e-02 9.58984673e-01 1.07634318e+00
2.94225067e-01 8.39036107e-01 4.97842878e-01 1.01874910e-01
9.47761014e-02 -4.98823151e-02 -3.21785919e-02 4.86246020e-01
6.06058955e-01 4.20031607e-01 -4.74426478e-01 6.51298603e-03
7.50730693e-01 1.29413307e-01 -2.55810142e-01 -6.67995095e-01
-1.25802767e+00 5.01686633e-01 5.28305590e-01 2.86792904e-01
-2.73487866e-01 4.51502681e-01 3.94730002e-01 5.95213883e-02
2.10328192e-01 -9.07578468e-02 -3.32253128e-01 -2.44427294e-01
-9.19396162e-01 1.91501617e-01 3.40395063e-01 9.20871198e-01
4.71693367e-01 3.59774590e-01 -5.03142774e-01 4.24024045e-01
5.84014356e-01 7.84162521e-01 3.15101713e-01 -5.96763015e-01
3.55706185e-01 5.68298519e-01 4.77486253e-01 -1.13847637e+00
-4.44905549e-01 -1.06544244e+00 -5.77610910e-01 2.06153721e-01
4.95608509e-01 -1.69873893e-01 -7.81086504e-01 1.87428510e+00
9.56231773e-01 5.56796372e-01 -4.13752377e-01 9.75005507e-01
5.50066829e-01 3.76751006e-01 1.96030468e-01 -4.71318454e-01
1.36063540e+00 -1.03337216e+00 -8.84855866e-01 -1.52453542e-01
4.54683632e-01 -7.74994195e-01 5.35233557e-01 2.11966977e-01
-8.35480034e-01 -9.41546023e-01 -1.15051436e+00 4.47877437e-01
-7.67241465e-03 3.61195922e-01 4.37414765e-01 6.43994153e-01
-7.93425500e-01 3.51569057e-01 -1.13565040e+00 -2.00607523e-01
4.36985314e-01 4.09549683e-01 -5.49781434e-02 2.60082483e-01
-9.18545008e-01 7.86309838e-01 4.29859728e-01 3.41266304e-01
-1.03226078e+00 -8.21994245e-01 -5.15790701e-01 5.76260500e-02
6.24581933e-01 -6.88722014e-01 1.06845641e+00 -6.78741872e-01
-1.48063421e+00 1.95778012e-01 -1.42615974e-01 -5.81401646e-01
9.10083830e-01 -5.72363615e-01 -5.34643888e-01 -9.96385068e-02
-1.33116066e-01 4.29158539e-01 1.26816976e+00 -1.14704502e+00
-1.03917873e+00 -2.21891358e-01 -1.05046816e-01 1.20333351e-01
-5.65103531e-01 -1.43718690e-01 -9.65832531e-01 -8.87245715e-01
-9.29466784e-02 -9.40025866e-01 -1.16156168e-01 3.42853278e-01
8.08105692e-02 -6.02817647e-02 1.22912586e+00 -5.07505596e-01
1.56490064e+00 -2.14046264e+00 7.99743682e-02 -1.49540022e-01
3.00484776e-01 7.61373580e-01 2.74277497e-02 -4.64019030e-02
2.16327816e-01 -5.76361418e-01 1.88590527e-01 -4.50717568e-01
-5.29470295e-02 7.97827020e-02 -2.42040351e-01 7.16262579e-01
-7.65223801e-03 9.98960674e-01 -1.04725909e+00 -5.17322779e-01
4.56929415e-01 6.87539697e-01 -4.15405184e-01 1.90108389e-01
-4.07445401e-01 6.54198349e-01 -5.97685575e-01 6.24225795e-01
8.24287236e-01 -3.01531285e-01 9.59040150e-02 -4.58933860e-01
-4.06640053e-01 -2.51827240e-01 -1.33703005e+00 1.70402777e+00
-8.79276842e-02 3.98238540e-01 1.79460570e-01 -6.79466844e-01
7.77312458e-01 2.65736759e-01 8.47695291e-01 -5.06475627e-01
2.82128155e-01 1.58576488e-01 6.24430291e-02 -1.93312764e-01
4.50989276e-01 2.50831336e-01 2.96587527e-01 1.03528403e-01
7.89826959e-02 6.02857113e-01 1.94237724e-01 6.82228953e-02
1.05353558e+00 3.84255022e-01 8.28756616e-02 3.46473828e-02
8.04762483e-01 -3.54204297e-01 7.75272310e-01 7.14320898e-01
-5.32119930e-01 2.35327885e-01 -6.77254796e-02 -5.92673957e-01
-8.01135957e-01 -9.37586963e-01 -9.67648923e-02 9.66952741e-01
6.04389250e-01 -4.33852434e-01 -5.51075518e-01 -7.79909134e-01
1.49696007e-01 7.18033165e-02 -4.86128956e-01 -3.03547204e-01
-7.84634411e-01 -7.67049849e-01 2.64175266e-01 6.02530837e-01
4.02276754e-01 -6.25934660e-01 -9.97694492e-01 5.33611298e-01
8.59971642e-02 -1.29193866e+00 -6.88188314e-01 -1.40256241e-01
-9.02193010e-01 -1.10158789e+00 -7.78244555e-01 -3.73239279e-01
4.92257386e-01 4.40611511e-01 7.17563868e-01 3.51944238e-01
-9.62882265e-02 4.65771854e-01 -2.48214185e-01 -1.85349777e-01
-2.39934847e-01 6.24697544e-02 3.02445799e-01 3.68396670e-01
1.97457299e-02 -4.22595590e-01 -8.28794003e-01 6.06733680e-01
-5.84371924e-01 -1.80445477e-01 5.66632390e-01 7.91842103e-01
5.09310484e-01 -1.15545683e-01 4.00858700e-01 -1.81224346e-01
-3.97040136e-02 -3.15460116e-01 -1.21629131e+00 3.97969127e-01
-3.55502814e-01 -8.02279338e-02 5.52253604e-01 -9.23801482e-01
-8.67619336e-01 3.50497603e-01 -2.92165112e-02 -8.82403374e-01
3.11113894e-01 2.68871505e-02 -2.89937884e-01 -4.98137623e-01
2.60625184e-01 3.05051267e-01 -6.79108575e-02 -5.12589872e-01
1.88258395e-01 2.12757289e-01 7.89992034e-01 -4.64229524e-01
1.28752184e+00 6.41005993e-01 3.60694602e-02 -5.67013502e-01
-8.83127451e-01 -5.79555094e-01 -4.47559774e-01 -7.62612641e-01
7.62903929e-01 -1.05240726e+00 -1.12227345e+00 5.60443342e-01
-1.04046392e+00 5.37235895e-03 -2.00897545e-01 6.95652962e-01
-2.82807499e-01 6.57795489e-01 -3.21637094e-01 -8.48480284e-01
-2.65600890e-01 -1.14665425e+00 1.08759892e+00 5.79256356e-01
2.70376533e-01 -9.76341844e-01 2.47868493e-01 -1.51358079e-02
4.68723863e-01 2.37210900e-01 1.12954244e-01 -4.74742532e-01
-9.15035903e-01 -2.90993690e-01 -2.10498109e-01 1.40745729e-01
1.43261775e-01 4.52536643e-02 -8.72130930e-01 -7.63432026e-01
6.05410784e-02 1.55826420e-01 7.82978475e-01 4.87883568e-01
8.25777769e-01 4.25545909e-02 -5.97318947e-01 7.50277519e-01
1.30615067e+00 1.47872135e-01 2.66454101e-01 3.84937346e-01
8.52261782e-01 1.09081427e-02 9.21453655e-01 5.94178438e-01
3.71159375e-01 1.15900242e+00 7.47088671e-01 2.65744060e-01
-3.21622163e-01 -1.48813859e-01 5.59729278e-01 7.72656620e-01
-4.30096313e-03 -1.52231902e-01 -5.27137995e-01 4.78252470e-01
-2.10851789e+00 -9.71777141e-01 -1.29923578e-02 2.41021276e+00
5.81054389e-01 3.92283082e-01 3.07094604e-01 -2.11784154e-01
8.97493720e-01 2.83130705e-01 -7.76932299e-01 5.01483679e-01
3.51614691e-02 -3.07601690e-01 4.96803433e-01 3.86196226e-01
-1.36294103e+00 9.24683213e-01 5.79823208e+00 9.51827705e-01
-1.22493494e+00 2.35350162e-01 -8.25343579e-02 -2.61404961e-01
1.06163599e-01 2.10694298e-02 -1.27714682e+00 9.41362619e-01
7.07580984e-01 2.14328319e-02 1.12047412e-01 8.64704788e-01
7.12374896e-02 4.70471494e-02 -8.44869792e-01 1.07338035e+00
-1.28105253e-01 -1.29174602e+00 -1.01806663e-01 1.11810751e-01
5.81131637e-01 7.56872967e-02 7.40937069e-02 3.05942982e-01
6.84034899e-02 -3.97905380e-01 1.03144491e+00 5.60201824e-01
3.43245775e-01 -5.11884689e-01 6.17807746e-01 3.26155365e-01
-1.76786780e+00 -2.25931704e-01 -4.07811880e-01 3.95026505e-01
4.43421543e-01 5.59619367e-01 -2.80913472e-01 6.75740778e-01
7.90286422e-01 9.06363130e-01 -6.52436376e-01 1.59534109e+00
2.78811827e-02 3.64304602e-01 -4.65102553e-01 1.92489833e-01
9.52232704e-02 2.83581410e-02 9.94745910e-01 9.88913298e-01
4.43166822e-01 -2.56133318e-01 4.03805703e-01 4.65049148e-01
1.75933108e-01 -2.56379306e-01 -1.38525799e-01 1.60244271e-01
5.69769561e-01 1.43634403e+00 -6.56519771e-01 -2.86740512e-01
-4.85130250e-01 5.99255860e-01 2.09789902e-01 2.49058738e-01
-1.47168624e+00 2.32019294e-02 8.96003187e-01 -7.81597197e-02
9.22190726e-01 -4.06859905e-01 8.87498781e-02 -1.36987174e+00
2.58834898e-01 -7.85176873e-01 3.59577686e-01 -3.10663670e-01
-9.14859712e-01 4.69459623e-01 -1.99732661e-01 -1.70744646e+00
9.67814401e-02 -3.93845677e-01 -4.84857470e-01 3.65677387e-01
-1.56913173e+00 -1.26161420e+00 -4.16906744e-01 5.15467584e-01
5.00474870e-01 -1.18224360e-01 4.33648042e-02 7.00149953e-01
-7.18528450e-01 8.32910120e-01 9.88423377e-02 1.82405338e-02
8.07260275e-01 -8.76336873e-01 2.00410277e-01 1.15338850e+00
2.08170742e-01 6.27258480e-01 8.24590147e-01 -8.18000257e-01
-1.66702783e+00 -1.15586257e+00 3.14232618e-01 -5.56781530e-01
8.05734992e-01 -2.55576015e-01 -1.01897466e+00 4.49169725e-01
-1.53453216e-01 4.62703437e-01 2.17861354e-01 -1.14466168e-01
-2.07452908e-01 -3.89814377e-01 -7.28469968e-01 4.90731210e-01
1.16355371e+00 -1.73537627e-01 -3.05030555e-01 9.69110206e-02
7.46409059e-01 -7.75133848e-01 -8.93673122e-01 5.78082383e-01
7.36612320e-01 -9.11950469e-01 1.12960780e+00 -3.45620006e-01
-4.37401533e-01 -9.66397405e-01 6.02687374e-02 -8.06839168e-01
-4.56771642e-01 -7.66438425e-01 -1.07341075e+00 1.25806201e+00
-2.52811283e-01 -3.76466632e-01 8.49830031e-01 1.28290832e-01
-1.53201953e-01 -6.73657656e-01 -1.27687335e+00 -9.87379789e-01
-5.33217251e-01 -4.01890576e-01 4.91824389e-01 4.83131617e-01
-6.38267398e-01 6.58558458e-02 -6.92626178e-01 5.18646955e-01
9.76372480e-01 5.97390570e-02 9.87172663e-01 -1.26627731e+00
-4.40326631e-01 -4.10042524e-01 -5.73011577e-01 -1.40084350e+00
-6.12863936e-02 -3.82200480e-01 4.41692919e-02 -1.07309031e+00
4.97308634e-02 -3.94576758e-01 -4.90250349e-01 1.69458076e-01
-5.19127190e-01 2.30572611e-01 4.95501816e-01 3.89383346e-01
-1.02256370e+00 8.82516682e-01 1.17797351e+00 -1.69595689e-01
-2.70656049e-02 2.85556108e-01 -3.26221049e-01 6.37337208e-01
2.78404444e-01 -5.91027617e-01 -2.66830772e-01 -5.58515489e-01
6.41300157e-02 -1.60882920e-01 5.14974236e-01 -1.40325463e+00
5.82104623e-01 7.25737289e-02 5.19187570e-01 -9.20993268e-01
3.05884361e-01 -1.20976353e+00 4.55291659e-01 7.10262656e-01
1.49350435e-01 1.31000057e-01 3.30225825e-01 1.00842321e+00
-1.12091623e-01 -1.11708231e-02 8.82313311e-01 3.29702020e-01
-8.12884808e-01 6.90413177e-01 -6.47638887e-02 -5.78962788e-02
1.17003477e+00 -3.45861286e-01 -1.56058446e-01 -1.39994249e-01
-5.71378708e-01 4.86063629e-01 6.29286528e-01 6.15146458e-01
2.60044783e-01 -1.52405405e+00 -4.08526242e-01 4.59064171e-02
-5.32152057e-02 -1.70811161e-01 4.92497712e-01 1.16368103e+00
-1.63609952e-01 2.69901425e-01 -9.68223736e-02 -1.02303255e+00
-1.16971827e+00 7.80378878e-01 3.72892976e-01 -3.65685076e-01
-7.32735455e-01 6.14443898e-01 2.91166633e-01 1.04207240e-01
5.00676572e-01 -2.72757441e-01 -2.48113245e-01 1.61165103e-01
8.03371608e-01 3.03670436e-01 -6.93999976e-02 -7.84150302e-01
-5.43922484e-01 9.59658384e-01 -1.07192852e-01 2.48292521e-01
1.04793179e+00 -3.28722626e-01 3.48655492e-01 2.08407015e-01
7.74747670e-01 1.17287241e-01 -1.94335449e+00 -3.76332194e-01
4.58397307e-02 -6.84145927e-01 -1.99385826e-02 -4.36244249e-01
-1.38344705e+00 7.15259373e-01 9.08835471e-01 -2.80786045e-02
1.03634906e+00 -3.49747270e-01 9.67455626e-01 8.07599500e-02
3.64161491e-01 -1.03953731e+00 2.11926669e-01 4.59225506e-01
4.33027744e-01 -1.16958284e+00 2.28924260e-01 -9.92846787e-02
-3.60354513e-01 1.01418102e+00 7.63708055e-01 5.43911047e-02
5.29632211e-01 3.09429109e-01 -1.47669222e-02 5.59301637e-02
-7.77590334e-01 -2.75362104e-01 5.19672751e-01 5.43296695e-01
2.22148187e-02 -3.92118365e-01 -1.58013567e-01 3.06908727e-01
3.33679497e-01 -3.23750749e-02 -6.65549040e-02 8.18818688e-01
-5.44505894e-01 -1.02117872e+00 -5.23122668e-01 -3.99887674e-02
-3.42724711e-01 3.80396932e-01 1.88298121e-01 8.31943095e-01
1.51087701e-01 5.92920840e-01 -9.28525403e-02 -4.42134857e-01
2.93967605e-01 -3.09055448e-01 4.76163775e-01 -5.88957332e-02
-7.17292070e-01 4.28375334e-01 -3.73545140e-01 -7.52442658e-01
-6.23350501e-01 -7.46260285e-01 -9.21312988e-01 -1.43311545e-01
-6.55709505e-01 7.67154470e-02 5.22492409e-01 9.80878592e-01
3.82351518e-01 6.29507959e-01 5.11562347e-01 -1.00260556e+00
-7.62119949e-01 -6.85274899e-01 -1.99384212e-01 1.60264000e-01
5.62735975e-01 -1.16973591e+00 -1.55755088e-01 -1.19899265e-01] | [6.340561389923096, -2.126520872116089] |
60c563f6-7e43-4e00-8711-79bd67c12c03 | the-ab-use-of-open-source-code-to-train-large | 2302.13681 | null | https://arxiv.org/abs/2302.13681v2 | https://arxiv.org/pdf/2302.13681v2.pdf | The (ab)use of Open Source Code to Train Large Language Models | In recent years, Large Language Models (LLMs) have gained significant popularity due to their ability to generate human-like text and their potential applications in various fields, such as Software Engineering. LLMs for Code are commonly trained on large unsanitized corpora of source code scraped from the Internet. The content of these datasets is memorized and emitted by the models, often in a verbatim manner. In this work, we will discuss the security, privacy, and licensing implications of memorization. We argue why the use of copyleft code to train LLMs is a legal and ethical dilemma. Finally, we provide four actionable recommendations to address this issue. | ['Maliheh Izadi', 'Ali Al-Kaswan'] | 2023-02-27 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 1.02421492e-01 7.26961792e-01 -3.14706624e-01 -3.24401826e-01
-7.08093286e-01 -6.84554935e-01 7.39811838e-01 3.59193861e-01
-3.24045718e-01 5.70542395e-01 2.88762361e-01 -8.28082919e-01
3.75661343e-01 -4.92517412e-01 -9.29825366e-01 1.56508118e-01
2.11085156e-01 -3.92258674e-01 -1.12974346e-01 -1.24956958e-01
7.04246283e-01 9.84468311e-02 -1.13397527e+00 5.65751374e-01
1.28625083e+00 3.17726672e-01 7.98638165e-02 3.98819774e-01
-4.43741262e-01 1.27411330e+00 -9.07602847e-01 -1.25236773e+00
2.18242794e-01 -4.54689473e-01 -9.27363217e-01 -2.27702409e-01
4.66875166e-01 -2.58744329e-01 1.34842312e-02 1.29010403e+00
2.00729266e-01 -2.86357194e-01 3.52700680e-01 -1.32436705e+00
-1.00835276e+00 1.09124947e+00 -7.49270678e-01 -1.76390812e-01
5.67402899e-01 3.38524617e-02 7.92610168e-01 -7.43155062e-01
8.52112949e-01 1.08725512e+00 4.30407524e-01 8.14764559e-01
-1.06205690e+00 -9.08599913e-01 -9.86526757e-02 -1.39987811e-01
-1.28695405e+00 -4.78390366e-01 5.29221237e-01 -7.55416870e-01
1.02569962e+00 3.18488747e-01 2.71856248e-01 1.45980942e+00
7.78538465e-01 8.01503003e-01 7.74171472e-01 -8.54210138e-01
2.39241377e-01 8.55811179e-01 -2.15319037e-01 6.51364386e-01
6.64292336e-01 -4.00267541e-01 -5.22692204e-01 -7.80125797e-01
2.70117044e-01 -1.93771496e-01 1.60308525e-01 -2.25777552e-01
-8.85263383e-01 1.19879007e+00 -4.82729338e-02 3.80469233e-01
6.46434277e-02 2.10946217e-01 5.85507095e-01 3.45183522e-01
5.64051926e-01 7.72102475e-01 -1.74430132e-01 -3.67519051e-01
-1.03166413e+00 3.27629969e-02 8.93337727e-01 1.25890243e+00
4.96519983e-01 5.86960725e-02 2.22877339e-01 7.72890091e-01
6.02030754e-01 4.15414304e-01 8.45866799e-01 -7.88844287e-01
7.58102655e-01 5.25662959e-01 6.69277981e-02 -1.12205374e+00
1.67070404e-01 -2.47171149e-01 -5.43185353e-01 -1.15001295e-02
4.51692417e-02 -2.01530561e-01 -3.22517842e-01 1.53696263e+00
-3.06274295e-01 -2.31756344e-01 -8.54204968e-02 -8.95186737e-02
4.00216281e-01 6.86574697e-01 4.50544536e-01 7.34479576e-02
7.84757853e-01 -5.96961498e-01 -6.20222569e-01 -7.02916443e-01
9.49601769e-01 -8.99993241e-01 1.19111145e+00 2.41170853e-01
-9.60593045e-01 -1.16562724e-01 -8.59599650e-01 1.16457999e-01
-2.89221644e-01 1.30832806e-01 5.93580723e-01 1.14684772e+00
-1.09574628e+00 3.81183356e-01 -6.39642060e-01 -3.47605288e-01
6.74421430e-01 2.83320644e-03 -4.31064248e-01 9.75223556e-02
-1.10800242e+00 8.67882013e-01 1.17597722e-01 -2.20455423e-01
-5.40976167e-01 -3.19298983e-01 -1.31612086e+00 1.35659948e-02
7.49771148e-02 -9.50991288e-02 1.33056831e+00 -1.05633473e+00
-1.25083745e+00 1.13951290e+00 -6.76252320e-02 -5.14147282e-01
5.03069162e-01 -3.66660386e-01 -5.06229937e-01 -2.85045236e-01
2.79387653e-01 2.17309996e-01 1.01890957e+00 -1.13468182e+00
-2.37712666e-01 1.68659568e-01 1.26513511e-01 -6.25176072e-01
-7.18157649e-01 5.19535482e-01 1.10516228e-01 -7.82784522e-01
-5.74565649e-01 -8.48823726e-01 -3.49728912e-01 -2.72932976e-01
-6.04643047e-01 -8.87808353e-02 3.97718608e-01 -8.38728547e-01
1.67460811e+00 -2.19090533e+00 -2.67964870e-01 3.63405734e-01
1.46324754e-01 5.41005969e-01 -3.18538547e-01 8.12886536e-01
8.17929506e-02 1.00965118e+00 -2.01762587e-01 -4.20657963e-01
1.83627337e-01 -4.36019719e-01 -8.19590926e-01 4.41484183e-01
1.49366975e-01 8.82515728e-01 -7.73984790e-01 -6.01968467e-01
-2.00112328e-01 1.60961315e-01 -7.84750998e-01 2.74591506e-01
-1.06997840e-01 5.78785501e-02 -3.89507622e-01 3.95017296e-01
6.32508814e-01 -2.67437935e-01 1.79252580e-01 6.94678307e-01
-9.19964463e-02 3.45989376e-01 -5.83491147e-01 1.71234000e+00
-7.78635561e-01 1.01389408e+00 -1.79352816e-02 -2.24947229e-01
8.51510525e-01 1.54673934e-01 -1.11242421e-01 -5.49612224e-01
3.34869069e-03 2.54529983e-01 6.92711323e-02 -7.41228521e-01
6.49263263e-01 -4.91186865e-02 -3.80283564e-01 1.22927070e+00
-1.21018954e-01 -3.03536475e-01 4.43884544e-02 7.35264242e-01
9.69950259e-01 -1.88100547e-01 4.28527117e-01 -9.75782722e-02
4.92710352e-01 -4.29973155e-01 1.76437020e-01 9.03196454e-01
7.16895089e-02 1.99994847e-01 5.76216638e-01 -1.34966627e-01
-8.29335034e-01 -6.20511353e-01 4.24180105e-02 1.04818273e+00
-3.31380904e-01 -9.22741473e-01 -1.15209711e+00 -8.80499005e-01
-9.57880765e-02 1.38170016e+00 -5.27168453e-01 -5.52686691e-01
-4.67919022e-01 -1.23910964e-01 9.18255210e-01 2.21909270e-01
1.01300359e-01 -1.25905979e+00 -6.82911396e-01 1.44115657e-01
-2.25744724e-01 -7.32132375e-01 -4.45585072e-01 -3.36762458e-01
-4.96073008e-01 -8.14346194e-01 -3.30877036e-01 -5.36300421e-01
1.06140721e+00 2.31924579e-02 9.44779575e-01 5.80523491e-01
-3.46915871e-01 4.22149062e-01 -4.57272679e-01 -8.64327073e-01
-1.36687267e+00 3.78328204e-01 -1.35962754e-01 -1.67133585e-01
4.18005884e-01 -2.85188138e-01 6.05722442e-02 -1.47788167e-01
-1.30922163e+00 3.36321890e-02 7.07143664e-01 4.36107486e-01
-2.86096126e-01 -3.92229706e-01 7.32149124e-01 -1.41613758e+00
1.20222449e+00 -7.33086407e-01 -5.27681470e-01 5.36425471e-01
-6.03055179e-01 3.29241417e-02 7.11404204e-01 -3.52958679e-01
-1.20645225e+00 -2.88919955e-01 -3.50137860e-01 2.73741841e-01
-1.76355690e-02 6.82643771e-01 -2.07693279e-02 -2.99753815e-01
8.89781475e-01 2.12641060e-01 -3.03969942e-02 -3.82947147e-01
4.29519057e-01 1.13895774e+00 -8.71222615e-02 -7.91597128e-01
1.21216476e+00 1.64278850e-01 -6.70197189e-01 -5.94701886e-01
-4.70007658e-01 2.48789728e-01 -2.97111332e-01 -1.37306437e-01
4.46139216e-01 -7.38707781e-01 -2.99034715e-01 3.04056704e-01
-1.39016771e+00 -3.25211108e-01 -1.45504415e-01 2.33775586e-01
-3.00824106e-01 7.11009204e-01 -7.46774673e-01 -8.36893320e-01
-4.03039217e-01 -1.03591430e+00 5.97639143e-01 2.04173565e-01
-7.83231258e-01 -8.33358943e-01 1.60183515e-02 3.60030174e-01
8.11019182e-01 1.21297777e-01 1.30913746e+00 -8.21811259e-01
-4.79833275e-01 -6.40455782e-01 9.55005437e-02 4.21949595e-01
1.61263019e-01 4.21929896e-01 -8.87136221e-01 -5.07105887e-01
6.37470558e-02 -6.07382536e-01 2.19872862e-01 -2.14240536e-01
1.10072994e+00 -7.44571805e-01 -1.85173869e-01 1.84189439e-01
1.17280471e+00 1.37122273e-01 5.98427773e-01 3.52577120e-01
3.34825248e-01 7.22787380e-01 4.32255358e-01 7.07484186e-01
1.92150354e-01 -7.32130706e-02 4.04092520e-01 3.16221356e-01
4.08809036e-01 -7.73252368e-01 6.47501469e-01 1.03503442e+00
4.14568663e-01 -2.84344047e-01 -1.12585461e+00 4.88551855e-01
-1.50021493e+00 -9.76357639e-01 3.15332189e-02 2.09267545e+00
9.74818528e-01 3.52932036e-01 -3.80850285e-01 -4.40447628e-01
6.19807065e-01 2.09186673e-01 -2.70855427e-01 -8.84199560e-01
9.99522582e-02 1.59511477e-01 3.96699011e-01 3.10316265e-01
-6.26610816e-01 6.91563427e-01 7.12690830e+00 9.53787506e-01
-1.09024429e+00 2.50588834e-01 6.75570607e-01 5.21241724e-02
-9.38412130e-01 2.87925631e-01 -4.10345376e-01 5.91898859e-01
1.31011868e+00 -8.11327219e-01 3.80174369e-01 1.31341410e+00
1.41558349e-02 -1.11408316e-01 -1.11061931e+00 7.82928646e-01
3.01400512e-01 -1.45141435e+00 2.12331235e-01 1.01135761e-01
7.57273197e-01 -7.51972869e-02 3.66158664e-01 4.83870000e-01
3.69542420e-01 -1.20389664e+00 1.13817132e+00 4.64293212e-02
9.29418325e-01 -7.47308195e-01 6.41193926e-01 8.41796517e-01
-6.42856836e-01 -1.85171202e-01 -5.96043527e-01 -2.60032892e-01
-1.10775009e-02 3.41261089e-01 -6.28899992e-01 7.43135363e-02
2.07667664e-01 4.28722650e-01 -1.09117174e+00 8.49488080e-01
-4.62126493e-01 7.62713671e-01 3.09364378e-01 -2.87877798e-01
-5.06907031e-02 -8.72949511e-03 1.44388378e-01 1.45149899e+00
4.17481840e-01 -4.01216209e-01 -2.41344512e-01 1.26202214e+00
-5.14526367e-01 3.03543031e-01 -9.76966381e-01 -7.02926755e-01
6.23249531e-01 1.09390461e+00 -6.04584754e-01 -8.42925608e-02
-8.20264578e-01 7.63485193e-01 4.16379571e-01 2.59165734e-01
-6.78246975e-01 -8.41759503e-01 5.19710004e-01 2.95323998e-01
-3.11465323e-01 -2.03050151e-01 -2.10922152e-01 -1.41935360e+00
2.51791120e-01 -1.23553586e+00 1.01470072e-02 -6.00237012e-01
-1.06048155e+00 7.52714515e-01 -1.14321560e-01 -1.19774950e+00
-5.85893989e-01 -2.54267573e-01 -7.43954360e-01 6.69689059e-01
-9.63837206e-01 -1.03342295e+00 2.00259686e-01 8.88392515e-03
3.70190501e-01 -4.19961423e-01 8.97218227e-01 2.87489951e-01
-4.32995111e-01 9.39352274e-01 1.76467985e-01 4.45403129e-01
8.52890491e-01 -7.46156931e-01 1.00139964e+00 1.14501190e+00
3.03588569e-01 1.48293209e+00 7.20773637e-01 -7.84341693e-01
-1.16299927e+00 -1.03860223e+00 1.15716040e+00 -8.02530944e-01
6.58147037e-01 -9.12426770e-01 -8.72065723e-01 1.12877202e+00
6.36005521e-01 -4.10300463e-01 1.23376310e+00 -3.56711715e-01
-5.85663021e-01 3.60585928e-01 -1.33502221e+00 5.92548907e-01
6.17778361e-01 -1.05052674e+00 -4.84553605e-01 3.34600508e-01
6.48507595e-01 -2.42959470e-01 -5.85036099e-01 -3.52099866e-01
4.08722371e-01 -9.39106107e-01 3.98279637e-01 -8.88804138e-01
8.03106844e-01 2.97606498e-01 1.54149130e-01 -1.25130403e+00
5.23647573e-03 -1.07787085e+00 7.77484775e-02 1.56574380e+00
6.00416005e-01 -8.11844826e-01 4.23357934e-01 1.40526783e+00
2.72784472e-01 -3.72028291e-01 -6.98665798e-01 -7.47989357e-01
4.28867102e-01 -5.36887765e-01 5.51814377e-01 1.12371218e+00
7.08693981e-01 -4.55727652e-02 -6.15285695e-01 -2.94306636e-01
3.77601922e-01 -2.15330552e-02 8.51336479e-01 -1.05398965e+00
-1.69974893e-01 -2.80060530e-01 -1.83087289e-01 -7.39417970e-01
8.11905265e-01 -1.16054213e+00 -1.25077352e-01 -1.28347826e+00
4.38045114e-01 -1.57527387e-01 2.22014207e-02 3.62235397e-01
1.75538957e-02 -2.93157339e-01 4.12101924e-01 7.40038902e-02
-6.00101352e-01 2.92214006e-01 4.29919124e-01 3.42374318e-03
1.03755347e-01 1.33863166e-02 -1.14213395e+00 8.24387133e-01
9.35527444e-01 -1.00810373e+00 -4.49817240e-01 -4.40501541e-01
8.31186295e-01 -1.27671167e-01 -1.22142835e-02 -6.80093825e-01
1.10908493e-01 -4.07638431e-01 -9.29938108e-02 1.63938746e-01
-2.06016421e-01 -7.26277053e-01 9.88244042e-02 6.46579504e-01
-8.72535825e-01 4.74877283e-02 3.08974087e-01 4.98293370e-01
-1.83691785e-01 -8.92103791e-01 5.38438320e-01 -3.23850840e-01
-3.27914268e-01 -1.92393996e-02 -1.04248309e+00 2.54455298e-01
1.08206129e+00 1.23739563e-01 -3.94939691e-01 -6.96735919e-01
-4.05115634e-02 -1.47334799e-01 8.37791324e-01 7.68584192e-01
5.57115316e-01 -1.05270350e+00 -5.71392059e-01 3.76913637e-01
3.66771221e-01 -5.27010739e-01 6.08638935e-02 2.81925023e-01
-4.60752308e-01 4.22347844e-01 -2.13492706e-01 1.12229213e-01
-1.09106088e+00 5.15201867e-01 -1.14551857e-01 -9.46952496e-03
-3.44033271e-01 7.98257232e-01 3.15535180e-02 -3.47940415e-01
1.05234146e-01 -1.46249279e-01 1.27535120e-01 -2.79907167e-01
8.04510891e-01 1.79384768e-01 -8.94794166e-02 -4.98370111e-01
-3.93751204e-01 -1.42713375e-02 -3.09636921e-01 -3.94307673e-02
1.27232873e+00 -4.29103151e-02 -6.44299090e-01 3.53969932e-01
1.15635884e+00 9.22092915e-01 -7.08842874e-01 -1.28246590e-01
4.30194169e-01 -7.35240698e-01 -7.91559756e-01 -5.05528808e-01
-8.14622998e-01 9.77100909e-01 1.64051335e-02 4.48995978e-01
4.27088082e-01 2.51340587e-02 9.22104120e-01 4.80117351e-01
7.56143630e-01 -1.28319621e+00 4.15592827e-02 3.30879688e-01
9.43281293e-01 -1.08573937e+00 -1.06238157e-01 -1.73844278e-01
-8.04345429e-01 1.24355340e+00 6.83208406e-01 2.53889829e-01
5.55368423e-01 4.06600118e-01 2.35887527e-01 1.89817965e-01
-6.45340621e-01 8.08012664e-01 -1.35103032e-01 6.35721445e-01
9.61031497e-01 -1.14746727e-01 -4.75848645e-01 8.68353784e-01
-2.66074866e-01 -3.10388170e-02 1.32995737e+00 1.27821612e+00
-3.19585383e-01 -1.41960967e+00 -3.29106569e-01 7.16783881e-01
-8.03356111e-01 -3.36478114e-01 -6.91372931e-01 4.85057950e-01
-2.07860783e-01 1.16860199e+00 -4.66474682e-01 -5.05059540e-01
-4.69526425e-02 2.64511287e-01 -1.54125988e-02 -1.10394382e+00
-8.47155869e-01 -5.47202170e-01 6.05266578e-02 -4.53994632e-01
7.01734573e-02 -5.06861985e-01 -1.20511699e+00 -4.50135499e-01
-2.33210623e-01 3.39419752e-01 7.33013391e-01 7.33559132e-01
7.17151999e-01 -9.29584056e-02 4.66864586e-01 -3.75047058e-01
-9.25294638e-01 -8.89655828e-01 -4.26870465e-01 5.12793124e-01
1.82605967e-01 -5.93690239e-02 -4.57140803e-01 1.09931424e-01] | [7.900961875915527, 7.7956061363220215] |
027e0c33-7b84-4b73-aef6-dc163d4052dd | clinical-outcome-prediction-from-admission | 2102.04110 | null | https://arxiv.org/abs/2102.04110v1 | https://arxiv.org/pdf/2102.04110v1.pdf | Clinical Outcome Prediction from Admission Notes using Self-Supervised Knowledge Integration | Outcome prediction from clinical text can prevent doctors from overlooking possible risks and help hospitals to plan capacities. We simulate patients at admission time, when decision support can be especially valuable, and contribute a novel admission to discharge task with four common outcome prediction targets: Diagnoses at discharge, procedures performed, in-hospital mortality and length-of-stay prediction. The ideal system should infer outcomes based on symptoms, pre-conditions and risk factors of a patient. We evaluate the effectiveness of language models to handle this scenario and propose clinical outcome pre-training to integrate knowledge about patient outcomes from multiple public sources. We further present a simple method to incorporate ICD code hierarchy into the models. We show that our approach improves performance on the outcome tasks against several baselines. A detailed analysis reveals further strengths of the model, including transferability, but also weaknesses such as handling of vital values and inconsistencies in the underlying data. | ['Alexander Löser', 'Felix A. Gers', 'Klemens Budde', 'Manuel Mayrdorfer', 'Jens-Michalis Papaioannou', 'Betty van Aken'] | 2021-02-08 | null | https://aclanthology.org/2021.eacl-main.75 | https://aclanthology.org/2021.eacl-main.75.pdf | eacl-2021-2 | ['medical-procedure', 'length-of-stay-prediction'] | ['medical', 'medical'] | [-1.43993691e-01 3.51248890e-01 -5.47242284e-01 -5.70247054e-01
-1.25287533e+00 -4.17992532e-01 1.79070830e-01 1.29178047e+00
-3.86170536e-01 9.12835240e-01 1.23504853e+00 -7.84963131e-01
-6.01178944e-01 -7.89435267e-01 -6.67160749e-02 -3.25297892e-01
-4.38105136e-01 1.07056832e+00 -3.47282529e-01 6.91255108e-02
-1.35436589e-02 5.73758364e-01 -8.82731140e-01 9.06495631e-01
7.36398339e-01 7.70793855e-01 -3.83156121e-01 7.26153970e-01
2.42633328e-01 1.48524487e+00 -2.94884205e-01 -5.05109966e-01
7.53161684e-02 -2.61191517e-01 -1.06809318e+00 -3.66578072e-01
-3.19356829e-01 -6.51984870e-01 -1.61607668e-01 2.21339807e-01
8.23292792e-01 -3.93701673e-01 1.11814535e+00 -9.41837966e-01
-3.36624086e-01 7.54697859e-01 1.83634967e-01 2.05865726e-01
5.54906249e-01 2.48727381e-01 8.88047338e-01 -4.37830895e-01
5.72608411e-01 9.07428145e-01 1.12997830e+00 5.34776926e-01
-1.28859544e+00 -3.28938842e-01 -2.65311986e-01 5.14735887e-03
-1.08958185e+00 -4.87093180e-01 -1.74046904e-02 -6.93746090e-01
1.07910454e+00 6.06853485e-01 4.06443179e-01 1.01943982e+00
2.67135829e-01 2.94106841e-01 7.87869513e-01 -2.01454878e-01
5.80551922e-02 1.47441119e-01 6.96750879e-02 7.20237792e-01
3.01906139e-01 1.57352947e-02 -8.54966268e-02 -8.56226802e-01
2.94683009e-01 5.49535871e-01 -1.80938959e-01 1.09162107e-01
-1.58640015e+00 5.92136383e-01 1.83756366e-01 -8.05785060e-02
-5.22801578e-01 -3.05221260e-01 6.48422301e-01 1.80517107e-01
4.03944314e-01 4.41694856e-01 -1.11004412e+00 -1.66046992e-01
-1.01046968e+00 3.16383511e-01 1.12202096e+00 6.42502308e-01
1.07598469e-01 -5.28295755e-01 -6.95224404e-01 7.20775306e-01
-1.47554409e-02 4.85921144e-01 5.59340239e-01 -9.80957866e-01
5.82044423e-01 7.44071841e-01 2.50591934e-01 -6.18937969e-01
-9.77638543e-01 -3.21696132e-01 -9.71496403e-01 -3.48322839e-01
4.66674984e-01 -5.41897178e-01 -7.65462816e-01 1.36170793e+00
-4.97599505e-02 8.16511177e-03 3.80874217e-01 4.60809946e-01
1.01970339e+00 2.14728549e-01 5.63309014e-01 -3.53780001e-01
1.27986825e+00 -4.29826796e-01 -4.38675016e-01 -3.03763952e-02
1.72510242e+00 -3.50935370e-01 4.74034518e-01 1.18270054e-01
-1.18746555e+00 2.05334164e-02 -2.42187968e-03 1.15964018e-01
-1.19815066e-01 2.20020190e-01 4.31839496e-01 3.28868598e-01
-8.89325380e-01 7.80392706e-01 -8.07165146e-01 -5.51659405e-01
4.93757725e-01 4.04515058e-01 -5.23671985e-01 7.23849759e-02
-1.12102115e+00 1.15119052e+00 5.29271901e-01 -5.67330182e-01
-6.11675858e-01 -1.11177599e+00 -9.28983629e-01 3.99991184e-01
2.07303077e-01 -1.18922615e+00 1.19965112e+00 -4.80256379e-01
-6.99809611e-01 1.17341018e+00 -1.66216284e-01 -5.45153141e-01
5.36464095e-01 -2.59573739e-02 -4.46305245e-01 1.74332038e-01
1.18279114e-01 3.68895590e-01 -2.36232772e-01 -8.31882596e-01
-8.08254123e-01 -3.14410478e-01 -2.79100269e-01 8.74966010e-02
-1.86439455e-01 3.11332464e-01 1.65806085e-01 -7.77229071e-01
-2.15717077e-01 -7.73126781e-01 -6.54383540e-01 -2.47930154e-01
-3.49863619e-01 1.44634368e-02 -2.47366697e-01 -8.83508682e-01
1.66938233e+00 -1.95579088e+00 -2.25821093e-01 2.02612400e-01
3.28595757e-01 -3.08463182e-02 1.89539641e-01 7.32785225e-01
-4.12996083e-01 5.87669313e-01 -1.78872749e-01 -2.93575227e-01
-3.51996690e-01 2.98204541e-01 -2.87255079e-01 2.49089673e-01
3.66199613e-01 1.11289108e+00 -7.58935094e-01 -9.95870948e-01
1.32459283e-01 1.87895596e-01 -1.12231231e+00 4.81453568e-01
2.41234571e-01 6.19524956e-01 -4.40150142e-01 6.43311977e-01
-7.32072303e-03 -6.77768409e-01 3.11025202e-01 2.90381998e-01
-3.89036052e-02 9.24237013e-01 -6.97211087e-01 1.17246783e+00
-3.17237735e-01 -1.98763266e-01 -2.09645480e-01 -1.01583004e+00
3.96431953e-01 8.53868544e-01 9.73257244e-01 -1.75288677e-01
6.20847158e-02 1.06239937e-01 2.68576860e-01 -9.03063715e-01
-3.16403180e-01 -4.98598218e-01 -2.21969366e-01 4.57475960e-01
-2.84361362e-01 9.35015306e-02 -1.72704175e-01 2.96639174e-01
1.32562160e+00 -3.24815720e-01 9.46811378e-01 -3.21740568e-01
3.15330148e-01 3.76027554e-01 8.02574873e-01 8.69011700e-01
-8.68493691e-02 8.39094400e-01 7.83650458e-01 -7.56365359e-01
-8.55040550e-01 -9.02923763e-01 -5.06974041e-01 7.50968456e-01
-6.18277311e-01 -5.50128222e-01 -2.07961783e-01 -7.66249657e-01
2.31653124e-01 7.95937657e-01 -7.12045491e-01 -2.62405366e-01
-3.38802606e-01 -1.06271958e+00 5.48264563e-01 7.23516583e-01
-2.26820692e-01 -1.04694104e+00 -7.37634301e-01 4.03534561e-01
-5.83687365e-01 -8.62091720e-01 3.81904244e-02 2.27756768e-01
-1.19278336e+00 -9.95344639e-01 -7.73364842e-01 -4.67219472e-01
2.80023277e-01 -7.21561491e-01 1.52328420e+00 4.33942676e-01
-3.14120978e-01 3.09114337e-01 -2.21669167e-01 -6.43981040e-01
-7.96572506e-01 1.96392357e-01 -1.10979236e-04 -3.83342355e-01
4.48495865e-01 -4.11266804e-01 -8.37653756e-01 -1.35643423e-01
-6.75663114e-01 1.76934466e-01 4.90432620e-01 7.89965034e-01
2.46690512e-01 -5.48587084e-01 8.01719725e-01 -1.26126111e+00
4.90297973e-01 -9.91211593e-01 1.57455280e-01 3.17595959e-01
-8.77837121e-01 9.49494243e-02 6.34814918e-01 1.32811278e-01
-7.31848955e-01 -7.32026994e-02 -4.58371818e-01 2.50984818e-01
-7.64083862e-01 7.66186476e-01 1.58482581e-01 7.53920555e-01
5.71031511e-01 -1.78648382e-01 2.66726897e-03 -5.55423498e-01
-2.68841237e-01 7.55110741e-01 2.86025614e-01 -5.28042376e-01
1.63937896e-01 5.22100985e-01 8.83114561e-02 -1.62494984e-02
-1.14165819e+00 -7.81602025e-01 -6.72521234e-01 3.86232495e-01
7.17653394e-01 -1.12387025e+00 -1.01336718e+00 -1.68798253e-01
-9.85614777e-01 -3.18653882e-01 -4.07495648e-01 8.35362554e-01
-5.27908862e-01 1.32536441e-01 -8.33169341e-01 -5.88390291e-01
-3.84090096e-01 -9.89669025e-01 9.43386078e-01 -3.82399887e-01
-6.97628796e-01 -1.43172085e+00 1.04335636e-01 1.20177001e-01
3.28144848e-01 5.40498018e-01 1.64007115e+00 -1.18464637e+00
5.39564528e-02 -3.20236117e-01 -3.05331618e-01 -1.26600862e-01
1.93627283e-01 -2.01322973e-01 -6.42246008e-01 -6.47394881e-02
-2.36035720e-01 -3.26742113e-01 7.26615429e-01 4.20727879e-01
1.44068694e+00 -5.49410939e-01 -6.40877426e-01 7.83435941e-01
1.36496437e+00 -1.85418203e-02 5.02772272e-01 2.83095062e-01
4.92690802e-01 8.44050288e-01 1.04433306e-01 9.81582224e-01
9.17951226e-01 2.95937538e-01 1.94842219e-01 -4.19992357e-01
2.56192833e-01 3.13429981e-02 -8.66463631e-02 6.68625236e-01
-3.63677144e-01 -1.48303956e-01 -1.71276653e+00 8.09985340e-01
-1.68505406e+00 -8.11237156e-01 -4.08797532e-01 2.14069247e+00
1.12334847e+00 1.44603267e-01 2.25462407e-01 5.80396280e-02
-4.60604532e-03 -5.36221147e-01 -2.56959915e-01 -6.29071236e-01
1.10510565e-01 2.40893960e-01 7.39357531e-01 3.54434699e-01
-8.57070446e-01 3.49074423e-01 7.44328928e+00 4.18748632e-02
-7.66952455e-01 1.16831340e-01 1.24136949e+00 -2.46992797e-01
-2.60927916e-01 -4.77518328e-02 -5.13677716e-01 4.53942031e-01
1.37330806e+00 -9.91291255e-02 -2.45163679e-01 6.39689684e-01
4.91016120e-01 1.13770418e-01 -1.48093796e+00 6.04040205e-01
-1.80897385e-01 -1.37906659e+00 1.23287350e-01 6.64098188e-02
4.97083545e-01 1.14515916e-01 -2.98470557e-01 3.62449467e-01
6.95280194e-01 -1.27211845e+00 3.01393539e-01 7.72831202e-01
1.16736722e+00 -4.31987435e-01 1.07139802e+00 5.22846520e-01
-6.57670319e-01 -5.77974379e-01 5.00762910e-02 -4.28718388e-01
2.51267403e-01 5.69482505e-01 -1.19823647e+00 5.46456873e-01
6.55199766e-01 8.81699681e-01 -5.54208398e-01 1.04218256e+00
1.21189304e-01 1.01595044e+00 -1.00623265e-01 6.04626715e-01
8.19763243e-02 2.29516000e-01 9.19953138e-02 1.55898798e+00
4.83729988e-01 7.49012887e-01 3.35020125e-01 4.03363675e-01
6.09539496e-03 4.04708117e-01 -7.95880556e-01 3.35768580e-01
-2.06649638e-04 7.97608078e-01 -3.54642868e-01 -8.73644054e-01
-5.31673789e-01 6.30796671e-01 3.34469169e-01 2.32388243e-01
-5.86309969e-01 1.42670527e-01 7.38835454e-01 6.85677111e-01
-2.19374701e-01 2.98051625e-01 -7.68643856e-01 -1.37349868e+00
-3.86779368e-01 -9.40370500e-01 9.93958890e-01 -6.40339255e-01
-1.11803949e+00 3.08476865e-01 7.86563940e-03 -1.11509323e+00
-5.34309149e-01 -5.81814945e-01 -3.05182397e-01 9.78967071e-01
-1.67938888e+00 -8.38299990e-01 2.63543949e-02 4.75857317e-01
1.41183630e-01 -5.95681369e-02 1.45799839e+00 4.76517051e-01
-4.73036408e-01 5.20648539e-01 -2.04960015e-02 5.74012458e-01
1.06563509e+00 -1.09032524e+00 1.53135872e-02 1.53586745e-01
-4.32541370e-01 5.56128383e-01 3.73339266e-01 -7.34328151e-01
-5.93530178e-01 -1.18571138e+00 1.75902271e+00 -1.05343962e+00
4.85890090e-01 1.28731027e-01 -8.77136469e-01 8.06743503e-01
-2.50268042e-01 -1.28553435e-01 1.23384535e+00 2.85925597e-01
-2.15294454e-02 2.17767626e-01 -1.02903187e+00 4.99355853e-01
1.09602284e+00 -4.06478912e-01 -8.79122317e-01 4.61527616e-01
4.36662465e-01 -2.09999338e-01 -1.24337292e+00 8.06333065e-01
5.78172743e-01 -8.63956153e-01 1.07380462e+00 -1.54624546e+00
1.00702035e+00 5.36302447e-01 1.28727704e-01 -1.21017420e+00
-6.10948920e-01 -3.05645019e-01 1.92546234e-01 8.33238482e-01
8.12007427e-01 -6.50335491e-01 4.54863250e-01 1.14127207e+00
1.61011338e-01 -8.63162160e-01 -5.50681174e-01 -1.75746426e-01
3.59155476e-01 -3.11681509e-01 7.34649003e-01 1.18585122e+00
3.46826643e-01 7.60399699e-02 -1.58173233e-01 3.65843236e-01
2.31564477e-01 1.72937140e-01 3.22537988e-01 -1.59915197e+00
-2.36434981e-01 -2.68781424e-01 -2.52907544e-01 -3.27551782e-01
3.64130177e-02 -1.16900790e+00 -4.93433923e-01 -2.26490664e+00
5.70828259e-01 -8.73230338e-01 -6.99019015e-01 8.86260986e-01
-5.18480003e-01 -3.72526012e-02 -1.01167656e-01 4.55192536e-01
-4.33827192e-01 -5.44269606e-02 6.97545230e-01 2.77931452e-01
-1.68468103e-01 1.06812239e-01 -8.65969837e-01 1.05716217e+00
8.86964083e-01 -1.07156324e+00 -6.97613275e-03 -5.25624216e-01
4.00616467e-01 8.86720419e-01 3.70599389e-01 -7.87414074e-01
-2.76984647e-02 -3.82169187e-01 4.41918522e-01 -2.12962687e-01
-9.02644545e-02 -1.01860666e+00 -3.02608982e-02 9.57767069e-01
-7.93745518e-01 5.14346242e-01 2.30861142e-01 3.11927378e-01
-1.17819563e-01 -3.44116386e-04 3.52504998e-01 -4.06538218e-01
-8.83820057e-02 1.74989000e-01 -7.09265530e-01 4.51235265e-01
9.35526907e-01 5.46272062e-02 -7.41895437e-02 -4.39894110e-01
-1.18194389e+00 4.68982577e-01 3.14449757e-01 5.79633415e-02
3.19906116e-01 -9.41023529e-01 -1.29472780e+00 5.64390793e-02
3.64860535e-01 -2.67864555e-01 4.22601961e-02 1.29763961e+00
-7.55808175e-01 5.56229770e-01 -1.62603389e-02 -2.70471692e-01
-1.25574684e+00 7.63349831e-01 3.52476925e-01 -7.62159646e-01
-7.76667237e-01 4.36936110e-01 2.64291078e-01 -3.75776857e-01
3.65182936e-01 -7.37630248e-01 -3.96321684e-01 5.61431348e-02
5.00382781e-01 1.80378661e-01 1.40107185e-01 -4.41995829e-01
-4.13445920e-01 9.96591225e-02 -5.36007099e-02 6.21997155e-02
1.53962398e+00 -1.04631774e-01 -2.12204427e-01 4.38448161e-01
1.04140127e+00 3.08308210e-02 -6.54435694e-01 -1.61834225e-01
2.88983971e-01 6.48350045e-02 -2.15793654e-01 -1.24330056e+00
-5.54392517e-01 9.38020587e-01 3.43833476e-01 2.13249736e-02
1.11050582e+00 5.67111447e-02 6.38003707e-01 3.25181276e-01
-4.18513678e-02 -6.26157224e-01 -5.40525913e-01 3.41221094e-01
4.46594656e-01 -1.42273557e+00 6.77702278e-02 -2.44463846e-01
-9.32682931e-01 7.85001218e-01 1.31588891e-01 2.18535453e-01
9.61273551e-01 4.41257030e-01 4.21708256e-01 -1.12996601e-01
-1.30878711e+00 -1.24252111e-01 2.27552786e-01 2.45447233e-01
7.56298780e-01 4.53307897e-01 -2.86241382e-01 8.39366794e-01
-4.44439836e-02 3.90602320e-01 5.41862845e-01 7.56396711e-01
-2.55388790e-03 -1.09713936e+00 -2.83021688e-01 1.09930420e+00
-1.40518689e+00 -8.31432223e-01 -2.89977372e-01 4.66610134e-01
4.14079279e-01 7.59401917e-01 4.27657403e-02 1.11037366e-01
5.05582988e-01 6.08292639e-01 -9.61329192e-02 -9.90084767e-01
-1.07404542e+00 -1.94362909e-01 2.52586693e-01 -3.98447782e-01
-2.98269242e-01 -7.63093889e-01 -1.39579272e+00 -1.19095609e-01
2.40021124e-01 2.91415185e-01 1.51934519e-01 8.94676566e-01
8.44675303e-01 6.91888690e-01 1.35661453e-01 -2.20503006e-02
-7.13702977e-01 -7.78969944e-01 -2.51244992e-01 7.76295304e-01
6.15277171e-01 -2.30351567e-01 -2.88340420e-01 4.77118403e-01] | [7.994041919708252, 6.492499351501465] |
63c46364-34f3-4d2b-a290-e8b7bfcd4d27 | whitenet-phishing-website-detection-by-visual | 1909.00300 | null | https://arxiv.org/abs/1909.00300v4 | https://arxiv.org/pdf/1909.00300v4.pdf | VisualPhishNet: Zero-Day Phishing Website Detection by Visual Similarity | Phishing websites are still a major threat in today's Internet ecosystem. Despite numerous previous efforts, similarity-based detection methods do not offer sufficient protection for the trusted websites - in particular against unseen phishing pages. This paper contributes VisualPhishNet, a new similarity-based phishing detection framework, based on a triplet Convolutional Neural Network (CNN). VisualPhishNet learns profiles for websites in order to detect phishing websites by a similarity metric that can generalize to pages with new visual appearances. We furthermore present VisualPhish, the largest dataset to date that facilitates visual phishing detection in an ecologically valid manner. We show that our method outperforms previous visual similarity phishing detection approaches by a large margin while being robust against a range of evasion attacks. | ['Sahar Abdelnabi', 'Katharina Krombholz', 'Mario Fritz'] | 2019-09-01 | null | null | null | null | ['phishing-website-detection'] | ['adversarial'] | [ 7.92184025e-02 -2.35192925e-01 3.29225231e-03 -8.84625763e-02
-3.01551640e-01 -1.36372447e+00 9.59044576e-01 1.47529751e-01
-7.06296787e-02 1.27548158e-01 -1.54534414e-01 -8.04701447e-01
1.89061239e-01 -7.33638585e-01 -6.13103211e-01 -3.13749641e-01
-5.84245138e-02 1.18493997e-01 4.41300273e-01 -3.53531480e-01
7.82483757e-01 2.84204990e-01 -1.06070042e+00 4.60000187e-01
7.63260126e-01 8.72828782e-01 -2.39674270e-01 7.37643421e-01
3.78820896e-01 1.40716046e-01 -6.26080036e-01 -8.46092045e-01
9.23871934e-01 4.34409045e-02 -5.92024922e-01 -5.52422762e-01
1.22805643e+00 -6.27246261e-01 -6.62607431e-01 1.47418618e+00
2.83801675e-01 -3.56990755e-01 7.63900042e-01 -1.40795529e+00
-1.61934340e+00 1.06428880e-02 -6.06215239e-01 5.00466704e-01
4.74873692e-01 5.62978327e-01 1.21056211e+00 -7.04934955e-01
8.69238794e-01 1.29156315e+00 1.03690386e+00 3.17778945e-01
-1.36544240e+00 -7.12412655e-01 -3.66805762e-01 4.35029179e-01
-1.03329527e+00 1.44081321e-02 5.39187074e-01 -5.17616749e-01
5.88994741e-01 1.40421569e-01 6.07259810e-01 1.79583049e+00
4.15499330e-01 6.10853076e-01 1.26521444e+00 1.99113768e-02
1.96878344e-01 3.92616451e-01 4.65892524e-01 6.10191524e-01
9.60205078e-01 4.84581262e-01 -3.54471765e-02 -6.71135962e-01
6.46583557e-01 1.12058677e-01 7.44201662e-03 -7.77249515e-01
-1.06393671e+00 1.36613321e+00 9.99486327e-01 -1.96721461e-02
-1.38648480e-01 1.10622592e-01 8.87893140e-01 4.67306882e-01
2.08640397e-01 7.46016085e-01 -3.02820325e-01 4.42547858e-01
-5.36131024e-01 1.91257924e-01 9.84676003e-01 5.07164538e-01
7.10436881e-01 -1.49552271e-01 1.35845020e-01 3.81114274e-01
4.95587319e-01 7.20611215e-01 1.42377570e-01 -7.04345405e-01
3.03514898e-01 5.65025508e-01 1.50694609e-01 -1.43952167e+00
3.01245395e-02 3.76108252e-02 -3.50654602e-01 9.12857771e-01
5.30853748e-01 3.42543200e-02 -8.05307925e-01 1.29007578e+00
2.68818259e-01 9.04179215e-02 -4.43322718e-01 6.58865213e-01
8.51702631e-01 5.24022341e-01 2.33560726e-01 6.74714148e-01
1.78409243e+00 -7.94262886e-01 -5.19424751e-02 -4.87590134e-01
5.43850482e-01 -5.90994596e-01 1.44496632e+00 3.61398190e-01
-4.99355137e-01 5.08823059e-02 -1.38020694e+00 1.13227010e-01
-1.12850142e+00 -5.52802503e-01 7.01301336e-01 1.44865012e+00
-1.36296344e+00 4.60243136e-01 3.65919992e-02 -9.07136023e-01
7.90509462e-01 1.18021220e-01 -5.84156990e-01 -1.34532601e-01
-1.49370205e+00 1.19061756e+00 4.18809652e-01 -5.26677430e-01
-1.20007181e+00 -5.80604553e-01 -7.92286396e-01 3.98093700e-01
3.68568569e-01 -5.25939405e-01 1.09075403e+00 -1.15450895e+00
-7.20498025e-01 1.08442247e+00 5.25021195e-01 -4.16327626e-01
3.43712479e-01 1.26752138e-01 -6.01105869e-01 3.06595474e-01
1.59330547e-01 3.59590381e-01 1.00803113e+00 -1.49380934e+00
-2.12327957e-01 -4.93138194e-01 8.70133340e-02 -1.16191052e-01
-6.66090786e-01 1.69905379e-01 3.02014589e-01 -3.69336009e-01
-5.19262791e-01 -9.48123634e-01 -7.10801631e-02 2.04862192e-01
-5.81468999e-01 -1.79297417e-01 1.19881034e+00 -1.09793496e+00
7.31305718e-01 -1.72502601e+00 -8.05042446e-01 4.87397254e-01
8.13695967e-01 1.19443274e+00 -4.81907368e-01 5.21705747e-01
-6.42532110e-02 6.17485881e-01 2.57351071e-01 2.31434733e-01
2.03465551e-01 -3.88765156e-01 -8.18112567e-02 6.90405548e-01
-1.49038285e-01 1.29287064e+00 -1.03366411e+00 -2.31738210e-01
2.46162593e-01 5.74822664e-01 -2.31634885e-01 -1.00950919e-01
9.23896506e-02 -3.17824006e-01 -2.83764929e-01 9.16209340e-01
9.92408931e-01 -6.90994561e-01 2.95313805e-01 1.16562992e-01
1.52929008e-01 2.79440492e-01 -3.99317682e-01 9.69117224e-01
-2.66577333e-01 1.03915501e+00 8.40910971e-02 -3.95835221e-01
8.53438199e-01 3.40847522e-02 -5.52118756e-02 -8.20391238e-01
2.23804504e-01 1.79736868e-01 -5.99467680e-02 -4.27364498e-01
5.03061891e-01 4.44130123e-01 4.05270867e-02 6.26023471e-01
-4.79046293e-02 7.96442747e-01 7.56251439e-02 8.07770133e-01
1.28384793e+00 -3.97444665e-01 6.26784563e-01 -3.53402525e-01
3.50489765e-01 1.22387499e-01 5.44513389e-02 9.85748470e-01
-9.43819582e-01 4.61622387e-01 5.45029640e-01 -8.81929994e-01
-1.74478006e+00 -1.47754014e+00 -1.42197218e-02 1.19796062e+00
3.85827392e-01 -1.77928925e-01 -8.56452405e-01 -1.53045809e+00
4.69361365e-01 4.35748488e-01 -1.09836376e+00 -2.61001438e-01
-2.52524018e-01 -6.55498326e-01 7.96403110e-01 3.02359104e-01
5.92405379e-01 -1.21784186e+00 -3.97286005e-02 -2.31651977e-01
8.81494507e-02 -6.51637912e-01 -5.63203514e-01 -3.61547858e-01
-2.57116973e-01 -1.79558027e+00 -7.85261750e-01 -8.19097817e-01
4.95339841e-01 1.13223994e+00 1.02714634e+00 1.56799391e-01
-5.86304665e-01 2.06382796e-01 -1.26718402e-01 -8.35984126e-02
-6.78509831e-01 1.43697158e-01 -1.25382826e-01 -1.34623036e-01
1.02007449e+00 -2.63362229e-01 -1.02259696e+00 4.20697421e-01
-7.68601537e-01 -7.62289345e-01 6.50125861e-01 5.41797221e-01
-6.16668224e-01 -3.48641634e-01 5.93636692e-01 -9.60712552e-01
9.90460098e-01 -1.09123492e+00 -5.48146129e-01 4.41204339e-01
-1.03636038e+00 -4.17547464e-01 5.08668661e-01 -6.08293951e-01
-9.94659901e-01 -2.10913688e-01 3.76605541e-01 -9.26085711e-02
-3.88795882e-01 4.41782223e-03 1.51075507e-02 -9.14949536e-01
1.41405416e+00 2.01126352e-01 1.34544909e-01 -4.02873158e-01
5.72593689e-01 6.67967200e-01 3.31092358e-01 2.53364801e-01
1.26355946e+00 4.95458931e-01 -2.39472881e-01 -5.95060170e-01
-2.59133071e-01 -9.42976952e-01 -6.31873071e-01 -3.35394800e-01
7.15688765e-01 -4.68689620e-01 -1.51253748e+00 4.47169006e-01
-1.21407616e+00 2.00496018e-01 6.61438107e-01 -2.09474280e-01
1.73228800e-01 1.36534417e+00 -8.98342431e-01 -4.52720731e-01
-5.21453381e-01 -6.51492357e-01 4.91662383e-01 1.91539377e-01
-1.09875239e-01 -1.38637972e+00 5.86069822e-01 5.94970107e-01
6.70141459e-01 2.83639431e-01 5.64914227e-01 -1.37002003e+00
-5.08584857e-01 -5.77053726e-01 -1.15449119e+00 2.25471258e-01
-1.49785336e-02 8.42413008e-02 -9.59176123e-01 -5.11721849e-01
-3.23114812e-01 -3.61332536e-01 9.60448503e-01 6.69638738e-02
4.72580761e-01 -7.59060383e-01 -4.92484123e-01 5.71011126e-01
1.54927349e+00 3.12498957e-01 8.85255396e-01 1.21223104e+00
7.94970274e-01 5.47282994e-01 5.16594499e-02 1.79498509e-01
1.76046401e-01 2.39966974e-01 1.12470603e+00 -2.61482924e-01
3.71496752e-02 -3.89310718e-01 3.89881641e-01 -2.42941648e-01
3.03671241e-01 -3.17282885e-01 -8.79487753e-01 4.15673524e-01
-1.51083076e+00 -1.17340910e+00 -4.61767942e-01 2.33702469e+00
5.32364137e-02 -2.08740890e-01 4.72396910e-01 -3.47289890e-01
1.30558419e+00 3.98913264e-01 -6.96875393e-01 -5.71796238e-01
-8.95474255e-02 -4.10823911e-01 9.24834907e-01 1.80758104e-01
-1.44321167e+00 8.74687493e-01 7.35992575e+00 4.05622244e-01
-8.00341904e-01 1.13398120e-01 3.70769024e-01 3.18506420e-01
-3.21743816e-01 1.23363033e-01 -6.15056396e-01 7.42204368e-01
7.62405097e-01 -7.25297108e-02 5.95257521e-01 1.34054804e+00
-1.81865826e-01 2.52737224e-01 -5.99350274e-01 7.88369536e-01
4.89450783e-01 -1.31818390e+00 -3.03091644e-03 4.50804353e-01
7.71282136e-01 2.61790127e-01 6.21877134e-01 7.22137541e-02
1.01547098e+00 -1.00228691e+00 1.34421900e-01 -3.74210715e-01
5.26735604e-01 -4.62536693e-01 6.83119774e-01 -3.88380259e-01
-7.76775897e-01 -2.67982602e-01 -6.30161047e-01 5.05785048e-01
1.60017103e-01 3.68455440e-01 -1.50035405e+00 -9.59455073e-02
9.14606988e-01 7.42343843e-01 -1.31426036e+00 1.35520053e+00
-1.86443433e-01 4.31666642e-01 8.88920054e-02 -2.65841901e-01
5.33044696e-01 1.32151293e-02 5.83962858e-01 1.25504243e+00
1.24108963e-01 -7.15463519e-01 -1.29941508e-01 1.03790057e+00
2.13055667e-02 -9.71109699e-03 -1.16685152e+00 -2.55035162e-01
4.76018012e-01 1.63651776e+00 -5.92125237e-01 -3.87259662e-01
-6.28147125e-01 1.23988295e+00 2.18910187e-01 3.98388386e-01
-7.15695143e-01 -6.26023233e-01 8.39389086e-01 1.63353249e-01
1.57642543e-01 -8.14023539e-02 -4.23783183e-01 -1.32270420e+00
-3.32950950e-01 -9.18741167e-01 6.94637895e-01 -8.43354404e-01
-1.89532840e+00 3.74183804e-01 -5.35439074e-01 -9.48807299e-01
1.13557197e-01 -1.14671087e+00 -7.24078596e-01 6.51906013e-01
-1.51459479e+00 -1.54217076e+00 -6.45177364e-01 6.95961177e-01
9.08026323e-02 -3.80560577e-01 4.94793117e-01 -4.09011962e-03
-3.15442264e-01 3.54902923e-01 3.24508041e-01 3.85220379e-01
1.12501812e+00 -1.29848540e+00 1.35245907e+00 7.11973608e-01
-2.68162519e-01 9.79714751e-01 5.71628690e-01 -1.10172045e+00
-1.07723951e+00 -6.92167401e-01 6.58292532e-01 -8.58667314e-01
1.12912989e+00 -5.40337980e-01 -1.07708979e+00 8.67404103e-01
4.96356398e-01 -2.29065120e-01 7.12138653e-01 2.14254931e-01
-1.70821774e+00 3.45508099e-01 -1.52619362e+00 7.40231752e-01
8.36737871e-01 -7.92064369e-01 -4.83881325e-01 5.04368365e-01
3.06554854e-01 6.95196450e-01 -5.96600950e-01 -2.77604908e-01
8.84778440e-01 -1.15477693e+00 1.53013670e+00 -9.24705863e-01
4.07409400e-01 -9.84362885e-02 1.83951497e-01 -1.27836382e+00
-1.04901719e+00 -4.71662670e-01 -4.10567597e-02 8.11198831e-01
2.16041237e-01 -1.00681829e+00 1.16066003e+00 3.71198624e-01
3.78263503e-01 3.22685301e-01 -4.74908412e-01 -1.14020717e+00
2.48270318e-01 4.39305604e-01 3.81209165e-01 1.46716607e+00
3.88141125e-01 2.27295160e-01 -7.19686747e-01 2.20623180e-01
1.34832072e+00 -5.07322013e-01 6.58212900e-01 -1.66338968e+00
-6.30676672e-02 -7.02917635e-01 -5.37598610e-01 -5.01325488e-01
-6.54378384e-02 -8.35047483e-01 -3.03739637e-01 -1.33523762e+00
6.91969156e-01 -8.88355374e-02 -3.21508348e-01 5.39294004e-01
-1.92983508e-01 8.07082415e-01 2.36364231e-01 4.34309065e-01
-6.78044558e-01 -1.71722233e-01 9.09585238e-01 -3.97470057e-01
2.79545665e-01 -2.89949745e-01 -9.73329782e-01 3.60215813e-01
1.14225709e+00 -2.84896702e-01 -2.44194478e-01 1.12311095e-01
6.06438696e-01 -6.48388743e-01 8.25680971e-01 -4.99160022e-01
-7.60368705e-02 -2.20874414e-01 6.23565495e-01 -1.99291646e-01
-8.17714352e-03 -5.75935662e-01 -2.91784644e-01 8.70157957e-01
-2.69539177e-01 1.83917671e-01 -2.90439297e-02 7.89192557e-01
5.24942696e-01 -2.83526361e-01 1.01211333e+00 -4.84874994e-01
-1.03431308e+00 3.88821512e-01 -7.48197019e-01 -1.48237139e-01
1.26493573e+00 -4.16468263e-01 -1.34880400e+00 -4.37235385e-01
-1.64214969e-01 -2.05173820e-01 9.37258482e-01 6.85328126e-01
5.14698029e-01 -1.27853537e+00 -6.86014175e-01 9.65975448e-02
4.75250900e-01 -1.27674949e+00 1.85734704e-01 2.97450602e-01
-8.64028275e-01 5.65136194e-01 -9.67499256e-01 -1.83322772e-01
-1.53830850e+00 1.40662456e+00 -3.59863192e-02 1.34311626e-02
-5.00669897e-01 3.62331033e-01 7.04917252e-01 -6.36750698e-01
-1.40861437e-01 7.43761003e-01 -3.25923860e-01 -3.07437301e-01
9.13025081e-01 7.73223519e-01 -9.20042694e-02 -6.39157951e-01
-3.06788832e-01 -3.26405764e-02 -5.27094245e-01 1.62825301e-01
9.10030246e-01 -1.74978942e-01 1.34951272e-03 -4.83541161e-01
1.55977690e+00 -3.02309960e-01 -1.08623910e+00 -3.20011526e-02
1.77631199e-01 -1.15070474e+00 -2.57609159e-01 -1.28891265e+00
-8.33748817e-01 8.58185351e-01 6.65321887e-01 8.91004324e-01
2.34248042e-01 -1.03942871e-01 1.03289604e+00 4.65805411e-01
1.62635460e-01 -7.46861339e-01 6.76122606e-01 3.15184444e-01
5.23311496e-01 -1.71907580e+00 -2.01117262e-01 -6.15132093e-01
-5.84647954e-01 1.09331572e+00 7.76494503e-01 -5.21267056e-01
3.81533712e-01 -2.39880398e-01 3.46668303e-01 -6.17975831e-01
-2.51357198e-01 1.52178630e-01 2.45834216e-01 1.62957168e+00
-1.48076847e-01 2.80888584e-02 -1.52011007e-01 -1.98257864e-01
3.32204342e-01 -6.59422755e-01 7.17006505e-01 5.71215868e-01
-7.64448225e-01 -9.79986429e-01 -6.71544254e-01 2.87522376e-01
-4.57786560e-01 -4.34501052e-01 -1.24476659e+00 8.35458636e-01
-4.07971859e-01 7.28676796e-01 -1.69276640e-01 -5.39450943e-01
-2.10533276e-01 -1.67208269e-01 -1.79620553e-02 -2.71742672e-01
-9.48979795e-01 -5.29351890e-01 1.37500450e-01 -4.26905006e-01
3.18258226e-01 -6.18344069e-01 -2.61288702e-01 -1.07832575e+00
-3.54393154e-01 -3.92706245e-01 8.97016287e-01 3.17834049e-01
5.60580850e-01 -3.91035557e-01 8.37648928e-01 -7.67058194e-01
-8.17579508e-01 -6.21452510e-01 -6.00497186e-01 1.05247092e+00
1.65689141e-01 -4.35899049e-01 -6.53804898e-01 -3.86850178e-01] | [7.793848037719727, 9.969390869140625] |
92b8c927-1d07-4875-8028-d9ea4fab3535 | complex-words-identification-using-word-level | null | null | https://aclanthology.org/2021.semeval-1.11 | https://aclanthology.org/2021.semeval-1.11.pdf | Complex words identification using word-level features for SemEval-2020 Task 1 | This article describes a system to predict the complexity of words for the Lexical Complexity Prediction (LCP) shared task hosted at SemEval 2021 (Task 1) with a new annotated English dataset with a Likert scale. Located in the Lexical Semantics track, the task consisted of predicting the complexity value of the words in context. A machine learning approach was carried out based on the frequency of the words and several characteristics added at word level. Over these features, a supervised random forest regression algorithm was trained. Several runs were performed with different values to observe the performance of the algorithm. For the evaluation, our best results reported a M.A.E score of 0.07347, M.S.E. of 0.00938, and R.M.S.E. of 0.096871. Our experiments showed that, with a greater number of characteristics, the precision of the classification increases. | ["Arturo Montejo-R{\\'a}ez", 'Jenny A. Ortiz-Zambrano'] | 2021-08-01 | null | null | null | semeval-2021 | ['lexical-complexity-prediction'] | ['natural-language-processing'] | [-3.10846329e-01 1.46784335e-01 -1.75753430e-01 -3.59203190e-01
-4.44477499e-01 -5.33994853e-01 7.32494116e-01 6.99210584e-01
-8.50777388e-01 6.74966276e-01 3.30094904e-01 -1.88077554e-01
-8.19493905e-02 -7.40797698e-01 -4.18721169e-01 -2.15090230e-01
-1.53090030e-01 2.54143000e-01 1.92082211e-01 -1.98675290e-01
7.08812296e-01 7.21315220e-02 -1.72690570e+00 3.53995830e-01
6.16811097e-01 1.18243885e+00 4.51859266e-01 6.75604880e-01
-2.68098891e-01 5.92892468e-01 -5.44358134e-01 -4.07824844e-01
1.05479151e-01 -2.31610239e-02 -7.17162788e-01 -3.23350459e-01
-7.98607096e-02 2.36105472e-01 1.99513257e-01 8.74106348e-01
9.80663821e-02 2.00865149e-01 7.92486370e-01 -8.95253539e-01
-2.03465261e-02 8.09053659e-01 -2.70957649e-01 2.58062094e-01
9.00910139e-01 -5.42834587e-02 1.27829075e+00 -1.29739904e+00
7.52580702e-01 9.06407952e-01 6.67472422e-01 -7.84962922e-02
-1.12028790e+00 -6.45271122e-01 9.92825478e-02 4.15123761e-01
-1.62836146e+00 -4.49590385e-03 3.55304480e-01 -5.66069245e-01
1.46718729e+00 1.15317419e-01 8.63246918e-01 7.60904253e-01
3.34377795e-01 -8.19153562e-02 1.46170080e+00 -8.70961905e-01
3.96496236e-01 4.45144176e-01 3.43571842e-01 7.62162983e-01
3.42156768e-01 -9.72563997e-02 -8.15230608e-01 -1.29954126e-02
1.42321780e-01 -4.16215688e-01 2.51689613e-01 -1.25477510e-02
-8.86440516e-01 1.06059742e+00 1.86115861e-01 5.71143150e-01
-4.60471362e-01 -2.97999978e-01 5.51236749e-01 5.10147929e-01
4.96256649e-01 5.43494225e-01 -9.48419094e-01 -4.84456807e-01
-6.67417407e-01 4.52772170e-01 1.10496140e+00 6.16842508e-01
6.49963200e-01 -5.26058018e-01 2.30197698e-01 9.18022752e-01
1.33220330e-01 2.99369186e-01 8.17515135e-01 -5.38981497e-01
3.73686194e-01 7.89559901e-01 2.19898447e-02 -9.64381754e-01
-7.30390728e-01 -2.60690451e-01 -2.64306694e-01 1.70150995e-02
5.21373332e-01 -1.77419975e-01 -6.00388050e-01 1.60626340e+00
-2.75695394e-03 -3.15756619e-01 -2.85740606e-02 5.17973959e-01
7.07589865e-01 6.86492741e-01 6.33234680e-01 -3.98940563e-01
1.53314638e+00 -4.86704290e-01 -7.15907812e-01 1.66841540e-02
9.51255262e-01 -1.05783200e+00 1.48086798e+00 6.89615369e-01
-8.71228695e-01 -6.10804617e-01 -9.95334804e-01 3.29995453e-01
-7.93190598e-01 -4.12214897e-04 4.64559793e-01 7.20365822e-01
-7.65575290e-01 5.36518276e-01 -4.91073042e-01 -3.65104645e-01
-4.30499732e-01 1.55942515e-01 -4.04483497e-01 1.82684541e-01
-1.51627123e+00 1.30708277e+00 7.21458137e-01 -4.99970406e-01
-1.04385488e-01 -3.75927359e-01 -4.79436040e-01 5.51570654e-02
1.54521599e-01 -2.48294011e-01 9.34879184e-01 -1.04879403e+00
-1.34632444e+00 1.01574779e+00 -1.04843872e-03 -4.42339927e-01
2.35688016e-01 -1.46799162e-01 -6.23546422e-01 -1.93229139e-01
3.14779490e-01 5.09221673e-01 4.19646561e-01 -8.30084801e-01
-8.51108313e-01 -3.67037654e-01 8.24486315e-02 3.82699966e-01
-4.31510895e-01 2.16298297e-01 1.65462978e-02 -5.95012307e-01
-6.70452639e-02 -9.34706926e-01 -4.99256775e-02 -6.47662401e-01
1.14561081e-01 -4.46327567e-01 9.79510974e-03 -8.60666215e-01
1.82843733e+00 -2.06690311e+00 -3.70069630e-02 5.47914624e-01
4.42840755e-02 7.47496029e-03 4.98783542e-03 6.52443647e-01
-1.45036891e-01 2.65658051e-01 5.77111878e-02 1.09536260e-01
-4.84426729e-02 -1.56195313e-01 1.71160713e-01 -1.11483760e-01
-1.99536264e-01 5.87609470e-01 -6.39077485e-01 -3.84907931e-01
-5.58382422e-02 6.56274781e-02 -6.17432177e-01 1.54562846e-01
-1.89901665e-01 -2.09299296e-01 -8.92063901e-02 2.29062110e-01
2.54494488e-01 -6.09007291e-03 3.37209314e-01 -4.38782386e-02
-3.78873199e-01 3.95284325e-01 -1.28180647e+00 1.32276583e+00
-8.68321717e-01 4.37107444e-01 -5.44830918e-01 -3.53075147e-01
1.08975124e+00 1.80247307e-01 2.37888157e-01 -8.43987048e-01
6.99990913e-02 4.54856753e-01 2.49065459e-01 -5.00441432e-01
7.43271947e-01 -1.76894769e-01 -3.66583556e-01 1.47038966e-01
1.32439770e-02 -1.04748935e-01 3.08293134e-01 8.45906883e-02
1.20649314e+00 -8.19888189e-02 9.20410335e-01 -6.12808585e-01
6.12902045e-01 1.72060691e-02 5.28979935e-02 2.78399616e-01
-1.25677645e-01 -4.37305309e-02 3.63217264e-01 -5.18203795e-01
-1.18104672e+00 -7.66120136e-01 -3.34521830e-01 1.25705016e+00
-1.81629732e-01 -1.09034252e+00 -7.02588677e-01 -4.24548417e-01
-1.81056872e-01 8.84617448e-01 -5.72892845e-01 -1.87839955e-01
-2.43675828e-01 -3.13013434e-01 2.76659548e-01 3.25723588e-01
3.47519442e-02 -1.20064259e+00 -7.45204270e-01 1.93927824e-01
-3.77046973e-01 -1.23526931e+00 -1.66573167e-01 4.54001456e-01
-6.48837149e-01 -8.53482902e-01 -3.02389879e-02 -8.28075230e-01
1.78558812e-01 -1.50638178e-01 1.20967007e+00 1.54133186e-01
-1.71856984e-01 1.22069687e-01 -7.73025453e-01 -4.94483739e-01
-3.51425707e-01 4.61602360e-01 1.09869219e-01 -5.35528898e-01
7.13855445e-01 -2.50088274e-01 -2.90378958e-01 2.44377866e-01
-7.48141348e-01 -2.26219520e-01 3.59598100e-01 6.86566591e-01
4.85567093e-01 2.20374301e-01 5.14644623e-01 -8.04995120e-01
8.85337651e-01 -5.85875034e-01 -4.96516317e-01 3.03706359e-02
-1.07884777e+00 6.60847947e-02 6.23810470e-01 -4.34006870e-01
-6.48632765e-01 1.14563413e-01 -2.43787676e-01 1.89099312e-01
-1.19456522e-01 9.04807210e-01 7.13798702e-02 2.93576092e-01
6.97110236e-01 -1.55931130e-01 -2.48982042e-01 -2.88801789e-01
3.28202814e-01 8.36717308e-01 3.35118622e-02 -4.31378037e-01
3.67933542e-01 -3.00385416e-01 -2.00000092e-01 -9.54310715e-01
-9.36987102e-01 -5.63146412e-01 -6.33423090e-01 -2.44696856e-01
7.98215568e-01 -9.00037825e-01 -4.31764096e-01 2.48591945e-01
-8.85683119e-01 -1.54817462e-01 -9.05764997e-02 7.27721989e-01
-5.08184254e-01 1.44416153e-01 -2.22014278e-01 -6.67458355e-01
-3.09840441e-01 -7.40698814e-01 5.21069765e-01 -2.66702436e-02
-1.03303146e+00 -9.86968517e-01 1.56779736e-01 4.07218814e-01
3.84186596e-01 7.37895316e-04 1.20089138e+00 -1.16782057e+00
2.14046881e-01 -3.63768011e-01 6.38074428e-02 2.02466831e-01
-1.35189697e-01 8.54603946e-02 -7.49061465e-01 1.32569090e-01
5.89616150e-02 -3.00642699e-01 4.51513857e-01 1.65638909e-01
1.11879420e+00 -2.22744092e-01 1.47086293e-01 -7.74674267e-02
1.47003615e+00 1.12018980e-01 4.30627435e-01 6.75845861e-01
3.16723913e-01 6.95309401e-01 9.72704530e-01 7.68672347e-01
3.87948304e-01 7.33046412e-01 2.14314714e-01 6.32745564e-01
2.02908158e-01 -3.24947417e-01 5.19782960e-01 1.00207388e+00
-1.01245213e-02 -4.10746783e-02 -1.27428937e+00 3.00834626e-01
-1.35100091e+00 -7.19301105e-01 -3.29106838e-01 2.18512869e+00
9.30849135e-01 8.16555917e-01 3.38501066e-01 5.65541625e-01
5.60948968e-01 3.42877954e-02 3.00341211e-02 -8.62739205e-01
1.17499940e-01 5.23638785e-01 3.65370154e-01 7.12952614e-01
-6.75476730e-01 1.12431586e+00 5.96358728e+00 8.99798036e-01
-9.19887960e-01 -3.91410030e-02 3.46999824e-01 5.15203513e-02
-1.02650940e-01 6.03272282e-02 -7.67969787e-01 6.76037729e-01
1.54847300e+00 -5.03479183e-01 4.86847669e-01 7.76205957e-01
4.99859713e-02 -6.09614134e-01 -7.47475207e-01 8.07751358e-01
7.87634701e-02 -8.07598829e-01 -6.01869263e-02 7.63359666e-02
5.13593912e-01 9.90979932e-03 -1.86236545e-01 5.60542047e-01
-1.53107032e-01 -1.00601661e+00 7.31481314e-01 6.51949942e-01
6.83850229e-01 -8.84004056e-01 1.06647801e+00 7.17567146e-01
-1.15491748e+00 -6.46215007e-02 -1.44935891e-01 -5.88676751e-01
-2.70330459e-01 9.18736160e-01 -9.04578805e-01 2.38359854e-01
6.93073869e-01 1.72046170e-01 -6.28566742e-01 5.88608861e-01
-2.19808817e-01 5.82082391e-01 -3.85171443e-01 -6.63752675e-01
7.93740433e-03 -3.72973420e-02 3.91910732e-01 1.39517927e+00
2.52285063e-01 -3.48866247e-02 9.10759345e-02 9.79138836e-02
3.93253099e-03 1.04012656e+00 -4.65782195e-01 2.89021805e-02
7.00752199e-01 1.28548300e+00 -9.36739445e-01 -3.78957659e-01
-3.39788496e-01 5.11368155e-01 4.01397467e-01 -3.65393579e-01
-8.45809102e-01 -4.70832139e-01 1.42261118e-01 4.07532811e-01
3.10060501e-01 -1.85991868e-01 -7.01506436e-01 -8.44211936e-01
6.62408993e-02 -8.29719067e-01 3.56481552e-01 -5.96958578e-01
-1.31018519e+00 7.69314051e-01 8.02522153e-02 -9.03520525e-01
-3.06094050e-01 -6.46901429e-01 -2.81129807e-01 1.06637478e+00
-9.68856871e-01 -5.49858391e-01 -2.75869101e-01 2.73482531e-01
5.30337751e-01 -3.40742469e-01 1.15879762e+00 -1.77000295e-02
-6.72057346e-02 4.89820719e-01 -3.03153366e-01 -2.59763122e-01
8.26434135e-01 -1.38672900e+00 1.32588029e-01 3.47412646e-01
7.98403621e-02 4.76974070e-01 8.99122536e-01 -7.67517388e-01
-7.33324349e-01 -7.44193077e-01 1.77102411e+00 -5.61514318e-01
1.09921789e+00 -9.53740329e-02 -7.90646732e-01 2.31914252e-01
5.44240400e-02 -2.92545706e-01 9.59425390e-01 2.47216448e-01
-3.99273634e-01 1.44253284e-01 -9.97667134e-01 2.89551914e-01
6.63713098e-01 -6.35670543e-01 -8.05655837e-01 2.07741290e-01
5.13762712e-01 -2.02650413e-01 -1.29268360e+00 1.11756839e-01
8.70538056e-01 -8.74275506e-01 5.62700987e-01 -3.58669579e-01
7.29349852e-01 -1.03168689e-01 -6.77177072e-01 -1.33679545e+00
-2.92256087e-01 5.45446835e-02 -8.92422721e-02 7.98285067e-01
8.04462969e-01 -4.53953981e-01 3.94794881e-01 4.04557347e-01
1.43927723e-01 -9.75796878e-01 -1.00135827e+00 -7.20183849e-01
2.37799749e-01 -6.06555521e-01 2.76294172e-01 8.71482313e-01
3.27564329e-01 7.36857593e-01 2.01418027e-02 -2.59163201e-01
4.93537970e-02 -2.08689317e-01 3.56239855e-01 -1.45996773e+00
-5.24376184e-02 -4.61407393e-01 -4.47187454e-01 -3.05204183e-01
2.83046186e-01 -9.51514900e-01 -3.78063500e-01 -1.22183943e+00
1.34289771e-01 -8.50627497e-02 -3.62566262e-01 2.07742438e-01
-3.91138971e-01 -8.88325721e-02 2.04529256e-01 1.33813798e-01
-4.02281910e-01 1.30384993e-02 6.88189507e-01 3.46310943e-01
-2.75481522e-01 7.61093386e-03 -4.36645269e-01 8.36922050e-01
1.49073768e+00 -6.47709966e-01 -2.90705621e-01 6.33590296e-02
6.64373219e-01 -8.60845000e-02 -1.04964070e-01 -1.01199687e+00
1.30250407e-02 -1.40694320e-01 3.69247347e-01 -5.04937589e-01
2.07772732e-01 -8.99030745e-01 2.84649312e-01 9.27562773e-01
-4.84697789e-01 6.30959213e-01 -1.13736244e-03 2.12672845e-01
-1.41979590e-01 -5.95327377e-01 6.74915135e-01 -5.96066192e-02
-7.28818655e-01 -3.69365662e-01 -5.37881494e-01 9.66144130e-02
1.17824972e+00 -1.65941283e-01 2.37515211e-01 -1.28441706e-01
-9.58018422e-01 -1.98831081e-01 3.31053287e-01 6.10975027e-01
3.23352516e-01 -1.06126785e+00 -5.32105982e-01 3.25973928e-02
5.59712648e-01 -6.96991146e-01 -1.40439153e-01 4.35237378e-01
-4.98208165e-01 4.77408469e-01 -2.76699156e-01 -1.12784974e-01
-1.50987303e+00 4.09106344e-01 4.18479145e-02 -6.58243358e-01
-2.78346986e-01 4.91036236e-01 -6.53612733e-01 -3.25421840e-01
1.53183743e-01 -4.60672051e-01 -5.16471744e-01 6.62629128e-01
3.32427442e-01 6.60211086e-01 4.12560046e-01 -6.91459954e-01
-4.11217272e-01 5.78748167e-01 7.56215081e-02 -4.37274128e-01
1.33123279e+00 -1.50528818e-01 -8.28244910e-02 1.03894091e+00
1.08738899e+00 3.25442880e-01 -3.57489347e-01 -2.49758467e-01
5.34627557e-01 -2.07790107e-01 6.10149316e-02 -1.15008020e+00
-3.43018889e-01 4.11183417e-01 7.43927658e-01 4.24319386e-01
8.85380208e-01 -2.15620890e-01 4.12013263e-01 4.97725129e-01
4.80649054e-01 -1.60850286e+00 -8.54197592e-02 1.05110443e+00
7.44632185e-01 -1.13641870e+00 7.69802034e-02 -4.12758529e-01
-8.09094012e-01 1.31626642e+00 4.23332572e-01 -7.75768608e-02
9.34535086e-01 1.87291369e-01 9.63297561e-02 -6.79234713e-02
-1.04288864e+00 -1.05324008e-01 1.73009127e-01 2.76879400e-01
9.19966578e-01 5.32751381e-01 -1.24062729e+00 9.36107457e-01
-8.56553257e-01 -4.91566993e-02 5.71816087e-01 8.33326876e-01
-8.56467187e-01 -1.11703682e+00 -2.23945007e-01 5.77887654e-01
-4.63596255e-01 -2.60270745e-01 -4.53749150e-01 7.66986549e-01
1.49194017e-01 9.45308506e-01 -3.40265781e-02 -6.21720910e-01
6.69440806e-01 5.91004074e-01 2.37135962e-01 -6.39326096e-01
-8.34058225e-01 -3.79729152e-01 2.50736713e-01 -4.54160064e-01
-2.80025840e-01 -5.95517874e-01 -1.35667145e+00 -3.90237898e-01
-3.68856996e-01 4.41335678e-01 1.15040767e+00 7.24536896e-01
-1.10004500e-01 4.14530694e-01 8.21578503e-01 -3.61991197e-01
-6.36062324e-01 -1.22970855e+00 -6.06456399e-01 5.05168378e-01
-3.55325550e-01 -6.12999797e-01 -6.08127534e-01 -4.60510924e-02] | [10.636628150939941, 10.427436828613281] |
a077b6db-5d04-416e-aeb2-05172eaa8e46 | clustop-an-unsupervised-and-integrated-text | 2301.00818 | null | https://arxiv.org/abs/2301.00818v1 | https://arxiv.org/pdf/2301.00818v1.pdf | ClusTop: An unsupervised and integrated text clustering and topic extraction framework | Text clustering and topic extraction are two important tasks in text mining. Usually, these two tasks are performed separately. For topic extraction to facilitate clustering, we can first project texts into a topic space and then perform a clustering algorithm to obtain clusters. To promote topic extraction by clustering, we can first obtain clusters with a clustering algorithm and then extract cluster-specific topics. However, this naive strategy ignores the fact that text clustering and topic extraction are strongly correlated and follow a chicken-and-egg relationship. Performing them separately fails to make them mutually benefit each other to achieve the best overall performance. In this paper, we propose an unsupervised text clustering and topic extraction framework (ClusTop) which integrates text clustering and topic extraction into a unified framework and can achieve high-quality clustering result and extract topics from each cluster simultaneously. Our framework includes four components: enhanced language model training, dimensionality reduction, clustering and topic extraction, where the enhanced language model can be viewed as a bridge between clustering and topic extraction. On one hand, it provides text embeddings with a strong cluster structure which facilitates effective text clustering; on the other hand, it pays high attention on the topic related words for topic extraction because of its self-attention architecture. Moreover, the training of enhanced language model is unsupervised. Experiments on two datasets demonstrate the effectiveness of our framework and provide benchmarks for different model combinations in this framework. | ['Yatong Zhou', 'Jingfei He', 'Siwei Duo', 'Chenghu Mi', 'Zhongtao Chen'] | 2023-01-03 | null | null | null | null | ['text-clustering'] | ['natural-language-processing'] | [-2.04007745e-01 7.49551505e-02 -2.56516099e-01 -4.04634416e-01
-8.04146230e-01 -2.13685766e-01 7.08853245e-01 2.77279913e-01
-3.48230809e-01 7.58109763e-02 3.81997049e-01 -1.61376402e-01
-2.03217015e-01 -9.72571075e-01 -1.23196051e-01 -9.97676015e-01
2.12538391e-02 5.85538149e-01 1.73644662e-01 2.33301163e-01
1.20728448e-01 -7.87357539e-02 -1.44931614e+00 1.29697829e-01
1.21273339e+00 7.05949664e-01 4.96719092e-01 2.82068789e-01
-6.81394637e-01 5.45042932e-01 -5.67707300e-01 7.35965976e-03
-1.80626839e-01 -2.05714509e-01 -6.72292233e-01 5.69438279e-01
-4.38760847e-01 -3.09968442e-01 -1.93318054e-01 9.38560605e-01
3.11275482e-01 2.69078553e-01 8.93589079e-01 -1.29388654e+00
-4.54603463e-01 1.01589501e+00 -1.01860189e+00 -2.56297469e-01
-7.87191391e-02 -2.80405432e-01 1.19107723e+00 -9.48819816e-01
2.82286376e-01 1.31712103e+00 1.64986372e-01 1.42794281e-01
-1.06844306e+00 -7.96789169e-01 5.31205356e-01 -1.62313625e-01
-1.49662971e+00 -3.91362607e-02 1.07889211e+00 -5.43465614e-01
6.27059579e-01 -6.98232278e-02 7.12682128e-01 7.19594002e-01
-2.12669104e-01 1.22826624e+00 5.43685973e-01 -2.20691711e-01
3.45180184e-01 4.76903200e-01 6.48608387e-01 1.05790719e-01
2.88395286e-01 -5.87613463e-01 -2.46753126e-01 -1.47912979e-01
2.45300680e-01 4.69159454e-01 -3.05674314e-01 -4.35584784e-01
-1.23181498e+00 1.17358351e+00 2.37895980e-01 5.18854558e-01
-5.13871372e-01 -1.05792902e-01 4.16750103e-01 -1.12187564e-01
6.74181759e-01 3.11648726e-01 -3.03437918e-01 1.25510260e-01
-1.20655823e+00 1.03901036e-01 6.74042284e-01 1.00079393e+00
1.01251471e+00 -2.71820486e-01 -2.06522807e-01 9.26518083e-01
6.29964769e-01 2.33387038e-01 8.72665048e-01 -3.78885210e-01
5.56585491e-01 1.06958222e+00 -1.23279981e-01 -1.33683765e+00
-4.42094117e-01 -2.05626890e-01 -9.58702087e-01 -3.75129402e-01
-1.03903800e-01 -3.57768476e-01 -7.78533578e-01 1.52689743e+00
5.50662279e-01 1.48124561e-01 2.02864278e-02 6.50228560e-01
8.04892480e-01 1.10362136e+00 2.21353903e-01 -5.03223002e-01
1.68166435e+00 -9.90295410e-01 -1.08585429e+00 1.13689136e-02
6.66597843e-01 -5.66043019e-01 9.91164565e-01 3.72014076e-01
-7.26782978e-01 -4.39391613e-01 -8.31155598e-01 -1.84251443e-01
-5.24652660e-01 3.13139617e-01 6.21242464e-01 5.01762331e-01
-6.90638065e-01 -3.40246148e-02 -1.12447453e+00 -3.15655380e-01
4.66649473e-01 4.94349837e-01 2.48895492e-02 1.54345766e-01
-1.25100791e+00 1.74548596e-01 8.75924230e-01 -2.09428594e-01
-3.99088770e-01 -5.19390464e-01 -8.16945016e-01 3.40363353e-01
6.56974614e-01 -4.44961727e-01 9.05975699e-01 -4.75805998e-01
-1.28683281e+00 4.74670351e-01 -4.38467026e-01 -3.64910394e-01
3.78681496e-02 -3.58639270e-01 -3.26799482e-01 3.18691045e-01
2.64780968e-01 8.80973279e-01 8.80838692e-01 -1.31886137e+00
-1.01712596e+00 -4.00352120e-01 -5.20840228e-01 5.05407870e-01
-1.26059377e+00 5.86192822e-03 -1.17756653e+00 -7.23510325e-01
2.75003791e-01 -6.38360620e-01 -2.13645205e-01 -5.88094771e-01
-8.02891612e-01 -8.07125628e-01 1.40261686e+00 -5.38231909e-01
1.73190176e+00 -2.24704790e+00 1.67597115e-01 2.95606047e-01
7.67372012e-01 -9.57229137e-02 3.46411556e-01 4.02800292e-01
-1.34207364e-02 3.32636505e-01 -1.67696446e-01 -5.66233456e-01
9.33214724e-02 -1.12314902e-01 -4.76474911e-01 2.21532539e-01
1.09373435e-01 8.45731497e-01 -7.44100809e-01 -1.03764129e+00
2.72503465e-01 5.44467866e-01 -6.22361183e-01 2.62078524e-01
-2.65254945e-01 -4.92798351e-02 -8.10334027e-01 1.73264816e-01
5.98097920e-01 -4.98322964e-01 3.15014839e-01 -1.15592614e-01
-1.68940932e-01 3.10596079e-01 -1.42882240e+00 1.46058381e+00
-1.66724533e-01 6.44486845e-01 2.01732423e-02 -1.11455369e+00
1.00147963e+00 4.08974916e-01 9.51024890e-01 -1.20263316e-01
2.87343889e-01 -2.43477881e-01 -1.47679791e-01 -4.05651063e-01
7.49838114e-01 -4.82469387e-02 -1.31483287e-01 9.78898168e-01
1.26082465e-01 4.45991717e-02 2.81804383e-01 6.22265935e-01
5.67539930e-01 -3.20015907e-01 2.45979220e-01 -3.59301418e-01
3.00329059e-01 1.03277907e-01 5.11471033e-01 2.17807159e-01
7.77776092e-02 3.27793658e-01 5.84129274e-01 -3.85604352e-02
-8.82730782e-01 -7.20625997e-01 -2.18233272e-01 1.20964146e+00
2.81861842e-01 -8.39529216e-01 -7.77848125e-01 -7.95842230e-01
-2.08648682e-01 5.93004584e-01 -7.02420115e-01 -1.72933578e-01
-3.05426806e-01 -1.13482332e+00 9.81566235e-02 4.21119124e-01
3.87930959e-01 -9.71241117e-01 -3.12508315e-01 2.20971629e-01
-4.35396403e-01 -8.54971945e-01 -5.27666450e-01 2.33975604e-01
-8.24930906e-01 -9.11262512e-01 -6.59704983e-01 -9.78313804e-01
6.96475804e-01 7.66863644e-01 7.94636071e-01 5.33201024e-02
1.06172308e-01 9.74353927e-04 -5.74671209e-01 -5.13534725e-01
-3.98052903e-03 6.12955272e-01 2.22202595e-02 9.66953561e-02
8.81971240e-01 -5.92865348e-01 -6.12096786e-01 2.80891657e-01
-1.16703999e+00 1.38339818e-01 5.80460608e-01 6.93715036e-01
2.49892548e-01 9.49617326e-01 5.72050571e-01 -9.78612602e-01
7.99563110e-01 -6.93892598e-01 -3.22357357e-01 -1.03385756e-02
-6.83271945e-01 -2.03911394e-01 5.93357265e-01 -3.79904211e-01
-1.13667333e+00 -6.72906488e-02 1.13740690e-01 -3.28048825e-01
-3.46884340e-01 6.97304070e-01 -5.43504179e-01 8.24951410e-01
2.01956376e-01 4.94893759e-01 -2.32419670e-01 -4.89378333e-01
5.89538872e-01 1.16588557e+00 1.29582256e-01 -4.47270900e-01
9.40705121e-01 7.05262780e-01 -7.01571763e-01 -1.12830889e+00
-7.61221230e-01 -8.94617081e-01 -8.29401672e-01 6.78604245e-02
9.81881022e-01 -1.08261275e+00 -6.06084287e-01 2.47987747e-01
-9.34056461e-01 -3.02907694e-02 -1.05727389e-01 6.37100816e-01
-1.59897000e-01 5.72786093e-01 -4.36644018e-01 -9.21885550e-01
-4.79172081e-01 -9.95336890e-01 1.25598943e+00 2.80943990e-01
-9.38322991e-02 -1.14996350e+00 -1.23495713e-01 1.97729114e-02
1.87668949e-02 -1.48533374e-01 1.04076409e+00 -1.05939472e+00
-3.46013010e-01 -9.36482847e-02 -4.20853764e-01 -2.39287373e-02
4.92663622e-01 1.29204482e-01 -8.50565493e-01 -3.16807181e-01
9.04123560e-02 -9.55887605e-03 1.14817786e+00 3.59799981e-01
1.03760850e+00 -2.68437237e-01 -7.50558257e-01 4.23598409e-01
1.19801307e+00 4.13196176e-01 3.52999240e-01 3.55828077e-01
9.65388000e-01 9.89224970e-01 5.78037262e-01 4.40414518e-01
7.51689732e-01 5.60917377e-01 -4.46819961e-02 -3.03011626e-01
3.34269077e-01 -1.99138656e-01 3.82575810e-01 1.22014415e+00
5.66881239e-01 -2.27902099e-01 -1.13336337e+00 7.05039263e-01
-1.88321900e+00 -7.48132825e-01 -2.84204662e-01 1.85974205e+00
9.52634275e-01 1.42103285e-01 4.15030807e-01 3.58424813e-01
9.31033373e-01 -1.35751003e-02 -3.66030931e-01 2.51976281e-01
1.96849063e-01 -4.34307933e-01 -1.53869405e-01 1.62736580e-01
-1.44508803e+00 1.24214208e+00 5.31518078e+00 1.05289364e+00
-8.93999398e-01 2.34210584e-02 6.15846574e-01 -1.84256788e-02
-1.73565865e-01 -5.17233126e-02 -1.05944610e+00 6.58108354e-01
5.22584617e-01 -3.40650618e-01 -1.44105583e-01 1.22180736e+00
2.98940003e-01 5.58160804e-02 -8.27804387e-01 8.11912537e-01
-6.46737963e-02 -8.75644565e-01 2.61628866e-01 4.94452238e-01
7.99091041e-01 -3.00560564e-01 -4.10486050e-02 5.53980172e-01
6.31200373e-01 -7.31537342e-01 2.35413283e-01 7.31886849e-02
2.48217538e-01 -1.05771816e+00 7.21193969e-01 4.66000378e-01
-1.47180688e+00 -4.85072620e-02 -5.86887300e-01 3.01968127e-01
1.71527952e-01 1.08789527e+00 -6.04218066e-01 6.74413741e-01
7.81534493e-01 9.25793350e-01 -4.13529158e-01 8.69278371e-01
-2.44779065e-02 7.78885663e-01 -4.48741913e-01 -6.37412891e-02
4.24755692e-01 -5.21241128e-01 3.71511459e-01 1.37482834e+00
1.70398563e-01 8.76083598e-02 6.25314772e-01 8.00087273e-01
-6.65050298e-02 5.43615043e-01 -3.86730313e-01 -7.19630942e-02
7.57862628e-01 1.55846488e+00 -1.22539628e+00 -7.07718849e-01
-4.51082915e-01 4.84248966e-01 2.94236153e-01 4.19495702e-01
-7.04451919e-01 -8.53843749e-01 5.38200080e-01 -6.68970793e-02
4.42743123e-01 -3.30069691e-01 -4.42464143e-01 -1.36175287e+00
-1.48781121e-01 -6.03165448e-01 3.52918833e-01 -3.09766740e-01
-1.09125340e+00 5.34327209e-01 8.41243640e-02 -1.23651350e+00
-4.32983786e-02 -1.94414467e-01 -7.88463295e-01 5.39475501e-01
-1.38869452e+00 -1.05736959e+00 -2.98965186e-01 6.29380643e-01
7.50223339e-01 -2.73230337e-02 4.42273587e-01 1.94664821e-01
-9.98016000e-01 2.95953602e-01 3.52018565e-01 3.55878323e-01
6.17415071e-01 -1.48693228e+00 1.62677854e-01 7.67995298e-01
9.26425979e-02 9.90832627e-01 2.17274114e-01 -7.64255643e-01
-1.05642259e+00 -1.50493133e+00 7.39528179e-01 -3.84428084e-01
6.71583176e-01 -6.86109662e-01 -1.21900690e+00 4.89081055e-01
3.33601564e-01 -9.37008500e-01 1.00285351e+00 4.49545294e-01
-9.65857357e-02 1.06192995e-02 -3.35818678e-01 6.93707049e-01
3.65805298e-01 -3.07811022e-01 -8.09405625e-01 4.07101989e-01
1.03306615e+00 2.43632674e-01 -8.60616386e-01 1.74588189e-02
1.76917598e-01 -6.60567164e-01 7.92872906e-01 -3.35638225e-01
4.88794923e-01 -4.61368382e-01 1.60346493e-01 -1.20149946e+00
-2.20503032e-01 -6.22117221e-01 -2.41425827e-01 1.91832483e+00
3.65075171e-01 -2.82760799e-01 7.84077644e-01 3.92521858e-01
1.85849324e-01 -7.18053937e-01 -3.08907837e-01 -5.11013567e-01
1.96941033e-01 -6.73080623e-01 7.01438546e-01 1.29898441e+00
4.10621613e-01 8.72850120e-01 -2.67926365e-01 1.43070936e-01
5.74880540e-01 4.34715927e-01 9.99662757e-01 -1.51570106e+00
1.45681113e-01 -7.35548079e-01 9.28926170e-02 -1.25167549e+00
1.64136738e-01 -6.77464008e-01 2.34654829e-01 -1.77576935e+00
4.88814771e-01 -4.36630577e-01 -3.90058570e-02 3.59653801e-01
-6.98617160e-01 -3.57055992e-01 1.10715598e-01 7.14724302e-01
-9.20335591e-01 8.96475315e-01 7.58595824e-01 -1.84877068e-02
-7.53903389e-01 -2.32452080e-02 -1.16930866e+00 8.40444922e-01
8.14360380e-01 -6.56076908e-01 -4.58551019e-01 -1.65191561e-01
5.10579534e-02 -3.23670745e-01 -2.57631749e-01 -7.82403708e-01
5.11600256e-01 7.99368545e-02 2.46291950e-01 -1.01322961e+00
5.61806932e-02 -9.30641711e-01 -5.42996526e-01 1.29146338e-01
-3.22580278e-01 -4.33210135e-01 -3.86484265e-02 6.95757568e-01
-4.75320637e-01 7.36662894e-02 6.01328075e-01 1.26279309e-01
-5.18015504e-01 4.90341067e-01 -6.31362319e-01 -1.49415582e-01
1.03752279e+00 4.46656346e-02 1.72729380e-02 -4.09950078e-01
-5.68432510e-01 9.58176255e-01 3.20056140e-01 5.41898012e-01
3.93373281e-01 -1.39871430e+00 -5.44148207e-01 3.29081975e-02
7.79315978e-02 5.42209208e-01 2.12043524e-01 8.59384060e-01
9.97072905e-02 6.50662005e-01 4.52922434e-01 -7.15506077e-01
-1.02468133e+00 1.05930364e+00 -2.76324481e-01 -5.67456424e-01
-6.72213674e-01 3.94237489e-01 8.21275413e-01 -3.72544229e-01
4.28696215e-01 -2.75657952e-01 -8.00318182e-01 5.83385885e-01
6.24007404e-01 1.77355230e-01 -2.21007124e-01 -4.59911406e-01
-1.49433926e-01 6.25617385e-01 -4.37227607e-01 -1.34247303e-01
1.31832087e+00 -5.01239657e-01 -2.39729375e-01 4.90442038e-01
1.33840275e+00 -2.11344913e-01 -8.82263541e-01 -4.52689409e-01
3.14340353e-01 -1.20579772e-01 3.80092412e-01 -1.75149515e-01
-1.02295542e+00 1.24390769e+00 6.55797198e-02 8.91976118e-01
1.25392759e+00 1.80663794e-01 8.81317317e-01 5.14388084e-01
-4.36043203e-01 -1.31819296e+00 2.68174231e-01 3.65851671e-01
3.93759489e-01 -1.23418438e+00 4.76221815e-02 -6.50344014e-01
-7.60249615e-01 9.79742050e-01 6.32107019e-01 5.82755655e-02
9.58495557e-01 1.67739600e-01 -6.80970028e-02 -4.11132157e-01
-6.98428988e-01 -4.83825743e-01 4.27307546e-01 4.98980016e-01
5.01645088e-01 -6.06698468e-02 -2.41933882e-01 1.10528386e+00
-3.29588234e-01 -4.03993517e-01 1.62970647e-01 7.39486754e-01
-9.36801374e-01 -1.03280675e+00 -4.10157353e-01 6.67786300e-01
-4.83737171e-01 -1.12178937e-01 -4.37015116e-01 8.69598806e-01
-2.01652363e-01 1.13502276e+00 1.76376030e-01 -5.03628314e-01
-2.62524616e-02 1.82446748e-01 -6.66334629e-01 -9.02345002e-01
-3.90904576e-01 9.30706799e-01 -3.72233242e-01 -1.11021064e-01
-3.15243095e-01 -5.79391658e-01 -1.37474990e+00 -1.13211803e-01
-7.42689669e-01 7.94288099e-01 6.98039711e-01 1.03399801e+00
4.97160882e-01 7.32138574e-01 9.66231465e-01 -5.63637137e-01
4.25072871e-02 -1.16135693e+00 -7.37153232e-01 2.78148562e-01
-4.54058200e-02 -5.49635351e-01 -1.95875898e-01 3.17834735e-01] | [10.37619686126709, 6.829369068145752] |
fa1e2d0d-34ab-42ff-a53b-2dedfac2794f | towards-open-intent-detection | 2203.05823 | null | https://arxiv.org/abs/2203.05823v3 | https://arxiv.org/pdf/2203.05823v3.pdf | Learning Discriminative Representations and Decision Boundaries for Open Intent Detection | Open intent detection is a significant problem in natural language understanding, which aims to identify the unseen open intent while ensuring known intent identification performance. However, current methods face two major challenges. Firstly, they struggle to learn friendly representations to detect the open intent with prior knowledge of only known intents. Secondly, there is a lack of an effective approach to obtaining specific and compact decision boundaries for known intents. To address these issues, this paper presents an original framework called DA-ADB, which successively learns distance-aware intent representations and adaptive decision boundaries for open intent detection. Specifically, we first leverage distance information to enhance the distinguishing capability of the intent representations. Then, we design a novel loss function to obtain appropriate decision boundaries by balancing both empirical and open space risks. Extensive experiments demonstrate the effectiveness of the proposed distance-aware and boundary learning strategies. Compared to state-of-the-art methods, our framework achieves substantial improvements on three benchmark datasets. Furthermore, it yields robust performance with varying proportions of labeled data and known categories. | ['Qianrui Zhou', 'Shaojie Zhao', 'Hua Xu', 'Hanlei Zhang'] | 2022-03-11 | null | null | null | null | ['open-intent-detection'] | ['natural-language-processing'] | [ 1.67459846e-01 -4.51260246e-02 -4.95144635e-01 -4.16408777e-01
-1.04673076e+00 -7.12556839e-01 4.03334409e-01 1.06977031e-01
-2.90622920e-01 5.41492105e-01 3.84301156e-01 -3.52568269e-01
-2.10703179e-01 -6.34123564e-01 -2.34267533e-01 -3.57396692e-01
-6.25701621e-04 3.61716449e-01 -2.90294103e-02 5.99612929e-02
6.14517450e-01 -5.76549722e-03 -1.44880116e+00 -8.67366269e-02
1.43582666e+00 1.30078042e+00 -5.63367531e-02 3.73440623e-01
-2.30482787e-01 7.61326194e-01 -4.23076689e-01 -1.64490372e-01
4.97150302e-01 -1.08076833e-01 -1.01131225e+00 -1.98254008e-02
5.78173637e-01 -4.92689133e-01 -5.30294597e-01 9.41764891e-01
2.39311069e-01 2.43885860e-01 1.04301441e+00 -1.36300731e+00
-9.83478844e-01 2.96236604e-01 -5.65189123e-01 6.01045728e-01
5.22081316e-01 4.24441025e-02 1.52061427e+00 -8.80286813e-01
3.09814304e-01 8.73729706e-01 7.24689007e-01 4.68998015e-01
-1.03882706e+00 -6.86360180e-01 5.13948143e-01 3.61840785e-01
-1.55625510e+00 -3.52070481e-01 9.49437618e-01 -5.44647098e-01
6.45641863e-01 2.56970495e-01 6.74966201e-02 1.12280214e+00
5.19638471e-02 9.75672543e-01 9.45920229e-01 -2.95493096e-01
2.76710421e-01 1.21990748e-01 9.65807915e-01 7.10379124e-01
3.79714996e-01 4.64649685e-02 -1.49244398e-01 -4.55903113e-01
3.68242979e-01 6.10884249e-01 -5.82569361e-01 -5.24032950e-01
-1.28075969e+00 1.12372768e+00 7.05291808e-01 1.55704811e-01
-1.97769746e-01 -2.48584330e-01 6.21910512e-01 2.75452048e-01
5.12719274e-01 4.70769674e-01 -3.83043677e-01 1.92172062e-02
-7.05056489e-01 -3.89164202e-02 8.71586621e-01 9.04087603e-01
9.58166003e-01 -4.20765132e-01 -3.37506294e-01 8.98273826e-01
4.73623246e-01 -8.78975820e-03 6.68141544e-01 -3.62459004e-01
7.42837131e-01 1.07108283e+00 -9.00176466e-02 -1.00569046e+00
-1.94503456e-01 -3.21951807e-01 -6.87577069e-01 -1.59580037e-01
1.98209882e-01 -1.54044956e-01 -1.00599360e+00 1.60835195e+00
3.45408797e-01 2.75419980e-01 3.15168768e-01 6.87243521e-01
8.94366264e-01 7.50799954e-01 -1.04150381e-02 -6.93883151e-02
1.42470813e+00 -1.19458830e+00 -7.45209754e-01 -5.91395199e-01
7.54068673e-01 -3.43821049e-01 1.12413538e+00 -7.29644345e-03
-3.90276045e-01 -4.26356256e-01 -1.21556425e+00 -2.49151081e-01
-4.27570909e-01 8.52318555e-02 7.92744219e-01 5.57779908e-01
-6.59527481e-01 3.91124561e-02 -4.24798191e-01 -2.76954681e-01
5.85010052e-01 2.48250186e-01 -1.55886903e-01 -2.34943792e-01
-1.23257732e+00 4.22403693e-01 4.89679158e-01 -2.86265463e-01
-8.72107565e-01 -8.86656046e-01 -1.28812742e+00 8.25033709e-02
5.94326496e-01 -6.21121466e-01 1.23655331e+00 -6.07089400e-01
-8.73594284e-01 8.38198364e-01 -6.87592775e-02 -5.65684974e-01
5.45255303e-01 -4.75051433e-01 -2.84379631e-01 -7.09416270e-02
5.01300931e-01 6.64388120e-01 9.36010122e-01 -1.17414296e+00
-7.69427598e-01 -5.65487027e-01 3.80672187e-01 3.94970983e-01
-7.47009277e-01 -2.56656766e-01 -2.17416465e-01 -8.27623725e-01
3.53758484e-01 -6.99350834e-01 -3.33739072e-01 1.36727706e-01
-4.04712498e-01 -6.95164680e-01 1.03422952e+00 -4.88438338e-01
1.57098269e+00 -2.29125690e+00 -2.68028211e-02 -8.65529627e-02
5.72490215e-01 1.06142364e-01 2.28319705e-01 1.52797475e-01
1.43386215e-01 2.42246687e-02 -4.76376683e-01 -2.41062358e-01
9.51097459e-02 2.07702965e-01 -8.40841472e-01 4.58034456e-01
-9.10355151e-02 6.97809041e-01 -7.12192237e-01 -5.70337415e-01
1.18466757e-01 1.78375468e-01 -6.50010824e-01 4.97626096e-01
-1.61266357e-01 3.79077867e-02 -6.42948270e-01 7.53799677e-01
6.77702010e-01 -4.09479618e-01 -1.63408503e-01 1.27567708e-01
-3.27032153e-03 5.55463016e-01 -1.15422666e+00 1.56730950e+00
-5.97670913e-01 4.53296125e-01 -7.56808445e-02 -9.33323860e-01
1.11011505e+00 8.47789943e-02 2.75204003e-01 -1.75813705e-01
1.68919936e-01 2.88575560e-01 -3.56899440e-01 -2.06427529e-01
4.69919950e-01 -2.77759843e-02 -4.76881742e-01 5.53112745e-01
-5.12611456e-02 3.12225789e-01 -1.30809963e-01 1.92622747e-02
1.11840451e+00 -3.79466385e-01 7.20731914e-01 -3.68615627e-01
8.09499204e-01 -4.88191433e-02 7.57673800e-01 6.73881352e-01
-7.52538443e-01 6.79580390e-01 3.32451046e-01 -8.81202340e-01
-4.25931782e-01 -8.61800194e-01 -2.88075119e-01 1.19080973e+00
6.35974705e-01 -2.99749553e-01 -5.32854974e-01 -1.52581024e+00
3.19848843e-02 8.09087515e-01 -7.40062058e-01 -2.84531027e-01
-4.12713528e-01 -4.10303652e-01 4.12373364e-01 7.16109693e-01
7.54886746e-01 -7.98557460e-01 -5.61623394e-01 -2.37060159e-01
-5.00500917e-01 -7.73934662e-01 -1.03468835e+00 3.68050814e-01
-8.19395363e-01 -1.14985120e+00 -6.04015172e-01 -1.14875627e+00
8.14718962e-01 6.72616482e-01 9.18112934e-01 3.65371257e-02
-3.40404302e-01 2.99681753e-01 -3.99339229e-01 -1.44392401e-01
-1.52853206e-01 3.42500120e-01 2.05600336e-01 5.73276617e-02
6.36143029e-01 -2.87417620e-01 -7.21061170e-01 5.60453057e-01
-8.34777832e-01 -2.67775625e-01 5.20442605e-01 9.85016823e-01
4.98636872e-01 9.19849426e-02 6.81576550e-01 -7.00700104e-01
7.84717977e-01 -7.63399422e-01 -4.46494341e-01 3.17145526e-01
-7.58914351e-01 1.69845939e-01 7.58890390e-01 -4.57571417e-01
-1.05723238e+00 -6.56639785e-02 -1.53612047e-01 -5.66519439e-01
-4.00870293e-01 2.83266097e-01 -2.61921644e-01 -1.36508271e-01
5.99014163e-01 3.49994779e-01 -1.49349302e-01 -3.59233826e-01
3.59645128e-01 1.19268191e+00 3.72986436e-01 -3.00260812e-01
7.26556897e-01 5.70544600e-01 -6.51075006e-01 -6.89090371e-01
-1.35905278e+00 -1.08155966e+00 -6.23098850e-01 2.89349675e-01
6.75345659e-01 -9.43004847e-01 -3.15312296e-01 2.89055020e-01
-8.37463975e-01 1.33913994e-01 -3.01962316e-01 2.06689715e-01
-5.93915164e-01 7.59955466e-01 -3.35505843e-01 -6.60905480e-01
-7.40113735e-01 -1.07040429e+00 1.06680810e+00 3.44231725e-01
-3.93562913e-01 -1.16748774e+00 2.52965659e-01 7.05565393e-01
6.75477907e-02 -2.18253613e-01 7.75073171e-01 -1.30090547e+00
-5.56473672e-01 -4.63478774e-01 -3.63818288e-01 2.51395166e-01
4.08864409e-01 -5.51692188e-01 -8.50937545e-01 -2.61342108e-01
2.04586774e-01 -7.10937619e-01 1.14924335e+00 9.56578478e-02
1.08783770e+00 -6.15632176e-01 -6.18220508e-01 8.69164109e-01
1.23387587e+00 2.05641463e-01 4.10117924e-01 3.69147480e-01
7.00367928e-01 3.32835615e-01 8.51498604e-01 6.25330925e-01
5.52977860e-01 4.04869407e-01 3.27411860e-01 1.65087834e-01
2.20407665e-01 -4.83972698e-01 2.33926579e-01 3.18497419e-01
5.80677807e-01 -9.49260667e-02 -9.65039253e-01 7.76105583e-01
-1.75573623e+00 -7.81480730e-01 3.00953031e-01 2.23944950e+00
6.24462068e-01 2.68497944e-01 -6.89535886e-02 2.95369089e-01
7.86361396e-01 4.40992743e-01 -8.42107892e-01 -1.44976258e-01
3.37393790e-01 -4.66736794e-01 2.57951498e-01 2.84463167e-01
-1.67075861e+00 6.10765934e-01 6.51675224e+00 7.95835793e-01
-7.78932273e-01 2.03661807e-02 5.83896816e-01 4.35574919e-01
-5.09336948e-01 4.24109101e-02 -1.17344046e+00 3.42121571e-01
3.48644942e-01 -1.17291592e-01 3.86032425e-02 1.43040705e+00
-5.08217514e-01 2.54812837e-01 -1.24552584e+00 1.09738469e+00
4.65155900e-01 -1.17560756e+00 5.45239523e-02 1.68969929e-01
7.22041249e-01 -1.39638022e-01 -2.72429436e-02 9.40157056e-01
2.71607161e-01 -6.72083557e-01 4.43555415e-01 6.28125668e-02
6.14622712e-01 -6.54302120e-01 6.54168367e-01 6.59201324e-01
-1.32895100e+00 -6.14235044e-01 -3.68542641e-01 -1.82599470e-01
-1.26623780e-01 2.52150834e-01 -7.52888024e-01 3.27313423e-01
3.90235573e-01 8.21975470e-01 -4.78018999e-01 1.12884212e+00
-2.04750419e-01 5.14833689e-01 -1.90546706e-01 -6.95016533e-02
4.14976209e-01 -6.95175305e-02 5.41314662e-01 7.85752296e-01
1.29623273e-02 -5.11456728e-02 6.83711350e-01 9.48684752e-01
-2.41793826e-01 8.16841498e-02 -9.38359559e-01 -4.61512469e-02
4.62104827e-01 1.06474769e+00 -5.71747720e-01 -6.48853555e-02
-6.84452891e-01 1.02461708e+00 6.69297218e-01 3.95730771e-02
-9.73638237e-01 -5.62656939e-01 8.52754295e-01 -1.78472307e-02
2.98735589e-01 -4.72904444e-02 -4.28101599e-01 -1.44954228e+00
3.06316856e-02 -6.19758010e-01 1.24747849e+00 -9.49497521e-02
-1.72360420e+00 5.77948987e-01 1.44700659e-02 -1.37133181e+00
-6.70177303e-03 -5.12989044e-01 -9.02743280e-01 4.96096969e-01
-1.65368450e+00 -1.10959530e+00 -4.32543516e-01 3.89741153e-01
1.17596865e+00 -2.67868321e-02 6.35056973e-01 8.34801570e-02
-6.91639125e-01 8.88609946e-01 1.81781501e-01 2.97774851e-01
4.81527597e-01 -1.19148302e+00 5.33424795e-01 7.81158149e-01
1.39397860e-01 7.61052787e-01 2.73301393e-01 -6.45937264e-01
-1.22854877e+00 -1.17892754e+00 6.29057884e-01 -5.71385562e-01
6.35688901e-01 -4.97365415e-01 -1.00021994e+00 8.67110014e-01
-6.51805028e-02 3.42329472e-01 8.81952047e-01 4.01693106e-01
-6.87015831e-01 -2.06228234e-02 -1.16701925e+00 5.65753162e-01
1.05682087e+00 -3.87373418e-01 -1.31316435e+00 2.32256636e-01
1.01048744e+00 -2.34992743e-01 -5.84297299e-01 7.24424064e-01
2.59119749e-01 -7.87488639e-01 1.10669959e+00 -4.81187522e-01
1.51898071e-01 -1.22881569e-01 -2.47488052e-01 -9.60886180e-01
-2.81002730e-01 -4.70313102e-01 -4.14286971e-01 1.35149884e+00
2.07398117e-01 -8.69387686e-01 7.37767875e-01 6.74948215e-01
-2.07884714e-01 -1.05008268e+00 -9.76228237e-01 -7.43467510e-01
-2.04267371e-02 -2.40840524e-01 2.85478026e-01 9.11470115e-01
1.90367743e-01 6.58326626e-01 -2.24832967e-01 2.08134994e-01
6.07835889e-01 6.94988608e-01 6.25280559e-01 -1.42108786e+00
1.28513083e-01 -2.10863143e-01 -5.59073567e-01 -1.44359744e+00
4.69846040e-01 -7.44703948e-01 1.78253144e-01 -1.64338648e+00
4.35720861e-01 -8.38347077e-01 -4.32759851e-01 5.40329754e-01
-6.22822881e-01 -9.84479934e-02 -1.93729177e-01 5.55778444e-01
-9.76643980e-01 9.03368354e-01 7.53652930e-01 -4.18699533e-01
-3.77921730e-01 2.28942543e-01 -1.07753265e+00 9.36812341e-01
8.04420292e-01 -3.23690742e-01 -7.40275085e-01 -1.78712923e-02
-1.80669814e-01 -2.99539179e-01 2.50410020e-01 -1.16641867e+00
1.39568165e-01 -1.90700218e-01 5.70470914e-02 -9.44495022e-01
1.36375157e-02 -8.28927159e-01 -5.16141772e-01 4.29754049e-01
-5.20411134e-01 -3.82459939e-01 -1.93563945e-04 1.05409563e+00
-2.22610399e-01 -3.75712484e-01 6.76730812e-01 6.97410153e-03
-1.10106266e+00 4.83938187e-01 -2.07836670e-03 6.73888743e-01
1.50007772e+00 -3.56570840e-01 -3.39420944e-01 -1.49306893e-01
-4.53328304e-02 6.34597540e-01 5.04414737e-01 6.89845026e-01
8.35389555e-01 -1.12634528e+00 -3.12031567e-01 4.78349388e-01
6.37460053e-01 2.20454946e-01 3.29245239e-01 4.66317147e-01
-3.28180343e-01 6.06403470e-01 2.56695598e-01 -5.31999052e-01
-1.13185406e+00 1.10869980e+00 1.58870503e-01 -4.48175371e-01
-8.21246505e-01 9.28207040e-01 7.29864657e-01 -7.99927175e-01
6.73156023e-01 -2.44818345e-01 -4.13462847e-01 -1.67665128e-02
7.52598822e-01 3.34032565e-01 -9.49473977e-02 -6.12933576e-01
-5.11472940e-01 4.35358077e-01 -5.92520475e-01 5.14792681e-01
9.28934634e-01 -2.87170351e-01 3.20162326e-01 2.98634738e-01
1.33375132e+00 -1.94442019e-01 -1.27773929e+00 -2.83114374e-01
-8.77410248e-02 -6.73521757e-01 1.30873650e-01 -4.08420861e-01
-5.72839677e-01 7.51130581e-01 4.95748937e-01 2.94946015e-01
9.84036982e-01 1.76526666e-01 1.07121885e+00 5.12719691e-01
3.62475336e-01 -7.76186228e-01 1.61182791e-01 6.68867171e-01
7.27767706e-01 -1.63560641e+00 -1.50088504e-01 -5.02407491e-01
-7.57465541e-01 7.70645022e-01 1.00313544e+00 3.00691910e-02
5.82975388e-01 -2.47681271e-02 1.79074883e-01 -1.69175729e-01
-6.68533564e-01 -1.94120705e-01 4.65048224e-01 4.92869705e-01
1.42567515e-01 -8.77473727e-02 -2.41102204e-01 7.15856612e-01
1.76248491e-01 -2.24873170e-01 2.38014296e-01 1.06292534e+00
-6.73434556e-01 -6.06055140e-01 -5.48855782e-01 6.88964128e-01
-3.57832402e-01 -1.13399521e-01 -4.65789199e-01 4.84060228e-01
-1.44780815e-01 8.58109772e-01 -3.94738577e-02 -4.27091062e-01
5.40502295e-02 2.31693596e-01 -2.06852719e-01 -6.82383716e-01
-2.00900435e-01 -5.31627297e-01 -1.90918878e-01 -2.66626567e-01
1.12011001e-01 -5.09211183e-01 -1.20009279e+00 1.18877813e-01
-8.39023709e-01 2.88020968e-01 1.60689026e-01 8.54646981e-01
5.42701960e-01 2.60924935e-01 9.91437674e-01 -4.42264646e-01
-1.20610809e+00 -7.85420716e-01 -4.47702348e-01 4.89478767e-01
5.86862922e-01 -1.00750732e+00 -7.83304513e-01 -3.32474768e-01] | [12.347663879394531, 7.398104190826416] |
7fa23c56-3c48-4ca8-8b47-ae0cc60411d8 | efficient-multi-stage-inference-on-tabular | 2303.11580 | null | https://arxiv.org/abs/2303.11580v1 | https://arxiv.org/pdf/2303.11580v1.pdf | Efficient Multi-stage Inference on Tabular Data | Many ML applications and products train on medium amounts of input data but get bottlenecked in real-time inference. When implementing ML systems, conventional wisdom favors segregating ML code into services queried by product code via Remote Procedure Call (RPC) APIs. This approach clarifies the overall software architecture and simplifies product code by abstracting away ML internals. However, the separation adds network latency and entails additional CPU overhead. Hence, we simplify inference algorithms and embed them into the product code to reduce network communication. For public datasets and a high-performance real-time platform that deals with tabular data, we show that over half of the inputs are often amenable to such optimization, while the remainder can be handled by the original model. By applying our optimization with AutoML to both training and inference, we reduce inference latency by 1.3x, CPU resources by 30%, and network communication between application front-end and ML back-end by about 50% for a commercial end-to-end ML platform that serves millions of real-time decisions per second. | ['Igor L Markov', 'Daniel S Johnson'] | 2023-03-21 | null | null | null | null | ['automl'] | ['methodology'] | [-2.15737134e-01 3.29981774e-01 -6.35861158e-01 -9.65464413e-01
-8.93240035e-01 -1.05410361e+00 2.08975092e-01 -4.55484241e-02
-1.83107585e-01 5.52531660e-01 -2.85536408e-01 -1.06356966e+00
1.04738608e-01 -8.56496811e-01 -8.90611410e-01 -1.79036885e-01
1.01469129e-01 6.28115773e-01 2.65487600e-02 3.79669964e-01
-1.34990215e-01 6.17132545e-01 -1.56532514e+00 6.62589729e-01
6.33310080e-01 1.25553119e+00 -1.88283622e-01 1.06158626e+00
-4.46475059e-01 1.18876398e+00 -8.21272969e-01 -6.30562544e-01
3.06119770e-01 2.25166529e-01 -7.26714313e-01 -3.65461171e-01
5.95146060e-01 -4.56367284e-01 8.58070776e-02 6.44594550e-01
6.06208146e-01 -1.15257032e-01 -1.27340034e-01 -1.67227852e+00
1.82689443e-01 1.00465930e+00 -2.31431946e-01 -2.77938992e-01
2.60345340e-01 5.60957849e-01 1.11705065e+00 -6.09701097e-01
5.13652861e-01 1.19221783e+00 8.33884478e-01 1.63227439e-01
-1.68557227e+00 -5.80653489e-01 1.16846807e-01 -3.05216044e-01
-1.42593169e+00 -1.09452116e+00 8.57281685e-02 -2.66972363e-01
1.52092147e+00 6.65485680e-01 2.19698906e-01 6.77937686e-01
1.43915817e-01 8.56907427e-01 4.12292749e-01 9.72765535e-02
3.99293542e-01 3.74538690e-01 1.93290100e-01 6.45414293e-01
2.12527111e-01 -3.77226025e-01 -4.04835254e-01 -3.51364553e-01
3.30568731e-01 -5.17731644e-02 4.47339304e-02 -1.18179180e-01
-1.10090876e+00 4.84820485e-01 6.15085401e-02 -1.50139153e-01
-3.39709163e-01 4.88172740e-01 6.29384339e-01 3.95946383e-01
2.28154927e-01 4.71792608e-01 -1.14414883e+00 -6.25595093e-01
-1.16992366e+00 2.83259749e-01 1.92315590e+00 1.32269287e+00
1.03441715e+00 -1.30288184e-01 -8.08910429e-02 5.89646041e-01
3.35116267e-01 6.86510146e-01 -1.40086249e-01 -1.46465981e+00
6.73447669e-01 6.11822367e-01 -1.79434657e-01 -8.06795835e-01
-2.90572912e-01 -3.66733670e-01 -5.86087465e-01 1.31864712e-01
4.29659814e-01 -6.83385193e-01 -2.94934988e-01 1.32947183e+00
2.78036147e-01 -2.12736219e-01 -7.67168701e-02 4.79607224e-01
3.93483698e-01 7.12223351e-01 -1.45423040e-01 1.12977564e-01
1.24340451e+00 -1.28785574e+00 -5.00414371e-01 -5.62409818e-01
9.53467667e-01 -8.58144224e-01 1.26570737e+00 6.73032284e-01
-1.27465367e+00 -4.88781214e-01 -1.05672598e+00 -3.92768830e-01
-4.40710783e-01 1.46123990e-01 1.12128389e+00 4.95691657e-01
-1.27530801e+00 6.38537526e-01 -1.08461249e+00 -1.18550673e-01
2.41123095e-01 7.75973201e-01 -7.59844854e-02 2.83267740e-02
-6.02422953e-01 5.44633746e-01 1.09755941e-01 2.50619352e-01
-5.42395175e-01 -1.05088365e+00 -6.58768475e-01 3.80969375e-01
6.79149866e-01 -7.61298418e-01 1.75565362e+00 -8.43873501e-01
-2.00166035e+00 3.22644055e-01 -2.55258322e-01 -4.36030060e-01
3.42376113e-01 -7.18841627e-02 -5.81056952e-01 -2.88209438e-01
-1.29898071e-01 3.72567952e-01 5.73178232e-01 -6.07428253e-01
-7.80158937e-01 -1.37549505e-01 3.05435807e-01 -1.68652698e-01
-2.20156871e-02 -1.47743404e-01 -7.70455658e-01 2.19684988e-01
-2.74050772e-01 -8.46795022e-01 -1.67414367e-01 1.85221583e-01
-4.58333552e-01 -2.91678645e-02 8.69628549e-01 -4.74964529e-01
1.18563783e+00 -2.23512483e+00 -3.75179917e-01 4.60394830e-01
4.73317027e-01 -2.45617256e-02 -1.11532062e-01 3.82154197e-01
2.88860679e-01 2.70732582e-01 1.12747557e-01 -4.77343202e-01
4.99316543e-01 5.22041440e-01 -2.55735695e-01 3.03888589e-01
2.92761121e-02 1.17775655e+00 -6.04510188e-01 -4.70405221e-01
1.94336504e-01 4.52167243e-01 -9.41297293e-01 2.74539620e-01
-7.64137805e-01 -5.52121550e-02 -2.49079064e-01 9.10277903e-01
6.60955787e-01 -7.02409029e-01 6.55811727e-01 -3.12516600e-01
-1.15181409e-01 8.97570491e-01 -1.21655190e+00 1.73377824e+00
-1.07623684e+00 7.60210276e-01 8.17788720e-01 -5.76051295e-01
5.25478840e-01 1.44125476e-01 2.22934946e-01 -4.22021478e-01
-1.12227671e-01 1.35832652e-01 -1.88414991e-01 -3.46487314e-01
2.38759637e-01 5.91814756e-01 -2.35060215e-01 6.39980853e-01
-1.23096660e-01 -2.39008963e-01 1.87135383e-01 4.17401046e-02
1.39715195e+00 1.97578505e-01 4.89183993e-04 -9.37769786e-02
4.76026177e-01 -1.19938612e-01 5.89273751e-01 8.07161689e-01
2.04421356e-01 -1.42525986e-01 1.04429162e+00 -4.69348788e-01
-6.95542097e-01 -9.94756460e-01 -8.55243281e-02 1.58833206e+00
-5.98612070e-01 -9.01623905e-01 -8.31530988e-01 -7.03826189e-01
4.12583679e-01 1.02723908e+00 1.76066145e-01 2.40365341e-01
-4.76844519e-01 -5.20836949e-01 7.04023719e-01 3.94123495e-01
5.06981373e-01 -8.55948389e-01 -4.02532965e-01 4.68335390e-01
6.51399419e-02 -1.26008666e+00 -3.18950444e-01 2.62183607e-01
-7.54425883e-01 -7.78920412e-01 3.83790284e-01 -8.75352770e-02
3.74194503e-01 -1.45087272e-01 1.77741921e+00 -5.32857627e-02
-3.77678692e-01 4.76800837e-02 4.58703429e-01 3.34598087e-02
-4.81655180e-01 5.83915234e-01 -3.63064170e-01 -1.87648609e-01
5.30565381e-01 -8.27745080e-01 -2.96600521e-01 2.02659130e-01
-5.17234981e-01 1.76122203e-01 7.54491031e-01 2.79368341e-01
4.94081676e-01 6.02285601e-02 1.46731183e-01 -1.19410563e+00
4.26002890e-01 -5.13004422e-01 -1.39223480e+00 4.66401428e-02
-9.33424890e-01 3.58859688e-01 8.84869456e-01 -2.64064789e-01
-9.74217892e-01 3.13963264e-01 -3.04995775e-01 -1.37110651e-01
-8.21375996e-02 5.40436924e-01 -4.47173297e-01 2.55521405e-02
5.21171093e-01 -2.13596791e-01 3.79257351e-01 -5.30731142e-01
7.46470690e-01 1.02179146e+00 5.01597047e-01 -7.37903535e-01
3.44158620e-01 1.02556787e-01 -1.94168583e-01 -5.19916177e-01
-8.32146406e-01 -1.28192350e-01 -2.01509163e-01 1.38345897e-01
4.06662017e-01 -9.54771698e-01 -1.92096889e+00 -3.90248746e-02
-1.21138918e+00 -8.03762197e-01 -2.86578536e-01 7.67740607e-02
-5.18593073e-01 -2.72108465e-01 -8.09624910e-01 -5.52721441e-01
-6.04276419e-01 -1.45740938e+00 1.27378225e+00 -2.44928747e-01
-5.51724017e-01 -7.13137746e-01 -4.86988842e-01 3.23511928e-01
9.17479336e-01 -1.70877963e-01 7.32435644e-01 -5.56018949e-01
-1.21873868e+00 -5.30848503e-01 -4.58282977e-01 5.65625727e-01
-1.85138837e-01 4.00091201e-01 -1.16705370e+00 -3.83420847e-02
-1.25844836e-01 6.87288344e-02 1.00478649e-01 1.13204569e-01
1.53033984e+00 -7.21695006e-01 -4.55182821e-01 1.01639152e+00
1.47068655e+00 -1.67455480e-01 3.28676999e-01 -3.34864557e-01
8.77032220e-01 3.51799220e-01 2.44404703e-01 4.07539934e-01
3.66738558e-01 6.02626860e-01 3.24110478e-01 -2.36875951e-01
2.03695089e-01 -1.81660965e-01 6.09299898e-01 7.94042051e-01
5.59582114e-01 -2.09851041e-01 -9.52507794e-01 1.36350900e-01
-2.06343389e+00 -7.11521447e-01 -1.22743994e-01 2.25839901e+00
1.04373574e+00 4.92584854e-01 2.30583046e-02 -2.07097545e-01
9.21734348e-02 -7.63650835e-02 -8.34932268e-01 -6.42924905e-01
3.16843390e-01 2.12604314e-01 1.00603664e+00 9.16015506e-01
-6.45775855e-01 8.25648129e-01 6.85714293e+00 7.99436867e-01
-1.18957269e+00 1.30609959e-01 7.26991177e-01 -5.35458863e-01
-3.99801321e-02 1.00916825e-01 -1.15736067e+00 5.65413356e-01
1.93999577e+00 -6.26457855e-02 9.79701459e-01 1.49975646e+00
1.87723085e-01 -1.05992459e-01 -1.79380691e+00 8.77421319e-01
-4.48898256e-01 -1.42622888e+00 -4.77778286e-01 1.90111920e-01
1.61716849e-01 4.10885483e-01 -3.04659426e-01 6.21098995e-01
7.60905147e-01 -9.31790233e-01 3.47788662e-01 4.37030733e-01
8.57098460e-01 -8.14933062e-01 7.57833540e-01 4.09394592e-01
-1.07063878e+00 1.03642881e-01 -3.93257551e-02 -3.91094834e-01
1.69274304e-02 1.06078196e+00 -1.11795473e+00 -6.48159608e-02
5.65522552e-01 5.76122701e-01 -6.15698159e-01 3.65203708e-01
-1.62529483e-01 5.86800694e-01 -7.75566399e-01 -2.19721287e-01
-6.17991537e-02 -6.54586256e-02 -1.25692829e-01 1.24097836e+00
6.16240175e-03 -5.76482058e-01 9.89013165e-02 1.03671134e+00
-2.41868228e-01 -1.70504674e-01 -4.59773272e-01 -1.71513155e-01
7.00319946e-01 1.58995426e+00 -3.16090405e-01 -6.90778613e-01
-5.92184186e-01 9.39208448e-01 2.56541699e-01 4.74844605e-01
-1.01642668e+00 -7.52353787e-01 1.21896577e+00 2.43479431e-01
2.01177672e-01 1.74057968e-02 -3.52280021e-01 -9.62550282e-01
2.55770922e-01 -1.16749394e+00 -8.09139237e-02 -5.91802716e-01
-1.21697009e+00 3.78025413e-01 -2.60254920e-01 -4.52102780e-01
-6.09542549e-01 -5.13035059e-01 2.43758541e-02 1.15275109e+00
-1.12057126e+00 -6.62868142e-01 -2.64177352e-01 2.63556421e-01
3.64574417e-02 1.74453899e-01 9.18020129e-01 8.27618122e-01
-7.38650918e-01 7.49630332e-01 6.10719621e-02 -4.52404283e-02
5.95482290e-01 -1.18616199e+00 8.75227034e-01 4.85136032e-01
-2.26734266e-01 9.72964227e-01 4.23351645e-01 -2.52335966e-01
-2.19813633e+00 -1.07598865e+00 1.23897445e+00 -6.24193668e-01
8.35478485e-01 -1.01155210e+00 -5.69782972e-01 1.11154556e+00
9.43337902e-02 4.74235684e-01 6.34328008e-01 5.60167730e-01
-5.28340578e-01 -8.24797571e-01 -1.15423942e+00 6.76718116e-01
8.02231193e-01 -7.21883833e-01 3.03007901e-01 6.93963647e-01
1.02131152e+00 -5.72130144e-01 -1.27592897e+00 5.94203651e-04
6.45376325e-01 -9.13634241e-01 8.96773279e-01 -2.67588496e-01
2.17263147e-01 -3.95659715e-01 -2.17176333e-01 -5.44625700e-01
1.73635602e-01 -1.14575589e+00 -6.34995222e-01 1.12077916e+00
7.69942701e-01 -9.67109442e-01 1.08595371e+00 9.90701735e-01
-2.51367223e-02 -6.82698548e-01 -3.78517896e-01 -3.97698581e-01
-5.65548182e-01 -8.22155893e-01 1.00698471e+00 5.34023821e-01
2.90698372e-02 8.24214399e-01 2.17849374e-01 1.79311872e-01
4.41067308e-01 2.36937329e-01 1.14961445e+00 -1.03173423e+00
-1.03770888e+00 -4.45490271e-01 5.01320027e-02 -1.46612573e+00
1.16777621e-01 -9.52949882e-01 1.05240338e-01 -1.19729590e+00
-3.92030507e-01 -5.46915829e-01 1.98168680e-01 9.25381124e-01
4.30473506e-01 -1.40280426e-01 9.79825631e-02 -5.76498844e-02
-7.90762424e-01 -3.86577755e-01 5.69389343e-01 1.14414215e-01
-1.54086873e-01 3.46273243e-01 -6.88235700e-01 6.51088297e-01
6.39458716e-01 -3.42462897e-01 -5.24677217e-01 -6.65574610e-01
7.98685789e-01 4.51538503e-01 1.41205803e-01 -9.97834504e-01
4.36532766e-01 2.00854708e-02 1.56404361e-01 -5.76462865e-01
4.06848907e-01 -1.14366484e+00 4.66698468e-01 2.21774280e-01
-3.07318091e-01 1.29905984e-01 3.42869729e-01 9.03319269e-02
-9.12258700e-02 -1.14206271e-02 4.11679506e-01 -1.41234091e-02
-1.40285283e-01 3.15516025e-01 -4.47381467e-01 -8.69715121e-03
7.85128653e-01 2.41028503e-01 -4.79512811e-01 -3.18176508e-01
-6.65299356e-01 2.38732263e-01 4.35636103e-01 5.46739623e-02
-1.80739149e-01 -8.60267341e-01 -1.94579586e-01 5.70820093e-01
-2.30464727e-01 2.51987100e-01 -2.00690717e-01 9.90075231e-01
-9.25957441e-01 6.48790121e-01 4.96577293e-01 -6.10563159e-01
-1.05069554e+00 5.26921093e-01 4.81807142e-01 -2.08250865e-01
-3.83762240e-01 8.63231540e-01 -3.15260172e-01 -8.77161264e-01
5.52767277e-01 -8.28080058e-01 9.27229702e-01 -1.75280556e-01
4.51839864e-01 4.53904003e-01 4.44279999e-01 3.08540553e-01
-3.43726277e-01 8.06576237e-02 7.85868615e-02 5.24783926e-03
1.10114038e+00 1.06740728e-01 -6.74244940e-01 6.15850806e-01
1.55491507e+00 2.19831079e-01 -1.17939854e+00 -8.56200829e-02
1.21294178e-01 -6.45649135e-02 1.68394297e-01 -1.10446417e+00
-1.28814161e+00 8.40685666e-01 3.43171805e-02 3.47303331e-01
1.04966104e+00 -2.40499243e-01 6.40816212e-01 8.23194444e-01
4.03676987e-01 -9.50605035e-01 -7.66770482e-01 2.56141305e-01
2.84337968e-01 -9.38477576e-01 1.07406773e-01 -7.13635325e-01
-1.03833355e-01 9.68303978e-01 4.40285802e-01 1.23615608e-01
7.42250800e-01 1.21664536e+00 -6.97470531e-02 -3.30817513e-02
-1.41339242e+00 4.38635081e-01 -2.68643141e-01 1.84317589e-01
6.22989893e-01 2.08105937e-01 1.46417648e-01 5.30226886e-01
-4.08518046e-01 3.90506566e-01 2.00892910e-01 9.05546129e-01
-3.30000557e-03 -1.18104994e+00 1.21985584e-01 7.10224569e-01
-5.85167170e-01 -3.89741659e-01 2.42143124e-02 8.41324270e-01
1.94716193e-02 8.80588651e-01 5.75367391e-01 -3.11179221e-01
3.40370208e-01 1.88582063e-01 1.84404328e-02 -6.20281100e-01
-9.92836833e-01 -9.60006639e-02 6.13296986e-01 -1.27876198e+00
4.79003757e-01 -3.47759575e-01 -1.18179786e+00 -1.12650180e+00
1.04414657e-01 -4.04728614e-02 1.15457904e+00 4.85833406e-01
1.22449446e+00 7.28718817e-01 2.72161543e-01 -6.43694401e-01
-8.64636898e-01 -7.07309783e-01 -2.48336717e-01 -3.68471980e-01
2.74170727e-01 7.68538937e-02 -5.51935852e-01 1.80094410e-02] | [8.74677848815918, 3.8009073734283447] |
645f933a-6b88-4059-ac03-58c20d872751 | rise-of-the-machines-intraday-high-frequency | 2009.04200 | null | https://arxiv.org/abs/2009.04200v1 | https://arxiv.org/pdf/2009.04200v1.pdf | Rise of the Machines? Intraday High-Frequency Trading Patterns of Cryptocurrencies | This research analyses high-frequency data of the cryptocurrency market in regards to intraday trading patterns related to algorithmic trading and its impact on the European cryptocurrency market. We study trading quantitatives such as returns, traded volumes, volatility periodicity, and provide summary statistics of return correlations to CRIX (CRyptocurrency IndeX), as well as respective overall high-frequency based market statistics with respect to temporal aspects. Our results provide mandatory insight into a market, where the grand scale employment of automated trading algorithms and the extremely rapid execution of trades might seem to be a standard based on media reports. Our findings on intraday momentum of trading patterns lead to a new quantitative view on approaching the predictability of economic value in this new digital market. | ['Wolfgang Karl Härdle', 'Raphael C. G. Reule', 'Alla A. Petukhina'] | 2020-09-09 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-8.83245051e-01 -1.77084133e-01 7.65888542e-02 -5.15093915e-02
-5.12747467e-01 -1.15604746e+00 1.25461233e+00 2.74100184e-01
-3.75673681e-01 6.68763578e-01 2.35795960e-01 -7.57655740e-01
-3.81298184e-01 -1.07941091e+00 -1.65725231e-01 -5.55892944e-01
-8.61560941e-01 5.56504786e-01 3.01288664e-01 -6.37134433e-01
6.08598292e-01 5.42528987e-01 -9.50759411e-01 1.59249231e-01
2.23659828e-01 1.61272991e+00 -6.19285882e-01 4.04141665e-01
-5.07197045e-02 7.37687051e-01 -7.08719492e-01 -6.08296871e-01
6.94838881e-01 -3.53009403e-01 -1.52100995e-01 -3.13479275e-01
-4.26721781e-01 -3.13297689e-01 3.60251293e-02 1.10333145e+00
-2.86220521e-01 -3.30700129e-01 4.32200581e-01 -7.53021002e-01
-2.43650571e-01 9.59945798e-01 -5.25070965e-01 8.94059241e-01
-5.13278395e-02 1.27196163e-01 1.43478167e+00 -6.23810947e-01
5.88758588e-01 5.37273765e-01 4.68216151e-01 -6.72984242e-01
-1.06307065e+00 -5.63667953e-01 -2.42415175e-01 -1.40746102e-01
-6.89724743e-01 1.65180862e-01 5.65237224e-01 -5.49136221e-01
6.85550153e-01 4.53216940e-01 1.16415489e+00 4.64490384e-01
7.99240530e-01 6.89636320e-02 1.49983358e+00 -1.70704708e-01
3.45614076e-01 1.35896742e-01 1.96029618e-01 -3.63381982e-01
8.32510769e-01 6.21661782e-01 -5.79684436e-01 -4.49830860e-01
6.15446270e-01 -2.26788014e-01 -8.45663175e-02 1.06263213e-01
-1.35232759e+00 1.03177238e+00 1.64958104e-01 7.42085576e-01
-7.95634687e-01 -8.81206393e-02 6.10602081e-01 1.17608941e+00
9.36330497e-01 4.54423428e-01 -1.00250924e+00 -6.25902414e-01
-9.65864837e-01 7.03821182e-01 1.37049007e+00 4.08656657e-01
3.39324027e-01 2.25601539e-01 1.96825951e-01 1.02810033e-01
2.77511209e-01 5.97388089e-01 5.75222909e-01 -6.39891207e-01
5.43335378e-01 2.80581504e-01 2.55594611e-01 -7.65300155e-01
-2.86749661e-01 -4.96410936e-01 -4.01509047e-01 5.30963063e-01
8.34746122e-01 -2.07093418e-01 9.70388576e-02 9.72720802e-01
1.72631815e-02 -3.75431776e-01 1.18646376e-01 5.49792945e-01
-9.50946808e-02 4.63158488e-01 -4.36351150e-01 -7.29233980e-01
1.52986062e+00 -2.11917266e-01 -7.26379097e-01 5.45478046e-01
3.23752016e-01 -1.12724614e+00 3.98131549e-01 8.41984808e-01
-1.13508654e+00 1.71719696e-02 -9.50760365e-01 7.48916388e-01
-4.58983779e-01 -5.58213830e-01 7.11505294e-01 7.09342003e-01
-8.64409924e-01 9.80151951e-01 -9.19628143e-01 3.28863412e-01
-3.07509869e-01 2.04553738e-01 2.40417778e-01 1.10951602e+00
-1.20476151e+00 9.89357054e-01 5.56002438e-01 2.90226609e-01
-2.16205135e-01 -7.25138009e-01 -3.71903181e-01 -5.57385897e-03
3.60673666e-01 2.81030923e-01 1.36326742e+00 -6.25854313e-01
-1.45468235e+00 5.89245021e-01 7.74075866e-01 -1.27515578e+00
9.27154541e-01 -2.54285205e-02 -9.22377825e-01 8.43854621e-02
-3.23055476e-01 -4.19633180e-01 3.18203539e-01 -4.72649306e-01
-7.45846748e-01 -3.07072461e-01 -4.25432175e-01 -3.86813909e-01
2.07023621e-01 2.16502205e-01 5.94469845e-01 -1.26799548e+00
4.78332520e-01 -6.56711876e-01 3.33101023e-03 -8.78887355e-01
1.24527246e-01 -7.74914846e-02 5.61659932e-01 -8.77078354e-01
1.42810273e+00 -1.76143408e+00 -7.13203013e-01 7.36693323e-01
1.05883650e-01 -4.12069470e-01 5.68228424e-01 8.97094846e-01
-2.43067443e-01 9.90315527e-02 1.95491403e-01 7.14160085e-01
3.59648168e-01 -2.22773254e-01 -9.78381753e-01 7.35117733e-01
9.40999687e-02 1.09636652e+00 -3.93892556e-01 1.93053111e-01
-3.58758345e-02 -1.82313100e-01 -3.46050829e-01 1.17530271e-01
-5.05220711e-01 2.48789579e-01 -4.84238446e-01 8.64605725e-01
6.51908040e-01 -1.56065866e-01 2.83615172e-01 1.79411881e-02
-9.21674848e-01 7.17920423e-01 -9.14775729e-01 5.17266810e-01
7.76619539e-02 6.17520034e-01 -1.62087500e-01 -4.48891073e-01
1.10854280e+00 2.91945249e-01 5.54285526e-01 -1.13670444e+00
2.46466115e-01 7.76727438e-01 4.30371970e-01 2.74282604e-01
7.88447917e-01 -2.37141520e-01 -1.59863874e-01 1.14396548e+00
-3.69969755e-01 -1.81879550e-01 4.15390998e-01 -2.06462324e-01
7.51370668e-01 1.78917069e-02 5.34262359e-01 -8.27961504e-01
1.89589068e-01 4.79854196e-02 1.23318665e-01 1.18624493e-01
1.98785871e-01 1.78585630e-02 1.14856625e+00 -7.75693834e-01
-1.06389606e+00 -9.64421630e-01 -6.53949261e-01 4.23707128e-01
-4.05761808e-01 -2.27322444e-01 -4.11182761e-01 -1.89478308e-01
3.44326138e-01 4.41251904e-01 -7.55158126e-01 3.49944383e-01
-5.47408462e-01 -1.35776877e+00 -7.87255839e-02 -6.54745996e-02
4.09226209e-01 -1.14876497e+00 -1.07853615e+00 5.07805824e-01
5.31431079e-01 -7.45473623e-01 -3.59998137e-01 5.07141829e-01
-1.16406333e+00 -1.24733496e+00 -7.01612592e-01 3.88547480e-02
6.91891648e-03 -4.13915515e-01 1.44854462e+00 -3.13698560e-01
8.20287913e-02 2.21963804e-02 -3.46284151e-01 -5.60835004e-01
-4.91815865e-01 -2.67379969e-01 -1.64344370e-01 1.56609148e-01
5.94191790e-01 -6.97011232e-01 -7.43593156e-01 3.65586668e-01
-8.95610392e-01 -7.09076881e-01 4.16081816e-01 3.94212216e-01
3.11797738e-01 1.51039705e-01 6.20184243e-01 -6.84849381e-01
9.25521135e-01 -5.57649791e-01 -1.59708655e+00 1.22117206e-01
-1.27402151e+00 1.16472527e-01 1.03017263e-01 -1.77887470e-01
-9.01487112e-01 -9.42517161e-01 5.08798897e-01 1.78011790e-01
5.71430266e-01 7.97609389e-01 5.57979286e-01 5.17322187e-05
9.28529799e-02 3.63680154e-01 8.44050348e-02 -7.60132015e-01
5.85875437e-02 2.50102669e-01 4.49970812e-01 -5.78621328e-01
8.25890303e-01 3.94476533e-01 -2.21352562e-01 -5.41425347e-01
-1.56277046e-01 -2.86153525e-01 -4.39076483e-01 -1.06402330e-01
5.30704796e-01 -8.05564761e-01 -1.28006899e+00 8.20482910e-01
-9.56956029e-01 -1.43809974e-01 -4.32072997e-01 8.99862170e-01
-4.38235611e-01 9.65804383e-02 -1.12121558e+00 -1.22536135e+00
-2.62692958e-01 -6.59179807e-01 2.93057829e-01 -1.06065944e-01
-5.67574084e-01 -1.24330902e+00 8.15993011e-01 -9.34336036e-02
7.27624297e-01 7.42743552e-01 7.81167209e-01 -1.40190494e+00
-8.35698664e-01 -1.68016583e-01 -1.18495747e-01 3.20859939e-01
1.45354465e-01 1.36720076e-01 -6.38047278e-01 -9.23439413e-02
8.19395125e-01 -2.58783367e-03 6.20628297e-01 2.96608716e-01
8.59527439e-02 -5.73269725e-01 5.67771196e-01 2.24864930e-01
1.49471712e+00 5.93933702e-01 3.03142697e-01 9.69900727e-01
-4.36487556e-01 7.89212346e-01 8.30744207e-01 1.00980842e+00
-7.32952403e-03 3.72632861e-01 2.87225455e-01 4.99417514e-01
6.01248384e-01 1.33710727e-01 5.56425571e-01 1.29045093e+00
-4.28494424e-01 4.45585787e-01 -7.57187188e-01 2.99424231e-01
-1.29122436e+00 -9.70776558e-01 -2.52606988e-01 2.35618734e+00
7.41353095e-01 7.87045062e-01 5.13192356e-01 3.95512022e-03
4.76463258e-01 5.71358144e-01 -1.29240051e-01 -3.97003859e-01
-3.59347284e-01 5.47093570e-01 9.27876949e-01 2.18775272e-01
-7.43050456e-01 3.00688177e-01 7.20963669e+00 4.62265521e-01
-1.17053127e+00 -1.47665948e-01 8.33637238e-01 -2.18957275e-01
-6.04685247e-01 5.66154495e-02 -6.54812038e-01 8.02381337e-01
1.45108175e+00 -7.57308543e-01 4.24676895e-01 5.47986746e-01
2.98622876e-01 -3.32273901e-01 -8.16153347e-01 5.99893868e-01
-5.84661365e-01 -1.63437593e+00 -1.05962522e-01 8.85840237e-01
5.62927246e-01 2.50396013e-01 2.03607738e-01 -3.99787240e-02
2.52287596e-01 -5.84632993e-01 1.12485576e+00 6.17517173e-01
-4.14010696e-02 -9.99755144e-01 1.04210365e+00 -1.82153761e-01
-1.01594770e+00 2.52309084e-01 -5.73046394e-02 -1.84131563e-01
1.35530382e-01 7.80679643e-01 -3.64422858e-01 5.97344935e-01
7.53092885e-01 5.24712265e-01 -3.43862206e-01 6.35496140e-01
3.84030789e-01 6.01846933e-01 -5.36445737e-01 -7.06714913e-02
5.49631357e-01 -1.20458865e+00 2.45454699e-01 7.96851635e-01
5.64669609e-01 8.60927068e-03 -6.10418379e-01 8.97605836e-01
1.42162398e-01 1.50642216e-01 -4.16979581e-01 -5.08723140e-01
2.53089249e-01 1.03172433e+00 -1.15361249e+00 -2.43020505e-01
-7.27334678e-01 2.40589529e-01 -5.19034147e-01 6.31202608e-02
-5.82134783e-01 -1.19010851e-01 5.90669870e-01 4.20438915e-01
5.62030375e-01 -3.41504633e-01 -6.48883522e-01 -1.11969209e+00
4.43880141e-01 -9.64465320e-01 5.19834876e-01 -9.08395126e-02
-1.15452862e+00 6.47519410e-01 -1.14533722e-01 -1.55492342e+00
-6.42679811e-01 -6.80049539e-01 -8.97699177e-01 8.38381648e-01
-1.49256194e+00 -1.97018564e-01 6.72126889e-01 5.53809166e-01
-1.16124064e-01 -6.45990312e-01 5.45099556e-01 -1.15979239e-01
1.92962348e-01 -1.27239972e-01 4.72579658e-01 1.56978011e-01
1.74501777e-01 -1.44478214e+00 8.88551295e-01 4.22835171e-01
6.34701014e-01 5.42965889e-01 9.65216756e-01 -9.44784760e-01
-1.18688178e+00 -2.55420893e-01 8.83237183e-01 -5.78149378e-01
1.85443938e+00 -4.53374237e-02 -8.42632174e-01 5.65525591e-01
6.58086836e-01 -3.72946024e-01 6.65890515e-01 -1.22941751e-02
-1.72528744e-01 -2.72225022e-01 -7.44069815e-01 3.18154991e-01
4.74199979e-03 -4.76499707e-01 -1.07422745e+00 3.96374129e-02
7.14165330e-01 -3.51966429e-03 -1.36436927e+00 2.24536210e-01
8.85865808e-01 -1.45081937e+00 3.65911216e-01 -2.03976065e-01
-5.84450550e-03 8.18799958e-02 -1.20392241e-01 -1.00517130e+00
2.90801108e-01 -1.24271905e+00 2.01588720e-02 8.54386866e-01
4.92737502e-01 -1.29493272e+00 6.63277686e-01 6.38980389e-01
5.29126346e-01 -4.04780269e-01 -1.22957134e+00 -9.12848294e-01
2.04980746e-01 -4.79974568e-01 8.38393986e-01 9.42838788e-01
2.05806270e-01 -4.85597074e-01 -1.09292850e-01 -4.12353426e-01
6.96499705e-01 6.94372237e-01 4.18479383e-01 -1.33029068e+00
-6.02977693e-01 -9.22770858e-01 -2.96282321e-01 -3.13518167e-01
-5.77141047e-01 -6.29561305e-01 -9.04651642e-01 -9.85946059e-02
-3.86647910e-01 1.57535486e-02 -7.32811809e-01 -4.10899341e-01
7.24386811e-01 2.05720440e-01 3.32651198e-01 5.94637394e-01
-9.73752439e-02 1.99648008e-01 1.01043904e+00 2.20091224e-01
-1.38382033e-01 4.16433126e-01 -2.44519636e-01 5.39515436e-01
6.94784760e-01 -2.16401383e-01 1.72348797e-01 4.28096712e-01
8.92086864e-01 5.56050718e-01 2.06041396e-01 -6.07788503e-01
1.00310996e-01 -1.69558316e-01 1.89410448e-01 -1.08901525e+00
-2.19672099e-01 -7.37147391e-01 4.88020718e-01 9.37220633e-01
-1.37402043e-02 8.01379681e-01 -4.95056398e-02 3.46322209e-01
-9.51288700e-01 3.16535272e-02 3.18736970e-01 -3.60861361e-01
-3.05125266e-01 1.88481048e-01 -3.33611727e-01 1.07469790e-01
1.18224895e+00 -1.21284217e-01 -2.06716150e-01 -2.61725694e-01
-9.26534951e-01 1.07871234e-01 4.28787768e-01 3.31583768e-01
1.18009271e-02 -1.18028843e+00 -8.34682345e-01 2.61673898e-01
-2.15536594e-01 -7.50202298e-01 -3.21008891e-01 1.06368375e+00
-1.08493924e+00 8.16179454e-01 -2.80991048e-01 -1.25611112e-01
-6.24928117e-01 2.34205335e-01 3.49410087e-01 -6.98380828e-01
-6.10121131e-01 1.10391691e-01 -7.89499581e-02 7.73466080e-02
-1.23372257e-01 -8.40486765e-01 -7.53398165e-02 1.02557814e+00
6.36332929e-01 4.21160996e-01 3.80709410e-01 -1.71219945e-01
-1.88516870e-01 2.92901725e-01 -3.39040570e-02 -4.72405314e-01
1.68119097e+00 1.20283537e-01 -6.99149668e-01 9.93419409e-01
9.43363726e-01 2.93421358e-01 -1.27681577e+00 5.14236391e-02
1.32078338e+00 -2.99829692e-01 -3.95700514e-01 -7.06952035e-01
-1.16331255e+00 3.28547925e-01 7.72114322e-02 1.08885229e+00
6.34596348e-01 -2.79809069e-02 6.58669889e-01 3.53447258e-01
6.62684083e-01 -1.33657420e+00 -4.40028340e-01 4.25190300e-01
9.92822051e-01 -9.50238705e-01 5.96143976e-02 2.33169377e-01
-4.13908392e-01 1.50214136e+00 -4.21596199e-01 -6.48648143e-01
1.52194238e+00 5.64368904e-01 2.39605367e-01 -4.16941285e-01
-1.06364143e+00 2.61875510e-01 3.75834256e-02 -2.58938462e-01
4.97592211e-01 3.56973737e-01 -9.64391112e-01 7.08640993e-01
-7.69945383e-01 -2.80824095e-01 5.94966471e-01 7.03048348e-01
-4.18939322e-01 -1.24260032e+00 -5.57683051e-01 5.04375458e-01
-1.17511559e+00 -2.93884486e-01 -3.96855086e-01 1.38035095e+00
-5.09384036e-01 3.91483188e-01 5.66495121e-01 6.01322576e-03
2.33758152e-01 8.53644609e-02 1.17063262e-01 -6.52211159e-02
-1.34713590e+00 6.71676993e-01 1.42495438e-01 -5.94189644e-01
-2.64629990e-01 -1.00756085e+00 -8.34024847e-01 -6.63707376e-01
-1.37035653e-01 7.35170126e-01 7.15628386e-01 7.20166326e-01
-2.96915889e-01 7.50039518e-02 9.18377280e-01 -5.19919932e-01
-1.19812417e+00 -9.90991592e-01 -1.55545640e+00 2.63279080e-02
4.74153250e-01 -1.10116333e-01 -7.19682813e-01 -3.02441120e-01] | [4.666437149047852, 4.12335729598999] |
178ee774-825b-4596-b0fb-eb8413e16a67 | the-nexus-between-job-burnout-and-emotional | 2208.04843 | null | https://arxiv.org/abs/2208.04843v1 | https://arxiv.org/pdf/2208.04843v1.pdf | The Nexus between Job Burnout and Emotional Intelligence on Turnover Intention in Oil and Gas Companies in the UAE | Currently, job satisfaction and turnover intentions are the significant issues for oil and gas companies in the United Arab Emirates (UAE). These issues need to be addressed soon for the performance of the oil and gas companies. Thus, the aim related to the current study is to examine the impact of job burnout, emotional intelligence, and job satisfaction on the turnover intentions of the oil and gas companies in the UAE. The goals of this research also include the examination of mediating the influence of job satisfaction alongside the nexus of job burnout and turnover intentions of the oil and gas companies in the UAE. The questionnaire method was adopted to collect the data from the respondents, and Smart-PLS were employed to analyse the data. The results show that job burnout, emotional intelligence, and job satisfaction have a positive association with turnover intentions. In contrast, job satisfaction positively mediates the nexus between job burnout and turnover intentions. These results provide the guidelines to the policymakers that they should enhance their focus on job satisfaction and turnover intentions of the employees that improve the firm performance. | ['Norziani Dahalan', 'Mohd Faiz Hilmi', 'Anas Abudaqa'] | 2022-08-06 | null | null | null | null | ['emotional-intelligence'] | ['natural-language-processing'] | [-8.01758945e-01 2.80610293e-01 -5.31564713e-01 -9.65959504e-02
2.81272769e-01 2.08108455e-01 -4.82207537e-02 -1.38120264e-01
-4.38184530e-01 3.05011898e-01 3.00628901e-01 -3.57053638e-01
-4.60101217e-01 -7.51876891e-01 2.03179076e-01 -7.50011265e-01
4.42799896e-01 1.99036598e-01 -4.74131823e-01 -5.80065429e-01
8.21467459e-01 5.30683637e-01 -8.93722296e-01 -4.75602895e-01
6.91891134e-01 4.34301257e-01 5.25192738e-01 -1.02235392e-01
2.16973528e-01 9.84766901e-01 -1.99651271e-01 -3.94386649e-01
-9.26741958e-02 -3.21971625e-01 -8.21854115e-01 6.78896427e-01
-6.05646193e-01 -4.41500068e-01 7.70249292e-02 6.39962971e-01
5.24364591e-01 1.99979514e-01 1.93229496e-01 -1.14232969e+00
-6.48353100e-01 3.13203841e-01 -1.21125713e-01 5.51854409e-02
-1.10086218e-01 1.18998468e-01 8.66153002e-01 -1.16611850e+00
2.48120859e-01 9.99342382e-01 6.11588717e-01 1.82437357e-02
-8.88015807e-01 -9.90776896e-01 -4.41108525e-01 2.76866525e-01
-1.09085357e+00 -6.55691803e-01 6.80889249e-01 -6.23929501e-01
8.71484041e-01 -2.58188605e-01 7.02041388e-01 3.60440135e-01
4.21406955e-01 -3.27311307e-01 1.46885002e+00 -5.01714885e-01
6.50204122e-02 8.48236084e-01 4.57211673e-01 4.97481115e-02
8.08986068e-01 8.43042582e-02 -1.79200456e-01 2.96270341e-01
8.64854693e-01 1.96856588e-01 8.82197469e-02 1.20980702e-02
-5.28272271e-01 1.22911334e+00 1.14524595e-01 7.83639729e-01
-8.70562315e-01 -1.90772757e-01 4.55426455e-01 2.98021406e-01
-5.66920079e-02 5.15423775e-01 -1.76081806e-01 -5.44308782e-01
-4.13179338e-01 -2.04746455e-01 7.38281488e-01 1.84011772e-01
6.50929511e-01 3.84244531e-01 1.68832079e-01 5.30480027e-01
4.87092763e-01 7.19198227e-01 1.46987513e-01 -1.05067539e+00
-2.41232723e-01 7.53231108e-01 4.84059900e-01 -1.30028558e+00
-2.77614921e-01 -8.44006062e-01 -3.01270217e-01 1.44283146e-01
-1.28416732e-01 -1.93538308e-01 -2.61747420e-01 1.10047054e+00
-2.95163572e-01 -5.22238076e-01 3.03168148e-01 9.43537772e-01
4.13046539e-01 4.29192573e-01 2.86276072e-01 -7.08893657e-01
1.32487130e+00 -9.43527341e-01 -1.27285028e+00 -6.80380464e-01
2.19580188e-01 -5.91879845e-01 1.25418532e+00 -3.54927063e-01
-1.38060021e+00 -1.10789728e+00 -4.71606761e-01 5.97196639e-01
1.11578919e-01 2.41241604e-01 5.38050950e-01 8.96361411e-01
-6.17690384e-01 3.81929614e-02 -7.76994228e-01 -4.88758385e-01
-9.79249999e-02 3.99148673e-01 -2.30524048e-01 9.14683416e-02
-1.16040695e+00 1.28981626e+00 4.41123903e-01 4.30899650e-01
-2.32679173e-01 1.16522148e-01 -4.77366030e-01 2.31760502e-01
3.63675207e-01 -4.67712253e-01 9.10852611e-01 -1.11543322e+00
-8.75199854e-01 5.66618741e-01 -1.06081381e-01 6.82478100e-02
-4.58423525e-01 -5.90398815e-03 -3.73962104e-01 -6.70329630e-02
6.39843464e-01 -1.05246022e-01 -9.23597291e-02 -1.28377604e+00
-9.34902608e-01 -7.83595741e-01 -3.90302926e-01 1.77365109e-01
-5.62626600e-01 2.69748449e-01 4.60776202e-02 1.06314898e-01
2.90748447e-01 -1.08001983e+00 -1.69441149e-01 -1.54415047e+00
4.92646307e-01 -5.84413588e-01 5.44440746e-01 -7.77065337e-01
1.23116326e+00 -2.38819790e+00 -6.13572538e-01 1.85745507e-01
-2.87135929e-01 -3.21973450e-02 3.21348965e-01 8.84756088e-01
-7.33956918e-02 1.51817009e-01 4.92516309e-01 3.58260602e-01
8.93675536e-02 7.00934708e-01 1.36430547e-01 3.59810054e-01
2.68049270e-01 7.37430453e-01 -7.16192782e-01 -1.04642503e-01
1.01260751e-01 5.44852465e-02 1.17870316e-01 3.54809225e-01
8.29460263e-01 2.85179138e-01 -6.07623875e-01 5.76957285e-01
6.79582596e-01 8.51801038e-03 2.95895010e-01 1.86496958e-01
-8.91283691e-01 5.55201888e-01 -5.64856291e-01 3.20612162e-01
-1.66856825e-01 2.40582913e-01 4.09198850e-01 -1.04509139e+00
1.72805178e+00 6.99195743e-01 5.56512654e-01 -1.19312716e+00
1.52094930e-01 6.72745824e-01 3.76851082e-01 -8.39528501e-01
7.13149548e-01 -8.64296198e-01 1.58407260e-02 1.16952680e-01
-6.52040958e-01 2.02299267e-01 2.91430831e-01 -1.81452304e-01
5.77133834e-01 -1.77944362e-01 4.52039272e-01 -1.16664797e-01
5.88197887e-01 1.27922446e-01 9.32546794e-01 2.42704064e-01
-5.11203527e-01 -7.44755924e-01 4.46468383e-01 -7.65053183e-02
-7.89880633e-01 -1.28615618e-01 -1.33991584e-01 7.63768792e-01
1.24620214e-01 3.23046088e-01 -2.14591220e-01 -1.08586229e-01
-2.91954368e-01 1.04154122e+00 1.49157643e-01 -1.28147349e-01
-1.43484682e-01 -3.87792408e-01 -4.28184420e-02 3.97793412e-01
9.40524817e-01 -1.16271460e+00 -1.12135470e+00 2.69939363e-01
-1.55235648e-01 -1.01233566e+00 2.35252529e-01 1.65404812e-01
-1.02448797e+00 -6.60077453e-01 -1.17398063e-02 -9.26627815e-01
6.72164023e-01 6.49224639e-01 8.41972470e-01 2.69621044e-01
3.42341870e-01 4.04559046e-01 -2.24424243e-01 -4.89066482e-01
-4.63412404e-01 -1.00276478e-01 6.66018575e-02 -5.62426925e-01
7.66849935e-01 -3.18009615e-01 -1.67670235e-01 6.05260074e-01
-5.55636346e-01 -5.14755957e-02 8.59946012e-01 7.64786005e-01
8.18244666e-02 1.05554771e+00 1.28178370e+00 -5.84909797e-01
8.27190638e-01 -6.59621954e-01 -4.70042489e-02 -1.26692995e-01
-1.53147423e+00 -8.08065355e-01 1.90532357e-02 2.98254400e-01
-1.39107132e+00 -5.06054699e-01 -1.41218781e-01 3.60342294e-01
-1.75839886e-01 1.16811931e+00 8.71339440e-03 3.71183753e-01
1.24641806e-01 -3.00051481e-01 5.12682080e-01 -1.12905674e-01
-6.31041229e-01 9.29019749e-01 5.15334606e-01 -1.61679965e-02
6.45431638e-01 4.54825424e-02 9.43213180e-02 -7.23592699e-01
-6.39211237e-01 -8.54842007e-01 -3.81805629e-01 -6.48048818e-01
8.93227935e-01 -1.02942550e+00 -1.11047316e+00 -1.87544718e-01
-4.94178027e-01 9.64366198e-02 1.53118595e-01 1.07471585e+00
-3.57369632e-01 6.10898854e-03 -3.48410755e-01 -1.43053746e+00
-1.20382413e-01 -1.17720723e+00 -2.03840248e-02 5.82207799e-01
-3.72904211e-01 -9.77412999e-01 -2.38518193e-01 1.10886693e+00
5.40720165e-01 7.98705071e-02 1.12586284e+00 -3.87852043e-01
-1.36118248e-01 -1.06408268e-01 -1.12830557e-01 3.75793815e-01
3.53099167e-01 -2.82605499e-01 -2.09595725e-01 -4.50731456e-01
8.07261050e-01 -4.09315020e-01 2.79841959e-01 2.42028281e-01
-3.63607407e-02 -3.52959931e-01 2.09356606e-01 -3.74244303e-01
1.59519041e+00 9.96807575e-01 8.03575993e-01 1.15810442e+00
9.52808000e-03 1.02399635e+00 1.48045707e+00 7.19374955e-01
5.18569231e-01 2.95430005e-01 4.86837834e-01 -2.82140166e-01
5.04648387e-01 2.75460482e-01 6.29528940e-01 9.76156235e-01
-4.59547698e-01 6.06903374e-01 -1.11438239e+00 6.29948556e-01
-1.72668397e+00 -1.21144056e+00 -7.28241861e-01 2.08402729e+00
3.66756231e-01 1.21188410e-01 2.06070483e-01 3.86407793e-01
4.54437762e-01 -2.84720659e-01 1.90061346e-01 -6.35596573e-01
3.58869016e-01 6.18636273e-02 5.74142754e-01 2.85911620e-01
-2.97818750e-01 5.05452573e-01 5.66594315e+00 -1.34441733e-01
-8.61714184e-01 -9.18170586e-02 8.91820431e-01 4.88585442e-01
-3.97685766e-01 7.28960633e-01 -7.20996737e-01 3.26605551e-02
1.20044708e+00 -2.82968611e-01 1.71218023e-01 8.37807178e-01
1.07351613e+00 -3.58136922e-01 -1.92778230e-01 2.06220090e-01
-2.72175491e-01 -5.00442922e-01 -1.19042695e+00 4.96742487e-01
5.70796013e-01 -7.19253957e-01 1.36733055e-01 5.96294224e-01
-3.59411597e-01 -1.00368536e+00 2.67911881e-01 6.23201787e-01
-1.77086983e-02 -1.20333219e+00 1.60446835e+00 5.42092562e-01
-3.54307652e-01 -5.69323421e-01 -4.72598255e-01 -1.07222366e+00
2.33106270e-01 3.24719310e-01 -1.39549470e+00 2.06119612e-01
6.13523126e-01 1.16970196e-01 -2.33010083e-01 8.44676346e-02
8.37042779e-02 8.57857943e-01 5.76663375e-01 2.77923252e-02
3.64757746e-01 -9.16034222e-01 -6.15204610e-02 5.87234795e-01
3.40268433e-01 1.11916410e-02 -9.21486225e-03 8.39258671e-01
4.89883840e-01 3.20587367e-01 -5.92306077e-01 -7.09388971e-01
7.35465050e-01 1.01243663e+00 -7.04403818e-01 -1.15702286e-01
-8.27732027e-01 2.92901605e-01 -1.25173509e-01 2.33504072e-01
-5.82896888e-01 -1.74809769e-01 3.52166325e-01 4.30157036e-01
9.12697911e-02 -4.61276591e-01 -7.87669599e-01 1.30890608e-01
-4.13477689e-01 -8.75263453e-01 -2.76653200e-01 -3.49161059e-01
-7.83052683e-01 -2.29531795e-01 -1.81608707e-01 -2.10790142e-01
8.73769522e-02 -3.49812716e-01 -3.77798617e-01 1.26694262e+00
-1.41259563e+00 -8.09032440e-01 -3.43971074e-01 -2.17287183e-01
2.60031223e-01 -3.08651030e-01 8.40255976e-01 2.01196857e-02
-8.15160632e-01 -3.92865360e-01 6.49293661e-02 -1.15589157e-01
4.82271701e-01 -7.58471370e-01 -2.59760857e-01 4.91398156e-01
-9.34924483e-01 9.44887996e-01 5.45798779e-01 -1.20060551e+00
-1.26807177e+00 -3.49690259e-01 1.83697569e+00 2.45351538e-01
5.23986578e-01 6.49097204e-01 -6.10236585e-01 6.53408766e-01
6.74722373e-01 -1.09243870e+00 9.97535586e-01 3.60730231e-01
8.47103298e-01 -2.70501226e-01 -6.54671729e-01 4.49423581e-01
1.33462235e-01 -3.71785492e-01 -2.96099275e-01 -9.49246585e-02
9.85890031e-02 3.14757377e-01 -1.52434194e+00 3.13980460e-01
3.08532655e-01 -1.32625902e+00 8.32640469e-01 -2.45164230e-01
4.80008006e-01 1.56413272e-01 1.41958401e-01 -7.03422070e-01
-1.06370139e+00 -1.45084321e-01 8.41364861e-01 1.22799265e+00
1.02616809e-01 -7.71532714e-01 7.62624919e-01 1.22270966e+00
-2.11755261e-01 -7.16061354e-01 -5.95231175e-01 -3.07297707e-01
-2.85509348e-01 1.34145409e-01 3.89011711e-01 1.00597680e+00
2.52956837e-01 5.83989024e-01 -3.79420698e-01 -2.27641943e-03
-2.02191193e-02 -1.56517699e-01 1.04211521e+00 -1.08091795e+00
-8.23425353e-02 -2.82288492e-01 9.36617628e-02 -3.18541408e-01
1.83130801e-01 -3.54602695e-01 -4.51702565e-01 -2.03080201e+00
4.33515579e-01 -4.15572762e-01 3.94388512e-02 4.47662920e-01
-3.12353104e-01 -5.24036109e-01 3.46806437e-01 5.95440149e-01
-1.58003978e-02 5.60511112e-01 1.23216581e+00 5.43946564e-01
-4.88154203e-01 4.10812438e-01 -1.15965474e+00 4.69200760e-01
1.34818971e+00 -2.90688008e-01 -2.87423611e-01 2.12801203e-01
3.60437095e-01 1.03467369e+00 1.62748218e-01 -4.73362118e-01
-4.19323891e-02 -1.03239346e+00 1.53319225e-01 -6.45279169e-01
1.28019229e-01 -1.55515409e+00 8.30387831e-01 9.08883154e-01
-6.27646595e-02 6.67090476e-01 -2.89802253e-01 -4.82519805e-01
-5.06198645e-01 -8.20314109e-01 4.66294199e-01 -2.01549515e-01
-4.30936635e-01 -3.98884714e-01 -6.15669906e-01 -5.58461607e-01
1.15672147e+00 -6.33404374e-01 -2.91317850e-01 -4.15346593e-01
-4.88468587e-01 -4.81777638e-02 4.55542326e-01 1.94799230e-01
5.98009050e-01 -1.21667457e+00 -5.58282316e-01 -9.80701204e-03
-3.47942144e-01 -5.45356035e-01 1.26649648e-01 1.57949245e+00
-1.51288778e-01 1.00853014e+00 -7.99679756e-01 2.76966900e-01
-1.22897518e+00 2.37696007e-01 -1.81577690e-02 -1.22387342e-01
-2.67288499e-02 1.59942597e-01 -1.16493687e-01 -2.72953673e-03
-1.58415258e-01 6.56116724e-01 -6.02397978e-01 4.05354425e-02
-2.49166057e-01 1.14110601e+00 -3.80565614e-01 -8.75467718e-01
-1.20660089e-01 -6.41452447e-02 4.48971212e-01 -1.38355255e-01
1.27874875e+00 -5.17513633e-01 -7.75528848e-01 7.28211939e-01
6.29922271e-01 3.79005134e-01 -6.24147594e-01 1.95984945e-01
4.38082874e-01 -1.16067600e+00 7.90756568e-02 -5.73875904e-01
-6.01208210e-01 4.51786280e-01 4.29300696e-01 4.68900710e-01
1.40831470e+00 -1.44362897e-01 5.40104210e-01 2.61438489e-02
3.73257510e-02 -1.60630774e+00 3.96855652e-01 5.28021038e-01
6.10583186e-01 -1.05535650e+00 -3.52741301e-01 -2.62809902e-01
-1.08729291e+00 1.15125656e+00 6.97966099e-01 3.26896787e-01
2.95458674e-01 -2.76136994e-02 2.23774865e-01 -4.60995615e-01
-5.41773438e-01 -9.25604105e-02 -3.74207705e-01 4.28638250e-01
1.30229390e+00 2.76269644e-01 -1.23466229e+00 7.81811535e-01
-1.16982400e-01 3.65587287e-02 5.48268676e-01 6.65905356e-01
-1.02820122e+00 -1.36346197e+00 -9.33516204e-01 3.20650667e-01
-7.00214624e-01 2.89859027e-01 -3.69370699e-01 1.15637350e+00
3.59184802e-01 1.37202668e+00 -4.83744591e-02 -2.76445657e-01
4.59938943e-01 6.82101473e-02 -4.22546595e-01 -3.11267644e-01
-8.64490509e-01 5.38205147e-01 4.60473806e-01 2.98981488e-01
-6.14297986e-01 -7.83685744e-01 -1.67108870e+00 -5.89770317e-01
-4.70824778e-01 5.64664841e-01 9.93314505e-01 8.62406015e-01
8.79720449e-02 5.50447583e-01 1.17777026e+00 1.52543038e-01
-8.70436013e-01 -1.18525743e+00 -9.69408512e-01 4.84376609e-01
-3.09839636e-01 -5.16652226e-01 -2.81899065e-01 -3.78789395e-01] | [9.012404441833496, 6.142904281616211] |
5a0bd13c-b909-41c4-9699-3cf8dcfdfaf3 | towards-end-to-end-open-conversational | 2210.07113 | null | https://arxiv.org/abs/2210.07113v1 | https://arxiv.org/pdf/2210.07113v1.pdf | Towards End-to-End Open Conversational Machine Reading | In open-retrieval conversational machine reading (OR-CMR) task, machines are required to do multi-turn question answering given dialogue history and a textual knowledge base. Existing works generally utilize two independent modules to approach this problem's two successive sub-tasks: first with a hard-label decision making and second with a question generation aided by various entailment reasoning methods. Such usual cascaded modeling is vulnerable to error propagation and prevents the two sub-tasks from being consistently optimized. In this work, we instead model OR-CMR as a unified text-to-text task in a fully end-to-end style. Experiments on the OR-ShARC dataset show the effectiveness of our proposed end-to-end framework on both sub-tasks by a large margin, achieving new state-of-the-art results. Further ablation studies support that our framework can generalize to different backbone models. | ['Hai Zhao', 'Zhuosheng Zhang', 'Siru Ouyang', 'Sizhe Zhou'] | 2022-10-13 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 3.90506089e-01 8.67438495e-01 1.94016285e-02 -7.92414606e-01
-1.51237833e+00 -6.84164762e-01 8.98652613e-01 1.44661620e-01
-4.29759324e-01 6.66593194e-01 4.72616941e-01 -7.70183802e-01
1.44546792e-01 -4.76334423e-01 -5.69738030e-01 -1.15870431e-01
3.70066166e-01 8.35041583e-01 2.55907118e-01 -6.40947640e-01
1.52833179e-01 -4.51425433e-01 -1.02976739e+00 8.98547232e-01
8.26108694e-01 8.81505728e-01 -1.91752594e-02 1.14561069e+00
-2.68818706e-01 1.48472071e+00 -3.21020961e-01 -1.03830874e+00
1.23832189e-01 -5.62784135e-01 -1.82995403e+00 3.45277004e-02
4.80226189e-01 -5.41341662e-01 -5.52303120e-02 5.13330400e-01
4.48236763e-01 2.39016086e-01 5.92376471e-01 -8.54214251e-01
-7.00816095e-01 8.23720396e-01 -2.22522855e-01 -1.38345107e-01
8.53675544e-01 1.92158058e-01 1.56336880e+00 -8.29918206e-01
4.11699593e-01 1.50182176e+00 4.75472718e-01 7.24701762e-01
-1.16820228e+00 1.35567874e-01 3.23668152e-01 3.91264170e-01
-9.01221931e-01 -5.11466801e-01 7.00608075e-01 -1.00532658e-01
1.34673822e+00 4.22154874e-01 4.13096137e-02 1.05998075e+00
-5.48180118e-02 1.18633139e+00 1.04115283e+00 -7.12839901e-01
1.20127961e-01 9.35459957e-02 4.53088909e-01 7.42493570e-01
-3.40013355e-01 -4.60514575e-01 -5.34116387e-01 -2.24523515e-01
-1.57031015e-01 -3.39633226e-01 -2.53858566e-01 1.07458636e-01
-1.02933311e+00 1.03747153e+00 9.43204835e-02 2.83570159e-02
-2.47121841e-01 -8.53799656e-02 4.90676969e-01 8.53457332e-01
6.77124500e-01 5.81520379e-01 -7.12665677e-01 -2.39448011e-01
-6.99194372e-01 6.85141742e-01 1.47743368e+00 8.95152092e-01
6.13352478e-01 -7.87515163e-01 -5.92724562e-01 1.04022586e+00
3.60272110e-01 2.35528246e-01 4.18368071e-01 -9.08926487e-01
8.55725884e-01 5.49970925e-01 1.28123134e-01 -4.93456990e-01
-4.53317314e-01 -1.30056158e-01 -4.60796922e-01 -3.57243657e-01
6.32945836e-01 -3.71191472e-01 -5.30308902e-01 1.80598724e+00
5.98487139e-01 -3.88334006e-01 4.59899157e-01 8.22546899e-01
8.75415802e-01 7.42620766e-01 1.47401923e-02 -3.06862686e-02
1.71096468e+00 -1.72988486e+00 -8.22158396e-01 -4.11950529e-01
9.06832993e-01 -8.37456226e-01 1.32772660e+00 2.81632096e-01
-1.28610706e+00 -2.84307390e-01 -9.25808787e-01 -8.69108617e-01
-1.56549454e-01 -5.41667594e-03 4.20463771e-01 3.33984464e-01
-1.07576847e+00 2.57502288e-01 -4.01743710e-01 -2.88028359e-01
-9.91339982e-03 -2.46802032e-01 6.70963898e-02 -1.57514587e-01
-1.50530255e+00 1.25406539e+00 1.70987651e-01 2.13152736e-01
-8.18833172e-01 -4.60485667e-01 -7.35864043e-01 -2.45882245e-03
7.20230520e-01 -1.09076798e+00 2.14257169e+00 -7.28513122e-01
-2.19786859e+00 1.07566655e+00 -4.39142615e-01 -5.86924732e-01
6.86238110e-01 -6.87093198e-01 6.15161434e-02 3.46628368e-01
3.17960531e-02 5.49650073e-01 6.89652801e-01 -1.02161849e+00
-6.17079198e-01 -3.64743173e-01 7.24625051e-01 6.33566797e-01
3.51266079e-02 -8.05119202e-02 -4.31178063e-01 -4.08816725e-01
-1.86926141e-01 -9.68394935e-01 -2.44435538e-02 -3.99946481e-01
-6.45651758e-01 -8.92197549e-01 4.36851799e-01 -9.44412053e-01
1.14814258e+00 -1.43075240e+00 2.84466505e-01 -3.62895399e-01
1.69910014e-01 1.55884922e-01 -3.78712595e-01 9.14949834e-01
1.84815347e-01 1.77537408e-02 -3.58762145e-01 -7.06216455e-01
1.91693500e-01 -5.50224446e-02 -4.25227195e-01 8.94102827e-02
4.57500279e-01 1.12042654e+00 -1.02896488e+00 -4.24038231e-01
-6.46667555e-02 -1.53367460e-01 -5.37303150e-01 7.16808319e-01
-8.26873779e-01 1.86081707e-01 -4.93385851e-01 4.00067449e-01
2.07665160e-01 -4.37187582e-01 2.50868678e-01 3.85889441e-01
3.72657180e-01 9.45483029e-01 -5.98706186e-01 1.95303643e+00
-7.81940997e-01 5.35107613e-01 1.65249586e-01 -8.95693779e-01
4.98749584e-01 5.67008376e-01 -1.99347943e-01 -5.57037830e-01
-6.61019050e-03 4.71224040e-02 -1.44727272e-03 -8.34400117e-01
7.21638918e-01 -4.28775072e-01 -1.63142517e-01 8.94659996e-01
2.27562666e-01 -1.58519208e-01 4.44429740e-02 5.70382059e-01
1.13453484e+00 1.05807662e-01 1.95499435e-01 -2.51789749e-01
7.33774900e-01 3.81280296e-02 -1.92884784e-02 1.02337682e+00
4.34419960e-02 5.08718133e-01 7.48208821e-01 -1.26632214e-01
-8.55160594e-01 -7.02026188e-01 -1.11779124e-02 1.50862062e+00
-3.47555652e-02 -4.26037908e-01 -1.01156998e+00 -1.07371962e+00
-2.42306054e-01 1.03613222e+00 -5.08126616e-01 -6.74307123e-02
-4.52050745e-01 -4.54456270e-01 7.91573524e-01 1.62110507e-01
4.73643333e-01 -1.01997793e+00 -3.91075104e-01 4.46248919e-01
-7.50862658e-01 -1.40289354e+00 -3.36706936e-01 -2.09847651e-02
-6.04157150e-01 -1.06180465e+00 -7.28461087e-01 -7.67878115e-01
2.50677794e-01 2.33891889e-01 1.61173451e+00 2.09537983e-01
9.79229435e-02 4.61461455e-01 -8.69942904e-01 -3.74097466e-01
-6.27316058e-01 5.64176559e-01 -4.44343358e-01 2.56786048e-01
2.56146610e-01 -3.72543097e-01 -6.27381325e-01 1.82280391e-01
-7.59849012e-01 4.78418916e-01 5.41626513e-01 9.21580911e-01
-1.77800991e-02 -6.18545711e-01 1.05991733e+00 -1.08805585e+00
1.14771104e+00 -5.98545194e-01 -9.47858915e-02 6.95235014e-01
-7.41829455e-01 3.27932686e-01 6.26698315e-01 -3.00175995e-02
-1.53156519e+00 -3.21393698e-01 -3.31669450e-01 2.55183637e-01
-1.26266927e-01 6.19609058e-01 -1.82056382e-01 5.80343664e-01
5.88070452e-01 2.76280731e-01 1.54775649e-01 -4.54620957e-01
9.29481387e-01 9.85513926e-01 3.94175589e-01 -7.75596917e-01
6.27837420e-01 6.76118210e-02 -5.39350986e-01 -7.51059473e-01
-1.47338104e+00 -6.58082128e-01 -4.02640134e-01 -2.03563586e-01
9.96554673e-01 -1.03298116e+00 -8.26810658e-01 6.06906235e-01
-1.44627368e+00 -7.45157063e-01 -3.94865014e-02 -2.60152481e-02
-5.24633646e-01 6.07527494e-01 -1.00103772e+00 -1.02196050e+00
-8.26800466e-01 -8.76933694e-01 1.29731047e+00 1.50195494e-01
-3.61000091e-01 -1.12219298e+00 2.44043946e-01 1.21126509e+00
3.20830643e-01 -2.51776576e-01 1.05005133e+00 -1.04945600e+00
-3.94305944e-01 -2.49290586e-01 3.08870617e-02 5.26251197e-01
-2.74093390e-01 -4.63009536e-01 -1.15214992e+00 -9.83004048e-02
2.44911075e-01 -1.17820561e+00 8.30710828e-01 -2.11404100e-01
9.02909815e-01 -6.70341671e-01 2.35492349e-01 -1.33577541e-01
1.07974994e+00 -3.64634871e-01 6.17820323e-01 1.15809366e-01
4.81409043e-01 1.13131440e+00 6.12974465e-01 5.80363870e-02
1.09112215e+00 6.53995514e-01 1.27217114e-01 8.20567235e-02
3.82605940e-02 -3.65687460e-01 4.78137732e-01 9.61962283e-01
3.00884843e-01 -5.49728572e-01 -8.94653499e-01 6.88562751e-01
-2.12221909e+00 -9.84937847e-01 -2.41157740e-01 1.72705960e+00
1.38265562e+00 3.04005761e-02 1.08630717e-01 -1.71150431e-01
4.03230667e-01 4.16471064e-01 -5.30784488e-01 -5.56773603e-01
1.57932177e-01 5.10575771e-02 -2.58081496e-01 9.28444088e-01
-8.91819179e-01 1.13186014e+00 5.74722767e+00 7.26462126e-01
-6.48181677e-01 2.90665746e-01 8.98902416e-01 -1.68871865e-01
-4.17332202e-01 3.04009855e-01 -7.11331785e-01 6.88947439e-02
9.88270342e-01 4.03922237e-02 4.21195596e-01 7.49608934e-01
4.29997444e-02 -2.62894958e-01 -1.43702590e+00 5.82903743e-01
3.02837729e-01 -1.00717950e+00 1.16661802e-01 -3.91061097e-01
4.89442915e-01 -9.60276648e-02 -3.87857586e-01 8.53841662e-01
5.37830710e-01 -9.20766175e-01 7.08576500e-01 4.15933669e-01
5.08895338e-01 -4.00199354e-01 8.36527348e-01 7.34393001e-01
-6.50135994e-01 6.23014048e-02 -5.46117276e-02 -4.58295763e-01
5.18536627e-01 4.68975782e-01 -1.05050039e+00 8.03874671e-01
2.38404319e-01 3.02426070e-01 -5.92494130e-01 2.39212617e-01
-5.52714050e-01 8.47312391e-01 -1.05249427e-01 -4.90834385e-01
4.92220759e-01 -1.20495923e-01 5.06241024e-01 1.29398155e+00
-3.95296097e-01 3.06743175e-01 1.99259087e-01 7.00584888e-01
-3.97777289e-01 2.17751130e-01 -1.58805311e-01 1.41551703e-01
2.46398091e-01 1.39055336e+00 -1.67699903e-01 -5.70772529e-01
-5.14156163e-01 1.26678085e+00 9.10245240e-01 3.29728842e-01
-6.41346395e-01 -4.44047362e-01 2.91189611e-01 -1.45964235e-01
1.55082509e-01 6.15735613e-02 -2.50773996e-01 -1.34398520e+00
4.31637615e-01 -1.32561958e+00 4.85550970e-01 -7.50190198e-01
-1.53359401e+00 6.77394271e-01 -7.93472975e-02 -6.85771585e-01
-6.36554420e-01 -4.89539981e-01 -6.50655210e-01 8.34176064e-01
-1.83545625e+00 -1.30427814e+00 1.15082059e-02 4.46255714e-01
1.09171367e+00 9.77116600e-02 8.58866453e-01 5.50144492e-03
-4.99556750e-01 7.33931243e-01 -1.40196592e-01 2.28430331e-01
8.36314023e-01 -1.58708549e+00 4.09412295e-01 8.27144623e-01
1.11166470e-01 5.07526577e-01 5.69954932e-01 -2.75526911e-01
-1.44759643e+00 -8.34399104e-01 1.30168307e+00 -8.93347323e-01
6.35726690e-01 -4.93904710e-01 -8.62553477e-01 7.23344564e-01
7.56808341e-01 -6.38132334e-01 8.17012668e-01 6.82493627e-01
-4.60244507e-01 9.26635489e-02 -7.46081650e-01 6.26396179e-01
8.55291307e-01 -9.41478789e-01 -1.21122193e+00 5.85691094e-01
9.86314952e-01 -4.73398060e-01 -7.10729957e-01 1.44617468e-01
3.85639042e-01 -8.04082155e-01 6.05893433e-01 -9.69381273e-01
1.05537689e+00 1.33861497e-01 -7.85676911e-02 -1.17168403e+00
8.56685415e-02 -9.77018237e-01 -2.02171788e-01 1.28204834e+00
8.99508774e-01 -4.26027805e-01 1.81834430e-01 8.03622723e-01
-2.07521617e-01 -1.08706903e+00 -8.30211043e-01 -3.32583129e-01
3.90980780e-01 -4.70116884e-01 2.50003874e-01 8.04747880e-01
4.64510769e-01 1.45624948e+00 -4.79232162e-01 -9.50876698e-02
3.23849529e-01 2.68520206e-01 9.23396885e-01 -8.22268486e-01
-5.55221975e-01 -4.15638834e-01 4.19973820e-01 -1.76089907e+00
3.71058494e-01 -9.92622077e-01 4.33003157e-01 -1.65123749e+00
3.49047095e-01 -9.21935588e-02 2.36672387e-01 4.93363559e-01
-8.18673551e-01 -2.31542185e-01 1.47524238e-01 4.30904180e-02
-1.21856928e+00 8.17156851e-01 1.11562073e+00 -8.26741159e-02
-4.98863906e-02 -2.79855393e-02 -9.28911388e-01 4.65042621e-01
6.12228632e-01 -2.51945972e-01 -5.93773365e-01 -5.84050357e-01
4.19012517e-01 5.64732313e-01 2.47749731e-01 -3.86746854e-01
1.74941063e-01 1.76013470e-01 -3.53006750e-01 -3.83789808e-01
2.00933129e-01 -3.62503558e-01 -7.34152019e-01 8.95325169e-02
-1.18123806e+00 -2.51801070e-02 -1.32702544e-01 7.20997691e-01
-2.97077447e-01 -4.36207950e-01 4.24578071e-01 -2.37246156e-01
-2.19202220e-01 -1.33342579e-01 -5.48458993e-01 6.24074459e-01
6.27428591e-01 2.83901185e-01 -4.30003613e-01 -1.07597113e+00
-4.69909340e-01 8.18243623e-01 -6.05621636e-02 6.35485709e-01
3.39732260e-01 -8.46930325e-01 -1.19862628e+00 -4.32566404e-01
2.90767163e-01 3.48131835e-01 3.66493255e-01 7.03494966e-01
-3.17698509e-01 6.55279875e-01 5.54857671e-01 -4.84174132e-01
-1.11268103e+00 3.03072184e-01 4.02505338e-01 -1.05938840e+00
-4.97068465e-01 8.99063587e-01 -1.10325562e-02 -9.85632718e-01
3.39177579e-01 -1.02259509e-01 -1.68427333e-01 -1.15630534e-02
6.92079484e-01 1.00750692e-01 3.05114508e-01 -1.42008096e-01
4.24767658e-02 2.10608561e-02 -4.84481633e-01 -4.98166889e-01
1.01665568e+00 -5.22806942e-01 -1.62474945e-01 5.98253787e-01
1.21722925e+00 -2.38695160e-01 -9.88600075e-01 -6.11678183e-01
2.51413316e-01 1.58971816e-01 -1.51988000e-01 -1.30973136e+00
-3.29059362e-01 9.42391992e-01 -3.78034636e-02 3.76114696e-01
7.63606727e-01 1.56004280e-01 1.05117917e+00 1.06501162e+00
9.97522101e-02 -1.11278081e+00 2.61805922e-01 1.01787853e+00
1.20509911e+00 -1.48593688e+00 -2.99615026e-01 -2.47520536e-01
-9.52124774e-01 1.04523230e+00 6.15854144e-01 2.97467709e-01
2.39040598e-01 -2.00622261e-01 4.10657376e-01 -2.84859151e-01
-1.34111011e+00 -1.93435878e-01 3.58233273e-01 3.63311619e-02
6.17539465e-01 -8.06958228e-02 -4.35117781e-01 6.07248008e-01
-3.66382688e-01 -1.88480392e-01 4.54760939e-01 9.56829906e-01
-1.70481116e-01 -1.12602890e+00 -7.78747373e-04 3.10593903e-01
-5.00540435e-01 -3.83221269e-01 -7.49340057e-01 4.51145440e-01
-7.58064628e-01 1.62928545e+00 -2.07377031e-01 -2.30287895e-01
2.71726549e-01 6.92846179e-01 3.08124781e-01 -6.46341860e-01
-9.08362865e-01 -3.39846134e-01 8.48668694e-01 -5.11387587e-01
-3.32135320e-01 -3.57351661e-01 -1.09816718e+00 -1.80358469e-01
-5.16165257e-01 3.32187414e-01 3.25431734e-01 1.36304438e+00
4.98525888e-01 3.28198045e-01 6.77593350e-01 -3.57232362e-01
-1.33540773e+00 -1.55010831e+00 -2.80018058e-02 7.74838746e-01
3.69539618e-01 2.15255786e-02 -3.12839061e-01 9.84209627e-02] | [11.88630199432373, 8.054312705993652] |
2a26a34d-87c3-4036-8b8f-f9990082730a | monash-summ-longsumm-20-scisummpip-an | null | null | https://aclanthology.org/2020.sdp-1.37 | https://aclanthology.org/2020.sdp-1.37.pdf | Monash-Summ@LongSumm 20 SciSummPip: An Unsupervised Scientific Paper Summarization Pipeline | The Scholarly Document Processing (SDP) workshop is to encourage more efforts on natural language understanding of scientific task. It contains three shared tasks and we participate in the LongSumm shared task. In this paper, we describe our text summarization system, SciSummPip, inspired by SummPip (Zhao et al., 2020) that is an unsupervised text summarization system for multi-document in News domain. Our SciSummPip includes a transformer-based language model SciBERT (Beltagy et al., 2019) for contextual sentence representation, content selection with PageRank (Page et al., 1999), sentence graph construction with both deep and linguistic information, sentence graph clustering and within-graph summary generation. Our work differs from previous method in that content selection and a summary length constraint is applied to adapt to the scientific domain. The experiment results on both training dataset and blind test dataset show the effectiveness of our method, and we empirically verify the robustness of modules used in SciSummPip with BERTScore (Zhang et al., 2019a). | ['Shirui Pan', 'Longxiang Gao', 'Ming Liu', 'Jiaxin Ju'] | null | null | null | null | emnlp-sdp-2020-11 | ['graph-clustering'] | ['graphs'] | [ 2.23217502e-01 3.80715847e-01 -1.81964207e-02 -1.76780134e-01
-9.70377505e-01 -7.63037920e-01 9.22342062e-01 6.36683464e-01
-8.28692988e-02 1.06536114e+00 1.08135760e+00 -1.73239201e-01
-4.08464670e-01 -6.92454398e-01 -5.06358922e-01 -1.26633346e-01
-1.18887685e-01 4.65923101e-01 3.34256262e-01 -3.36022347e-01
1.21214020e+00 4.03713584e-01 -9.70071137e-01 7.60559857e-01
1.26030493e+00 2.10564613e-01 5.80077589e-01 1.38950586e+00
-4.62387323e-01 1.15119171e+00 -1.08191788e+00 -5.61033309e-01
-1.20479330e-01 -6.99646533e-01 -1.30196607e+00 -2.11512253e-01
8.02069902e-01 2.12742552e-01 -3.94473553e-01 1.06678736e+00
1.05474806e+00 1.53325051e-01 7.88951099e-01 -8.41604471e-01
-8.40871215e-01 1.07476890e+00 -7.45139062e-01 4.63811398e-01
6.88047171e-01 -2.49044612e-01 1.13900948e+00 -5.92471778e-01
1.26530874e+00 1.30563509e+00 5.44903219e-01 3.14773977e-01
-8.39953482e-01 -2.10914165e-01 -8.90838802e-02 4.39218014e-01
-9.16702986e-01 -5.21018565e-01 7.40112901e-01 -1.13386728e-01
1.27375984e+00 7.22180307e-01 4.63191956e-01 1.03000677e+00
7.33699858e-01 9.11990762e-01 8.09253573e-01 -4.61701274e-01
2.84431875e-01 -2.60325193e-01 6.91822886e-01 6.06423676e-01
4.39711541e-01 -9.77331817e-01 -9.29909170e-01 -1.96933851e-01
1.52178347e-01 -6.11757994e-01 -1.34734571e-01 1.25400394e-01
-1.32356358e+00 6.02260470e-01 2.32953634e-02 5.32432914e-01
-4.18571830e-01 -1.54694691e-01 8.92959118e-01 6.83201194e-01
8.52053881e-01 9.01936054e-01 -1.13823898e-01 -3.02347187e-02
-1.50566101e+00 6.09251082e-01 1.19712007e+00 1.16477787e+00
2.05840647e-01 -1.38334781e-01 -1.07366025e+00 9.59166348e-01
6.34010285e-02 2.98794389e-01 9.26163971e-01 -1.16495335e+00
5.81674635e-01 6.16720438e-01 -2.54011244e-01 -1.10237229e+00
-4.80096906e-01 -6.00340009e-01 -1.07049060e+00 -2.96590418e-01
-3.60135943e-01 -2.43590966e-01 -6.63571298e-01 1.03711474e+00
-1.91523045e-01 7.13831931e-02 3.32382858e-01 4.28445429e-01
1.73731518e+00 1.02104974e+00 -2.05714896e-01 -6.01409912e-01
1.41052020e+00 -1.27018547e+00 -9.96037841e-01 3.90451588e-02
6.79557681e-01 -1.06452751e+00 6.31903648e-01 6.60976112e-01
-1.44874597e+00 -3.19177181e-01 -1.09857798e+00 -4.43151742e-01
-4.85134184e-01 2.31154040e-01 3.28956455e-01 3.05706441e-01
-1.52453780e+00 8.14758301e-01 -2.16244727e-01 -9.05723572e-01
3.63515079e-01 9.15037319e-02 -3.86927158e-01 1.81334853e-01
-9.22897816e-01 8.08060944e-01 5.35009980e-01 -5.10651529e-01
-2.83967167e-01 -6.77687168e-01 -4.41373646e-01 2.44266540e-01
2.79452980e-01 -1.17407501e+00 1.20474160e+00 -4.91489142e-01
-1.87071669e+00 1.08146095e+00 -8.38819593e-02 -8.49730611e-01
4.30386275e-01 -2.11906821e-01 -2.19725460e-01 4.95886713e-01
3.34120661e-01 2.81155020e-01 3.29061270e-01 -1.10881293e+00
-2.41814837e-01 -3.48054320e-01 -4.05652761e-01 3.33582848e-01
-1.39653295e-01 4.34291571e-01 -4.55042273e-01 -8.55472863e-01
-2.34107584e-01 -2.97027022e-01 -2.95613948e-02 -8.36568356e-01
-9.54430461e-01 -4.91486222e-01 5.64880490e-01 -9.93098199e-01
1.42325115e+00 -1.48333681e+00 4.11261290e-01 -5.11287637e-02
3.98401052e-01 2.56902188e-01 -3.92360032e-01 1.22506201e+00
7.07442313e-02 3.75751406e-01 -4.87371176e-01 -5.11954725e-01
4.76854481e-03 -1.49108380e-01 -5.67673504e-01 1.97065040e-01
-2.41661698e-01 8.82526398e-01 -1.03368652e+00 -9.05741096e-01
-1.60087451e-01 -3.00973922e-01 -2.61873454e-01 1.50354698e-01
-4.07337546e-01 1.93986725e-02 -4.93362725e-01 3.15259337e-01
6.50390685e-01 -7.11978152e-02 1.69165477e-01 -1.74199358e-01
-5.44031501e-01 3.24052364e-01 -8.90484333e-01 2.07847881e+00
-7.42432252e-02 8.70722830e-01 -1.13715855e-02 -1.06829989e+00
9.15260613e-01 2.76734769e-01 4.16410029e-01 -5.46889424e-01
6.81933621e-03 1.66242421e-01 -2.23456845e-01 -8.34436893e-01
1.16942775e+00 5.64531386e-01 -6.21989332e-02 6.09060049e-01
4.87470120e-01 -5.80775499e-01 1.10167646e+00 1.22690845e+00
1.42492080e+00 4.39787954e-02 4.18525666e-01 -8.02080333e-01
7.29976058e-01 2.40019336e-01 2.76391625e-01 1.19471037e+00
1.07461460e-01 9.50777173e-01 8.38565767e-01 1.20321825e-01
-1.12962079e+00 -7.15152144e-01 1.32370321e-03 9.81634259e-01
-1.79113150e-01 -1.01661885e+00 -8.16824496e-01 -6.33108914e-01
-1.36711806e-01 1.01006830e+00 -3.78984004e-01 -1.41743090e-04
-5.64365923e-01 -9.76952136e-01 7.73732662e-01 -1.59177884e-01
6.06370866e-01 -1.14456153e+00 -1.57599047e-01 2.19514102e-01
-2.40494102e-01 -8.92236948e-01 -4.75962251e-01 -3.35953206e-01
-7.89550602e-01 -9.94736314e-01 -9.13939476e-01 -8.33793819e-01
3.86267900e-01 4.14595515e-01 1.13479054e+00 -2.25241765e-01
-3.40669006e-01 5.94032288e-01 -6.08068824e-01 -5.93660593e-01
-6.77305579e-01 3.61244619e-01 -4.36999381e-01 -4.50182438e-01
-1.16049446e-01 -5.35431147e-01 -4.87588584e-01 -5.61498582e-01
-9.89306509e-01 4.06561404e-01 6.92950606e-01 5.36839664e-01
2.06766203e-01 -4.17940944e-01 1.04177880e+00 -1.26870072e+00
1.52606547e+00 -5.39908886e-01 -1.25503004e-01 6.53555274e-01
-6.48901641e-01 1.28377408e-01 5.28152406e-01 2.04199538e-01
-1.20478702e+00 -6.66534424e-01 -4.28245366e-02 3.57667238e-01
3.61855000e-01 8.89282346e-01 7.91920722e-02 1.41089652e-02
7.12552786e-01 5.76590657e-01 -3.14211518e-01 -7.21330464e-01
5.09043515e-01 9.22638416e-01 6.53903067e-01 -3.98801982e-01
1.64233297e-01 5.11288717e-02 1.05067603e-01 -9.50288653e-01
-7.67322481e-01 -5.82832515e-01 -4.62784410e-01 -4.40092742e-01
6.51797056e-01 -7.55248308e-01 -4.80144262e-01 2.32004941e-01
-1.55928147e+00 1.26093328e-01 -2.55046815e-01 1.88521340e-01
-4.48154032e-01 1.03919446e+00 -7.15048850e-01 -5.53192914e-01
-1.30609727e+00 -5.92073619e-01 1.05892766e+00 2.41669789e-01
-4.04808879e-01 -8.51475358e-01 6.53042316e-01 4.59577918e-01
3.09360236e-01 3.45588624e-01 7.93586373e-01 -9.88780260e-01
-1.05293743e-01 -5.01629971e-02 -3.18943501e-01 2.22224236e-01
-2.41466656e-01 2.30425507e-01 -6.30294383e-01 -1.97965235e-01
-1.05626307e-01 -2.89447844e-01 1.34576190e+00 4.91406173e-01
1.28296542e+00 -5.79832673e-01 -7.95994028e-02 2.33593911e-01
1.16172981e+00 -1.89853206e-01 8.15644681e-01 4.59687680e-01
6.33397758e-01 6.07262492e-01 1.68568790e-01 3.91903639e-01
4.28602040e-01 1.03500254e-01 -1.22265808e-01 3.31039429e-01
-5.83162904e-01 1.06580608e-01 4.74362075e-01 1.72913325e+00
-3.20956600e-03 -1.03161180e+00 -7.18915403e-01 4.14869130e-01
-2.21502161e+00 -1.24827302e+00 -8.41829002e-01 1.83379114e+00
1.04354298e+00 -4.16880362e-02 -1.09392956e-01 -2.78339863e-01
7.34087706e-01 4.10813421e-01 -1.76942900e-01 -8.85425031e-01
-6.28688037e-01 1.77157283e-01 5.87623060e-01 5.56005239e-01
-6.93051100e-01 1.00747228e+00 5.68302870e+00 1.21044397e+00
-5.79488456e-01 4.49125953e-02 4.02712435e-01 -1.81268185e-01
-4.34106648e-01 -1.31685729e-03 -3.39961857e-01 4.78439271e-01
1.07449412e+00 -9.60303545e-01 -1.97948404e-02 1.73489168e-01
3.17129493e-01 -3.19424480e-01 -6.29241884e-01 7.14306653e-01
7.85075366e-01 -1.83848011e+00 5.14068842e-01 -4.42923605e-01
9.42008018e-01 2.04381645e-01 -5.71845651e-01 3.04634780e-01
6.24697030e-01 -7.58377790e-01 5.64367652e-01 1.01801121e+00
2.68260270e-01 -5.92492282e-01 8.57494414e-01 3.25570822e-01
-5.34184873e-01 4.50687468e-01 -5.90615571e-01 1.70275256e-01
-1.62365753e-02 9.90475178e-01 -6.25356078e-01 1.37549412e+00
3.85285079e-01 9.67821300e-01 -1.00166750e+00 1.19227195e+00
-7.55822808e-02 7.92650700e-01 1.88456416e-01 -4.49340105e-01
2.94093285e-02 -3.65040660e-01 1.18211818e+00 1.86070442e+00
3.73498261e-01 -6.17948920e-02 3.10079139e-02 4.64987457e-01
-6.88112319e-01 6.02385223e-01 -5.00296235e-01 -1.14952028e-01
2.89634556e-01 1.48449469e+00 -8.20447624e-01 -8.03542912e-01
1.17105111e-01 1.19466281e+00 2.08204091e-01 2.07248926e-01
-1.12845808e-01 -9.97322679e-01 -4.07946616e-01 -3.45485479e-01
7.85596017e-03 -6.57512844e-02 -6.52494550e-01 -1.21063125e+00
-1.40729710e-01 -8.40164602e-01 5.00509918e-01 -1.12734151e+00
-1.26716757e+00 4.52953756e-01 3.01007386e-02 -9.24392879e-01
1.44960254e-01 -1.73761994e-01 -1.24554718e+00 6.70701802e-01
-1.34047878e+00 -1.29374349e+00 -1.76531747e-02 2.24606484e-01
8.16975951e-01 -4.48187381e-01 5.78242421e-01 -9.80904922e-02
-6.90221727e-01 1.11382075e-01 7.03060150e-01 -2.74344057e-01
1.16923594e+00 -1.54628313e+00 4.99973089e-01 8.29859078e-01
-2.22958857e-03 5.53961217e-01 1.06916499e+00 -9.71622527e-01
-1.69850492e+00 -1.21382642e+00 1.34022081e+00 -3.24167430e-01
7.80694604e-01 -3.43738168e-01 -8.44380558e-01 3.84930342e-01
1.36581945e+00 -9.97394681e-01 5.95288694e-01 -1.75984457e-01
1.18115574e-01 -5.33907861e-03 -1.03806579e+00 6.44947231e-01
1.03119099e+00 4.24622633e-02 -9.38091397e-01 1.10517788e+00
7.12033212e-01 -3.24559242e-01 -8.07050228e-01 2.37450138e-01
1.31033316e-01 -7.49626577e-01 5.89255869e-01 -5.01687467e-01
9.77077961e-01 -5.52077889e-02 3.85846823e-01 -1.56130540e+00
-4.87697989e-01 -8.21051717e-01 -1.30682558e-01 1.64860618e+00
2.66773641e-01 -3.55757356e-01 3.44362497e-01 -1.14817604e-01
-7.33995259e-01 -3.68775278e-01 -7.36842334e-01 -6.23245358e-01
2.67282516e-01 3.02822590e-02 2.47594088e-01 8.72304201e-01
5.52510500e-01 8.70887041e-01 -1.22894235e-01 -2.87646979e-01
4.71629322e-01 3.77524763e-01 7.72934914e-01 -1.28651738e+00
-6.05424913e-03 -8.71951580e-01 -1.85725808e-01 -6.95921242e-01
1.33444503e-01 -1.39816284e+00 -2.08073840e-01 -2.56147480e+00
6.45164371e-01 6.09626234e-01 -2.50794850e-02 9.74413976e-02
-1.47769585e-01 -1.92354918e-01 2.58365273e-01 3.83066654e-01
-1.28232372e+00 6.11949563e-01 1.29421055e+00 -3.44828576e-01
8.11207574e-03 -3.95979315e-01 -9.69968855e-01 3.65131319e-01
9.59271431e-01 -4.34513092e-01 -2.59253114e-01 -3.94138128e-01
6.28093183e-01 1.08249739e-01 4.19933237e-02 -8.68434906e-01
8.38181674e-01 -6.56824261e-02 1.36603028e-01 -9.72114563e-01
-4.16706502e-01 2.54366726e-01 -3.27690750e-01 4.41805929e-01
-8.22347403e-01 2.25017518e-01 2.00570181e-01 3.45720589e-01
-1.23365127e-01 -6.80223525e-01 3.55080575e-01 -3.58242631e-01
-3.50160033e-01 -9.01655555e-02 -7.61439264e-01 2.10023627e-01
7.01327503e-01 1.76667973e-01 -9.53935683e-01 -3.23054403e-01
-1.52458429e-01 6.61770821e-01 5.12713678e-02 2.26127744e-01
5.47068417e-01 -7.64496565e-01 -1.55341911e+00 -4.53866571e-01
-1.15793131e-01 -1.97681248e-01 2.93759733e-01 7.42305398e-01
-9.80263948e-01 5.37876248e-01 -1.26098469e-01 4.04532328e-02
-1.34358156e+00 3.23895484e-01 -2.31526375e-01 -6.10605478e-01
-7.35129774e-01 6.50330424e-01 -3.42870086e-01 -3.23738188e-01
1.27357161e-02 -8.08001980e-02 -7.47634590e-01 9.87312496e-02
7.17158735e-01 9.17222619e-01 3.12550694e-01 -1.57834858e-01
-2.70057637e-02 2.24365100e-01 -4.38475043e-01 -1.38890132e-01
1.59783626e+00 -1.15023367e-01 -9.53811586e-01 5.09460270e-01
1.08410108e+00 4.08307642e-01 -7.12849796e-02 -7.35959411e-02
2.98315465e-01 1.06146656e-01 3.23374599e-01 -1.16133153e+00
-5.99584043e-01 4.49007750e-01 -9.47501976e-04 6.92306221e-01
7.70902336e-01 -2.06908867e-01 8.08221161e-01 6.50182545e-01
-2.50625134e-01 -1.46204126e+00 -7.93559179e-02 7.78562367e-01
1.57944226e+00 -9.31758583e-01 7.78730214e-01 -2.85659343e-01
-6.27092540e-01 1.48291588e+00 2.63293535e-01 -2.62792110e-01
4.82289083e-02 -1.93238258e-02 -4.24321771e-01 -5.48740387e-01
-9.60139453e-01 2.10650548e-01 4.59564000e-01 2.06337422e-01
7.06968606e-01 -1.57074690e-01 -1.34113133e+00 6.43250823e-01
-4.51313019e-01 3.84679623e-02 1.09837615e+00 9.44560289e-01
-8.44408095e-01 -1.00602591e+00 -2.82282710e-01 9.43250179e-01
-5.15011966e-01 -2.86900789e-01 -1.17717862e+00 2.79123545e-01
-3.18046987e-01 1.13340414e+00 -2.33555749e-01 3.29698399e-02
3.92865419e-01 -1.78820326e-03 4.92863685e-01 -9.45811212e-01
-1.05758655e+00 -2.14384809e-01 5.42859554e-01 -9.94015783e-02
-4.44778264e-01 -6.63615584e-01 -1.19267786e+00 -4.86976415e-01
6.27140030e-02 6.43536925e-01 8.13760221e-01 6.80269957e-01
9.07423794e-01 1.04301977e+00 4.57827061e-01 -5.81736743e-01
-3.30359101e-01 -1.48849988e+00 -6.18998528e-01 1.57328919e-01
6.27310351e-02 1.91389889e-01 -1.50839135e-01 3.05927694e-01] | [12.470355987548828, 9.5418701171875] |
a38d80cf-8a24-417e-829f-11eedcf5c496 | scai-qrecc-shared-task-on-conversational | 2201.11094 | null | https://arxiv.org/abs/2201.11094v1 | https://arxiv.org/pdf/2201.11094v1.pdf | SCAI-QReCC Shared Task on Conversational Question Answering | Search-Oriented Conversational AI (SCAI) is an established venue that regularly puts a spotlight upon the recent work advancing the field of conversational search. SCAI'21 was organised as an independent on-line event and featured a shared task on conversational question answering. Since all of the participant teams experimented with answer generation models for this task, we identified evaluation of answer correctness in this settings as the major challenge and a current research gap. Alongside the automatic evaluation, we conducted two crowdsourcing experiments to collect annotations for answer plausibility and faithfulness. As a result of this shared task, the original conversational QA dataset used for evaluation was further extended with alternative correct answers produced by the participant systems. | ['Maik Fröbe', 'Johannes Kiesel', 'Svitlana Vakulenko'] | 2022-01-26 | null | https://aclanthology.org/2022.lrec-1.525 | https://aclanthology.org/2022.lrec-1.525.pdf | lrec-2022-6 | ['conversational-search'] | ['natural-language-processing'] | [ 2.68220067e-01 7.68144667e-01 2.79149741e-01 -4.76833642e-01
-1.26452947e+00 -8.61800015e-01 1.25450444e+00 3.58372927e-01
-4.44265157e-01 9.94140506e-01 8.99075806e-01 -4.89508033e-01
-5.70160039e-02 -3.51977527e-01 -1.43374175e-01 1.17468335e-01
3.21756512e-01 1.23147225e+00 5.48665822e-01 -7.29732394e-01
7.07296193e-01 -2.90996850e-01 -1.33322275e+00 9.05663371e-01
9.48917210e-01 6.08129263e-01 -3.24524730e-01 1.06750154e+00
-3.01273048e-01 1.53787744e+00 -9.19759989e-01 -9.84158099e-01
-2.39856131e-02 -6.97114170e-01 -1.98325861e+00 -3.09395820e-01
2.97552675e-01 -8.01419318e-02 -9.79720429e-02 5.28626382e-01
6.55286133e-01 4.06833857e-01 1.93299770e-01 -1.66878748e+00
-5.27064741e-01 5.85391879e-01 4.69633400e-01 2.65439272e-01
1.28857386e+00 4.96549606e-01 1.39130461e+00 -9.12248790e-01
8.38006377e-01 1.38973105e+00 5.87042093e-01 6.94154918e-01
-9.33053911e-01 -1.62985355e-01 -2.30291113e-01 4.79957938e-01
-9.37141597e-01 -6.68617070e-01 4.55148995e-01 -3.06467623e-01
1.21986449e+00 7.58818984e-01 4.40005004e-01 1.07649732e+00
-3.72081041e-01 1.00496864e+00 1.30851209e+00 -6.69977963e-01
2.79382050e-01 4.20787036e-01 4.00112212e-01 3.14931154e-01
-3.26909035e-01 -1.68760538e-01 -8.31904054e-01 -7.16124356e-01
3.84451449e-02 -7.96831369e-01 -4.86721963e-01 -1.41577786e-02
-1.23180330e+00 1.01596665e+00 9.79201421e-02 4.27087873e-01
-1.28212512e-01 -8.74515325e-02 4.97449398e-01 6.02974892e-01
4.97240156e-01 1.00609434e+00 -2.22854793e-01 -7.62979269e-01
-3.51959020e-01 1.08395195e+00 1.85619128e+00 6.36727810e-01
4.52558309e-01 -7.32652962e-01 -7.39894152e-01 1.00193667e+00
2.63869584e-01 1.84759885e-01 3.27337801e-01 -1.53758895e+00
6.53703690e-01 1.12447071e+00 8.62663448e-01 -9.42176998e-01
-3.00774902e-01 1.95003808e-01 -4.55213301e-02 -1.24236129e-01
8.33422840e-01 -1.59168109e-01 -6.35990947e-02 1.40051103e+00
4.76900458e-01 -4.38043445e-01 2.13633493e-01 9.71303344e-01
1.10196924e+00 3.28863233e-01 -5.19167930e-02 3.59420776e-02
1.41462207e+00 -1.22877157e+00 -7.32687175e-01 -1.59090146e-01
6.42547429e-01 -1.12227106e+00 1.10655093e+00 9.85005498e-02
-1.40055656e+00 -2.20205158e-01 -7.41693377e-01 -2.96026438e-01
-2.03224272e-01 -4.13754016e-01 4.82590497e-01 4.08298373e-01
-1.44970787e+00 9.85105857e-02 -9.68281627e-02 -5.21418273e-01
-1.09268717e-01 -7.15586692e-02 -5.85734770e-02 -2.08422527e-01
-1.69916046e+00 1.57787383e+00 -1.08526513e-01 1.47791862e-01
-5.72033226e-01 -4.07715768e-01 -5.33112109e-01 -3.27906162e-01
5.44453204e-01 -6.54582918e-01 2.04327750e+00 -6.49785936e-01
-1.65056527e+00 1.21416461e+00 -4.25791264e-01 -5.97974896e-01
6.15895450e-01 -3.40965018e-02 -3.99362296e-01 6.88270256e-02
5.79957187e-01 8.12501013e-01 2.43206441e-01 -1.02602804e+00
-7.26029456e-01 -1.59364477e-01 7.25258291e-01 5.13571262e-01
3.48746628e-01 4.49029773e-01 -2.87786312e-02 -2.85579991e-02
-2.48303697e-01 -1.09391916e+00 1.32383868e-01 -4.84720528e-01
-1.72518134e-01 -1.05695939e+00 3.74028236e-01 -6.16186798e-01
1.15719151e+00 -1.56586695e+00 6.06441870e-02 -6.60963282e-02
3.54933530e-01 2.82333463e-01 -1.69026051e-02 1.08088851e+00
3.98796201e-01 1.80174515e-01 -1.86244175e-01 -1.83742166e-01
2.86863059e-01 6.80421367e-02 -4.14798230e-01 -9.61124748e-02
1.70646593e-01 1.14927864e+00 -1.28494990e+00 -5.16913414e-01
-1.84320837e-01 -2.12050110e-01 -3.57504547e-01 3.92314851e-01
-7.12404072e-01 6.74204767e-01 -2.92440891e-01 2.80679405e-01
8.95530581e-02 -3.31090122e-01 1.42052146e-02 3.35211456e-01
-2.01053798e-01 9.36270416e-01 -8.01542640e-01 1.57110918e+00
-3.82974684e-01 7.33460426e-01 2.99456120e-01 -3.32645714e-01
8.79873514e-01 8.05409312e-01 1.55035645e-01 -6.77845359e-01
-1.12886839e-01 5.39132893e-01 2.86619544e-01 -7.34939277e-01
8.20917070e-01 1.75383314e-01 -2.01577649e-01 9.68167663e-01
-1.05302207e-01 -7.29494274e-01 1.76797301e-01 5.75388014e-01
1.12130177e+00 -3.89824691e-03 1.80320442e-01 -3.78917217e-01
9.41481531e-01 7.80967414e-01 3.66673656e-02 1.00871778e+00
-5.90870738e-01 4.07082707e-01 3.77868980e-01 -7.67069221e-01
-8.41202617e-01 -4.61348295e-01 6.80568963e-02 1.15861344e+00
1.23473167e-01 -5.94175160e-01 -8.33781779e-01 -7.25688100e-01
-4.33590412e-01 8.35474432e-01 -4.67803389e-01 3.10629517e-01
-5.51616967e-01 -1.40653640e-01 7.26323903e-01 1.23426415e-01
6.67062700e-01 -1.50479543e+00 -7.24984348e-01 3.80430192e-01
-1.07461774e+00 -9.75606561e-01 -3.78564626e-01 -2.29885459e-01
-2.84823179e-02 -1.41879499e+00 -5.37508190e-01 -6.68623865e-01
-1.60940923e-02 3.66580039e-01 1.80113947e+00 5.31258821e-01
-8.63712560e-03 8.77645969e-01 -6.05474591e-01 -6.73323154e-01
-8.06702793e-01 1.91976875e-01 -5.27900100e-01 -2.93503404e-01
7.15087712e-01 -2.72790045e-01 -8.68027866e-01 7.32693493e-01
-6.35932148e-01 -2.11273301e-02 -2.70130783e-01 6.88265204e-01
-3.35000336e-01 -8.74722481e-01 8.49751472e-01 -6.81388021e-01
1.64751887e+00 -4.46900457e-01 -1.92863554e-01 4.88061756e-01
-4.52936321e-01 -1.05302878e-01 -8.37017745e-02 1.45452738e-01
-1.42725968e+00 -5.31938791e-01 -1.72307834e-01 7.67022729e-01
-1.06302775e-01 4.05170590e-01 1.73983961e-01 -8.90178159e-02
1.17980933e+00 -2.51720667e-01 1.19244523e-01 4.03423384e-02
4.89503860e-01 9.33012486e-01 4.52540278e-01 -8.08213830e-01
4.24657315e-01 4.47763316e-02 -6.47836447e-01 -5.33653677e-01
-9.69664097e-01 -7.17330754e-01 -2.13396803e-01 -5.09826779e-01
8.16701710e-01 -7.49038815e-01 -1.11747575e+00 2.63967603e-01
-1.60410619e+00 -6.13356173e-01 -1.47217035e-01 -1.03620626e-01
-3.84107292e-01 3.92792255e-01 -4.23205256e-01 -1.02780342e+00
-7.35858202e-01 -9.23803985e-01 7.99463868e-01 1.20557643e-01
-1.24930608e+00 -9.56896901e-01 6.75156176e-01 1.16021025e+00
8.19605768e-01 -1.41620943e-02 7.84604669e-01 -1.26087165e+00
-3.70220244e-01 -3.92772377e-01 -1.38864979e-01 -2.63593905e-02
-4.24302816e-01 -2.86470503e-01 -1.07799470e+00 2.62514114e-01
2.16107413e-01 -1.01659131e+00 2.01172054e-01 -3.28647494e-01
2.39650756e-01 -5.05510390e-01 6.62541091e-02 -7.36584842e-01
6.90678954e-01 1.77563459e-01 4.39397037e-01 5.19722104e-01
1.46773830e-01 1.11775696e+00 5.32790601e-01 1.77443355e-01
9.89410579e-01 9.35691297e-01 1.11217342e-01 5.21554112e-01
-6.45291507e-02 -2.12339163e-01 2.77671423e-02 6.91185296e-01
6.45987391e-02 -2.94806093e-01 -1.28315699e+00 9.90613997e-01
-2.19260716e+00 -1.05281627e+00 -6.67809308e-01 2.01872635e+00
1.05207610e+00 -8.08548629e-02 2.95262575e-01 -6.95138946e-02
5.97444177e-01 1.00474283e-01 -1.74962804e-01 -6.39230847e-01
1.21216699e-02 2.33191162e-01 -5.09648323e-01 1.17525828e+00
-5.73711216e-01 8.18806827e-01 6.69083977e+00 3.65053594e-01
-1.71032295e-01 2.40583375e-01 5.40513694e-01 2.64949858e-01
-7.01039970e-01 3.83483142e-01 -4.70330715e-01 4.47834373e-01
1.01297498e+00 -4.57670480e-01 5.77486634e-01 5.93410909e-01
-1.31591067e-01 -6.30938351e-01 -1.29359806e+00 4.35175300e-01
1.52955368e-01 -1.38946867e+00 -3.08309942e-01 -4.13807482e-01
6.85138762e-01 1.48154542e-01 -6.35345519e-01 6.38327599e-01
7.86707878e-01 -1.13415384e+00 4.34873432e-01 4.14863884e-01
2.16036037e-01 -2.62651026e-01 1.04510605e+00 5.64645529e-01
-6.01535797e-01 6.47572726e-02 1.68640316e-01 -7.49787211e-01
4.41947103e-01 -4.75046560e-02 -1.35971177e+00 1.87814474e-01
5.21329701e-01 -1.92560837e-01 -6.68139994e-01 9.41545010e-01
-5.31575084e-01 7.24563003e-01 -1.88321739e-01 -6.46579146e-01
2.97139406e-01 -4.62889113e-02 7.51846135e-01 1.08912098e+00
-5.13743162e-01 2.96291858e-01 3.41203064e-01 6.54576182e-01
-1.49374068e-01 -1.06341697e-01 -5.92558503e-01 1.97751641e-01
9.40681338e-01 1.20759773e+00 -3.05386275e-01 -3.65036190e-01
-1.71794116e-01 9.21257973e-01 3.62015575e-01 1.98801681e-01
-3.97858143e-01 -2.93663383e-01 2.27313951e-01 -2.58041352e-01
-3.14832687e-01 6.65918589e-02 -4.37497109e-01 -8.80169451e-01
2.47189164e-01 -1.44771135e+00 4.28254187e-01 -1.00461316e+00
-1.29073048e+00 8.15014482e-01 2.60547036e-03 -6.67808473e-01
-1.03039181e+00 -1.81336462e-01 -9.15957451e-01 1.04931569e+00
-1.15923810e+00 -9.10766184e-01 -6.93931282e-01 5.35673082e-01
7.30702162e-01 1.02478035e-01 1.15769267e+00 -5.94177619e-02
-4.07470874e-02 2.69942939e-01 -5.28223991e-01 -6.46049306e-02
8.86899054e-01 -1.23837781e+00 7.18762696e-01 2.95662522e-01
2.93261617e-01 7.52954543e-01 9.05427277e-01 -4.43504632e-01
-1.02639830e+00 -3.23806107e-01 1.81097078e+00 -1.39692187e+00
9.50444996e-01 -1.86155915e-01 -9.36259091e-01 4.91062164e-01
9.54008698e-01 -6.49260342e-01 7.86969543e-01 1.44102022e-01
1.32144559e-02 3.92137438e-01 -9.75905120e-01 5.93280256e-01
8.30483079e-01 -1.06259906e+00 -1.08646512e+00 7.90379286e-01
9.59179223e-01 -5.46138585e-01 -6.28575623e-01 9.49802101e-02
2.05025971e-01 -9.97254252e-01 6.27204180e-01 -4.39772844e-01
5.41878879e-01 -2.23897442e-01 4.15871218e-02 -1.17095375e+00
2.21222907e-01 -1.19090128e+00 2.52514929e-01 1.48553967e+00
7.10629046e-01 -4.57632571e-01 4.91654992e-01 1.57237363e+00
-1.49857670e-01 -4.88880575e-01 -1.00604284e+00 -1.18497863e-01
-2.35347403e-03 -3.87188613e-01 4.01877046e-01 6.65260732e-01
5.75184822e-01 9.75227535e-01 -1.05872013e-01 -4.30125713e-01
3.82984579e-02 3.16126458e-02 1.11889076e+00 -1.21367455e+00
-1.26181930e-01 -4.26280886e-01 1.04835346e-01 -9.24141526e-01
-4.39858772e-02 -9.04306769e-01 2.49850810e-01 -1.71591580e+00
3.59989814e-02 -1.01876967e-01 5.65126657e-01 2.43575171e-01
-5.38735271e-01 2.05253735e-01 2.09368795e-01 4.89018321e-01
-1.03155518e+00 3.24184448e-01 9.50945258e-01 -4.34959047e-02
-1.94881961e-01 3.31990153e-01 -9.45231080e-01 5.71214736e-01
7.28560984e-01 -7.71505609e-02 -5.80742061e-01 -5.16048074e-01
8.11274052e-01 2.99195766e-01 5.82693577e-01 -7.21745729e-01
6.89402938e-01 7.18143359e-02 -5.13903141e-01 -4.49809164e-01
3.28338265e-01 -3.92320126e-01 -2.73004800e-01 2.23964676e-02
-1.00228536e+00 1.77313089e-01 4.86172922e-02 3.21574390e-01
-5.33175409e-01 -5.67648113e-01 -2.24144962e-02 -3.57518375e-01
-4.77580726e-01 -3.31969619e-01 -5.60063541e-01 8.59881103e-01
6.58293426e-01 -2.04948425e-01 -6.77290738e-01 -9.76638615e-01
-3.41229886e-01 7.04085708e-01 1.23344354e-01 4.32258457e-01
3.54008377e-01 -1.01996732e+00 -1.19225121e+00 -5.69387197e-01
4.25177217e-01 4.17613201e-02 6.65484741e-02 7.37140954e-01
-3.56855482e-01 7.51837432e-01 1.36510924e-01 -3.43088925e-01
-1.14525414e+00 8.91081691e-02 3.07826936e-01 -5.15906572e-01
-2.26152003e-01 9.51243639e-01 -6.03919566e-01 -9.13382888e-01
2.89050430e-01 1.32463336e-01 -3.93796921e-01 -4.63511758e-02
7.31757998e-01 5.90122163e-01 9.36897174e-02 -1.87082663e-01
-2.51582354e-01 -2.79382646e-01 3.55669372e-02 -9.09503102e-01
8.84108663e-01 -3.14090639e-01 -4.43263173e-01 5.38019955e-01
7.85934269e-01 -4.46413942e-02 -6.77340031e-01 -4.41139996e-01
6.02351308e-01 -2.67268568e-01 -6.03517354e-01 -1.51569569e+00
2.45429948e-01 4.67016786e-01 -2.32581254e-02 7.58516967e-01
3.25320274e-01 2.03261659e-01 7.20775962e-01 7.78677464e-01
4.60726768e-01 -1.13150978e+00 2.24168256e-01 9.34044719e-01
1.60604572e+00 -1.53686607e+00 -3.28654647e-01 -2.14952782e-01
-1.03518212e+00 9.54297423e-01 7.58370161e-01 3.22486728e-01
-1.05243646e-01 -2.88331985e-01 2.88509279e-01 -6.21273577e-01
-1.01074946e+00 -9.96432453e-02 3.02419931e-01 5.72574615e-01
7.30578661e-01 -1.31621122e-01 -7.62004733e-01 5.08091450e-01
-4.80147123e-01 9.76463780e-02 3.72214407e-01 9.48727667e-01
-3.90244961e-01 -1.11157334e+00 -3.12616616e-01 2.19854321e-02
-2.16847345e-01 -1.71562701e-01 -1.40363967e+00 6.98001266e-01
-4.60032582e-01 1.95048642e+00 -2.10754707e-01 -2.28398845e-01
4.49749857e-01 4.44336027e-01 1.22778006e-01 -5.54817796e-01
-1.37464869e+00 -9.45843160e-01 1.09109724e+00 -5.75318635e-01
-5.23815751e-01 -5.08046150e-01 -8.52678955e-01 -2.97935635e-01
-3.20594788e-01 8.29905629e-01 5.36864460e-01 1.26304924e+00
5.27318597e-01 -1.32639527e-01 3.85321558e-01 -2.99617052e-01
-6.88501656e-01 -1.41082501e+00 4.28085625e-01 8.31342936e-01
-1.88363343e-03 -2.29772419e-01 -3.98110330e-01 -2.34698012e-01] | [12.089581489562988, 7.973697662353516] |
a93e4dbc-79ea-446d-abe4-5dba0474bceb | f2-nerf-fast-neural-radiance-field-training | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_F2-NeRF_Fast_Neural_Radiance_Field_Training_With_Free_Camera_Trajectories_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_F2-NeRF_Fast_Neural_Radiance_Field_Training_With_Free_Camera_Trajectories_CVPR_2023_paper.pdf | F2-NeRF: Fast Neural Radiance Field Training With Free Camera Trajectories | This paper presents a novel grid-based NeRF called F^2-NeRF (Fast-Free-NeRF) for novel view synthesis, which enables arbitrary input camera trajectories and only costs a few minutes for training. Existing fast grid-based NeRF training frameworks, like Instant-NGP, Plenoxels, DVGO, or TensoRF, are mainly designed for bounded scenes and rely on space warping to handle unbounded scenes. Existing two widely-used space-warping methods are only designed for the forward-facing trajectory or the 360deg object-centric trajectory but cannot process arbitrary trajectories. In this paper, we delve deep into the mechanism of space warping to handle unbounded scenes. Based on our analysis, we further propose a novel space-warping method called perspective warping, which allows us to handle arbitrary trajectories in the grid-based NeRF framework. Extensive experiments demonstrate that F^2-NeRF is able to use the same perspective warping to render high-quality images on two standard datasets and a new free trajectory dataset collected by us. | ['Wenping Wang', 'Christian Theobalt', 'Taku Komura', 'Ziwei Liu', 'Lingjie Liu', 'Zhaoxi Chen', 'YuAn Liu', 'Peng Wang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['novel-view-synthesis'] | ['computer-vision'] | [-1.78578883e-01 -4.62755919e-01 9.24172178e-02 -2.65132897e-02
-5.14171124e-01 -1.01560867e+00 8.03494692e-01 -6.09739542e-01
-2.30050012e-01 5.52787662e-01 2.36296967e-01 -5.83688796e-01
-1.58861250e-01 -7.35164404e-01 -7.33516216e-01 -6.00799859e-01
-1.97740062e-03 1.38099223e-01 4.09810901e-01 -4.20600504e-01
1.44931033e-01 9.14031148e-01 -1.26920187e+00 1.33679330e-01
6.12792969e-01 4.65346038e-01 1.99007496e-01 9.90492463e-01
5.41638397e-03 6.24674737e-01 -2.81735241e-01 -3.53338540e-01
5.72024167e-01 -2.83877969e-01 -3.69087666e-01 1.19557887e-01
5.69705844e-01 -6.67973042e-01 -8.36166739e-01 5.63208520e-01
5.08356571e-01 6.19580686e-01 4.52060819e-01 -1.29125464e+00
-3.35886478e-01 -4.70928513e-02 -2.34875798e-01 2.51646876e-01
4.96914148e-01 2.90033191e-01 4.01156873e-01 -1.17440009e+00
1.04815316e+00 1.09882510e+00 6.49460196e-01 6.98988855e-01
-8.77920866e-01 -4.52670485e-01 1.61652893e-01 1.62202165e-01
-1.24100697e+00 -2.08095297e-01 7.17128038e-01 -3.40112507e-01
1.09180272e+00 6.02453589e-01 8.41028690e-01 1.35560179e+00
4.14571345e-01 5.46902359e-01 1.08985686e+00 -2.59927325e-02
2.08264709e-01 -5.50226331e-01 -3.41478229e-01 3.86079043e-01
-3.05902481e-01 4.85687703e-01 -2.99132943e-01 6.60660267e-02
1.38465464e+00 1.26094446e-01 -5.66501260e-01 -3.80363405e-01
-2.21384001e+00 6.44800961e-01 4.04954731e-01 -2.29259711e-02
-1.12518044e-02 9.78768095e-02 4.85798955e-01 1.84309021e-01
4.76569921e-01 4.58460689e-01 -2.79482394e-01 -4.76764590e-01
-9.01137769e-01 5.17179370e-01 4.91600603e-01 1.39440084e+00
3.59430283e-01 2.29304329e-01 -2.81081140e-01 3.25706333e-01
-2.80283868e-01 5.43505371e-01 1.80738598e-01 -1.33608997e+00
5.80526888e-01 1.02539018e-01 4.06083316e-01 -1.00243950e+00
-4.53573406e-01 -3.81614044e-02 -9.48358297e-01 3.43196094e-01
1.80455104e-01 -1.87095508e-01 -8.53908300e-01 1.35823703e+00
5.49333155e-01 5.30145943e-01 7.65143707e-02 1.09933376e+00
7.37948596e-01 1.13458931e+00 -5.11773288e-01 -3.20019066e-01
1.03435850e+00 -1.32696116e+00 -7.29718268e-01 3.34985644e-01
5.45896828e-01 -8.35864842e-01 1.38939881e+00 5.15226245e-01
-9.90839064e-01 -4.99722153e-01 -8.89193892e-01 -2.77746826e-01
-2.26940349e-01 1.95482761e-01 5.70333958e-01 1.42375246e-01
-1.06902599e+00 7.50911772e-01 -8.74152005e-01 -4.10519838e-01
-6.86519668e-02 -3.46307494e-02 -6.43196046e-01 -2.21663028e-01
-8.06486189e-01 6.56676412e-01 3.02374423e-01 8.19446072e-02
-9.23895121e-01 -5.92544973e-01 -6.69748425e-01 -4.76433039e-02
4.09735948e-01 -9.83187258e-01 1.02476895e+00 -4.04045016e-01
-1.88604057e+00 2.37605438e-01 -9.82643440e-02 -9.31567252e-02
9.82296765e-01 -2.90165424e-01 -4.25450265e-01 1.26317352e-01
1.95792858e-02 7.40805924e-01 9.11887348e-01 -1.17357457e+00
-2.71076798e-01 3.49535607e-02 2.94001967e-01 5.53977132e-01
-1.32123798e-01 7.19981454e-03 -4.16747421e-01 -1.02851677e+00
-1.28416419e-02 -1.06315112e+00 -3.22369903e-01 1.74444109e-01
-4.46635395e-01 5.67931421e-02 1.33507943e+00 -3.55635941e-01
1.14571702e+00 -2.23354435e+00 4.05962080e-01 -2.65385538e-01
1.51570901e-01 2.59955883e-01 -2.27874652e-01 8.49362791e-01
-2.30966523e-01 -5.23950830e-02 -2.28101071e-02 -4.03822958e-01
-1.38760775e-01 4.43725228e-01 -7.44861305e-01 4.41854566e-01
-2.75283337e-01 7.21024394e-01 -1.20010006e+00 -3.02007169e-01
6.80168390e-01 6.57092810e-01 -5.38904607e-01 2.85665512e-01
-1.57322735e-02 6.95926070e-01 1.33949919e-02 3.47444147e-01
7.76714265e-01 1.00580923e-01 -1.57642215e-01 -4.32047158e-01
-6.78473711e-01 -2.01394260e-01 -1.05378759e+00 2.12520027e+00
-4.74770248e-01 8.11901510e-01 -1.66194901e-01 -1.43287659e-01
6.78442121e-01 2.57176489e-01 4.14432794e-01 -3.35021138e-01
1.20772175e-01 1.53963536e-01 -4.97630566e-01 -4.80191469e-01
7.80807436e-01 -1.18850313e-01 1.22959733e-01 1.38817251e-01
-1.38873877e-02 -2.43784368e-01 1.81965023e-01 2.28511065e-01
1.09275353e+00 5.72001517e-01 9.41459090e-02 -1.16493158e-01
3.67017299e-01 -4.79421616e-02 6.44020796e-01 2.72609770e-01
8.76382738e-03 1.10167682e+00 1.97698712e-01 -7.50960588e-01
-1.39082491e+00 -1.23504531e+00 9.04566944e-02 5.92906952e-01
3.85002643e-01 -8.47232759e-01 -7.52682328e-01 -7.69846201e-01
-5.22948623e-01 8.53761256e-01 -3.61253142e-01 2.48969570e-01
-7.90525973e-01 -2.81464040e-01 2.75922894e-01 6.94358706e-01
6.77225590e-01 -6.78789496e-01 -6.61548316e-01 1.46166444e-01
-3.06642234e-01 -1.36614060e+00 -1.01078868e+00 -4.69315618e-01
-9.31856453e-01 -9.69648004e-01 -1.06544173e+00 -4.05657291e-01
6.68310940e-01 9.02785361e-01 7.17168629e-01 -3.45767051e-01
1.45735934e-01 4.32739198e-01 -4.66490597e-01 2.99625043e-02
-7.19532296e-02 -1.99435860e-01 2.95101434e-01 -9.17036831e-02
-3.33539963e-01 -7.44051278e-01 -7.78772414e-01 7.67654777e-01
-8.71590018e-01 6.32725477e-01 1.15243673e-01 7.81861722e-01
6.10883892e-01 6.92270920e-02 5.66366240e-02 -4.56924349e-01
2.77697802e-01 -2.41963431e-01 -6.04560971e-01 1.43338330e-02
-2.15443850e-01 -1.44840255e-01 1.07934463e+00 -7.50346303e-01
-9.84307349e-01 -3.34238820e-02 -2.77300954e-01 -1.25147688e+00
7.37077370e-02 1.21883467e-01 -1.68640271e-01 -2.38965720e-01
6.16612911e-01 3.55943978e-01 -4.05859768e-01 -2.97201097e-01
6.68493509e-01 1.67957321e-01 8.03185344e-01 -4.82136250e-01
1.13296402e+00 9.07221437e-01 1.18348137e-01 -8.04199219e-01
-4.82257992e-01 -3.16587150e-01 -7.00849593e-01 -5.77665031e-01
1.10140800e+00 -8.19984674e-01 -6.44436955e-01 3.94718558e-01
-1.39016271e+00 -4.70945954e-01 -4.23433542e-01 6.78719103e-01
-9.29826260e-01 5.18219650e-01 -4.06589150e-01 -4.28836107e-01
-2.00086400e-01 -1.13372052e+00 1.34019876e+00 -2.02046428e-02
1.09738827e-01 -1.02706945e+00 2.74527222e-01 3.14304382e-02
3.94862890e-01 7.37932861e-01 4.91006345e-01 -3.03970072e-02
-8.33660066e-01 1.16823902e-02 -1.42785460e-01 1.02411203e-01
-6.51827082e-02 2.92066872e-01 -5.73620379e-01 -4.59276825e-01
6.37179241e-02 1.57801136e-01 5.60678363e-01 2.01518074e-01
1.10737765e+00 -3.61746490e-01 -2.93744802e-01 1.33939719e+00
1.42765784e+00 2.85716414e-01 9.26748097e-01 3.19034606e-01
1.10331094e+00 2.83113599e-01 8.17371011e-01 3.04293662e-01
4.09884900e-01 9.17165995e-01 5.79971433e-01 -1.06393427e-01
-2.39087064e-02 -5.02583265e-01 4.45925474e-01 1.08165884e+00
-5.67598343e-01 -4.16707456e-01 -7.20773816e-01 4.48057353e-01
-1.98988664e+00 -1.14677358e+00 -1.84037879e-01 2.13027573e+00
1.84588060e-01 -1.17059700e-01 1.19121067e-01 4.27697599e-02
5.60883343e-01 4.63547498e-01 -5.01850247e-01 -3.35161179e-01
9.76317097e-03 -1.18386537e-01 6.16821408e-01 5.21736085e-01
-9.46348727e-01 9.49771106e-01 6.28885937e+00 9.37600970e-01
-1.21807528e+00 1.56083375e-01 6.88770115e-02 -3.33891809e-01
-4.97491539e-01 1.71852693e-01 -5.55637836e-01 3.73318225e-01
7.95711398e-01 -2.52475262e-01 6.73303127e-01 8.75055254e-01
4.94410098e-01 2.85654694e-01 -9.74982440e-01 1.43252599e+00
8.41350779e-02 -1.64722300e+00 1.94254428e-01 5.45140356e-02
7.93263972e-01 -5.70325293e-02 -7.38260448e-02 3.40207875e-01
1.34744510e-01 -8.57933581e-01 6.84928119e-01 6.08179033e-01
1.23160160e+00 -8.79158258e-01 3.50692958e-01 5.56733608e-01
-1.36047506e+00 5.20742089e-02 -4.74786609e-01 1.70177482e-02
6.24638498e-01 4.53765571e-01 -5.10266542e-01 8.96003187e-01
4.68165457e-01 9.04834569e-01 -1.07884690e-01 7.85001576e-01
-9.59324017e-02 2.86958843e-01 -3.42767417e-01 4.28254932e-01
4.19491768e-01 -2.82708615e-01 1.00536835e+00 9.84610319e-01
1.03582466e+00 3.80517602e-01 1.05507649e-01 5.23820162e-01
2.40829095e-01 -1.74890205e-01 -9.70871031e-01 2.71080047e-01
2.39183843e-01 1.41763985e+00 -7.82105625e-01 -1.43189281e-01
-3.08116853e-01 1.42162979e+00 -4.26141825e-03 5.23943067e-01
-1.27078950e+00 -2.96866506e-01 6.71652377e-01 2.97060132e-01
2.49276906e-01 -9.28625405e-01 1.50398329e-01 -1.76724613e+00
-6.40171394e-02 -6.52802885e-01 -6.77132830e-02 -1.24759185e+00
-9.87992525e-01 8.35987628e-01 3.52544874e-01 -1.98585737e+00
-1.42340630e-01 -6.61454678e-01 -8.53273273e-01 5.15400231e-01
-1.38551354e+00 -1.39393890e+00 -7.20314503e-01 9.84736323e-01
8.38841319e-01 -7.58503675e-02 7.27951825e-01 1.92790896e-01
-4.79599684e-01 2.17119932e-01 2.82046050e-01 -1.93872988e-01
7.42769957e-01 -1.21283734e+00 9.15710032e-01 9.78660345e-01
1.10020734e-01 4.18877780e-01 5.97142518e-01 -4.78171289e-01
-1.75364852e+00 -1.42291784e+00 6.02060199e-01 -6.35315061e-01
3.79689693e-01 -4.40456718e-01 -6.53535724e-01 8.57683837e-01
2.84157157e-01 3.70992243e-01 4.11802620e-01 -4.24624234e-01
-2.90885985e-01 -4.46172394e-02 -1.04031014e+00 9.75895107e-01
1.43664849e+00 -3.32733333e-01 -3.59414637e-01 5.85432827e-01
9.90348279e-01 -9.90039110e-01 -8.97511125e-01 3.25965941e-01
3.77558768e-01 -1.27452254e+00 1.07166421e+00 -3.56131613e-01
4.40580904e-01 -7.05045938e-01 -2.18358740e-01 -1.51265192e+00
-4.91076022e-01 -1.18297434e+00 -1.75353110e-01 8.26414704e-01
-3.21500689e-01 -5.95020533e-01 6.16266787e-01 -3.34152319e-02
-5.33216119e-01 -8.69865716e-01 -1.09022498e+00 -1.18203533e+00
-4.43682596e-02 -5.47401726e-01 7.56302655e-01 1.06003451e+00
-2.66433954e-01 8.52347165e-02 -5.53755283e-01 1.91593468e-01
5.28564751e-01 1.36359679e-02 1.15667832e+00 -6.22118294e-01
-2.11110353e-01 4.93889973e-02 -3.23159605e-01 -1.27752709e+00
-7.46965855e-02 -5.94752252e-01 -2.01236531e-01 -1.57654154e+00
-1.98537737e-01 -1.32968068e-01 3.27689320e-01 1.65517613e-01
6.91717193e-02 1.58049777e-01 5.45174778e-01 3.94105047e-01
-2.10241780e-01 7.71328211e-01 1.70277441e+00 2.64855981e-01
-1.49204478e-01 -2.91173816e-01 -2.72936020e-02 7.21863210e-01
4.21681464e-01 -3.00962273e-02 -9.07722533e-01 -6.07132435e-01
3.05124491e-01 4.72576618e-01 5.50925314e-01 -1.25524318e+00
1.72330588e-01 -5.16024709e-01 2.16121107e-01 -8.97143960e-01
5.29147744e-01 -9.08241510e-01 5.88044703e-01 -1.34794533e-01
1.89879462e-01 6.81358099e-01 1.86044097e-01 5.21276474e-01
-1.26180619e-01 3.12264711e-01 5.73860466e-01 -1.46525338e-01
-7.06010759e-01 7.24695981e-01 -1.45206749e-01 -1.07488178e-01
1.51604283e+00 -3.51331055e-01 -4.73536491e-01 -5.58831453e-01
-8.38692546e-01 1.45279899e-01 8.79763246e-01 3.67059201e-01
9.68958557e-01 -1.84006000e+00 -5.69843709e-01 2.75150031e-01
-4.81861494e-02 2.87693501e-01 6.61053061e-01 7.43147612e-01
-1.06943130e+00 3.35198581e-01 -2.80223966e-01 -6.90218270e-01
-1.14441645e+00 9.49280381e-01 2.82734454e-01 -2.21292034e-01
-1.08437479e+00 6.23967469e-01 3.84008557e-01 -5.34355521e-01
-2.75378644e-01 -5.58626711e-01 1.72090054e-01 -3.15131485e-01
6.18691444e-01 5.30860782e-01 -9.69267115e-02 -6.19816899e-01
-1.48439944e-01 8.41508567e-01 1.70387089e-01 -3.06230307e-01
1.22406244e+00 -1.31434619e-01 2.21977264e-01 4.27656353e-01
1.15189493e+00 1.31530121e-01 -1.75486624e+00 3.40071529e-01
-8.99388254e-01 -8.72597814e-01 -5.02718240e-02 -3.67172927e-01
-9.14369166e-01 1.04787624e+00 2.64984667e-01 -3.36535797e-02
1.31099367e+00 -6.41816616e-01 1.19324613e+00 1.80412218e-01
7.20008671e-01 -6.55964017e-01 9.01163928e-03 7.17589557e-01
1.17212987e+00 -7.43164480e-01 -7.69562870e-02 -5.34832597e-01
-7.21297324e-01 1.52833140e+00 6.53362989e-01 -2.74009556e-01
3.35680664e-01 3.39879572e-01 -1.27871007e-01 -2.86087282e-02
-8.22743833e-01 2.47548103e-01 1.39415324e-01 6.48355901e-01
-6.15668111e-02 -7.60354102e-02 -5.70610352e-02 3.67245860e-02
-3.93064290e-01 1.24224164e-01 8.25195789e-01 7.84555674e-01
6.09599091e-02 -9.68605697e-01 -5.19771516e-01 -1.38735911e-02
2.01687738e-01 6.07643388e-02 -5.29499948e-02 9.49036181e-01
4.84961681e-02 4.13967222e-01 1.86253458e-01 -7.45205402e-01
5.87699354e-01 -2.79282689e-01 6.86231375e-01 -3.58936787e-01
-3.98970813e-01 8.19745809e-02 1.01146318e-01 -1.01832783e+00
-4.45223302e-01 -6.12626851e-01 -1.12200284e+00 -7.96510816e-01
3.36728804e-03 -1.88568220e-01 4.91512567e-01 6.29733980e-01
4.75284636e-01 4.72049981e-01 7.89134145e-01 -1.51634502e+00
-2.12572366e-01 -5.09795010e-01 -3.01212788e-01 2.05379263e-01
3.85002047e-01 -7.42857575e-01 -3.84719223e-01 4.80917431e-02] | [10.131444931030273, -1.833882451057434] |
38f1d8d3-c0e8-4889-a00d-b2be7fd92ab8 | post-hoc-concept-bottleneck-models | 2205.15480 | null | https://arxiv.org/abs/2205.15480v2 | https://arxiv.org/pdf/2205.15480v2.pdf | Post-hoc Concept Bottleneck Models | Concept Bottleneck Models (CBMs) map the inputs onto a set of interpretable concepts (``the bottleneck'') and use the concepts to make predictions. A concept bottleneck enhances interpretability since it can be investigated to understand what concepts the model "sees" in an input and which of these concepts are deemed important. However, CBMs are restrictive in practice as they require dense concept annotations in the training data to learn the bottleneck. Moreover, CBMs often do not match the accuracy of an unrestricted neural network, reducing the incentive to deploy them in practice. In this work, we address these limitations of CBMs by introducing Post-hoc Concept Bottleneck models (PCBMs). We show that we can turn any neural network into a PCBM without sacrificing model performance while still retaining the interpretability benefits. When concept annotations are not available on the training data, we show that PCBM can transfer concepts from other datasets or from natural language descriptions of concepts via multimodal models. A key benefit of PCBM is that it enables users to quickly debug and update the model to reduce spurious correlations and improve generalization to new distributions. PCBM allows for global model edits, which can be more efficient than previous works on local interventions that fix a specific prediction. Through a model-editing user study, we show that editing PCBMs via concept-level feedback can provide significant performance gains without using data from the target domain or model retraining. | ['James Zou', 'Maggie Wang', 'Mert Yuksekgonul'] | 2022-05-31 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [ 3.20111930e-01 4.96778965e-01 -1.51079178e-01 -4.52752203e-01
-2.94029355e-01 -6.51709378e-01 2.82040566e-01 5.52864909e-01
-5.62562108e-01 5.50938070e-01 1.39384583e-01 -5.10445058e-01
-1.20893165e-01 -5.36905885e-01 -9.81386602e-01 -1.03776976e-01
-7.71329701e-02 7.66809404e-01 1.01453230e-01 -3.45642149e-01
1.62765622e-01 6.90410435e-02 -1.60883832e+00 5.97317755e-01
1.12162769e+00 5.66616476e-01 5.12249827e-01 6.67387426e-01
-3.32559019e-01 5.50464034e-01 -7.43642449e-01 -3.98921013e-01
2.99375039e-02 -2.88073212e-01 -9.19909418e-01 -3.16952229e-01
3.51186275e-01 -3.09851617e-01 3.60517323e-01 7.25501239e-01
7.56926239e-02 3.27839434e-01 4.87261385e-01 -1.54378164e+00
-8.07670176e-01 6.99799180e-01 -2.55596220e-01 -2.74459422e-02
2.75555074e-01 -1.49235129e-02 1.15775514e+00 -9.07328963e-01
5.58369696e-01 1.54221177e+00 7.16131330e-01 9.39215720e-01
-1.68915427e+00 -6.40960634e-01 6.99188530e-01 9.67240781e-02
-1.11052704e+00 -3.34974706e-01 3.16366345e-01 -4.63385671e-01
1.15382206e+00 4.13097173e-01 6.32340610e-01 1.08187640e+00
-1.40258864e-01 7.20722258e-01 6.55630052e-01 -5.38284242e-01
3.97487402e-01 4.58082259e-01 3.93425435e-01 6.65690780e-01
2.54604727e-01 1.36980921e-01 -7.61394441e-01 -1.08468696e-01
5.32413900e-01 1.29586831e-01 -2.11820021e-01 -4.17405069e-01
-1.00957119e+00 8.65277588e-01 6.10500395e-01 -4.71733091e-03
-5.35074770e-02 1.71155080e-01 3.45084041e-01 5.56181908e-01
3.30337852e-01 1.06958866e+00 -8.09775174e-01 -1.83193177e-01
-4.52996224e-01 1.95097789e-01 8.00189197e-01 1.13177574e+00
8.19970787e-01 -3.73429507e-01 -2.56302208e-02 8.29358816e-01
2.17070356e-01 4.80026364e-01 4.18081641e-01 -1.04667747e+00
4.40561414e-01 7.74871945e-01 1.32609487e-01 -7.38870144e-01
-4.33105797e-01 -2.74764687e-01 -5.00989318e-01 1.59513533e-01
2.28890598e-01 -1.70051754e-01 -9.49345469e-01 2.02841735e+00
4.74731736e-02 -1.06070325e-01 4.98483032e-02 6.85579538e-01
3.34342420e-01 5.89740038e-01 3.59278470e-01 2.48266682e-01
1.06569767e+00 -8.15406501e-01 -3.99255514e-01 -5.48954129e-01
1.02694023e+00 -3.39947760e-01 1.59074986e+00 4.34763193e-01
-7.20884681e-01 -5.53872526e-01 -1.01087785e+00 -9.65640843e-02
-5.14494359e-01 -1.64153054e-01 8.58586609e-01 4.95713651e-01
-1.02464199e+00 6.05892003e-01 -9.46368277e-01 -4.95481700e-01
4.32396293e-01 5.78851819e-01 -3.56671989e-01 -3.37045729e-01
-1.20611441e+00 1.14582074e+00 5.88950634e-01 -1.14412367e-01
-9.82620597e-01 -1.12228656e+00 -9.45212185e-01 3.39291871e-01
4.61870879e-01 -8.19718599e-01 1.50366676e+00 -1.65250242e+00
-1.13844073e+00 2.49730334e-01 -3.55886400e-01 -2.90283591e-01
2.89288849e-01 -3.74785066e-01 -1.85511351e-01 -1.19040199e-01
-3.30630690e-02 1.28840744e+00 6.86610579e-01 -1.28376722e+00
-7.82462656e-01 7.23060668e-02 5.76669991e-01 2.26384833e-01
-5.67073405e-01 -2.41492301e-01 -5.54581583e-01 -3.50734562e-01
-1.55355662e-01 -9.47959840e-01 -2.68380076e-01 2.07921937e-01
-3.57793063e-01 -2.22715642e-02 7.16829062e-01 -6.33465588e-01
1.11640573e+00 -2.08256483e+00 2.72894263e-01 3.34663957e-01
3.66056979e-01 1.71053201e-01 -5.16490817e-01 3.80061686e-01
-2.30710432e-01 5.74374795e-01 -4.02022570e-01 -5.75247586e-01
-8.99788365e-02 5.08766413e-01 -2.58572102e-01 -1.90473884e-01
5.69107234e-01 7.95314848e-01 -9.37447608e-01 1.40416948e-02
4.18449156e-02 3.27399492e-01 -1.18898618e+00 1.16630062e-01
-6.64024651e-01 4.16096419e-01 2.50025594e-04 2.38366738e-01
4.80530590e-01 -2.33446538e-01 2.55617917e-01 1.09763332e-01
3.20183963e-01 2.37613037e-01 -9.67909634e-01 1.59108293e+00
-7.87638009e-01 7.17926741e-01 -6.12977371e-02 -1.04315138e+00
5.80688238e-01 4.24706414e-02 3.97023559e-02 -6.94235682e-01
-4.83675689e-01 -1.75652325e-01 3.46512586e-01 -2.80796558e-01
4.15683687e-01 -3.63246977e-01 5.68504483e-02 4.57581311e-01
7.83791989e-02 2.20967904e-01 2.36054048e-01 3.11613649e-01
1.01582146e+00 -8.96371156e-03 1.06134869e-01 -2.70333409e-01
1.12948187e-01 1.94006532e-01 6.99702263e-01 9.87588763e-01
2.89330214e-01 3.28758150e-01 7.57271290e-01 -3.10052395e-01
-9.50363755e-01 -1.22946239e+00 1.41230598e-01 1.56553972e+00
-9.47760567e-02 -6.79305971e-01 -6.65963590e-01 -7.01727509e-01
2.69143790e-01 1.13565624e+00 -8.10053289e-01 -5.93265295e-01
-1.80742726e-01 -4.23227847e-01 3.83901775e-01 8.58792543e-01
1.23925604e-01 -8.23689520e-01 -5.43785751e-01 8.91038999e-02
-1.56921402e-01 -6.64592862e-01 -4.15536165e-01 2.50626981e-01
-1.15963578e+00 -1.10020030e+00 -1.38721988e-01 -4.96895492e-01
1.28939521e+00 9.44926441e-02 1.29866660e+00 4.39313501e-01
-2.06938654e-01 4.15927887e-01 -3.70363712e-01 -6.96186841e-01
-7.40992785e-01 2.36307934e-01 6.33521900e-02 -3.90500963e-01
3.74750465e-01 -4.58244681e-01 -2.37963423e-01 4.06854421e-01
-9.06954944e-01 5.32322228e-01 5.48012853e-01 1.00751519e+00
1.45877510e-01 4.86426502e-02 6.30524397e-01 -1.36567450e+00
8.09618473e-01 -4.90569443e-01 -6.50896668e-01 2.87703753e-01
-9.31862175e-01 4.10431147e-01 6.65456414e-01 -5.37127614e-01
-1.12375474e+00 1.15666352e-02 1.18276007e-01 -3.22819114e-01
-2.20744219e-02 9.75176215e-01 -4.26562317e-02 3.56773853e-01
9.13070858e-01 -2.42516905e-01 1.30079389e-01 -5.87467492e-01
5.55935264e-01 3.36090297e-01 2.49346241e-01 -8.22414875e-01
6.61914229e-01 1.54951721e-01 -4.05018955e-01 -6.07573926e-01
-8.36332440e-01 -3.63612235e-01 -7.70149112e-01 1.21855102e-01
5.11795700e-01 -8.29037368e-01 -8.67792726e-01 -9.42488536e-02
-1.12257743e+00 -6.05188906e-01 -1.30207315e-01 2.69574165e-01
-2.30248481e-01 -4.51051891e-02 -3.75787258e-01 -6.84598446e-01
1.98467746e-02 -1.07000244e+00 6.43039107e-01 1.43517673e-01
-7.96291709e-01 -1.32682765e+00 -3.15295428e-01 1.45264700e-01
6.16523981e-01 -1.61672011e-01 1.58466554e+00 -7.30149627e-01
-5.29117167e-01 -1.04171842e-01 -4.32739593e-02 4.30195183e-01
1.38239861e-01 -3.73781659e-02 -1.06002605e+00 -4.16041464e-01
-5.23003280e-01 -3.36260617e-01 8.85890365e-01 1.43892080e-01
1.35104275e+00 -4.14757162e-01 -5.71704030e-01 4.07737613e-01
1.15916491e+00 2.00372651e-01 3.66822600e-01 9.79264304e-02
7.23550141e-01 7.00107276e-01 5.02680242e-01 1.32935554e-01
3.02527308e-01 4.88649160e-01 2.64288932e-01 -2.30443060e-01
3.19818974e-01 -5.17585933e-01 5.66403329e-01 4.50933754e-01
1.64285913e-01 -1.48729548e-01 -1.20972466e+00 3.99788141e-01
-1.90478849e+00 -5.94138145e-01 3.05195093e-01 2.46546435e+00
9.18682396e-01 2.49346197e-01 -1.17672957e-01 -2.48018995e-01
5.42353332e-01 -5.43520451e-01 -8.65754962e-01 -5.79701304e-01
2.89169937e-01 2.54226416e-01 3.17672879e-01 9.71619904e-01
-9.06076849e-01 9.19526219e-01 6.68497610e+00 3.27542156e-01
-9.67049062e-01 2.41086990e-01 5.02375305e-01 -4.13930804e-01
-6.40912950e-01 1.64146230e-01 -6.44919395e-01 1.88054353e-01
1.11853790e+00 -1.37025774e-01 6.65309548e-01 8.70084107e-01
2.26167515e-01 -2.69287139e-01 -1.84051371e+00 6.36581242e-01
-3.78822647e-02 -1.23768890e+00 3.54180336e-01 -1.00792147e-01
3.78475130e-01 -5.47993854e-02 -3.26951630e-02 7.13803291e-01
6.14476860e-01 -1.42015338e+00 6.48370564e-01 5.44596136e-01
6.46321833e-01 -8.38478506e-01 4.98290956e-01 5.67091584e-01
-7.38481581e-01 -3.82507324e-01 -4.56706583e-01 -2.71963298e-01
-3.34671348e-01 1.95676416e-01 -1.25251126e+00 1.44047633e-01
5.59255183e-01 5.85199893e-01 -9.48835850e-01 8.94744992e-01
-2.46446997e-01 7.46114671e-01 -3.49343777e-01 1.19372795e-03
-3.49343494e-02 2.95659490e-02 2.60859132e-01 1.09511828e+00
1.65163666e-01 -2.50235170e-01 1.79108515e-01 1.34162354e+00
-3.58839221e-02 -1.61036447e-01 -5.48676848e-01 -3.12953770e-01
6.84193194e-01 9.65660214e-01 -5.52713990e-01 -4.08184379e-01
-3.95198166e-01 1.07203162e+00 5.80836236e-01 5.61530113e-01
-5.85764229e-01 -2.69372165e-01 1.02491641e+00 1.09287791e-01
1.23741545e-01 -4.53433320e-02 -3.10039341e-01 -1.11016333e+00
-9.16791111e-02 -9.38297451e-01 4.48237658e-01 -8.99261475e-01
-1.19527304e+00 2.00098068e-01 2.92161554e-01 -7.77391613e-01
-2.66004145e-01 -8.54248583e-01 -4.71963108e-01 1.03670681e+00
-1.10158050e+00 -9.12000954e-01 -2.43561640e-01 2.47444659e-01
4.75917250e-01 8.50050598e-02 1.13247538e+00 1.14792928e-01
-4.60232019e-01 7.67581046e-01 1.13185264e-01 5.26440889e-02
7.44590878e-01 -1.49657094e+00 5.92107415e-01 5.93137741e-01
-4.43099113e-03 1.26339710e+00 6.68884754e-01 -6.71641409e-01
-1.04975629e+00 -1.16936839e+00 5.96459985e-01 -8.31660211e-01
2.57742763e-01 -8.66727889e-01 -1.25488794e+00 1.03213239e+00
-3.02872341e-02 -3.73508692e-01 9.39693034e-01 9.24723148e-01
-5.01272440e-01 -1.55978397e-01 -8.02046180e-01 7.80948281e-01
1.12755191e+00 -5.78868926e-01 -6.14799500e-01 2.07480639e-01
1.10952938e+00 -2.93444991e-01 -6.89831138e-01 -2.80597191e-02
4.29378897e-01 -6.80249810e-01 9.21824515e-01 -1.23487782e+00
5.03031015e-01 -1.64033934e-01 7.56775290e-02 -1.78191233e+00
-2.10382164e-01 -2.66687721e-01 4.44488078e-02 1.16347373e+00
1.01254213e+00 -7.27428317e-01 6.30355835e-01 1.19543040e+00
-1.62349954e-01 -4.18030947e-01 -6.67032659e-01 -6.91379249e-01
2.23964021e-01 -7.97455549e-01 6.17584467e-01 1.01040387e+00
2.52339065e-01 5.12720585e-01 -1.62545487e-01 2.00163543e-01
2.19988540e-01 -2.07646459e-01 6.43345416e-01 -1.43023813e+00
-4.87146616e-01 -3.92838925e-01 -1.90521345e-01 -9.65091705e-01
1.80525407e-01 -1.05256474e+00 1.13105416e-01 -1.40598476e+00
2.54414439e-01 -6.90357327e-01 -3.68256658e-01 1.05826962e+00
-4.05952632e-01 -4.21205997e-01 3.94521207e-01 1.36680663e-01
-4.43298757e-01 4.25471783e-01 9.17407155e-01 -1.89284414e-01
-3.56100321e-01 -6.22036941e-02 -1.13047326e+00 6.99726880e-01
7.28892982e-01 -6.02071583e-01 -1.11686957e+00 -7.32364774e-01
4.12495047e-01 -2.34920457e-01 4.20658588e-01 -8.28840196e-01
1.16309144e-01 -1.62208900e-01 6.16837978e-01 -1.38080940e-01
3.14477026e-01 -8.33270371e-01 5.94576076e-02 4.94127363e-01
-7.78825998e-01 2.77694225e-01 7.78927922e-01 7.60854244e-01
-7.69692734e-02 -3.26500595e-01 5.84233165e-01 -1.03195116e-01
-8.68306100e-01 -4.99901101e-02 -5.47090411e-01 9.82318819e-02
7.99789667e-01 -1.68885607e-02 -3.60666454e-01 -5.43184817e-01
-7.55226731e-01 5.91385484e-01 5.50720930e-01 7.22110689e-01
6.86291099e-01 -1.07387996e+00 -2.41019547e-01 3.97850692e-01
4.92669702e-01 2.23128587e-01 -4.72626463e-02 5.72505653e-01
-3.85993242e-01 4.05393124e-01 -1.16622567e-01 -7.10040152e-01
-1.12349212e+00 6.59023046e-01 6.04694247e-01 6.68393821e-02
-4.54246968e-01 8.51641476e-01 5.99921107e-01 -1.05842841e+00
3.70231748e-01 -7.55243719e-01 9.72750969e-03 -8.44071731e-02
4.81425077e-01 -7.96285421e-02 -5.71812764e-02 5.08917011e-02
-3.00831556e-01 -1.30146397e-02 -4.30452019e-01 -3.21569860e-01
1.22813070e+00 9.70374327e-04 -1.05967270e-02 5.21257639e-01
9.80374157e-01 -3.24912190e-01 -1.38873088e+00 -1.77896410e-01
3.36047351e-01 -3.12057406e-01 -1.21698268e-01 -1.37515390e+00
-7.50861764e-01 1.01199543e+00 4.56101418e-01 -2.15269729e-01
1.14384842e+00 5.45484805e-03 1.40263706e-01 1.00598323e+00
1.38882902e-02 -1.14470530e+00 6.90869093e-02 5.50952852e-01
9.85642195e-01 -1.25049686e+00 -3.51887733e-01 -3.47726673e-01
-8.38867724e-01 1.01972437e+00 1.00326967e+00 3.91019136e-01
4.27789271e-01 1.35167807e-01 8.55415538e-02 2.49955077e-02
-1.17445469e+00 7.31077865e-02 2.17693865e-01 9.66564059e-01
3.70209306e-01 4.01925519e-02 4.02203232e-01 7.76862860e-01
-1.63991600e-01 -1.49382260e-02 4.40666676e-01 8.14371586e-01
-3.66939604e-01 -1.21217144e+00 -2.50977427e-01 6.07811213e-01
2.12748915e-01 -4.83012736e-01 -4.94228810e-01 9.75435376e-01
1.57365978e-01 9.42366838e-01 3.50405425e-01 -5.09693205e-01
5.08307695e-01 4.50934321e-01 2.35731021e-01 -1.04256117e+00
-3.52960140e-01 -4.26354051e-01 2.34529227e-01 -6.35307014e-01
-2.08871122e-02 -6.14234984e-01 -1.40260220e+00 -3.69825691e-01
-1.96120024e-01 2.81637490e-01 5.07919848e-01 1.07409060e+00
6.63873851e-01 5.55960357e-01 1.17503710e-01 -3.85503680e-01
-4.34878081e-01 -1.04217803e+00 -1.29172042e-01 5.58272600e-01
3.88881028e-01 -6.98002040e-01 -3.88733223e-02 3.45470011e-01] | [9.635565757751465, 7.123847007751465] |
caa958d0-0698-4d3c-8f9f-6a3d28c146f9 | towards-efficient-discriminative-pattern | 1908.06801 | null | https://arxiv.org/abs/1908.06801v1 | https://arxiv.org/pdf/1908.06801v1.pdf | Towards Efficient Discriminative Pattern Mining in Hybrid Domains | Discriminative pattern mining is a data mining task in which we find patterns that distinguish transactions in the class of interest from those in other classes, and is also called emerging pattern mining or subgroup discovery. One practical problem in discriminative pattern mining is how to handle numeric values in the input dataset. In this paper, we propose an algorithm for discriminative pattern mining that can deal with a transactional dataset in a hybrid domain, i.e. the one that includes both symbolic and numeric values. We also show the execution results of a prototype implementation of the proposed algorithm for two standard benchmark datasets. | ['Yoshitaka Kameya'] | 2019-08-15 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 4.92471844e-01 -3.27977598e-01 -6.27125263e-01 -5.80456674e-01
9.21340510e-02 -2.40412071e-01 2.59785116e-01 4.38554674e-01
-1.32372722e-01 7.65763938e-01 -1.99534521e-02 -3.90706182e-01
-5.15805721e-01 -1.12171483e+00 -1.87999502e-01 -7.64938951e-01
-3.57771188e-01 9.76916790e-01 5.20531595e-01 -9.54044610e-02
4.09621477e-01 6.30622923e-01 -1.92450345e+00 8.02987814e-01
4.66973603e-01 1.21488738e+00 -3.12502891e-01 2.47706741e-01
-1.60507262e-01 5.56368291e-01 -4.44246501e-01 6.01867326e-02
4.54297572e-01 -5.40793955e-01 -6.69158697e-01 2.19894677e-01
5.85287996e-02 4.94780988e-02 2.04230681e-01 7.65948892e-01
-2.84928501e-01 1.00409232e-01 6.79530621e-01 -1.66085708e+00
2.34459758e-01 7.02782631e-01 -8.87473226e-01 5.42608082e-01
2.77806967e-01 -2.81011492e-01 1.09820187e+00 -9.24164712e-01
6.29489601e-01 9.33100104e-01 2.38832518e-01 -7.04196319e-02
-1.58144176e+00 -7.01472342e-01 -9.10608992e-02 6.07146442e-01
-1.48537326e+00 -1.08490065e-01 5.94553947e-01 -3.61393988e-01
1.21336389e+00 7.16907203e-01 6.30857527e-01 4.09181058e-01
2.86573380e-01 5.41966259e-01 1.04006183e+00 -3.89811128e-01
2.99410433e-01 1.25766650e-01 8.99314046e-01 5.97281381e-02
8.67981791e-01 1.90459266e-01 -4.90636528e-01 -6.80800557e-01
6.42710030e-02 4.43039536e-01 2.72113420e-02 -4.78083193e-01
-9.98857796e-01 8.07081819e-01 -4.17792678e-01 6.10014081e-01
-5.29670238e-01 -2.79817373e-01 4.84549314e-01 1.01612985e+00
6.05227873e-02 1.60264134e-01 -6.65181220e-01 -3.41909885e-01
-8.57972503e-01 3.93732369e-01 1.25702894e+00 1.05292261e+00
1.03413653e+00 -2.70147473e-01 3.76934069e-03 7.17072606e-01
4.02023606e-02 2.26196870e-01 4.69358951e-01 -1.96828440e-01
2.57621557e-01 1.27801764e+00 -1.18725589e-02 -1.01123035e+00
-3.59215885e-01 -1.07584089e-01 -6.12331688e-01 -1.08784042e-01
3.74651700e-01 3.53581347e-02 -5.76097250e-01 1.26134551e+00
5.05342543e-01 1.25425339e-01 -5.47894568e-04 4.91982669e-01
4.13546234e-01 6.17368996e-01 -3.27858686e-01 -8.17327976e-01
1.24043441e+00 -5.18101692e-01 -3.29865217e-01 2.29608312e-01
7.30539620e-01 -6.22691631e-01 6.30021334e-01 7.84482241e-01
-4.32562977e-01 -3.47740948e-01 -9.67444241e-01 4.38912541e-01
-2.67497927e-01 -7.32494593e-02 4.12117839e-01 6.16584122e-01
-4.07411158e-01 3.44806105e-01 -4.79545146e-01 -4.82361883e-01
-2.30270281e-01 6.07920766e-01 -5.21876693e-01 1.82655826e-01
-7.56220579e-01 4.10054207e-01 9.93433297e-01 -3.60602230e-01
-1.35731354e-01 -4.83962685e-01 -4.90992934e-01 2.17696548e-01
7.24416435e-01 -4.10496667e-02 7.81027436e-01 -1.12629998e+00
-6.87339008e-01 8.37056100e-01 -4.79642421e-01 -6.08110368e-01
9.01924372e-02 2.07577348e-01 -1.03423619e+00 -4.20042992e-01
-3.09248548e-02 -2.34036654e-01 7.01788962e-01 -6.95091844e-01
-1.17388093e+00 -4.70715970e-01 -4.79079366e-01 -3.02426130e-01
-2.23349169e-01 -8.14007688e-03 -1.88036636e-01 -6.66720450e-01
5.66696338e-02 -8.98733675e-01 -1.62414029e-01 -5.81957698e-01
-3.71481657e-01 -6.91631615e-01 1.43185532e+00 -1.63364470e-01
1.50985038e+00 -2.37648201e+00 -1.16406709e-01 1.37861395e+00
-3.00526340e-02 -2.46060461e-01 3.09009910e-01 6.38300896e-01
-2.49515146e-01 -9.43148881e-02 -2.94396877e-01 4.11210567e-01
-2.02091083e-01 7.69837916e-01 -4.88138497e-01 3.75452757e-01
-1.56106040e-01 3.54947537e-01 -4.57596868e-01 -5.15613019e-01
-1.27158076e-01 -5.18781781e-01 -4.73372102e-01 8.16829056e-02
-3.54854017e-01 2.78013814e-02 -2.60312647e-01 7.85699964e-01
7.86846459e-01 -6.27246499e-02 9.51430976e-01 -3.66363116e-02
-4.87961590e-01 7.48234838e-02 -1.68009150e+00 5.97183943e-01
2.17008054e-01 2.19949841e-01 -3.00734907e-01 -1.38325071e+00
1.30613053e+00 -2.81459372e-02 7.75078535e-01 -8.37367952e-01
-6.77354336e-02 4.32473004e-01 6.12801552e-01 -3.01157504e-01
5.90473235e-01 -9.46239606e-02 -3.37133557e-01 9.89069164e-01
-4.69823390e-01 6.29154086e-01 7.55108178e-01 -3.05352092e-01
1.30003679e+00 -5.97281337e-01 1.02661192e+00 -6.08491182e-01
6.99950337e-01 4.59967762e-01 1.14128673e+00 7.07701862e-01
2.95035511e-01 6.00014478e-02 8.89773786e-01 -9.14100707e-01
-8.48330975e-01 -8.04353058e-01 -3.19878459e-01 1.09123325e+00
2.43705675e-01 -5.95921814e-01 2.01019496e-01 -7.93205142e-01
5.75909913e-01 4.40840095e-01 -6.62845194e-01 -3.38158046e-04
-8.89850199e-01 -6.90063059e-01 1.44385725e-01 4.03729796e-01
3.14644277e-01 -1.04758215e+00 -7.33256638e-01 4.33113664e-01
2.63426244e-01 -7.08177567e-01 -3.12906891e-01 5.31108379e-01
-9.43928063e-01 -1.54047358e+00 2.19919562e-01 -1.00912023e+00
5.70903540e-01 2.16016293e-01 1.09816086e+00 2.38016248e-01
-1.97212160e-01 -1.77064553e-01 -6.58352315e-01 -5.07631958e-01
-3.82262349e-01 8.86648446e-02 8.51840675e-02 4.73198533e-01
1.03033531e+00 -6.67780936e-01 -1.91708460e-01 9.37239528e-01
-8.47136915e-01 -2.50299662e-01 6.82340682e-01 8.26032996e-01
1.09720290e+00 5.35267055e-01 5.26904881e-01 -1.33664751e+00
6.06519520e-01 -9.81497228e-01 -4.57996070e-01 5.49275935e-01
-9.24870670e-01 1.76357493e-01 6.41353071e-01 -3.83546621e-01
-7.26244569e-01 2.40256801e-01 1.19956478e-01 3.93344723e-02
-2.67047912e-01 9.74610865e-01 -2.94946581e-01 2.39299878e-01
5.76092720e-01 6.82744324e-01 1.50830567e-01 -6.09689832e-01
-3.50132614e-01 8.03234518e-01 3.18019032e-01 -5.79438388e-01
5.83959877e-01 3.61364305e-01 2.60102689e-01 -1.16179192e+00
-1.91700971e-03 -9.70996439e-01 -4.97592717e-01 1.45631790e-01
2.84868642e-03 -2.48387009e-01 -5.34161925e-01 1.86815724e-01
-5.06772757e-01 1.12214617e-01 -4.45875645e-01 3.77958417e-01
-4.30155963e-01 4.94481742e-01 -1.43575773e-01 -4.04049605e-01
-2.89907724e-01 -6.27055466e-01 3.51196080e-01 -1.99524444e-02
-7.55312800e-01 -6.54343188e-01 3.78996849e-01 -2.72693038e-01
8.33120011e-03 4.27052289e-01 1.36680484e+00 -1.60021269e+00
-1.41395241e-01 -1.78077847e-01 1.31076321e-01 -7.10519403e-02
4.18601185e-01 -2.25245999e-03 -1.25300616e-01 -2.85309672e-01
2.68979073e-02 1.32357046e-01 5.76983094e-01 -1.35220125e-01
1.30657291e+00 -4.88802195e-01 -7.75355041e-01 2.90121645e-01
1.09460688e+00 8.88696373e-01 5.88746428e-01 1.91053972e-01
2.00009555e-01 5.29072404e-01 1.41351819e+00 7.08092868e-01
-1.34911150e-01 1.03109443e+00 -1.90715402e-01 1.76810220e-01
6.44565940e-01 4.57092114e-02 7.81065524e-02 2.28984177e-01
1.52562968e-02 1.08389920e-02 -9.91819859e-01 6.00246847e-01
-2.25183892e+00 -1.34610689e+00 -5.54596722e-01 1.96296883e+00
8.63878727e-01 1.74169272e-01 7.53499031e-01 8.01907241e-01
7.66988575e-01 -5.10140300e-01 -5.77759027e-01 -1.05665994e+00
6.61985949e-02 3.39381784e-01 2.89367467e-01 3.14384177e-02
-1.13037884e+00 3.21888715e-01 6.29509306e+00 7.04315484e-01
-1.07809925e+00 -4.61796612e-01 4.41727072e-01 1.13722362e-01
-3.81940901e-01 -3.25361900e-02 -7.81602919e-01 5.73822796e-01
8.23082030e-01 -6.31226003e-01 -2.48919204e-01 1.22321081e+00
9.19722766e-02 -4.08879757e-01 -1.55018580e+00 1.00334120e+00
-1.43996015e-01 -1.24152493e+00 1.77183226e-01 4.47612464e-01
5.15426219e-01 -4.30451781e-01 -2.29784772e-01 1.45853385e-01
-1.33359045e-01 -7.91144490e-01 4.12727565e-01 1.73990205e-01
4.05437797e-01 -9.13470209e-01 8.37112367e-01 6.01297379e-01
-1.38065231e+00 -1.98860794e-01 -1.61298737e-01 -1.09734342e-01
-1.89624995e-01 7.66464889e-01 -1.10463881e+00 5.32758534e-01
7.36762583e-01 9.89760101e-01 -1.51314422e-01 1.30803490e+00
3.02307129e-01 7.31598973e-01 -6.24737442e-01 -1.40333712e-01
-1.93639733e-02 -4.05555308e-01 4.84000713e-01 1.37201989e+00
4.97782499e-01 1.21405542e-01 2.83063233e-01 3.63927096e-01
3.46643507e-01 3.63668054e-01 -8.12669277e-01 2.40877882e-01
5.27477384e-01 8.12961400e-01 -9.84390259e-01 -2.89328992e-01
-6.01586223e-01 3.73782307e-01 -2.18088225e-01 -4.60794382e-02
-5.78178346e-01 -4.89114821e-01 7.60467649e-01 3.53869349e-01
5.91403306e-01 -1.43314496e-01 -6.10699117e-01 -8.98672402e-01
2.83175141e-01 -1.12399244e+00 1.16773975e+00 3.51783901e-01
-1.13328075e+00 3.08976620e-01 3.74802709e-01 -1.57319415e+00
-4.70116407e-01 -5.84680796e-01 -6.98277771e-01 6.33319676e-01
-7.76530623e-01 -6.72031522e-01 -2.24906921e-01 9.14745569e-01
3.10266793e-01 -3.22968304e-01 6.53952718e-01 7.99154937e-02
-3.43056798e-01 8.54497910e-01 2.05602974e-01 -4.34464850e-02
3.35697055e-01 -9.56981838e-01 1.72874153e-01 8.47506166e-01
4.08040106e-01 7.18308985e-01 6.82817400e-01 -7.20121324e-01
-1.40677726e+00 -9.36494827e-01 1.06725323e+00 8.37878585e-02
5.00121534e-01 -2.30896458e-01 -1.02528214e+00 7.04324424e-01
-2.83188075e-01 -1.80825278e-01 1.25351501e+00 3.49413157e-01
-4.12467957e-01 -4.76065308e-01 -1.53874803e+00 2.37672269e-01
1.05607927e+00 5.35249226e-02 -8.98654521e-01 -3.92506987e-01
6.94061369e-02 1.34994969e-01 -8.72978151e-01 5.01645148e-01
8.28437984e-01 -9.15548444e-01 4.87806082e-01 -5.89274228e-01
-7.35735968e-02 -6.78532243e-01 -2.55076140e-01 -8.15315127e-01
-3.24849755e-01 -3.00491005e-01 -3.05141389e-01 8.21371853e-01
5.79918146e-01 -8.57769310e-01 1.06635380e+00 2.97504663e-01
2.12943524e-01 -5.93947172e-01 -1.07264471e+00 -1.02442265e+00
-4.70355541e-01 -4.09917980e-01 9.51433957e-01 9.50217485e-01
4.90998507e-01 1.81784511e-01 -3.91714156e-01 -2.59064078e-01
6.48632348e-01 1.23854744e+00 9.33455050e-01 -1.53132725e+00
-5.49982190e-01 -1.83098078e-01 -8.44495058e-01 -8.36828172e-01
-1.89033091e-01 -7.49692082e-01 -2.59213686e-01 -7.85409927e-01
1.84721828e-01 -8.50766778e-01 -3.28803569e-01 7.19364524e-01
4.44159180e-01 1.13989227e-01 -3.12363982e-01 5.35384595e-01
-3.98556411e-01 -7.48699903e-02 4.46750939e-01 -2.99616516e-01
-5.40077865e-01 2.67528206e-01 -6.48636699e-01 2.77432889e-01
9.73928988e-01 -6.52044356e-01 -4.66914982e-01 4.65426654e-01
1.93394333e-01 -1.40852062e-02 -1.83500364e-01 -6.15095198e-01
1.42162308e-01 -9.83510256e-01 1.30794302e-01 -1.07132375e+00
-3.13613892e-01 -1.28472042e+00 9.05076325e-01 9.89709496e-01
-4.63939756e-02 1.92623604e-02 -3.22831841e-03 5.24716318e-01
-4.65431929e-01 -1.66922897e-01 7.19197094e-01 4.59714234e-01
-1.26757729e+00 1.93314120e-01 -5.77191889e-01 -2.69906849e-01
1.68159759e+00 -7.49080062e-01 -2.00019687e-01 1.07159555e-01
-9.11413193e-01 2.77733177e-01 5.13911724e-01 3.82468909e-01
8.57404053e-01 -1.39410996e+00 -6.11198485e-01 6.65202856e-01
7.03156054e-01 -3.38990420e-01 -2.75331467e-01 1.04223835e+00
-4.56556588e-01 3.12695891e-01 -4.82337505e-01 -5.64964771e-01
-1.95145309e+00 6.89878762e-01 5.45424744e-02 -4.85684335e-01
-6.62777960e-01 -1.89261436e-02 -1.71108633e-01 1.34136930e-01
-2.80348007e-02 -4.38227296e-01 -3.99475515e-01 1.18386962e-01
7.07541108e-01 5.73745489e-01 2.08769575e-01 -4.04102713e-01
-7.70665526e-01 2.80970007e-01 -2.53422409e-01 3.72500300e-01
1.36227548e+00 3.53051156e-01 -5.54059565e-01 4.93032545e-01
1.02536237e+00 8.45041126e-02 -3.28472227e-01 -3.27801317e-01
8.89475763e-01 -6.29784346e-01 -7.83968568e-01 -5.06422818e-01
-7.98294544e-01 6.11191755e-03 4.30074483e-01 7.72696972e-01
1.34244335e+00 1.72080666e-01 4.99909252e-01 7.77954757e-01
7.46369183e-01 -1.18385279e+00 -3.53360951e-01 6.57453120e-01
8.87558162e-01 -8.03400517e-01 9.16161463e-02 -5.88698864e-01
-5.54976225e-01 1.30456877e+00 6.31977856e-01 -3.40164512e-01
1.18779361e+00 5.08655071e-01 -4.25894499e-01 -1.55392140e-01
-1.05734396e+00 -2.06655875e-01 4.33763675e-02 4.96115148e-01
1.95010617e-01 3.21483463e-01 -1.10903621e+00 7.58814156e-01
-7.17335790e-02 7.03240046e-04 3.26759607e-01 1.41804147e+00
-5.97914815e-01 -1.73668289e+00 -3.98506731e-01 1.06913292e+00
-3.44969392e-01 2.93204665e-01 -7.77392983e-01 7.76165843e-01
4.13037360e-01 8.94622266e-01 4.82875913e-01 -9.25824821e-01
5.98687708e-01 7.90006667e-02 2.02040970e-01 -7.10671246e-01
-6.45802557e-01 -1.25085190e-01 3.79648000e-01 -4.18052018e-01
-2.47393742e-01 -8.84717643e-01 -1.31951547e+00 -5.50239563e-01
-1.24392686e-02 6.59270704e-01 -4.79906239e-02 6.65870965e-01
2.93750525e-01 -3.36787701e-01 1.00866973e+00 2.63090849e-01
-2.73851126e-01 -7.00082898e-01 -1.11307287e+00 4.67592567e-01
1.35245789e-02 -6.99358046e-01 -1.34407550e-01 -7.21311495e-02] | [8.302258491516113, 6.295173645019531] |
cb6130dc-a5f7-4b97-8dde-413f0a80a905 | early-and-in-season-crop-type-mapping-without | 2110.10275 | null | https://arxiv.org/abs/2110.10275v1 | https://arxiv.org/pdf/2110.10275v1.pdf | Early- and in-season crop type mapping without current-year ground truth: generating labels from historical information via a topology-based approach | Land cover classification in remote sensing is often faced with the challenge of limited ground truth. Incorporating historical information has the potential to significantly lower the expensive cost associated with collecting ground truth and, more importantly, enable early- and in-season mapping that is helpful to many pre-harvest decisions. In this study, we propose a new approach that can effectively transfer knowledge about the topology (i.e. relative position) of different crop types in the spectral feature space (e.g. the histogram of SWIR1 vs RDEG1 bands) to generate labels, thereby support crop classification in a different year. Importantly, our approach does not attempt to transfer classification decision boundaries that are susceptible to inter-annual variations of weather and management, but relies on the more robust and shift-invariant topology information. We tested this approach for mapping corn/soybeans in the US Midwest and paddy rice/corn/soybeans in Northeast China using Landsat-8 and Sentinel-2 data. Results show that our approach automatically generates high-quality labels for crops in the target year immediately after each image becomes available. Based on these generated labels from our approach, the subsequent crop type mapping using a random forest classifier reach the F1 score as high as 0.887 for corn as early as the silking stage and 0.851 for soybean as early as the flowering stage and the overall accuracy of 0.873 in Iowa. In Northeast China, F1 scores of paddy rice, corn and soybeans and the overall accuracy can exceed 0.85 two and half months ahead of harvest. Overall, these results highlight unique advantages of our approach in transferring historical knowledge and maximizing the timeliness of crop maps. Our approach supports a general paradigm shift towards learning transferrable and generalizable knowledge to facilitate land cover classification. | ['Zhenong Jin', 'David B. Lobell', 'Jinwei Dong', 'Xiao-Peng Song', 'Liheng Zhong', 'Chenxi Lin'] | 2021-10-19 | null | null | null | null | ['crop-classification'] | ['miscellaneous'] | [ 4.16881263e-01 -1.52959883e-01 -4.62157696e-01 -2.92519093e-01
-3.61891419e-01 -1.11771882e+00 3.39269310e-01 5.79510450e-01
-1.29627377e-01 1.01848888e+00 -3.14765126e-01 -9.71106648e-01
-2.37570301e-01 -1.54067206e+00 -7.62581408e-01 -7.59931624e-01
-3.49412709e-01 -2.10568868e-02 2.02264339e-02 -5.09972692e-01
-2.09334910e-01 7.64376760e-01 -1.79485500e+00 9.86020491e-02
1.12850845e+00 8.16701114e-01 7.65171707e-01 5.32790363e-01
-7.38532320e-02 3.41179222e-02 -2.31643662e-01 3.60170215e-01
3.93678159e-01 -1.28120914e-01 -5.62518060e-01 -1.93063933e-02
2.15804890e-01 -5.79724610e-01 3.22124422e-01 1.14086699e+00
1.98856592e-02 -2.72331178e-01 5.83872259e-01 -9.94823754e-01
-4.27027494e-01 6.97996318e-01 -9.87264454e-01 -2.56935149e-01
-9.27375853e-02 3.32533754e-02 7.68790483e-01 -4.67482865e-01
5.16912520e-01 6.94132388e-01 8.35593104e-01 -2.27083296e-01
-1.51881111e+00 -8.06187510e-01 4.13048714e-01 -1.06694549e-01
-1.48249078e+00 -1.20954491e-01 1.99531749e-01 -7.26649642e-01
4.39868927e-01 4.23179984e-01 9.19554174e-01 2.26234555e-01
3.57683569e-01 4.17851001e-01 1.26947391e+00 -5.62188447e-01
2.49000609e-01 -8.78292769e-02 -9.72600803e-02 4.58780259e-01
5.43745875e-01 4.93061572e-01 -7.39291161e-02 -2.61554471e-03
8.21024120e-01 3.62979144e-01 -6.84910715e-02 -3.07036281e-01
-1.12178075e+00 8.97758365e-01 1.09888089e+00 3.26575160e-01
-7.48931944e-01 -2.97997922e-01 2.69741789e-02 1.65997714e-01
5.62394619e-01 4.37307656e-01 -6.80728614e-01 5.54539502e-01
-1.30476320e+00 1.56414732e-01 5.15094340e-01 9.43561256e-01
1.35415256e+00 -3.57976817e-02 1.81071058e-01 5.70340037e-01
6.21733405e-02 1.10706413e+00 -1.58983663e-01 -6.44942701e-01
3.52544687e-03 5.81389129e-01 5.03059924e-01 -1.04541445e+00
-5.04341722e-01 -3.94383818e-01 -8.85523081e-01 4.71960366e-01
3.23357105e-01 -3.23519140e-01 -1.27489471e+00 1.63381135e+00
2.11191386e-01 -2.22990274e-01 8.82603675e-02 6.23640835e-01
2.81818718e-01 8.75707448e-01 4.10564125e-01 -2.45687515e-01
1.48413587e+00 -2.30098471e-01 -4.57241893e-01 -3.99900556e-01
7.40147889e-01 -6.35163426e-01 6.61168456e-01 -1.70692623e-01
-2.90218085e-01 -6.08170986e-01 -1.14239049e+00 8.08176756e-01
-9.36766922e-01 5.77842295e-01 9.69979823e-01 4.95863080e-01
-9.39136088e-01 5.62122226e-01 -9.21608388e-01 -7.15342343e-01
3.19201261e-01 6.15390614e-02 -5.11195779e-01 -8.88862833e-02
-1.06014526e+00 9.81347263e-01 7.64512241e-01 6.37633860e-01
-5.53652883e-01 -8.44535828e-01 -8.06902409e-01 4.77012098e-02
2.17096522e-01 3.50326039e-02 7.95386970e-01 -1.00983810e+00
-1.13951004e+00 8.92731726e-01 -6.18556142e-02 -3.55943978e-01
-3.57846217e-03 -6.34333566e-02 -4.88054812e-01 -1.21214345e-01
7.48318851e-01 8.66334975e-01 1.85508519e-01 -1.34704578e+00
-1.20856249e+00 -6.47092938e-01 7.26734102e-02 9.64478329e-02
9.80868042e-02 -1.93809494e-01 6.80827498e-01 -5.32274485e-01
6.28685653e-01 -1.22114825e+00 -3.21563005e-01 8.93804580e-02
1.02874339e-01 5.28715312e-01 6.16056025e-01 -9.34387386e-01
8.97155285e-01 -2.00367332e+00 -4.70486164e-01 3.28875542e-01
-2.69326121e-01 3.43334019e-01 -4.99550924e-02 4.58922178e-01
3.12608890e-02 3.64740491e-01 -5.29015422e-01 1.10141945e+00
-3.97532910e-01 3.98612529e-01 -3.65944684e-01 2.98980057e-01
4.49112236e-01 8.02717566e-01 -1.13214827e+00 6.15082681e-02
4.38755035e-01 1.34253010e-01 1.97948068e-01 1.62602231e-01
-8.23269635e-02 3.54279190e-01 -2.87968814e-01 9.24460590e-01
1.27427983e+00 6.05301075e-02 5.03237367e-01 -1.78158265e-02
-8.84657800e-01 -6.91351965e-02 -1.00785935e+00 1.28343809e+00
-4.92266625e-01 5.64851463e-01 2.82108992e-01 -1.00881565e+00
1.25931776e+00 1.48508742e-01 3.69123489e-01 -4.74916846e-01
-4.74055201e-01 3.89040649e-01 3.43132578e-02 -3.14806774e-02
5.31145871e-01 -2.20463984e-02 1.66034754e-02 1.57561556e-01
-1.79767415e-01 -1.62954777e-01 4.79940884e-02 -4.27491635e-01
4.55903590e-01 5.43181896e-01 5.56639314e-01 -8.72987568e-01
1.01596594e-01 7.30560064e-01 5.03983438e-01 6.70146525e-01
-1.28095910e-01 2.09450245e-01 2.15339303e-01 -5.50797820e-01
-9.70748186e-01 -1.07195568e+00 -5.55068254e-01 1.31391597e+00
1.32031813e-02 2.83873260e-01 -3.49162340e-01 -2.93333769e-01
3.74292105e-01 9.33828354e-01 -6.92966819e-01 3.85058522e-02
-6.17479756e-02 -1.11512852e+00 4.49159950e-01 6.02863133e-01
8.58780861e-01 -1.05755794e+00 -8.73152316e-01 3.92949820e-01
-2.40220740e-01 -8.43968868e-01 2.90629029e-01 5.27045786e-01
-9.40196335e-01 -7.84159064e-01 -9.04044569e-01 -6.28123522e-01
7.48344421e-01 6.85973287e-01 9.44575846e-01 -3.97865653e-01
-1.81660026e-01 -2.22332031e-01 -5.82881987e-01 -5.78137338e-01
-4.03639048e-01 4.37625468e-01 -3.99348378e-01 -2.00938106e-01
4.69957829e-01 -5.02894044e-01 -4.24965203e-01 4.43254977e-01
-7.37453640e-01 3.73394758e-01 8.17520678e-01 7.32093155e-01
5.07008135e-01 1.82687208e-01 7.49004006e-01 -7.67093182e-01
-2.97383130e-01 -7.64612436e-01 -1.02519572e+00 6.31897211e-01
-6.98732376e-01 -1.94399133e-01 1.55659243e-01 -3.85571048e-02
-1.00838709e+00 6.02733791e-01 3.71133834e-01 5.48582911e-01
-5.97202957e-01 1.13472068e+00 -6.10061996e-02 2.31272914e-02
7.46399820e-01 6.78319559e-02 1.38093354e-02 -2.42476583e-01
2.64321446e-01 8.09728622e-01 7.59140551e-01 -3.70150030e-01
9.04677033e-01 4.37933743e-01 6.15336038e-02 -1.07044685e+00
-7.13812888e-01 -5.50020576e-01 -1.15939021e+00 -2.94948518e-01
6.33060396e-01 -1.14439178e+00 -2.65498817e-01 7.36373007e-01
-7.84416318e-01 -5.00327706e-01 -7.43087456e-02 6.67723179e-01
-2.40765691e-01 -1.34144679e-01 -2.04417314e-02 -7.55943716e-01
-3.58732551e-01 -6.93731844e-01 8.45152974e-01 5.00372350e-01
6.17163964e-02 -8.22701454e-01 -1.74367666e-01 -2.60552526e-01
6.29035413e-01 8.67098451e-01 9.10254717e-01 -1.03007950e-01
-3.63292217e-01 -2.77788699e-01 -7.49032557e-01 1.43374756e-01
9.30152476e-01 2.13399738e-01 -8.80863726e-01 -2.93330282e-01
-6.49083197e-01 2.71208473e-02 8.25701118e-01 7.81850636e-01
4.21819031e-01 1.30248144e-01 -4.27110463e-01 5.86501598e-01
1.70625055e+00 4.07413244e-01 4.49976444e-01 4.72730339e-01
3.75020921e-01 8.72478247e-01 1.19995654e+00 2.55115420e-01
2.96984047e-01 3.81393880e-01 5.19833863e-01 -6.38975084e-01
1.90075591e-01 -2.84392983e-01 2.47927710e-01 -7.75945187e-02
-2.30193660e-01 1.85335621e-01 -1.25915062e+00 8.28337729e-01
-1.77911031e+00 -1.03135824e+00 -2.16833398e-01 2.50250316e+00
8.48515987e-01 -2.95678705e-01 -1.88775584e-01 3.45619321e-02
9.30146337e-01 1.07425690e-01 -5.87733090e-01 -1.28069252e-01
-3.89106512e-01 2.35768124e-01 1.45290148e+00 5.39128304e-01
-1.37133813e+00 1.33885002e+00 5.92777014e+00 2.35827535e-01
-1.34084070e+00 -4.75632906e-01 4.99675483e-01 6.54774189e-01
-4.95286006e-03 4.78435844e-01 -7.63928711e-01 -1.11215241e-01
8.48157108e-01 -1.37655973e-01 3.98390681e-01 6.01619065e-01
3.64118487e-01 -7.03538120e-01 -5.43540716e-01 1.52387321e-01
-6.00331068e-01 -1.28796494e+00 -2.49177545e-01 9.56367180e-02
1.03261805e+00 -1.91239659e-02 -2.94421345e-01 1.34305526e-02
7.50622332e-01 -8.30095947e-01 4.98218596e-01 3.84202927e-01
1.20784771e+00 -6.32591724e-01 9.71166790e-01 3.89950663e-01
-1.57281482e+00 5.89860752e-02 -4.96504188e-01 -3.40515971e-01
-2.65092552e-01 6.24657631e-01 -9.39606071e-01 9.03729200e-01
7.81229556e-01 7.28570938e-01 -5.65191269e-01 6.94315434e-01
-2.99617499e-01 7.22364366e-01 -6.40485704e-01 3.20835024e-01
4.33061898e-01 -1.90587118e-01 -2.54218578e-02 1.16786349e+00
7.26757169e-01 2.15291634e-01 3.96451682e-01 8.27483475e-01
5.20239651e-01 7.23387077e-02 -8.00050139e-01 1.01295620e-01
7.08484948e-01 1.19591582e+00 -1.00432932e+00 -1.35274697e-02
1.88535638e-02 5.49169183e-01 -2.31598258e-01 4.82741922e-01
-3.67793679e-01 -4.90664542e-01 4.99249816e-01 1.37182340e-01
3.18321288e-01 -3.86472613e-01 -3.44479650e-01 -7.93709338e-01
-4.59290855e-02 -4.68902171e-01 1.17859967e-01 -7.93255985e-01
-8.06017995e-01 1.79701671e-01 3.04187179e-01 -1.08348262e+00
-3.46454710e-01 -6.67355180e-01 -4.58366364e-01 1.41396201e+00
-1.80519402e+00 -1.59241426e+00 -7.59533763e-01 -4.68871966e-02
-2.68764648e-04 3.82266864e-02 1.36404157e+00 -3.20014507e-01
-1.76791787e-01 8.43809247e-02 5.77501535e-01 7.10016713e-02
4.92228925e-01 -1.06089842e+00 4.12271947e-01 9.02860224e-01
-2.99905419e-01 2.54430175e-01 3.53548706e-01 -8.25826466e-01
-8.33779991e-01 -1.53650486e+00 8.33229661e-01 3.48839611e-01
6.33869708e-01 -3.39362882e-02 -8.11983466e-01 7.74431169e-01
-3.31772774e-01 -5.27968034e-02 5.64245582e-01 2.79204130e-01
-1.95364386e-01 -5.68270147e-01 -1.21895552e+00 2.19314158e-01
6.32129073e-01 -4.02964085e-01 8.40978622e-02 2.85176635e-01
2.99005955e-01 -1.32673100e-01 -1.00580239e+00 6.87970161e-01
9.87810791e-01 -6.64606392e-01 7.35509396e-01 -1.60272360e-01
2.62282699e-01 -4.50706303e-01 -6.24651015e-01 -1.49823880e+00
-7.06268251e-01 -1.28168553e-01 9.08308923e-01 1.38141656e+00
5.92137754e-01 -7.81983912e-01 4.73854572e-01 1.48510158e-01
1.93712950e-01 6.73954084e-04 -4.46685135e-01 -8.74620676e-01
3.96212131e-01 9.19734314e-02 9.19430792e-01 9.08697724e-01
-5.36835492e-01 -2.88494974e-01 1.02830686e-01 9.69668686e-01
4.83438194e-01 4.73233253e-01 7.03494549e-01 -1.69030368e+00
3.48497897e-01 -2.47026846e-01 -4.40362126e-01 -4.12632078e-01
3.21341567e-02 -7.93446362e-01 1.49501413e-01 -1.44626582e+00
8.96203052e-03 -1.11103368e+00 -2.49277353e-01 1.06596935e+00
-3.68766636e-01 -4.53178845e-02 -1.75953154e-02 2.26654068e-01
6.85204268e-01 8.22991207e-02 6.82070792e-01 -2.07771987e-01
-5.29487789e-01 2.03074768e-01 -5.85963130e-01 4.87774014e-01
1.07085001e+00 -4.44745094e-01 -1.69795901e-01 -4.51991618e-01
4.50667879e-03 8.85161087e-02 6.00820541e-01 -9.10264850e-01
-3.87676597e-01 -8.58728409e-01 4.97340888e-01 -8.68501067e-01
-3.64995837e-01 -9.72676694e-01 5.82128525e-01 6.27713263e-01
-4.13954072e-02 -2.10918128e-01 7.50803769e-01 1.93736941e-01
1.59637127e-02 -2.68342406e-01 6.54725730e-01 -5.31255156e-02
-1.08158863e+00 1.51384398e-01 -4.76391047e-01 -5.43892741e-01
1.08661389e+00 -1.90799952e-01 -2.50025600e-01 -1.64810881e-01
-6.60338283e-01 3.05529684e-01 6.31277561e-01 4.51943040e-01
-9.36800241e-02 -1.18367088e+00 -1.21825445e+00 5.96232474e-01
4.41403955e-01 -1.83663100e-01 9.53158457e-03 3.10006261e-01
-9.36489701e-01 4.22068596e-01 -8.64340961e-01 -9.29133713e-01
-9.79137301e-01 1.45759776e-01 2.88206637e-01 -1.94833055e-01
-1.84168547e-01 3.26102406e-01 8.51610452e-02 -6.23602748e-01
-4.56717849e-01 -4.70366806e-01 -2.04010546e-01 4.03113246e-01
2.51976997e-01 -2.13648807e-02 4.31443751e-02 -7.64056742e-01
-3.40767354e-01 4.88129824e-01 2.03801632e-01 -1.25806883e-01
1.44349754e+00 -2.08224710e-02 -2.40654141e-01 7.10627854e-01
5.46580613e-01 -2.97620118e-01 -1.33638144e+00 -2.17316553e-01
2.44165689e-01 -5.07088959e-01 3.84288251e-01 -1.12656164e+00
-9.22220945e-01 5.86992741e-01 1.04364347e+00 3.08406085e-01
1.38092220e+00 -3.24848980e-01 9.48558599e-02 3.67179185e-01
5.94558835e-01 -8.62033844e-01 -1.14176142e+00 3.14329267e-01
7.37367392e-01 -1.46739364e+00 1.96276546e-01 -3.69781733e-01
-1.77168310e-01 1.22551465e+00 1.77326277e-01 1.88989297e-01
7.16634870e-01 1.85655221e-01 3.51630569e-01 1.10761888e-01
-1.69686630e-01 -6.65703237e-01 -7.93240685e-03 1.02922690e+00
6.54379964e-01 1.02257299e+00 -7.47760087e-02 2.11954471e-02
-2.99741089e-01 1.30879536e-01 4.01556224e-01 1.17469490e+00
-7.91244328e-01 -8.97727549e-01 -6.71787798e-01 5.69944561e-01
1.45443603e-01 -1.68105498e-01 -1.75389931e-01 8.63278687e-01
1.77546233e-01 9.30167139e-01 7.00739920e-02 -2.30114311e-01
2.00494960e-01 2.48762593e-02 1.71173781e-01 -6.82408571e-01
-2.24632755e-01 3.41465101e-02 -1.86633825e-01 -1.03547163e-01
-6.30920112e-01 -9.67719555e-01 -9.26090419e-01 -5.63288927e-01
-7.15696752e-01 -1.18981590e-02 9.87155020e-01 6.65039420e-01
2.67167836e-01 1.09252900e-01 8.77306879e-01 -9.71543133e-01
-3.64869505e-01 -1.03148532e+00 -1.02275419e+00 -2.83878475e-01
4.25859749e-01 -6.47708476e-01 -5.23381755e-02 1.67073652e-01] | [9.404327392578125, -1.5779999494552612] |
1df5f036-b7e7-45eb-bc1a-80561d0a19a2 | real-time-dense-stereo-embedded-in-a-uav-for | 1904.06017 | null | http://arxiv.org/abs/1904.06017v1 | http://arxiv.org/pdf/1904.06017v1.pdf | Real-Time Dense Stereo Embedded in A UAV for Road Inspection | The condition assessment of road surfaces is essential to ensure their
serviceability while still providing maximum road traffic safety. This paper
presents a robust stereo vision system embedded in an unmanned aerial vehicle
(UAV). The perspective view of the target image is first transformed into the
reference view, and this not only improves the disparity accuracy, but also
reduces the algorithm's computational complexity. The cost volumes generated
from stereo matching are then filtered using a bilateral filter. The latter has
been proved to be a feasible solution for the functional minimisation problem
in a fully connected Markov random field model. Finally, the disparity maps are
transformed by minimising an energy function with respect to the roll angle and
disparity projection model. This makes the damaged road areas more
distinguishable from the road surface. The proposed system is implemented on an
NVIDIA Jetson TX2 GPU with CUDA for real-time purposes. It is demonstrated
through experiments that the damaged road areas can be easily distinguished
from the transformed disparity maps. | ['Huaiyang Huang', 'Jie Pan', 'Jianhao Jiao', 'Rui Fan', 'Shaojie Shen', 'Ming Liu'] | 2019-04-12 | null | null | null | null | ['stereo-matching'] | ['computer-vision'] | [ 7.79355764e-01 -1.28037006e-01 3.67911547e-01 -5.02169132e-02
1.37339756e-01 -3.90923291e-01 6.14313364e-01 -4.68208134e-01
-5.25392413e-01 7.01037586e-01 -5.99670231e-01 -3.97922367e-01
1.07594896e-02 -1.15030730e+00 -3.88904929e-01 -7.83308148e-01
4.00017589e-01 1.34288192e-01 4.65767145e-01 -2.62735263e-02
4.71161366e-01 6.70088887e-01 -2.43638563e+00 -1.76351041e-01
1.21809924e+00 9.96702731e-01 6.44531488e-01 5.85383475e-01
2.56494939e-01 2.12753117e-01 -1.48692384e-01 -1.68209285e-01
8.09979677e-01 -2.51450809e-03 -6.09405518e-01 4.90634471e-01
7.78239548e-01 -6.03130460e-01 -2.60119051e-01 1.46049094e+00
1.72562093e-01 1.25279635e-01 6.22479200e-01 -1.02837825e+00
4.10980195e-01 -8.12284350e-01 -6.61691129e-01 -1.05732903e-01
3.48657519e-01 2.66196579e-01 5.28108835e-01 -7.71175325e-01
6.94672227e-01 1.13534606e+00 2.43119970e-01 8.41425210e-02
-9.80702341e-01 -4.32702780e-01 -1.32583693e-01 2.94241697e-01
-1.43517566e+00 -2.06827149e-01 5.23893774e-01 -7.61237442e-01
7.61083305e-01 5.21898568e-01 9.26544547e-01 1.54958293e-01
7.28184104e-01 1.00207575e-01 1.38690305e+00 -4.48161930e-01
1.62869602e-01 -2.86547765e-02 -2.86113955e-02 7.22204506e-01
5.46463311e-01 6.63139522e-01 -9.55255404e-02 -9.68112275e-02
9.09738302e-01 8.19609463e-02 -5.97948194e-01 -6.71186090e-01
-1.07376027e+00 5.01872659e-01 2.84432650e-01 -1.75122470e-01
-4.25274968e-01 -2.00758353e-01 -3.25104147e-02 1.80666089e-01
4.74650979e-01 -1.11972280e-01 -1.52262405e-01 1.98356435e-01
-6.73667848e-01 9.81910825e-02 5.36047816e-01 7.24850237e-01
1.07305372e+00 1.86587945e-01 5.89504361e-01 3.84912103e-01
5.72507203e-01 1.08095491e+00 -1.29784808e-01 -1.15176344e+00
3.55011851e-01 7.18682945e-01 2.78497666e-01 -1.24790001e+00
-8.98129717e-02 -9.15951282e-02 -6.63376093e-01 1.26780295e+00
2.56471962e-01 -1.40156269e-01 -7.88421869e-01 9.98448074e-01
6.64086044e-01 -3.53331231e-02 8.31235722e-02 8.55773866e-01
5.11268258e-01 7.82876551e-01 -4.57961887e-01 -2.10153148e-01
1.09367180e+00 -4.42976624e-01 -5.25532901e-01 -2.80639350e-01
2.08415478e-01 -1.02359533e+00 4.59456325e-01 5.92928112e-01
-8.45494092e-01 -5.68375111e-01 -1.09844160e+00 6.65218011e-02
-1.03613637e-01 3.32441688e-01 1.16424948e-01 4.94770288e-01
-1.04561210e+00 4.63430226e-01 -6.28559709e-01 -2.42426783e-01
-1.50816634e-01 4.03894037e-01 -4.04927909e-01 -1.66025534e-01
-9.88664269e-01 1.14767516e+00 1.89248532e-01 5.39737046e-01
-4.19564575e-01 -2.79778272e-01 -9.63765860e-01 -3.15474331e-01
1.57084078e-01 -7.88049459e-01 6.72697365e-01 -1.01486766e+00
-1.50955677e+00 1.14862251e+00 -2.27243796e-01 -2.12885171e-01
8.15580368e-01 -2.71515921e-02 -5.04790768e-02 2.16086224e-01
8.80112313e-03 4.45161968e-01 1.07793891e+00 -1.17116344e+00
-1.15161824e+00 -5.66321552e-01 7.24260807e-02 7.02779710e-01
3.14248174e-01 -3.62152576e-01 -9.05527323e-02 -1.04791261e-01
2.02577353e-01 -9.79791760e-01 -2.93397844e-01 2.32365578e-01
-1.73321635e-01 4.17883247e-01 1.06262505e+00 -7.85844386e-01
8.64445865e-01 -2.02676105e+00 3.23831290e-02 4.65562314e-01
-2.83312052e-02 2.91729003e-01 2.86975831e-01 7.82718435e-02
1.60943940e-01 -6.34663284e-01 -5.03502846e-01 2.07547143e-01
-6.53683126e-01 5.51684611e-02 -3.92538123e-02 8.59439969e-01
-1.81360945e-01 1.21517822e-01 -5.68622887e-01 -3.60539198e-01
7.31587052e-01 5.45736134e-01 -2.33084962e-01 3.61877270e-02
1.35851890e-01 4.45401758e-01 -2.44547695e-01 3.68248552e-01
1.28505313e+00 6.72882080e-01 -1.39636502e-01 -1.05671488e-01
-7.79159784e-01 -2.05005512e-01 -1.50409925e+00 9.82870758e-01
-5.22939205e-01 9.54861581e-01 5.32624245e-01 -4.41096455e-01
1.25483835e+00 2.15966720e-02 3.24624598e-01 -6.00815415e-01
2.01486617e-01 3.28973442e-01 -9.54920202e-02 -2.98980415e-01
6.74500108e-01 -1.55142574e-02 4.19149399e-01 -1.31982595e-01
-6.26793623e-01 -5.87476492e-01 -9.70390812e-02 -2.87085414e-01
4.84712005e-01 1.84657782e-01 3.78892839e-01 -6.43979669e-01
9.18152034e-01 3.33284706e-01 5.48470497e-01 -1.04699768e-01
1.84546024e-01 4.58472311e-01 6.36876523e-02 -5.03802538e-01
-9.02715981e-01 -9.99365270e-01 -5.35548031e-01 -6.11139201e-02
8.20339859e-01 2.68780172e-01 -8.47240806e-01 -1.38446718e-01
8.73057917e-02 4.90722835e-01 -2.07252577e-01 5.52563556e-02
-3.35859656e-01 -5.56653202e-01 -1.73463315e-01 -1.42387329e-02
9.17743444e-01 -6.71840429e-01 -1.43221533e+00 2.29494218e-02
-2.48246327e-01 -9.34575260e-01 8.67644232e-03 -3.21102500e-01
-1.03369522e+00 -1.48057532e+00 -6.84768915e-01 -8.09793413e-01
9.07423794e-01 9.55404520e-01 7.16792941e-01 -3.00018638e-02
-6.27435982e-01 5.25835276e-01 1.56021714e-01 -1.51183441e-01
-1.48528516e-01 -5.43377459e-01 4.00002254e-03 2.83586919e-01
1.91171721e-01 -1.48115352e-01 -7.26480484e-01 8.17195773e-01
-6.18991733e-01 3.16478819e-01 2.15632632e-01 6.99924350e-01
7.39536166e-01 4.21350479e-01 -2.85503298e-01 -3.68067294e-01
1.27581134e-01 -2.25718226e-02 -1.42407906e+00 -1.29084885e-01
-5.26335835e-01 -5.04343748e-01 2.61215895e-01 1.84255227e-01
-1.32656717e+00 4.00148898e-01 5.63073233e-02 -7.08795860e-02
-2.61128396e-01 -5.23479283e-02 -3.63827765e-01 -5.06635070e-01
3.42846602e-01 8.00902247e-02 4.99484837e-01 -7.61080235e-02
-2.69120522e-02 8.24910223e-01 5.49752235e-01 6.68871179e-02
8.35172236e-01 1.00951433e+00 5.00482917e-01 -1.34502149e+00
-8.42362791e-02 -6.16333961e-01 -8.61150622e-01 -7.86602199e-01
1.00113368e+00 -9.76383209e-01 -7.17854440e-01 8.18671048e-01
-1.14048004e+00 -8.20577964e-02 3.90672423e-02 5.94247282e-01
-6.77082181e-01 6.54364288e-01 -1.38296276e-01 -8.76306653e-01
-2.98266411e-01 -1.26823521e+00 7.79272556e-01 3.50095332e-01
1.07669190e-01 -1.03712487e+00 -7.66326413e-02 2.74093181e-01
1.97016299e-01 4.62657094e-01 5.38819134e-01 5.84585130e-01
-9.27268386e-01 -2.35511988e-01 -2.18901575e-01 5.10732710e-01
9.12630484e-02 5.02954781e-01 -1.03712738e+00 -2.29880556e-01
3.92715573e-01 4.05415684e-01 8.00226033e-01 7.10910857e-01
3.26562792e-01 9.79112312e-02 -3.69165093e-01 6.64087772e-01
1.73239851e+00 4.53842014e-01 8.89018893e-01 6.28466547e-01
6.08220577e-01 1.08797050e+00 1.32474828e+00 2.29068667e-01
9.52662975e-02 7.20611572e-01 9.68419254e-01 -3.06873381e-01
3.33719589e-02 1.93272740e-01 5.25994182e-01 4.41647679e-01
-3.23528588e-01 -4.50545270e-03 -8.60585034e-01 3.92007738e-01
-1.56753469e+00 -8.58720660e-01 -1.01420653e+00 2.73452258e+00
3.02979182e-02 -9.50428918e-02 -3.40361238e-01 3.46355230e-01
1.00275040e+00 -5.40006021e-03 -3.33998770e-01 -5.37819505e-01
-2.63988990e-02 -1.08894385e-01 7.97429979e-01 1.06241906e+00
-1.13723528e+00 6.82325959e-01 5.24472189e+00 6.44169867e-01
-1.27398050e+00 -3.68767291e-01 2.49097556e-01 3.10689539e-01
-1.22940801e-01 2.26317853e-01 -7.87632048e-01 4.27577674e-01
4.61342752e-01 -2.73896456e-01 1.10450685e-01 7.78568804e-01
3.48630488e-01 -1.05856836e+00 -5.00138044e-01 1.10618675e+00
-3.58208790e-02 -7.47762203e-01 2.39956230e-02 5.72870731e-01
6.46973372e-01 -1.07805720e-02 1.66332394e-01 -6.01896286e-01
-1.03961870e-01 -6.50559008e-01 4.81665045e-01 5.22961855e-01
1.08983362e+00 -1.03803408e+00 9.53567863e-01 4.56046551e-01
-1.25779283e+00 1.65672347e-01 -7.11906314e-01 -4.01723415e-01
3.08737248e-01 6.88375890e-01 -7.26156294e-01 5.81750214e-01
4.89473462e-01 6.13081276e-01 -2.37513930e-01 1.18666303e+00
-7.23040998e-02 -1.79912850e-01 -2.75969923e-01 3.34534109e-01
1.75540060e-01 -1.04131091e+00 9.89720225e-01 6.14566445e-01
6.27574205e-01 8.14506114e-02 -1.34301344e-02 5.33678532e-01
6.91454887e-01 7.60272592e-02 -1.18180788e+00 5.94853878e-01
-8.30270573e-02 1.17987883e+00 -6.64805949e-01 -1.57674283e-01
-3.74395937e-01 9.26358283e-01 -5.23875654e-01 1.05856977e-01
-5.71885169e-01 -5.25844455e-01 8.22070420e-01 4.80521917e-01
6.62775338e-03 -1.95513844e-01 -3.18142772e-01 -1.08672905e+00
1.26034826e-01 -4.72823560e-01 -9.57911760e-02 -8.26764464e-01
-6.30753279e-01 7.68522024e-01 -7.75436535e-02 -1.79736137e+00
-1.40902191e-01 -7.67338455e-01 -3.86596680e-01 1.35238373e+00
-1.63287497e+00 -6.70435727e-01 -7.32016385e-01 6.43751502e-01
4.99167472e-01 -2.29615048e-02 7.02621460e-01 1.27688006e-01
-2.44102031e-01 -3.42164099e-01 1.69828773e-01 -4.41568702e-01
3.12974572e-01 -8.44494224e-01 3.31465036e-01 1.27427197e+00
-5.66092491e-01 1.90268382e-01 6.71596110e-01 -9.14576828e-01
-1.15990627e+00 -1.04736054e+00 7.69695520e-01 -1.10498793e-01
2.06730455e-01 7.92365596e-02 -7.21844733e-01 1.91337973e-01
2.24878967e-01 -2.87837267e-01 7.71472454e-02 -7.84314454e-01
2.76879787e-01 -1.74261123e-01 -1.37922323e+00 5.00222921e-01
8.13819706e-01 -3.67428064e-01 -4.83277798e-01 6.62130769e-04
-3.04155257e-02 -4.38264400e-01 -5.01253903e-01 5.32185376e-01
6.20964468e-01 -1.35494840e+00 8.20566893e-01 2.19206333e-01
1.82277203e-01 -7.62459159e-01 -7.22042546e-02 -1.24777639e+00
-8.45777318e-02 -5.66126406e-01 4.83299792e-01 8.42530549e-01
-1.04886023e-02 -1.05053401e+00 6.50036097e-01 4.23219413e-01
-5.33683598e-02 -4.09434944e-01 -1.13224316e+00 -7.66112030e-01
-4.07286465e-01 -8.96909684e-02 1.09575137e-01 4.28712636e-01
-3.14530075e-01 -8.71642604e-02 -1.91344678e-01 4.88543332e-01
1.09424925e+00 2.72753656e-01 8.08652639e-01 -1.71716571e+00
2.82599986e-01 -1.62802249e-01 -9.07784224e-01 -1.01296675e+00
-5.31216525e-02 -5.59036076e-01 1.17730863e-01 -1.57816839e+00
-1.45280227e-01 -3.74037296e-01 5.81463575e-01 -3.86419654e-01
2.93745752e-02 2.01979771e-01 8.42100475e-04 1.45234600e-01
3.83651108e-01 3.73011082e-01 1.25338030e+00 9.52890366e-02
-9.28729326e-02 5.03409088e-01 8.62946510e-02 1.19350672e+00
7.16402054e-01 -7.33139962e-02 -5.05404532e-01 -2.67813712e-01
1.27165824e-01 1.39270723e-01 4.90329802e-01 -1.26281273e+00
1.06191665e-01 -1.61172315e-01 1.45047143e-01 -9.23833072e-01
5.38685203e-01 -1.25132167e+00 4.67459559e-01 7.81032979e-01
5.15067756e-01 2.76485384e-01 1.45253286e-01 5.95167756e-01
-4.15672928e-01 -2.50164509e-01 1.26786780e+00 1.37200207e-02
-9.83338475e-01 9.38882455e-02 -7.08841562e-01 -4.18202817e-01
1.63154685e+00 -8.63146782e-01 -1.47241473e-01 -2.39286587e-01
-4.29237098e-01 8.30765963e-02 1.20277536e+00 -7.74226040e-02
9.54663813e-01 -9.37876225e-01 -4.90971923e-01 7.34469056e-01
-1.00255571e-01 2.06867129e-01 5.15515149e-01 7.61583328e-01
-1.43161571e+00 4.19969559e-01 -6.37845159e-01 -9.13267195e-01
-1.78868628e+00 3.98421437e-01 4.45928127e-01 2.72895962e-01
-7.63127744e-01 4.68118310e-01 4.26375955e-01 -2.12204903e-01
1.49332862e-02 -3.24659914e-01 -3.62876922e-01 5.98924756e-02
5.52870989e-01 9.05338049e-01 1.15063274e-02 -1.01834047e+00
-3.24123681e-01 1.37790918e+00 4.02367443e-01 -3.68953407e-01
8.55430782e-01 -4.85730141e-01 -1.90856397e-01 -1.05559044e-02
9.16766703e-01 1.78510368e-01 -1.44111848e+00 2.39238232e-01
-3.93594623e-01 -9.85365212e-01 3.84493172e-01 -1.62173688e-01
-1.00227237e+00 1.04006064e+00 7.42773712e-01 1.12450324e-01
1.17293966e+00 -8.05106938e-01 6.80591941e-01 1.39739648e-01
8.77474487e-01 -1.10316122e+00 -7.21661866e-01 4.96780694e-01
6.25561714e-01 -1.07557845e+00 1.57510638e-01 -9.28363919e-01
-7.05578506e-01 1.47119832e+00 4.38279599e-01 -2.82190055e-01
6.26953304e-01 2.87366331e-01 1.12482481e-01 4.71523255e-02
-3.50916088e-01 -5.50803244e-01 2.11840883e-01 7.71836877e-01
-1.44536823e-01 9.29292366e-02 -2.91392982e-01 -5.78465343e-01
-1.71209842e-01 -8.30542073e-02 8.78639162e-01 8.64651263e-01
-8.77268076e-01 -6.62230670e-01 -6.85123205e-01 2.31685460e-01
-1.24736197e-01 9.10350010e-02 -1.81758642e-01 7.39013076e-01
2.60036737e-01 1.07297277e+00 3.83382291e-01 -3.13792169e-01
5.70865512e-01 -4.22206908e-01 4.16032404e-01 -4.28711176e-01
-1.83590114e-01 2.27184105e-03 1.57168940e-01 -6.20457828e-01
-4.63151276e-01 -7.99601316e-01 -1.00608528e+00 -2.61824787e-01
-3.29164296e-01 2.91275196e-02 8.63328338e-01 4.74819720e-01
2.59304464e-01 6.81354925e-02 9.32573497e-01 -9.41418648e-01
-1.99517086e-01 -3.59054834e-01 -6.78533077e-01 -1.94384322e-01
3.08661729e-01 -1.13092983e+00 -7.26557255e-01 1.14385501e-01] | [8.908501625061035, -2.4322566986083984] |
8690cdf9-efd3-45e0-85c1-58d8d49092de | unsupervised-learning-of-stereo-matching | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Zhou_Unsupervised_Learning_of_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Zhou_Unsupervised_Learning_of_ICCV_2017_paper.pdf | Unsupervised Learning of Stereo Matching | In recent years, convolutional neural networks have shown its strong power for stereo matching cost learning. Current approaches learn the parameters of their models from public datasets with ground truth disparity. However, due to the limitations of these datasets and the difficulty of collecting new stereo data, current methods fail in real-life cases. In this work, we present a framework for learning stereo matching cost without human supervision. Our method updates the network parameter in a iterative manner. It starts with randomly initialized network. Correct matchings are carefully picked and used as training data in each round. In the end, the networks converges to a stable state, which performs comparably with supervised methods on various benchmarks.
| ['Chao Zhou', 'Xiaoyong Shen', 'Jiaya Jia', 'Hong Zhang'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['stereo-matching'] | ['computer-vision'] | [ 7.26088434e-02 -6.28285259e-02 -5.16587853e-01 -6.32813931e-01
-6.07617199e-01 -3.38080466e-01 4.28394794e-01 -2.79407613e-02
-9.65825915e-01 8.80156994e-01 2.92705476e-01 2.92529371e-02
1.78784356e-01 -9.54015195e-01 -9.08871174e-01 -3.75843287e-01
1.79557860e-01 5.99235892e-01 3.62465024e-01 -1.13236956e-01
4.21502471e-01 3.08714181e-01 -1.87655771e+00 2.79223323e-02
9.90760922e-01 1.12863159e+00 7.76957870e-02 4.11426991e-01
7.11343586e-02 8.31124246e-01 -2.70612329e-01 -5.78794241e-01
8.11247647e-01 -1.99109260e-02 -9.54004943e-01 4.17302130e-03
9.28455472e-01 -8.01645577e-01 -6.16380930e-01 1.09526730e+00
5.53761244e-01 -2.75229872e-03 2.40583360e-01 -1.16394114e+00
-4.82497504e-03 3.91443640e-01 -3.63468140e-01 -8.49361252e-03
2.46173799e-01 3.54602307e-01 1.11222708e+00 -6.87330127e-01
6.44671738e-01 1.11044753e+00 9.75056350e-01 6.15956306e-01
-1.29942596e+00 -9.57450390e-01 -8.62891879e-03 6.73434436e-02
-1.18018341e+00 -5.26158035e-01 6.63637698e-01 -4.34534907e-01
1.03741479e+00 -2.43668169e-01 9.61892962e-01 9.37236786e-01
6.13391399e-03 7.55130768e-01 9.63319540e-01 -2.99454182e-01
1.56004310e-01 -2.40890846e-01 -8.38362128e-02 7.11395621e-01
3.38245779e-01 6.85803831e-01 -6.98870838e-01 -2.36945264e-02
8.87313783e-01 -8.80188942e-02 -7.49249011e-02 -5.91351926e-01
-1.10560250e+00 8.46238315e-01 8.19768429e-01 -5.35431914e-02
-2.55477905e-01 4.74169463e-01 4.31438446e-01 4.20057774e-01
5.33761680e-01 4.17498410e-01 -4.66968030e-01 -1.16444483e-01
-9.10977006e-01 5.46990156e-01 6.07459486e-01 8.26667547e-01
1.12118328e+00 -4.49300051e-01 8.50180462e-02 9.53279018e-01
9.04518366e-03 3.21196198e-01 4.98565853e-01 -1.40834689e+00
6.96667135e-01 7.62355983e-01 9.67742652e-02 -7.86688745e-01
-2.78952152e-01 -2.72956908e-01 -8.67622018e-01 4.12463456e-01
7.20337033e-01 -2.17721254e-01 -9.87049639e-01 1.71076810e+00
1.54977679e-01 3.84095460e-01 -8.21407363e-02 8.38216603e-01
7.41373420e-01 2.76401132e-01 -2.32049003e-01 4.03918445e-01
7.32268393e-01 -9.35139716e-01 -2.32796952e-01 -4.82728571e-01
4.82432425e-01 -6.77679300e-01 9.18288350e-01 3.31705332e-01
-1.11592555e+00 -5.62481821e-01 -9.79671597e-01 -3.68772522e-02
-2.17786759e-01 -1.05171040e-01 1.01251268e+00 4.49249029e-01
-1.24432433e+00 8.92493606e-01 -6.39323473e-01 -2.62599438e-01
7.47364879e-01 6.53557539e-01 -2.71040738e-01 -1.81947842e-01
-1.05327988e+00 7.07456708e-01 5.55358529e-01 1.09679237e-01
-7.55865097e-01 -7.68351376e-01 -9.72762823e-01 -2.50179559e-01
2.45811522e-01 -1.02324545e+00 1.57609534e+00 -1.16799104e+00
-1.58712912e+00 1.35093403e+00 -8.16026889e-03 -7.14892328e-01
9.96950805e-01 -4.06389117e-01 2.22570658e-01 -3.66127729e-01
5.96553311e-02 1.39389801e+00 6.57846928e-01 -1.04447067e+00
-1.09314144e+00 -1.53025597e-01 5.48859656e-01 2.14737594e-01
-9.11845788e-02 -2.79787749e-01 -8.38359356e-01 -3.96615148e-01
3.47492546e-01 -1.05832493e+00 -5.14415383e-01 2.04548717e-01
-2.63414621e-01 -8.82679820e-02 2.13727549e-01 -1.35437310e-01
8.43125939e-01 -2.07085562e+00 -1.77802239e-02 2.01894477e-01
2.79885173e-01 9.24119875e-02 -5.62024936e-02 6.93742260e-02
-8.83383770e-03 -1.99864075e-01 -3.86171676e-02 -5.05180478e-01
-2.33610556e-01 2.71041423e-01 -5.27412035e-02 4.16695952e-01
6.37463331e-02 8.44081283e-01 -1.06621397e+00 -6.30606711e-01
4.03936505e-01 9.58827361e-02 -9.78943646e-01 4.08528388e-01
-7.22943768e-02 4.00658756e-01 -1.90389361e-02 5.75953603e-01
6.47678912e-01 -4.28237081e-01 -9.31095053e-03 -2.42995679e-01
1.54198091e-02 3.62028241e-01 -1.35307610e+00 1.90930128e+00
-5.58869660e-01 9.49923694e-01 -3.54299664e-01 -8.62817466e-01
7.50818789e-01 -9.99527574e-02 5.58252573e-01 -9.56585586e-01
2.79121965e-01 4.13520455e-01 -7.76775330e-02 -2.05039814e-01
4.69486833e-01 3.68845873e-02 2.09906444e-01 3.25021923e-01
4.68891673e-02 -4.24040854e-01 2.96185881e-01 -2.31835738e-01
8.02020073e-01 1.18209139e-01 2.26719797e-01 -9.53159928e-02
4.79804605e-01 9.79269817e-02 7.65768051e-01 7.42486835e-01
-1.92687571e-01 9.33352053e-01 9.60624814e-02 -1.08534062e+00
-1.36753237e+00 -9.24205303e-01 -2.32864544e-01 8.08401585e-01
3.35547507e-01 -1.15620665e-01 -8.94281447e-01 -5.71381450e-01
2.71217734e-01 -3.66953835e-02 -7.18638837e-01 -6.09075725e-02
-7.67676890e-01 -5.61690927e-01 3.87416393e-01 7.18792081e-01
9.67465818e-01 -1.12665534e+00 -7.11539090e-01 2.61856556e-01
-3.28305423e-01 -1.26002896e+00 -4.44488913e-01 6.72673732e-02
-1.22764587e+00 -1.28439748e+00 -7.98002899e-01 -8.92648816e-01
8.84742796e-01 2.25679234e-01 1.45527279e+00 3.99219602e-01
-2.38052249e-01 -8.95901844e-02 2.63995200e-01 -4.59915131e-01
-2.69363016e-01 5.34440756e-01 -1.97579071e-01 -2.66950607e-01
4.02835608e-01 -4.49500769e-01 -8.87570858e-01 4.62497950e-01
-7.22781241e-01 3.27922970e-01 4.19541925e-01 8.38493884e-01
6.23637497e-01 -2.08138585e-01 4.67316881e-02 -9.08232212e-01
4.28430825e-01 -8.80629197e-02 -1.18281758e+00 -9.08950195e-02
-8.21191370e-01 1.61966577e-01 3.69842291e-01 -2.23986059e-01
-7.33779430e-01 2.93673128e-01 -7.84583688e-02 -3.80683094e-01
-1.64022803e-01 2.17168763e-01 3.35354716e-01 -3.67230564e-01
7.81899154e-01 -1.48474172e-01 1.03432715e-01 -2.27229655e-01
-1.13672741e-01 3.80993605e-01 6.93745613e-01 -5.73716581e-01
8.62645745e-01 6.38540387e-01 -1.08124048e-01 -2.17115164e-01
-1.12676764e+00 -3.36431563e-01 -7.54059315e-01 -3.48012507e-01
4.28539634e-01 -1.15300941e+00 -7.30956674e-01 1.03113067e+00
-1.00811875e+00 -5.68423688e-01 -3.65571767e-01 4.53198493e-01
-6.96998358e-01 1.24806598e-01 -5.39422989e-01 -3.20638984e-01
-2.67678112e-01 -1.04916108e+00 1.06957674e+00 3.21617246e-01
-4.40025590e-02 -1.11682057e+00 2.76862144e-01 4.16559964e-01
4.27908808e-01 2.07353219e-01 4.07631546e-01 1.52087137e-02
-8.68154526e-01 -3.08566839e-01 -4.09664929e-01 4.95543092e-01
2.34770611e-01 -4.97389548e-02 -1.14969742e+00 -3.48720610e-01
-5.24411976e-01 -6.81516767e-01 9.64270413e-01 7.28702307e-01
1.41483092e+00 6.96464628e-02 -2.21239790e-01 1.10691929e+00
1.64644408e+00 -2.18255430e-01 7.03331769e-01 9.32200491e-01
3.32191557e-01 4.82077926e-01 5.87294698e-01 3.46244961e-01
5.58063030e-01 5.82305253e-01 6.78566396e-01 -3.28012079e-01
-8.94264802e-02 -4.45293367e-01 -2.44678631e-02 3.25742781e-01
-2.94857979e-01 1.00940369e-01 -1.11724067e+00 5.32350004e-01
-1.99191284e+00 -7.58042037e-01 2.38116995e-01 2.39654446e+00
8.51286829e-01 6.43294036e-01 1.20577224e-01 4.53812666e-02
6.08283639e-01 1.09975815e-01 -7.60540545e-01 8.69341344e-02
-2.48661265e-01 2.28907894e-02 1.04437411e+00 4.82635438e-01
-1.17475832e+00 1.08689356e+00 7.46075916e+00 2.70414114e-01
-1.25781846e+00 -2.32469484e-01 7.97469974e-01 -1.12719439e-01
-2.12188497e-01 1.25747956e-02 -7.88283527e-01 4.45910484e-01
4.43927884e-01 -1.50447935e-01 5.70802808e-01 6.63153172e-01
5.27615622e-02 -3.47663492e-01 -1.23952305e+00 1.36813378e+00
-2.20434189e-01 -1.68430805e+00 -6.03622198e-02 4.30779569e-02
1.03804982e+00 6.29648745e-01 2.41004005e-02 5.84536567e-02
8.19625199e-01 -1.05176651e+00 7.57914841e-01 3.27803671e-01
8.95560324e-01 -7.50380814e-01 1.12962389e+00 3.14073741e-01
-9.04941320e-01 -1.78716809e-01 -5.66870987e-01 -2.97885269e-01
4.97300103e-02 6.62536979e-01 -6.26636863e-01 2.82137603e-01
9.99313712e-01 9.60826516e-01 -5.85505605e-01 1.43871665e+00
-1.51015639e-01 2.00334176e-01 -5.15431166e-01 2.19248667e-01
3.57930183e-01 -2.44192854e-01 -2.40993891e-02 9.47743833e-01
6.33881912e-02 -3.13424826e-01 6.05263412e-02 5.63643932e-01
-3.73919755e-01 1.33795217e-02 -7.92986155e-01 4.55093712e-01
4.19254363e-01 8.05276930e-01 -4.99403328e-01 -2.83530444e-01
-4.25171137e-01 6.50494516e-01 6.72316611e-01 1.76441833e-01
-7.42719769e-01 -9.15450677e-02 7.47207642e-01 6.46763146e-02
1.97735317e-02 -1.48961088e-02 -3.66517395e-01 -1.21996653e+00
1.91197991e-01 -8.48909020e-01 4.72321004e-01 -5.16735554e-01
-1.29303133e+00 4.53682542e-01 -9.80285779e-02 -1.46935630e+00
-3.71040732e-01 -5.53921402e-01 -5.92156172e-01 5.82970560e-01
-1.94725263e+00 -7.15563476e-01 -7.91038990e-01 6.56025469e-01
5.37294745e-01 -2.06111550e-01 5.61059177e-01 4.99196291e-01
-3.65484864e-01 7.25682020e-01 -2.80555971e-02 3.68666261e-01
8.00133765e-01 -1.19448578e+00 7.81201303e-01 4.27942216e-01
-4.23159935e-02 2.94972479e-01 6.53765023e-01 -2.34436527e-01
-8.70492578e-01 -8.77045155e-01 8.63424838e-01 -1.42087817e-01
4.96205449e-01 -1.88524589e-01 -5.96054673e-01 6.03332400e-01
1.47145912e-01 2.67807990e-01 2.32410878e-01 2.35251561e-01
-2.49113113e-01 -4.26342607e-01 -1.23587644e+00 6.18912756e-01
1.40125453e+00 -3.36704999e-01 -3.18057775e-01 2.61782080e-01
4.38390046e-01 -1.02861917e+00 -5.59595406e-01 5.87392569e-01
6.82477832e-01 -1.38268733e+00 9.96884942e-01 -4.63731498e-01
6.71374023e-01 -6.83314130e-02 -5.56568317e-02 -1.29935646e+00
-3.70172560e-02 -4.74676371e-01 2.64734983e-01 6.80961311e-01
3.41426492e-01 -7.52960682e-01 1.27455223e+00 5.31440854e-01
1.71182543e-01 -6.86783969e-01 -1.04399562e+00 -7.27104366e-01
-5.70116490e-02 -4.39722449e-01 8.49468470e-01 8.25893879e-01
-4.80969548e-01 8.47809017e-02 -3.08645308e-01 -4.26410474e-02
1.01552570e+00 1.89699486e-01 1.14217222e+00 -1.38118529e+00
-9.16647092e-02 -5.85227966e-01 -6.64211154e-01 -1.27065361e+00
3.29382777e-01 -5.74160278e-01 9.49611962e-02 -1.33527672e+00
4.23414111e-01 -7.97772229e-01 -1.63032502e-01 3.09684068e-01
-2.40656793e-01 3.68216753e-01 -6.79032207e-02 1.43937424e-01
-4.39821810e-01 2.60166079e-01 1.16458488e+00 -2.58090109e-01
-1.34427264e-01 2.06251115e-01 -2.53279865e-01 8.16947103e-01
1.07590711e+00 -3.64120752e-01 -3.67305547e-01 -7.92133987e-01
3.99143606e-01 -3.25107425e-01 3.49515229e-01 -1.30412352e+00
2.75585234e-01 -3.46878380e-01 2.67932087e-01 -5.96911728e-01
1.32709175e-01 -9.62380826e-01 1.73712075e-01 6.81661963e-01
-3.25404376e-01 1.24896318e-01 1.17611229e-01 2.30768919e-01
-4.16713029e-01 -2.90420473e-01 9.59217548e-01 -1.81308687e-01
-8.48819613e-01 6.30128205e-01 1.95063323e-01 3.76480997e-01
8.10707450e-01 -6.56245828e-01 -2.62160897e-02 -4.88554955e-01
-2.90108919e-01 4.46916312e-01 9.13337827e-01 5.04781067e-01
3.87753874e-01 -1.40639496e+00 -6.38879597e-01 2.39522323e-01
3.52226824e-01 5.23623407e-01 -1.37226462e-01 3.72977197e-01
-8.08570206e-01 1.88340858e-01 -4.41260070e-01 -9.60239291e-01
-1.06145799e+00 2.42222950e-01 8.64918113e-01 -1.83066159e-01
-6.40255272e-01 7.15725541e-01 8.66377801e-02 -5.14420331e-01
6.54381573e-01 -4.54434633e-01 -9.92918983e-02 -2.14620933e-01
2.63615966e-01 1.15034245e-01 9.80324000e-02 -3.01765352e-01
-2.00699996e-02 7.41524100e-01 -7.77114406e-02 1.44368619e-01
1.51912999e+00 -5.78613281e-02 7.42776841e-02 4.38670635e-01
1.21066380e+00 -3.68601322e-01 -1.69688690e+00 -5.41581631e-01
2.81661395e-02 -8.62495899e-01 8.58576596e-02 -2.86323607e-01
-1.56450999e+00 6.00065827e-01 8.49738300e-01 -3.32441211e-01
9.04973090e-01 -1.66185945e-01 9.84808087e-01 5.79424441e-01
7.02391207e-01 -1.46561205e+00 6.25042319e-02 6.73106313e-01
4.23223764e-01 -1.73922205e+00 -8.83303359e-02 -3.45087618e-01
-2.04551727e-01 1.11096048e+00 9.03419673e-01 -3.21163297e-01
6.04824007e-01 2.72031814e-01 3.42066497e-01 -9.05424207e-02
-7.37594664e-01 -2.30595887e-01 1.67737260e-01 5.03540516e-01
3.73907715e-01 -1.92284137e-01 3.11897956e-02 -3.97982359e-01
-5.56753635e-01 4.02076453e-01 3.44146401e-01 9.39126313e-01
-3.91338199e-01 -1.17580700e+00 -9.77533013e-02 5.16202569e-01
-2.84159929e-01 -2.67822091e-02 -1.29716933e-01 7.34476626e-01
1.22658461e-01 5.70815921e-01 3.29755455e-01 -2.66480267e-01
7.22668409e-01 -2.88996428e-01 5.45809865e-01 -4.50103879e-01
-6.50777698e-01 -4.06049877e-01 1.56348124e-02 -9.85212982e-01
-5.57261467e-01 -5.45081198e-01 -1.11437297e+00 -6.52485132e-01
-2.25485638e-01 -2.82850534e-01 4.55097824e-01 8.58303964e-01
1.72186002e-01 1.61916390e-01 8.55198324e-01 -1.01649547e+00
-4.59805310e-01 -6.23842359e-01 -1.17579557e-01 7.32406378e-01
4.73893106e-01 -6.80458546e-01 -3.61794472e-01 1.18806604e-02] | [8.755315780639648, -2.2914960384368896] |
c688061e-ffc1-46e8-b9bd-def096f29030 | embeddia-project-cross-lingual-embeddings-for | null | null | https://aclanthology.org/2022.eamt-1.36 | https://aclanthology.org/2022.eamt-1.36.pdf | EMBEDDIA project: Cross-Lingual Embeddings for Less- Represented Languages in European News Media | EMBEDDIA project developed a range of resources and methods for less-resourced EU languages, focusing on applications for media industry, including keyword extraction, comment moderation and article generation. | ['Andraž Pelicon', 'Senja Pollak'] | null | null | null | null | eamt-2022-6 | ['keyword-extraction'] | ['natural-language-processing'] | [-1.83230400e-01 7.58196592e-01 -9.39732611e-01 5.58592498e-01
-8.64805937e-01 -9.54680800e-01 1.17131507e+00 5.87674975e-01
-5.59724569e-01 1.18087947e+00 1.22462726e+00 -7.02180266e-01
5.46722040e-02 -5.65139353e-01 -1.97322704e-02 8.16653967e-02
-8.61050040e-02 2.15921357e-01 -1.91017136e-01 -2.92741686e-01
5.12165427e-01 -3.24418932e-01 -1.20187724e+00 4.11436528e-01
9.94030595e-01 4.63001549e-01 2.39209250e-01 5.75190246e-01
-7.55734921e-01 9.73346710e-01 -1.08882928e+00 -7.22958088e-01
-4.69905853e-01 -3.73311460e-01 -8.17267179e-01 -4.13630717e-02
-1.43724039e-01 3.44992727e-01 -1.27824485e-01 1.11305761e+00
4.81977224e-01 -4.71284151e-01 7.19721735e-01 -1.08069551e+00
-4.21576500e-01 2.03984356e+00 -1.99496582e-01 1.46766290e-01
1.00539994e+00 -5.28561771e-01 1.64767051e+00 -1.07805610e+00
1.76261926e+00 1.05462706e+00 4.46910918e-01 1.08975537e-01
-7.91139424e-01 -7.34268367e-01 -4.50798422e-01 -1.41585693e-01
-1.12325799e+00 -3.62543464e-01 8.76537919e-01 -8.57129157e-01
1.34646809e+00 6.73231959e-01 7.34126627e-01 1.42801023e+00
4.39715892e-01 6.57983005e-01 1.19974613e+00 -7.75398970e-01
-1.94714919e-01 6.49301469e-01 -3.42515588e-01 3.80674452e-01
4.71486837e-01 -4.56319869e-01 -6.92218125e-01 -6.92698002e-01
8.07763338e-02 -6.08951151e-01 -1.25532255e-01 6.66608572e-01
-1.69898188e+00 1.11057651e+00 -2.27582097e-01 7.03889072e-01
-7.68643677e-01 -4.06321585e-01 1.20080173e+00 5.07313251e-01
1.22437203e+00 1.17402935e+00 -6.87862217e-01 -5.21417737e-01
-1.17769098e+00 1.02304924e+00 1.41235065e+00 1.29006195e+00
2.54244328e-01 -1.66363060e-01 -2.60036826e-01 7.25034714e-01
6.24235034e-01 6.43294573e-01 5.08242786e-01 -7.17164099e-01
9.27098513e-01 7.05946803e-01 1.72381118e-01 -1.18995631e+00
-4.76198643e-01 -4.22306657e-01 -6.25869572e-01 -5.67466617e-01
-1.40722975e-01 -1.11092436e+00 -4.31837469e-01 1.10368001e+00
7.50514045e-02 -6.19336545e-01 1.81380615e-01 3.03960592e-01
1.76912272e+00 9.33519244e-01 3.85525227e-02 -8.62793863e-01
1.31546867e+00 -9.86188293e-01 -1.14338505e+00 -1.81564093e-01
7.96785355e-01 -1.52057576e+00 6.06130421e-01 4.17962521e-01
-1.50760913e+00 1.97271496e-01 -9.31138277e-01 -7.34909549e-02
-9.48447287e-01 1.24834113e-01 6.89695179e-01 3.88629198e-01
-6.73216879e-01 1.52584761e-01 -9.87983570e-02 -1.52334034e-01
2.61911273e-01 -2.57193148e-01 -3.14329594e-01 6.56725287e-01
-1.73670292e+00 7.12318242e-01 3.88558537e-01 -5.11981428e-01
3.15077789e-02 -1.26749134e+00 -8.56613040e-01 -3.72796595e-01
3.92313302e-01 -7.05479145e-01 1.42384982e+00 -3.60587209e-01
-1.68472147e+00 1.32597363e+00 -6.00300096e-02 -6.17048264e-01
5.01665115e-01 -3.23226064e-01 -9.07489061e-01 -1.98166952e-01
4.89131778e-01 -3.70504856e-02 7.66517341e-01 -1.16324723e-01
-7.16119945e-01 7.79982880e-02 -4.15448189e-01 -1.00363307e-02
-5.58363855e-01 1.09737933e+00 -3.56134862e-01 -1.12010860e+00
-6.44997418e-01 -6.40135646e-01 -2.93762296e-01 -8.64734769e-01
-1.04291034e+00 -4.37677145e-01 6.69064581e-01 -1.31205213e+00
2.14979672e+00 -1.74021053e+00 5.60082257e-01 3.55158150e-01
5.82251966e-01 -3.64097029e-01 3.42720985e-01 1.10425031e+00
1.25538245e-01 9.04749393e-01 4.95316744e-01 3.70877460e-02
5.90464234e-01 -3.87556106e-01 -5.83874643e-01 2.12710276e-01
-2.52054751e-01 9.10347581e-01 -7.72868633e-01 -7.53889859e-01
-1.73297763e-01 -4.97530624e-02 -4.80019063e-01 -8.75179619e-02
-5.75659394e-01 -1.73608631e-01 -6.26283407e-01 6.23304784e-01
-3.31139900e-02 -1.90567762e-01 4.82837170e-01 1.93890825e-01
-7.79952288e-01 1.25254846e+00 -1.03540528e+00 1.63107741e+00
-7.43631840e-01 1.09351575e+00 3.35841268e-01 9.28741843e-02
8.66739452e-01 5.55564702e-01 5.29619336e-01 -4.03073967e-01
3.70587558e-01 6.47317111e-01 -1.22066438e-02 -1.69817239e-01
1.06986594e+00 4.02259380e-01 -6.99941814e-01 5.40341318e-01
2.01077208e-01 -4.36219662e-01 7.98595905e-01 7.18736529e-01
8.57814312e-01 -3.20118606e-01 1.08985877e+00 -5.62518299e-01
4.87825751e-01 3.40701789e-01 1.17277570e-01 2.85185903e-01
4.98026907e-01 1.00396775e-01 6.98250413e-01 8.92793238e-02
-1.31284249e+00 -2.48715222e-01 -6.14990950e-01 1.00860906e+00
-5.95547259e-01 -1.66630816e+00 -9.46193099e-01 -4.32796866e-01
8.30839500e-02 6.74093723e-01 -5.68332314e-01 7.37292826e-01
-5.68592072e-01 -3.60239416e-01 3.74413669e-01 -3.95823538e-01
6.00795820e-02 -1.49440300e+00 -3.76821578e-01 5.98458469e-01
-3.23079973e-01 -9.58745480e-01 -4.98239577e-01 2.48639554e-01
-2.35743925e-01 -6.89941883e-01 -1.05821609e+00 -4.32964534e-01
6.31190315e-02 -6.26268864e-01 1.65818405e+00 -1.50595263e-01
-2.36090943e-01 -8.66327658e-02 -4.59703505e-01 -9.73675370e-01
-9.54334855e-01 7.54004240e-01 -1.92507997e-01 -1.00518715e+00
7.03606680e-02 -1.79994762e-01 3.18861566e-02 -3.58531415e-01
-6.53634310e-01 -1.38544038e-01 2.04599306e-01 7.78101742e-01
7.01762810e-02 -2.66362906e-01 7.53038406e-01 -1.72501230e+00
1.46646249e+00 -7.83993185e-01 -6.27345264e-01 1.37772396e-01
-9.07387912e-01 -2.65273452e-01 1.94767773e-01 -4.23973590e-01
-1.01910162e+00 -9.67518628e-01 -3.24753791e-01 8.74693036e-01
4.13420081e-01 1.61586344e+00 3.90841141e-02 4.55080420e-01
7.53019929e-01 -5.05227089e-01 -1.28570989e-01 -4.99553591e-01
7.57653117e-01 1.06233585e+00 1.48866311e-01 -2.64654279e-01
5.47679842e-01 -1.26207545e-01 -6.16741657e-01 -9.90639329e-01
-3.46418023e-01 -5.08559465e-01 -8.93318839e-03 -4.37597781e-01
4.60974693e-01 -1.27212870e+00 -1.15124352e-01 2.05233872e-01
-1.31443775e+00 1.40958861e-01 -6.28639460e-01 3.22725028e-01
3.39947678e-02 -1.26814738e-01 -7.49033570e-01 -2.86632925e-01
-9.98755634e-01 -6.63348556e-01 6.32110775e-01 2.92147249e-01
-1.38389564e+00 -1.40933836e+00 7.17213929e-01 -1.77924260e-01
3.30883533e-01 4.88134742e-01 7.39515126e-01 -7.35351264e-01
2.94329733e-01 -7.39434436e-02 1.54681727e-01 -2.83534497e-01
-2.05693051e-01 6.12639010e-01 -1.72210082e-01 9.38083977e-02
-6.13017619e-01 -9.89930779e-02 4.91108298e-01 6.27047122e-01
5.07732689e-01 -1.18303180e+00 -7.98929811e-01 1.61112428e-01
9.17687178e-01 -2.99874067e-01 6.99573934e-01 9.83770669e-01
3.47100496e-01 4.07722682e-01 5.80528259e-01 1.02691042e+00
2.75731593e-01 5.73748589e-01 -1.16240099e-01 -3.55251506e-02
-6.31832480e-02 -4.11991119e-01 7.97087491e-01 1.68195760e+00
-9.23193172e-02 -5.42769313e-01 -8.15551281e-01 7.83536375e-01
-1.64644766e+00 -8.90157878e-01 -6.18783832e-01 1.05688465e+00
1.34698057e+00 3.68086547e-01 5.17355680e-01 -1.47423558e-02
3.76260519e-01 5.99997342e-01 2.42246151e-01 -8.61930072e-01
-4.60604727e-01 5.36507428e-01 6.99492455e-01 2.64614165e-01
-6.89629078e-01 1.06086195e+00 7.83036089e+00 1.06158543e+00
-6.77015483e-01 2.81369686e-01 1.21383682e-01 3.66158843e-01
-1.11524737e+00 9.30321962e-02 -1.15077031e+00 2.81800717e-01
1.32766092e+00 -1.25303090e+00 -1.94045737e-01 1.28165877e+00
2.80171335e-01 1.01395033e-01 -4.08266664e-01 5.09750426e-01
-1.01628363e-01 -2.48151612e+00 -2.92842478e-01 3.60895365e-01
1.25497222e+00 5.03368080e-01 -3.56580913e-02 3.04284573e-01
6.83226347e-01 -6.60155118e-01 1.06766737e+00 2.36746892e-01
7.45690942e-01 -6.87290370e-01 7.98743486e-01 -1.37349188e-01
-7.91332304e-01 1.84963629e-01 -3.17740887e-01 -4.03034613e-02
6.36852443e-01 1.15831828e+00 -8.93513322e-01 6.97498798e-01
5.19340038e-01 1.27053797e+00 -5.24691105e-01 8.53291988e-01
-5.60765803e-01 8.12363982e-01 -2.94938684e-02 -8.97663891e-01
3.33332628e-01 -2.84038126e-01 1.29752636e+00 1.95542550e+00
4.60047871e-01 -3.66258830e-01 2.01980472e-01 6.64354146e-01
-5.58587849e-01 9.54940200e-01 -5.76015294e-01 -9.25696611e-01
7.29269385e-01 1.53212774e+00 -6.70410872e-01 -3.86387140e-01
-4.68581140e-01 5.06013095e-01 -1.63036391e-01 2.78133541e-01
-4.24263000e-01 -8.62532675e-01 4.51286912e-01 3.10234874e-01
2.53398508e-01 -3.16087186e-01 -2.82893330e-01 -9.78090167e-01
-3.87048692e-01 -1.43371975e+00 6.25146747e-01 -4.86137301e-01
-1.07725883e+00 8.06437314e-01 -7.03005567e-02 -1.08886909e+00
-1.02808213e+00 -4.22583669e-01 -6.09778285e-01 8.55369091e-01
-8.98552954e-01 -1.06486154e+00 6.19718075e-01 2.49840811e-01
5.37382066e-01 -5.28519452e-01 9.18765724e-01 5.80355763e-01
-1.29770562e-01 2.33566239e-01 1.60480291e-02 -3.53749283e-02
9.88887668e-01 -1.48816645e+00 7.28371382e-01 4.50479865e-01
3.26272249e-01 5.29635370e-01 1.25779128e+00 -9.62450027e-01
-1.20339715e+00 -7.44861305e-01 1.74307895e+00 -3.31785351e-01
1.79437053e+00 -3.45693350e-01 -1.27804533e-01 4.15378362e-01
1.08304060e+00 -1.17035532e+00 6.83449388e-01 6.66704297e-01
-1.19009547e-01 6.26500368e-01 -5.72418392e-01 9.39991176e-01
5.78322053e-01 -6.28310561e-01 -7.02523291e-01 7.20180690e-01
9.57162678e-01 -8.06926012e-01 -1.33126187e+00 -9.01253819e-02
3.77918094e-01 -9.67083424e-02 6.81998909e-01 -4.24110055e-01
7.96948671e-01 3.17272209e-02 4.03607845e-01 -1.54206431e+00
-1.54388681e-01 -1.61627913e+00 -2.77565807e-01 1.83160484e+00
1.63434577e+00 -4.71356630e-01 1.21275410e-01 8.75641406e-02
-2.41479576e-01 -1.97436623e-02 -5.67731082e-01 -2.61529058e-01
1.50355563e-01 -5.36171436e-01 3.60640824e-01 1.06596720e+00
1.04693580e+00 7.85276115e-01 -3.16372246e-01 -7.31985211e-01
-5.53597771e-02 1.00517787e-01 7.61303008e-01 -1.18360722e+00
-3.43886483e-03 -6.03288174e-01 -7.11835250e-02 -5.10891140e-01
3.96638036e-01 -9.26342785e-01 -2.22716913e-01 -1.51722705e+00
3.15611273e-01 -1.17122941e-01 5.54877996e-01 4.18930799e-02
4.16011035e-01 7.99836516e-02 -3.45884264e-01 1.93984464e-01
-3.99635196e-01 3.86040024e-02 9.70766246e-01 -2.14362919e-01
-3.75591427e-01 -1.80296041e-02 -8.67245018e-01 8.01746070e-01
4.00909096e-01 -6.23244524e-01 6.89255521e-02 3.41103941e-01
1.55232656e+00 -3.37577313e-01 -1.66904643e-01 -2.39581108e-01
8.55399445e-02 -1.83555886e-01 -2.65887648e-01 -8.43745410e-01
-5.17787278e-01 -2.48384371e-01 3.91570002e-01 1.07137427e-01
-7.20140040e-01 7.12241828e-01 2.24345833e-01 2.18572989e-02
-5.43482065e-01 -5.75731456e-01 2.50104610e-02 -2.76787251e-01
-2.86788553e-01 4.15698811e-02 -1.21160662e+00 7.35708356e-01
4.40780282e-01 3.97731543e-01 -5.23401856e-01 -6.61878228e-01
-5.16765594e-01 -1.03342950e-01 2.58049130e-01 5.08054197e-01
3.47153723e-01 -1.20537603e+00 -1.32914424e+00 -5.25278568e-01
1.25366807e-01 -8.12859654e-01 -4.64901626e-01 7.36530423e-01
-4.93125200e-01 6.43703282e-01 6.14281408e-02 1.29567653e-01
-1.39655161e+00 1.90255251e-02 -4.66997445e-01 -9.48027551e-01
-9.55351770e-01 4.90451425e-01 -8.66711378e-01 -3.42389733e-01
8.74594152e-02 -5.71322203e-01 -7.13461339e-01 7.06318378e-01
9.00467098e-01 4.50724393e-01 -3.06758564e-02 -7.15905845e-01
-2.40761057e-01 -1.33151725e-01 -3.27863321e-02 -7.99565494e-01
1.36616325e+00 -4.09001231e-01 -7.19955325e-01 7.58142471e-01
9.51443493e-01 9.34589982e-01 2.31821127e-02 4.18258505e-03
7.96270370e-01 4.71675068e-01 4.76124972e-01 -8.80070329e-01
-4.05050039e-01 -1.84555687e-02 -4.65091348e-01 7.98308671e-01
4.38652188e-01 3.45736414e-01 6.54444933e-01 1.88247964e-01
-2.18737721e-01 -1.77204418e+00 -4.90504831e-01 7.65701115e-01
1.35576284e+00 -7.59266913e-01 7.34214187e-01 -4.85102594e-01
-6.44830287e-01 1.22934115e+00 -1.69277757e-01 2.13835761e-01
1.26475513e+00 7.89602160e-01 2.61367142e-01 -7.57541955e-01
-9.27205920e-01 2.44260713e-01 9.29890096e-01 2.46993020e-01
1.05863380e+00 2.44925857e-01 -1.28220642e+00 9.09137249e-01
-8.63427997e-01 -4.62510765e-01 8.14010143e-01 5.64130187e-01
6.47628531e-02 -1.44272268e+00 -1.03930250e-01 3.95758510e-01
-1.23974109e+00 -6.24933243e-01 -8.75678301e-01 1.18007720e+00
-1.80617034e-01 8.00675988e-01 -8.42592940e-02 -1.78915620e-01
6.24455027e-02 -2.79089063e-01 3.50146592e-02 -8.11283708e-01
-1.20340693e+00 5.16207218e-01 1.32446945e+00 -1.64181992e-01
-9.41081822e-01 -1.17143726e+00 -8.05507779e-01 -4.67177689e-01
-5.38970172e-01 6.42343700e-01 9.44195986e-01 8.77894044e-01
2.99465090e-01 7.60623753e-01 4.12131906e-01 -5.25499105e-01
3.11959445e-01 -1.47111976e+00 -7.57006586e-01 -3.42470258e-01
-1.55738309e-01 -1.23033814e-01 -6.16349101e-01 2.28706360e-01] | [12.035922050476074, 9.493075370788574] |
1077f29b-d368-42f9-8c5f-62a4f78f9fd0 | efficient-search-of-comprehensively-robust | 2305.07308 | null | https://arxiv.org/abs/2305.07308v1 | https://arxiv.org/pdf/2305.07308v1.pdf | Efficient Search of Comprehensively Robust Neural Architectures via Multi-fidelity Evaluation | Neural architecture search (NAS) has emerged as one successful technique to find robust deep neural network (DNN) architectures. However, most existing robustness evaluations in NAS only consider $l_{\infty}$ norm-based adversarial noises. In order to improve the robustness of DNN models against multiple types of noises, it is necessary to consider a comprehensive evaluation in NAS for robust architectures. But with the increasing number of types of robustness evaluations, it also becomes more time-consuming to find comprehensively robust architectures. To alleviate this problem, we propose a novel efficient search of comprehensively robust neural architectures via multi-fidelity evaluation (ES-CRNA-ME). Specifically, we first search for comprehensively robust architectures under multiple types of evaluations using the weight-sharing-based NAS method, including different $l_{p}$ norm attacks, semantic adversarial attacks, and composite adversarial attacks. In addition, we reduce the number of robustness evaluations by the correlation analysis, which can incorporate similar evaluations and decrease the evaluation cost. Finally, we propose a multi-fidelity online surrogate during optimization to further decrease the search cost. On the basis of the surrogate constructed by low-fidelity data, the online high-fidelity data is utilized to finetune the surrogate. Experiments on CIFAR10 and CIFAR100 datasets show the effectiveness of our proposed method. | ['Xiaoqian Chen', 'Tingsong Jiang', 'Wen Yao', 'Jialiang Sun'] | 2023-05-12 | null | null | null | null | ['architecture-search'] | ['methodology'] | [-2.39863724e-01 -6.52608871e-01 3.63876313e-01 -4.60565597e-01
-9.09636259e-01 -6.21179700e-01 5.04230969e-02 -3.60766262e-01
-6.70388758e-01 5.43935478e-01 8.03576317e-03 -2.90927202e-01
-4.91502881e-01 -6.68077290e-01 -6.78702593e-01 -8.75540435e-01
2.99487174e-01 -1.78716853e-01 2.29350224e-01 -3.08777690e-01
7.59208649e-02 6.09674573e-01 -1.18684530e+00 -2.41067499e-01
8.29883873e-01 1.32738376e+00 7.27737173e-02 1.09662727e-01
1.15667626e-01 3.19536269e-01 -8.17460060e-01 -5.22871673e-01
6.77909315e-01 -2.21271634e-01 -2.64600009e-01 -4.55034554e-01
3.42365861e-01 -3.73736054e-01 -7.89921582e-01 1.64888799e+00
9.50387657e-01 4.68621939e-01 1.96769565e-01 -1.20779061e+00
-5.49521208e-01 6.50321960e-01 -3.64730626e-01 4.56147999e-01
-3.06846440e-01 4.58544850e-01 7.92474985e-01 -8.39824140e-01
3.07759512e-02 1.34610927e+00 6.14258170e-01 5.83307743e-01
-8.88065159e-01 -1.27431226e+00 6.67030215e-01 2.61919975e-01
-1.51649928e+00 -5.53612471e-01 1.05237806e+00 -4.33169827e-02
5.11548579e-01 1.90901354e-01 1.43784970e-01 9.30346191e-01
-1.86323255e-01 5.23191154e-01 8.37927222e-01 7.47836474e-03
2.89231628e-01 -2.46588558e-01 1.71725228e-01 6.15305662e-01
3.48697484e-01 3.87889057e-01 -9.08812881e-03 -8.12560990e-02
7.77326882e-01 1.41997680e-01 -4.31965739e-01 1.25942796e-01
-9.03189182e-01 6.86958730e-01 7.00396836e-01 1.43319800e-01
-2.03094199e-01 1.74319595e-01 5.62132001e-01 3.02053064e-01
5.30053191e-02 4.11575019e-01 -4.84142542e-01 1.02853149e-01
-7.49221742e-01 1.14042111e-01 2.17868760e-01 7.56209850e-01
5.48359573e-01 6.74116135e-01 -2.61648834e-01 1.06233227e+00
5.89596212e-01 5.37729383e-01 5.18345177e-01 -8.79180372e-01
8.18024099e-01 5.71369052e-01 -2.37271890e-01 -1.32976353e+00
-3.75136554e-01 -8.57925057e-01 -1.26057243e+00 2.12136984e-01
1.84089437e-01 -3.78965080e-01 -8.61529648e-01 2.22642374e+00
2.12511480e-01 3.07388425e-01 8.06626379e-02 1.17993689e+00
7.87208200e-01 6.02640331e-01 -1.14141308e-01 -3.25439692e-01
1.00805330e+00 -8.03336859e-01 -5.97673595e-01 4.54113260e-02
1.74713790e-01 -4.96039003e-01 1.02640533e+00 2.64275700e-01
-8.57379258e-01 -6.15742624e-01 -1.26117909e+00 2.53944635e-01
-1.30676821e-01 1.28459528e-01 1.79210454e-01 8.49002540e-01
-5.62128305e-01 4.48683351e-01 -6.65812492e-01 4.75115269e-01
6.09723389e-01 5.10101914e-01 -2.06441030e-01 -7.85542428e-02
-1.54565096e+00 5.96107006e-01 4.92690533e-01 6.25104547e-01
-1.05451691e+00 -7.26179004e-01 -5.91569364e-01 8.08423236e-02
3.54036152e-01 -2.98803717e-01 7.63001382e-01 -7.03869224e-01
-1.52025771e+00 2.59797610e-02 2.60732383e-01 -3.92319620e-01
4.66180414e-01 -4.81511019e-02 -8.59954536e-01 3.36937867e-02
-3.36228043e-01 3.33539367e-01 7.97328472e-01 -1.04856896e+00
-2.15693772e-01 -4.12168771e-01 1.38925567e-01 4.12544161e-02
-6.31880760e-01 1.66082025e-01 -5.32926440e-01 -1.05991852e+00
1.90229699e-01 -8.51578891e-01 -4.45317924e-01 -7.00457469e-02
-4.28830922e-01 5.47987781e-02 5.27477086e-01 -7.32867837e-01
1.44599438e+00 -2.33784652e+00 1.04874223e-01 7.15536296e-01
1.92340568e-01 6.85983360e-01 -4.66532290e-01 -3.40046853e-01
-4.81582023e-02 2.64162272e-01 -2.82473445e-01 -6.58580735e-02
2.02179313e-01 1.46517649e-01 -3.10304016e-01 2.56084114e-01
1.37709919e-02 6.40062809e-01 -5.28317332e-01 -2.69326359e-01
5.00403941e-02 5.82383692e-01 -4.30285782e-01 -5.21750655e-03
6.24564253e-02 2.08084077e-01 -7.43179500e-01 5.60072064e-01
8.86304498e-01 1.09736182e-01 -2.70787477e-01 -6.22742593e-01
2.05756146e-02 -1.79029897e-01 -1.55333781e+00 1.41648161e+00
-3.16349268e-01 1.13767095e-01 1.47847220e-01 -9.57771838e-01
1.12907267e+00 2.54265070e-01 1.79890543e-01 -7.86584914e-01
4.47557420e-01 3.27157736e-01 1.36772394e-01 -1.28829822e-01
-7.38439858e-02 -2.86722984e-02 -1.38034165e-01 1.97665900e-01
-1.13178320e-01 3.97068024e-01 -3.17072198e-02 -1.36573017e-01
1.19649589e+00 -2.38948509e-01 -2.36267060e-01 -1.48318484e-01
8.57517958e-01 -4.84986454e-01 1.13712525e+00 5.87633729e-01
-4.77302641e-01 6.50284588e-01 2.34039798e-01 -4.73349243e-01
-7.78469682e-01 -1.02464890e+00 -2.01356784e-02 7.39411473e-01
1.60133630e-01 2.07776111e-02 -7.51633286e-01 -7.36991227e-01
-2.29406416e-01 4.82975006e-01 -3.13261658e-01 -6.62652433e-01
-7.07510769e-01 -1.03262770e+00 9.72281456e-01 5.20588636e-01
9.98856843e-01 -8.95563662e-01 -2.02627853e-01 1.06382690e-01
-2.00771168e-02 -8.75573993e-01 -6.98051155e-01 -6.20365478e-02
-8.03083301e-01 -7.76710570e-01 -7.68061876e-01 -7.21519411e-01
7.67431676e-01 8.20370317e-02 5.13467312e-01 2.44778991e-01
-1.04286313e-01 -1.47362560e-01 -1.94607049e-01 -1.70861065e-01
-3.98375057e-02 -1.64819762e-01 4.93476629e-01 1.41849115e-01
-1.50343895e-01 -6.22478306e-01 -6.76503003e-01 4.85608160e-01
-1.03606009e+00 -5.93768239e-01 6.95012033e-01 8.79568398e-01
8.70681524e-01 5.63938379e-01 9.09217954e-01 -3.16805124e-01
8.69298637e-01 -3.76439482e-01 -9.29747224e-01 4.73623395e-01
-6.01930261e-01 2.54077852e-01 1.13493156e+00 -8.11634958e-01
-8.64298940e-01 -1.20080471e-01 -4.01089907e-01 -1.22187519e+00
1.80052072e-01 5.82189202e-01 -7.01699913e-01 -3.79755110e-01
5.87494314e-01 2.48395517e-01 -2.60383755e-01 -6.62795782e-01
2.81337738e-01 2.41776720e-01 6.00550413e-01 -5.28115749e-01
1.14701843e+00 -7.55022094e-02 9.90988463e-02 -2.96138287e-01
-6.32204115e-01 8.38781297e-02 -2.77223042e-03 -3.28650996e-02
5.72083712e-01 -7.25988925e-01 -8.81728172e-01 5.72345018e-01
-8.83048236e-01 7.97547996e-02 1.26146063e-01 6.65865779e-01
7.13858604e-02 4.29740906e-01 -4.76548851e-01 -7.04096079e-01
-7.34769106e-01 -1.51285195e+00 2.66298652e-01 4.55555439e-01
4.79357004e-01 -5.91021299e-01 -1.64097264e-01 2.31533334e-01
6.29012585e-01 2.72947252e-01 8.32587481e-01 -6.89099133e-01
-5.51781178e-01 -3.48857105e-01 -3.51761967e-01 7.47713506e-01
3.54325175e-02 -1.51160613e-01 -7.01576114e-01 -2.89374948e-01
3.00016373e-01 -1.61814049e-01 7.68121183e-01 2.16466472e-01
1.47443330e+00 -7.47610986e-01 5.58968149e-02 9.16891038e-01
1.38184392e+00 4.73862857e-01 6.57380223e-01 3.96442056e-01
7.44844198e-01 1.23187549e-01 4.10022944e-01 3.43849361e-01
-2.17028949e-02 4.92911339e-01 4.41805571e-01 2.45122671e-01
1.00487940e-01 -1.26360178e-01 3.68164510e-01 9.76129949e-01
2.55197822e-03 -5.72304614e-02 -6.72334611e-01 2.83733785e-01
-1.63969219e+00 -9.01832163e-01 3.99608701e-01 2.22537518e+00
8.91822815e-01 3.08984429e-01 -1.08494505e-01 2.81711340e-01
8.91980648e-01 3.99157926e-02 -9.62723613e-01 8.49733502e-03
-2.70250767e-01 1.00255862e-01 3.38013738e-01 1.82843223e-01
-9.18668449e-01 5.08812666e-01 5.24896908e+00 1.30043948e+00
-1.21307230e+00 1.19232565e-01 7.11610734e-01 -3.39161009e-01
-4.24390644e-01 -1.41017348e-01 -7.76825249e-01 8.47787261e-01
5.39959967e-01 -6.43380955e-02 6.81092978e-01 9.15840447e-01
3.40998955e-02 5.96036613e-01 -6.47104561e-01 1.16714823e+00
-4.39948887e-02 -1.21654332e+00 7.33621642e-02 -3.15285832e-01
7.04375684e-01 3.67841050e-02 2.37243548e-01 4.74673867e-01
1.72550544e-01 -8.79365861e-01 5.46464384e-01 5.53545892e-01
5.57904303e-01 -1.07124257e+00 7.36724675e-01 1.61779761e-01
-1.37017035e+00 -4.09316361e-01 -5.55720508e-01 5.63542724e-01
-6.40746579e-02 5.84564805e-01 -4.54420829e-03 6.32698059e-01
7.55285621e-01 2.16950655e-01 -5.33569455e-01 1.19884491e+00
-1.27905399e-01 4.62301821e-01 -4.97833431e-01 -4.95574176e-02
1.01921842e-01 -1.31557032e-01 6.43860042e-01 6.96270108e-01
4.10281092e-01 3.49666238e-01 1.69560686e-01 8.54052246e-01
-4.21191990e-01 -1.84931699e-02 -1.15171179e-01 1.24033049e-01
8.67682040e-01 1.15841866e+00 -3.71703148e-01 5.07810004e-02
-9.50240344e-02 5.47346294e-01 2.39555776e-01 4.66420263e-01
-1.00832784e+00 -7.10811436e-01 6.29155219e-01 -4.45087790e-01
8.84683654e-02 -1.43629029e-01 -2.16832355e-01 -1.04068065e+00
2.81759501e-01 -1.18601859e+00 4.10926640e-01 -5.01566648e-01
-1.50239158e+00 8.91684175e-01 -1.70436114e-01 -1.30315673e+00
2.97146201e-01 -2.87057966e-01 -7.41722822e-01 9.21016514e-01
-1.44939339e+00 -8.63957584e-01 -2.08408833e-01 8.85264874e-01
2.63473690e-01 -5.58955491e-01 5.26802838e-01 7.57130861e-01
-1.21214879e+00 1.33285785e+00 1.39237836e-01 3.27778786e-01
3.82888287e-01 -4.53312069e-01 2.43778601e-01 1.22321367e+00
-5.17176315e-02 9.27514374e-01 4.11290795e-01 -4.14259493e-01
-1.34448969e+00 -1.31736672e+00 1.26429617e-01 9.95958522e-02
5.61044276e-01 2.39278991e-02 -1.12282217e+00 2.86747575e-01
-3.30004543e-01 3.06912959e-01 3.16369832e-01 -3.06042820e-01
-6.54115856e-01 -7.58945405e-01 -1.23098242e+00 8.96554649e-01
1.04256451e+00 -3.69659752e-01 -3.81599486e-01 2.49492340e-02
1.09878290e+00 -2.31004849e-01 -9.47985053e-01 9.00607705e-01
4.23014402e-01 -5.13849318e-01 1.16523147e+00 -3.95705014e-01
-7.20576122e-02 -6.06252968e-01 -3.80755186e-01 -1.19790721e+00
-2.81197578e-01 -5.92783272e-01 -1.09656365e-03 1.38222539e+00
4.57801908e-01 -8.71536314e-01 5.72265267e-01 7.92705417e-01
-3.38606477e-01 -9.40534234e-01 -1.18497062e+00 -1.01023102e+00
1.07727759e-03 -4.04535502e-01 9.07073379e-01 7.25919783e-01
-5.35290182e-01 -4.40484509e-02 -2.15560660e-01 4.31118011e-01
6.95302427e-01 -1.70309067e-01 4.39062804e-01 -9.59290326e-01
-3.51588577e-01 -7.91200519e-01 -4.09066796e-01 -7.63286054e-01
2.12550953e-01 -6.30784750e-01 1.20125949e-01 -1.12007010e+00
-2.06117451e-01 -8.09443474e-01 -1.02809715e+00 6.29684567e-01
-3.09552312e-01 1.19082620e-02 2.39466146e-01 2.66480267e-01
-5.85952461e-01 8.83145928e-01 9.98014688e-01 -1.71434879e-01
-7.00742230e-02 7.01798350e-02 -7.22455382e-01 6.83757603e-01
9.58426774e-01 -5.01057744e-01 -5.34849286e-01 -6.38098419e-01
2.38400698e-01 -8.58312249e-02 3.12673777e-01 -1.19287324e+00
4.00305957e-01 -1.51130423e-01 3.92812401e-01 -4.88915294e-01
3.35280806e-01 -9.00207520e-01 1.85251057e-01 4.57285583e-01
-3.96228999e-01 2.09474534e-01 3.80969286e-01 5.79448223e-01
-1.98657647e-01 -2.78060377e-01 9.38495934e-01 6.35815635e-02
-5.40397346e-01 6.83888257e-01 2.88918942e-01 1.51342332e-01
6.86946273e-01 4.42712419e-02 -4.74381536e-01 7.05627352e-02
-4.79225129e-01 4.32996511e-01 7.07144514e-02 4.33233172e-01
9.24336553e-01 -1.55452180e+00 -6.55549765e-01 2.80960143e-01
-1.60786077e-01 9.40071195e-02 6.88979805e-01 5.92486978e-01
-2.57092237e-01 4.25081030e-02 -2.67450780e-01 -1.49305195e-01
-9.59476411e-01 5.79297900e-01 5.93813837e-01 -1.58205464e-01
-2.28452578e-01 1.06413293e+00 1.58878699e-01 -3.91134828e-01
5.29406309e-01 -3.15561034e-02 -2.41661370e-01 -3.33904564e-01
5.31621933e-01 5.44050992e-01 1.89126074e-01 -4.44420993e-01
-5.70797443e-01 6.23718023e-01 -3.05611752e-02 -1.84076861e-01
1.17619622e+00 -4.75126803e-02 -4.92244735e-02 6.72643483e-02
1.29993713e+00 -1.26772642e-01 -1.40588081e+00 -3.86947751e-01
-3.09791058e-01 -3.14175993e-01 2.48469964e-01 -7.11316049e-01
-1.75032640e+00 7.69798279e-01 8.55239213e-01 -1.97710410e-01
1.54437017e+00 -6.47251308e-01 9.51066554e-01 5.68751037e-01
3.07486087e-01 -1.14404142e+00 1.12163216e-01 4.90629673e-01
9.09157097e-01 -1.01192522e+00 -1.31265819e-01 9.92096066e-02
-4.72829282e-01 8.87641191e-01 8.48964095e-01 -2.39123121e-01
8.66142571e-01 1.07845403e-01 2.66024359e-02 2.18138158e-01
-4.39415604e-01 2.65417546e-01 3.42364222e-01 2.92937189e-01
-3.04481953e-01 -1.83130860e-01 -2.52272069e-01 1.11146259e+00
-2.89876089e-02 -3.29180062e-01 1.11730797e-02 5.41397929e-01
-3.04660887e-01 -9.57726896e-01 -3.52751553e-01 3.89555365e-01
-6.06673121e-01 -2.06012905e-01 7.28750825e-02 3.47427160e-01
1.64800838e-01 1.08601332e+00 -3.64480704e-01 -8.32324088e-01
5.21598816e-01 -2.27741241e-01 1.70652211e-01 -3.62128466e-02
-6.26946330e-01 -1.21806435e-01 -2.58425653e-01 -4.24609125e-01
-4.80824970e-02 -2.80583888e-01 -1.41527891e+00 -2.20159724e-01
-5.71473002e-01 1.31353825e-01 7.12780952e-01 9.15499985e-01
3.75376195e-01 5.59152126e-01 1.04011571e+00 -4.14995790e-01
-1.07808328e+00 -7.37450540e-01 -2.49208853e-01 3.08049828e-01
1.86242327e-01 -6.75406635e-01 -5.41049123e-01 -4.11238968e-01] | [5.547145366668701, 7.9163594245910645] |
a68b8cf5-259c-44a3-b94d-9acdb55d8b88 | robust-regression-via-model-based-methods | 2106.10759 | null | https://arxiv.org/abs/2106.10759v4 | https://arxiv.org/pdf/2106.10759v4.pdf | Robust Regression via Model Based Methods | The mean squared error loss is widely used in many applications, including auto-encoders, multi-target regression, and matrix factorization, to name a few. Despite computational advantages due to its differentiability, it is not robust to outliers. In contrast, l_p norms are known to be robust, but cannot be optimized via, e.g., stochastic gradient descent, as they are non-differentiable. We propose an algorithm inspired by so-called model-based optimization (MBO) [35, 36], which replaces a non-convex objective with a convex model function and alternates between optimizing the model function and updating the solution. We apply this to robust regression, proposing SADM, a stochastic variant of the Online Alternating Direction Method of Multipliers (OADM) [50] to solve the inner optimization in MBO. We show that SADM converges with the rate O(log T/T). Finally, we demonstrate experimentally (a) the robustness of l_p norms to outliers and (b) the efficiency of our proposed model-based algorithms in comparison with gradient methods on autoencoders and multi-target regression. | ['Edmund Yeh', 'Stratis Ioannidis', 'Khashayar Kamran', 'Armin Moharrer'] | 2021-06-20 | null | null | null | null | ['multi-target-regression'] | ['miscellaneous'] | [ 2.82308497e-02 -1.37071595e-01 1.22132279e-01 -3.82837087e-01
-1.05136597e+00 -2.84682900e-01 2.13758603e-01 1.96059495e-01
-6.13554120e-01 8.32327902e-01 -3.43801305e-02 -1.55320168e-01
-3.90786767e-01 -3.83307785e-01 -1.09204650e+00 -7.24056005e-01
-4.32622619e-02 3.64974529e-01 -2.05358908e-01 -1.88788146e-01
2.78285086e-01 4.97816086e-01 -1.32809412e+00 -2.97724247e-01
1.11055350e+00 1.21095288e+00 -1.10247627e-01 5.74861169e-01
2.62685061e-01 6.47534847e-01 -3.26273143e-01 -6.02501273e-01
3.65073472e-01 -4.05626565e-01 -4.51192886e-01 3.05027485e-01
5.91521323e-01 -1.45264670e-01 -2.72353828e-01 1.14575064e+00
4.16898727e-01 5.68731189e-01 6.21528447e-01 -1.40476024e+00
-6.32674515e-01 2.82786757e-01 -6.68024182e-01 4.42796312e-02
1.35131821e-01 -3.01291496e-01 1.06085229e+00 -1.45807362e+00
3.35275739e-01 1.19436467e+00 9.00865853e-01 3.44658703e-01
-1.31124520e+00 -5.09662390e-01 2.12487444e-01 8.54670107e-02
-1.37536573e+00 -5.50501347e-01 5.98922908e-01 -5.63313782e-01
6.40316010e-01 3.07669193e-01 2.39211082e-01 7.99170911e-01
1.18757337e-01 8.82889628e-01 1.00338256e+00 -3.04592848e-01
2.63608247e-01 2.30660602e-01 -8.42015743e-02 8.67795885e-01
2.95446366e-01 8.86297897e-02 -7.49776721e-01 -2.76458025e-01
4.96116400e-01 2.29398720e-02 -3.35587531e-01 -3.54054272e-01
-1.22300780e+00 1.06705034e+00 2.39608303e-01 8.27634409e-02
-4.77149010e-01 3.61615032e-01 3.99877578e-01 5.27607203e-01
8.38272154e-01 3.03817749e-01 -5.88784397e-01 -1.20596290e-01
-1.08967102e+00 2.61517465e-01 8.62626731e-01 7.51275897e-01
7.93508351e-01 4.24232721e-01 8.90963003e-02 9.01523530e-01
5.23226976e-01 4.97321844e-01 7.26240933e-01 -8.76560390e-01
6.46762729e-01 1.84747815e-01 2.44825333e-01 -1.19268131e+00
-4.45892543e-01 -5.29692054e-01 -1.07491791e+00 2.18038112e-01
4.57778990e-01 -3.09458137e-01 -3.55006307e-01 1.77475333e+00
4.00311023e-01 3.29148531e-01 -5.09071015e-02 8.49591970e-01
2.32437566e-01 5.25414467e-01 -4.12788332e-01 -5.24077177e-01
6.66193485e-01 -1.11143064e+00 -7.62205601e-01 -2.54288524e-01
5.06590188e-01 -7.67438650e-01 7.74988592e-01 6.93381310e-01
-1.14762402e+00 -3.72986227e-01 -1.03142881e+00 1.37558319e-02
-1.57942459e-01 1.99485585e-01 3.94067913e-01 5.34215748e-01
-8.97535443e-01 1.01727223e+00 -9.97716665e-01 -1.10750712e-01
2.09523082e-01 4.40479338e-01 -4.66858774e-01 1.77527532e-01
-9.69264984e-01 9.03867006e-01 2.50631571e-01 3.32549602e-01
-8.89143109e-01 -7.02491701e-01 -9.84741390e-01 -5.83748296e-02
5.50774574e-01 -4.74973828e-01 9.54816282e-01 -1.02838492e+00
-1.87966597e+00 5.13089001e-01 -4.36726481e-01 -6.28642201e-01
8.20744574e-01 -7.61946261e-01 -3.11729997e-01 -1.88767031e-01
-5.72450040e-03 7.62676299e-02 1.47771347e+00 -9.21839297e-01
-5.88350594e-01 -3.50759983e-01 -1.43158495e-01 2.57806852e-02
-5.89987218e-01 7.50556141e-02 -1.28685713e-01 -8.56764615e-01
3.27316880e-01 -9.62420225e-01 -4.58329767e-01 1.36086598e-01
-4.63528603e-01 6.99775070e-02 6.00308001e-01 -8.98175955e-01
1.40983951e+00 -2.08625078e+00 4.30296540e-01 4.03193295e-01
6.72241524e-02 -3.46153998e-03 -1.31696969e-01 4.57963258e-01
-1.88535243e-01 -4.57320400e-02 -6.25650287e-01 -6.10072553e-01
1.14704080e-01 1.44394055e-01 -3.25739861e-01 9.38032925e-01
1.89006075e-01 4.54162717e-01 -9.32229221e-01 -2.72940040e-01
1.74028933e-01 5.15500009e-01 -6.75414503e-01 -2.07102764e-03
1.19200669e-01 4.27329391e-01 -1.97853684e-01 4.01870728e-01
6.33591413e-01 -1.68815181e-01 -7.08007291e-02 -1.85966492e-01
-1.67493731e-01 -9.16259363e-02 -1.71783471e+00 1.43852377e+00
-7.52471864e-01 7.55579293e-01 3.82898211e-01 -1.31657088e+00
9.46546197e-01 2.40848199e-01 7.33917773e-01 -1.74857497e-01
1.64520573e-02 6.24052644e-01 -2.98889369e-01 -3.22076023e-01
4.89351273e-01 -8.71796310e-02 2.14914799e-01 3.00057352e-01
8.43549240e-03 3.16957536e-04 4.07446921e-01 5.20369224e-03
8.55964363e-01 1.79859176e-01 4.43467736e-01 -2.21327648e-01
7.75710344e-01 -3.44102710e-01 8.51518512e-01 6.94340944e-01
-1.76558346e-02 7.28823721e-01 2.90547669e-01 -3.20577204e-01
-9.04436409e-01 -8.31579328e-01 -1.91936135e-01 1.06831813e+00
-1.94835022e-01 -3.19609284e-01 -6.14409745e-01 -4.83084112e-01
2.93517858e-01 4.94844079e-01 -4.89636064e-01 -1.33379087e-01
-6.67032897e-01 -9.65330064e-01 2.23563820e-01 2.86229610e-01
3.46901983e-01 -4.52256352e-01 -3.31950635e-01 5.15020907e-01
-1.22384511e-01 -9.37657714e-01 -7.71023035e-01 1.40002608e-01
-1.14187300e+00 -1.01431000e+00 -7.32661247e-01 -4.48594809e-01
7.48197436e-01 1.05296612e-01 9.76273715e-01 -8.43183473e-02
3.82974707e-02 5.03632307e-01 -1.42709628e-01 -2.88391888e-01
-3.67803931e-01 -7.83735290e-02 6.54053450e-01 5.12585759e-01
-3.63064297e-02 -6.57259047e-01 -4.71142292e-01 5.23687005e-01
-1.00548899e+00 -3.83199930e-01 5.98066926e-01 1.15769315e+00
8.74444604e-01 -1.26936529e-02 5.70568442e-01 -1.00802708e+00
7.23150313e-01 -6.26226068e-01 -9.35885310e-01 2.36962825e-01
-8.96063864e-01 1.59014478e-01 9.88489330e-01 -4.64621305e-01
-6.72854304e-01 2.00775832e-01 -2.55580973e-02 -7.10418105e-01
4.49245870e-01 6.85359836e-01 1.04092956e-01 -3.76538575e-01
7.10571051e-01 2.33847618e-01 6.93551078e-02 -5.11681318e-01
4.29319382e-01 4.93873209e-01 4.12877232e-01 -4.43644196e-01
1.14733005e+00 5.05670130e-01 2.04967156e-01 -9.07045543e-01
-1.01615846e+00 -6.03337348e-01 -5.32694399e-01 -8.33089873e-02
5.42987287e-01 -8.51588964e-01 -7.83192277e-01 2.95313239e-01
-1.01537907e+00 -1.59806609e-01 -3.03477764e-01 7.70892620e-01
-8.07844043e-01 4.96741444e-01 -4.42950934e-01 -8.99647057e-01
-3.51810366e-01 -8.66199553e-01 8.23660672e-01 -3.05104591e-02
5.41204028e-02 -1.04494524e+00 1.96329411e-02 1.77398190e-01
3.78287882e-01 4.62061465e-01 4.36891735e-01 -6.93463087e-01
-1.48728371e-01 -3.68260741e-01 5.45951501e-02 8.62227261e-01
3.59205082e-02 6.73289075e-02 -7.72154391e-01 -6.36718929e-01
3.19870710e-01 -2.75067627e-01 8.39653432e-01 4.68761951e-01
1.12797785e+00 -7.11010516e-01 -2.26515848e-02 8.43157053e-01
1.48389649e+00 -1.63919255e-01 2.38112599e-01 3.42670679e-01
6.63573682e-01 3.35513711e-01 7.76899755e-01 8.80038857e-01
1.13635294e-01 6.04730308e-01 4.03894573e-01 -9.57940333e-03
4.61502850e-01 -7.36304373e-02 7.15263784e-01 1.25309730e+00
-4.32006940e-02 1.25547767e-01 -5.71896553e-01 4.29774284e-01
-2.16715693e+00 -7.79359937e-01 -2.09712446e-01 2.68627095e+00
8.01284552e-01 5.51961325e-02 1.99074596e-01 2.87419796e-01
6.85211122e-01 1.30146042e-01 -6.92463100e-01 -4.56300110e-01
-2.09169641e-01 2.29397193e-01 8.39550674e-01 7.98807263e-01
-1.16509926e+00 6.15706444e-01 6.09450769e+00 8.30766559e-01
-1.04913056e+00 2.90875524e-01 3.84406567e-01 -1.52192309e-01
-8.45646411e-02 -5.85327332e-04 -6.47451639e-01 4.40407753e-01
9.82368588e-01 -2.65024394e-01 7.79247999e-01 1.09675753e+00
4.00368959e-01 1.01776943e-01 -1.08292437e+00 1.08955097e+00
2.58519143e-01 -9.91741836e-01 -3.47721577e-01 -1.59658656e-01
1.02574241e+00 5.40395081e-02 1.35551110e-01 2.45409697e-01
1.13634348e-01 -8.97799909e-01 8.52866113e-01 4.91780877e-01
3.37233633e-01 -8.80397022e-01 5.64935267e-01 2.97474861e-01
-1.04833066e+00 -2.86087811e-01 -7.16125309e-01 7.17662349e-02
1.58505991e-01 1.19889677e+00 -4.00806874e-01 5.25184512e-01
5.63243151e-01 7.40361035e-01 -2.09751502e-01 1.11199379e+00
-2.38887578e-01 5.63940108e-01 -5.77208936e-01 2.86449566e-02
2.87942648e-01 -7.12691665e-01 7.87344873e-01 1.04047811e+00
5.31832933e-01 -3.60152870e-01 1.96582764e-01 5.95103920e-01
-2.13740572e-01 4.14742857e-01 -3.47684234e-01 -4.19558771e-03
3.41136307e-01 1.09051359e+00 -2.80150086e-01 -1.93388045e-01
-6.39108658e-01 1.14456201e+00 3.31682593e-01 4.98365998e-01
-7.84455657e-01 -5.94531775e-01 7.14173555e-01 -1.66560575e-01
4.95496303e-01 -3.91508698e-01 -7.16171935e-02 -1.35919666e+00
3.60352278e-01 -1.11412728e+00 3.79752874e-01 -3.33350152e-01
-1.37713993e+00 5.10569215e-01 -3.25565279e-01 -1.33343887e+00
-2.72076756e-01 -6.82385027e-01 -2.10191026e-01 7.19372392e-01
-1.55949903e+00 -5.99544466e-01 -1.03237659e-01 7.34341621e-01
4.88132447e-01 -4.22647856e-02 6.19010687e-01 5.07792711e-01
-8.91996443e-01 6.62892044e-01 9.48834598e-01 -7.28728399e-02
7.67218530e-01 -1.47260857e+00 1.16549298e-01 9.39762533e-01
2.02556491e-01 4.73090261e-01 9.25833881e-01 -4.04078394e-01
-1.56139553e+00 -1.23883379e+00 7.57967830e-01 -7.72736594e-02
9.62972283e-01 -2.25494027e-01 -8.50257814e-01 8.68521869e-01
-2.91970849e-01 3.81695956e-01 5.23206651e-01 6.03391044e-02
-2.18922928e-01 -4.23659742e-01 -1.09930837e+00 3.86578858e-01
6.42078578e-01 -4.61083561e-01 -1.85951605e-01 6.60493672e-01
5.53255439e-01 -5.40496886e-01 -1.24126709e+00 1.86770037e-01
4.36839461e-01 -9.04632866e-01 9.07958388e-01 -7.53799617e-01
3.10385257e-01 -2.96031415e-01 -3.68597925e-01 -1.55102754e+00
-1.94331840e-01 -1.13057804e+00 -4.90383208e-01 9.89793360e-01
4.96576339e-01 -9.04194772e-01 6.01214468e-01 4.51826751e-01
-1.78563520e-01 -9.13905799e-01 -1.19756866e+00 -1.13948023e+00
8.84881243e-02 -5.86959541e-01 2.22998396e-01 8.59771132e-01
-1.90859288e-01 7.39656314e-02 -8.30935299e-01 2.14658424e-01
7.18716443e-01 -8.56044590e-02 7.81177700e-01 -1.13251054e+00
-5.90856671e-01 -3.70380759e-01 -3.88043702e-01 -1.17798066e+00
1.77391291e-01 -7.81865597e-01 2.44670913e-01 -1.09306657e+00
-2.87715793e-01 -3.55357200e-01 -4.05866414e-01 1.57452226e-01
-2.54221559e-01 1.52965412e-01 1.15933709e-01 1.93138644e-01
-4.67688948e-01 7.90525913e-01 8.24913740e-01 -6.37765974e-02
-3.74177337e-01 2.72237390e-01 -5.24505079e-01 9.54222083e-01
5.55636585e-01 -7.20834076e-01 -1.48834690e-01 -3.92941833e-01
5.35826147e-01 8.15142691e-02 -3.27923484e-02 -1.04050183e+00
2.03592673e-01 -9.90240052e-02 1.39828458e-01 -2.53395528e-01
3.32818657e-01 -7.85204172e-01 -1.22368544e-01 4.94097948e-01
-2.38466144e-01 3.01187813e-01 -7.48725235e-02 6.99312150e-01
-4.65684444e-01 -4.87209737e-01 8.57067168e-01 3.58748063e-02
-2.73844540e-01 5.84978878e-01 -1.92251831e-01 2.54485577e-01
7.18284011e-01 -2.46006530e-03 1.15212560e-01 -4.93950903e-01
-7.15253472e-01 1.60750419e-01 1.45330608e-01 2.43008379e-02
7.16535687e-01 -1.43937635e+00 -9.80837345e-01 -7.23070651e-02
-2.09266797e-01 7.46528134e-02 3.55425589e-02 1.46997726e+00
-4.17761952e-01 9.60595310e-02 3.51512045e-01 -4.72442746e-01
-9.81820941e-01 5.55098891e-01 3.92709106e-01 -3.95831972e-01
-3.76879007e-01 8.36238086e-01 -4.33004573e-02 -4.37889248e-01
3.82689089e-01 -1.73568502e-01 2.45314375e-01 1.17567748e-01
5.45944333e-01 9.09290373e-01 2.22993299e-01 -6.01576626e-01
-3.40921223e-01 6.48227692e-01 3.87189239e-02 -1.67493016e-01
1.53295946e+00 -1.80194378e-01 -2.59068310e-01 6.17513657e-01
1.49260533e+00 1.32036999e-01 -1.29750574e+00 -4.24625427e-01
2.55700380e-01 -5.42410612e-01 6.34839013e-02 -2.28306875e-01
-1.27840388e+00 7.97669351e-01 4.83974040e-01 1.46473646e-01
1.25677514e+00 -6.83180153e-01 8.52061629e-01 7.77520418e-01
1.12018719e-01 -1.40884304e+00 9.49572101e-02 5.48723936e-01
1.14541388e+00 -1.34044540e+00 2.08448559e-01 -1.91551492e-01
-4.71919119e-01 1.29109967e+00 3.27904940e-01 -4.60187852e-01
8.23101163e-01 2.45556813e-02 1.80164762e-02 2.70336032e-01
-6.29469693e-01 -5.97682260e-02 2.64983505e-01 2.60750830e-01
3.09004128e-01 -1.74081609e-01 -4.01853889e-01 2.11101815e-01
-1.12405770e-01 6.00113422e-02 3.89960378e-01 6.71777368e-01
-2.23812625e-01 -9.30479467e-01 -5.78486860e-01 5.58194935e-01
-5.64631879e-01 -1.44430667e-01 -3.09592690e-02 6.87762797e-01
-1.57838553e-01 1.10173786e+00 -2.62297601e-01 -3.75083923e-01
3.23333353e-01 -8.27943310e-02 3.36069465e-01 -2.32020184e-01
-4.56744522e-01 -7.42068142e-02 -5.82766235e-02 -7.06827521e-01
-3.52984786e-01 -8.67970407e-01 -8.26110065e-01 -3.10140252e-01
-5.96173048e-01 2.01564312e-01 8.76438379e-01 1.02694285e+00
1.37862042e-01 2.92300969e-01 1.14345694e+00 -6.43360555e-01
-1.12848854e+00 -1.01703191e+00 -6.41246736e-01 5.16809940e-01
5.38744807e-01 -6.73073649e-01 -7.34371006e-01 -1.48860946e-01] | [7.163823127746582, 4.414854526519775] |
d7afff57-5bfe-47d9-9b1c-c8b6d04af8b7 | language-model-analysis-for-ontology | 2302.06761 | null | https://arxiv.org/abs/2302.06761v3 | https://arxiv.org/pdf/2302.06761v3.pdf | Language Model Analysis for Ontology Subsumption Inference | Investigating whether pre-trained language models (LMs) can function as knowledge bases (KBs) has raised wide research interests recently. However, existing works focus on simple, triple-based, relational KBs, but omit more sophisticated, logic-based, conceptualised KBs such as OWL ontologies. To investigate an LM's knowledge of ontologies, we propose OntoLAMA, a set of inference-based probing tasks and datasets from ontology subsumption axioms involving both atomic and complex concepts. We conduct extensive experiments on ontologies of different domains and scales, and our results demonstrate that LMs encode relatively less background knowledge of Subsumption Inference (SI) than traditional Natural Language Inference (NLI) but can improve on SI significantly when a small number of samples are given. We will open-source our code and datasets. | ['Ian Horrocks', 'Hang Dong', 'Ernesto Jiménez-Ruiz', 'Jiaoyan Chen', 'Yuan He'] | 2023-02-14 | null | null | null | null | ['ontology-subsumption-inferece'] | ['knowledge-base'] | [-2.03958854e-01 6.85792446e-01 -6.44002378e-01 -6.27317369e-01
-2.75044411e-01 -5.19048452e-01 5.47269106e-01 2.01906756e-01
-1.12375073e-01 1.22553873e+00 1.87147602e-01 -6.64086223e-01
-5.00689089e-01 -1.48921132e+00 -1.04940736e+00 2.53070265e-01
-4.09325123e-01 9.06768084e-01 7.68603742e-01 -4.02195245e-01
-1.64274827e-01 3.31031024e-01 -1.75539160e+00 5.90303123e-01
9.07302320e-01 1.08010077e+00 -3.45988989e-01 6.93952367e-02
-5.88974357e-01 1.48871589e+00 -4.43516523e-01 -7.37055540e-01
-2.12826476e-01 1.60608292e-01 -1.50604415e+00 -5.89371264e-01
4.70233828e-01 -2.52511799e-01 -3.57092708e-01 1.15620661e+00
1.23875804e-01 -1.52419657e-02 4.47373182e-01 -1.65391016e+00
-9.05606270e-01 1.25609028e+00 2.78683633e-01 4.05849889e-02
7.54317403e-01 -2.81557143e-01 1.23962212e+00 -5.62646329e-01
7.53339648e-01 1.79421401e+00 8.65415275e-01 4.06837046e-01
-1.20468676e+00 -8.39229405e-01 3.57144177e-02 7.20393836e-01
-1.73308396e+00 -5.63495100e-01 1.82734087e-01 -2.76853833e-02
1.60409117e+00 2.08612531e-01 4.35616016e-01 9.09217656e-01
-1.39165595e-01 8.29265714e-01 9.71729636e-01 -6.83024824e-01
3.30645382e-01 7.45934665e-01 3.87131482e-01 1.03383791e+00
8.20217252e-01 -3.01522255e-01 -5.81486940e-01 -4.07032371e-01
5.83456933e-01 -5.47000229e-01 -8.68567005e-02 -3.30315799e-01
-8.36564958e-01 8.07618797e-01 2.46520817e-01 1.43127471e-01
7.53772026e-03 3.55249643e-03 2.46684149e-01 5.62564135e-01
8.05005357e-02 6.38729811e-01 -1.05821669e+00 1.11463591e-01
-4.96069938e-01 3.44794035e-01 1.43839693e+00 1.63761246e+00
1.09046507e+00 -4.73090887e-01 3.38402480e-01 8.99346352e-01
3.31717342e-01 4.00267154e-01 3.90867472e-01 -1.05832982e+00
4.46064353e-01 1.14202440e+00 1.72301739e-01 -7.71110415e-01
-4.23744440e-01 1.62010938e-01 -2.81883091e-01 -4.62187499e-01
1.61948338e-01 -3.53054665e-02 -4.26650107e-01 1.62848008e+00
2.44428456e-01 1.08918257e-01 5.73994994e-01 3.94025862e-01
1.23909640e+00 3.86994183e-01 1.41971424e-01 1.16248831e-01
1.81066561e+00 -7.02621460e-01 -6.02728665e-01 -4.88519311e-01
9.56306994e-01 8.16049203e-02 1.11176789e+00 3.26934338e-01
-1.16876709e+00 -3.22346658e-01 -1.07933533e+00 -3.05101275e-01
-1.10003030e+00 -3.31343770e-01 1.20701623e+00 6.04674220e-01
-9.33751762e-01 2.54208326e-01 -5.56811094e-01 -7.34311521e-01
5.45210302e-01 2.83771902e-01 -5.09731412e-01 -4.85312670e-01
-1.93533599e+00 1.18842280e+00 1.16664481e+00 -2.56167144e-01
-6.94595337e-01 -8.55342746e-01 -1.05949843e+00 1.32845268e-01
1.02700019e+00 -7.85805583e-01 1.08588982e+00 -3.87272239e-01
-1.19219995e+00 7.86679506e-01 -1.97440982e-01 -6.13105059e-01
2.54551433e-02 -1.61583781e-01 -9.90364075e-01 -6.02065362e-02
9.86066759e-02 6.92159355e-01 -6.89941272e-02 -1.12206185e+00
-8.85913372e-01 -2.53735989e-01 1.12990332e+00 -2.59804666e-01
-6.45684421e-01 2.40521003e-02 -4.85882461e-01 9.36994478e-02
-4.01635468e-02 -5.92410743e-01 8.94511938e-02 1.37583865e-02
-5.42115569e-01 -6.20808601e-01 2.43122801e-01 -2.33359501e-01
1.44355941e+00 -1.61063218e+00 -1.81556448e-01 2.27788150e-01
-6.09771796e-02 -5.86993098e-02 1.28661776e-02 4.39517289e-01
1.13230452e-01 4.71297950e-01 1.27223268e-01 4.63723898e-01
5.14851272e-01 8.56413662e-01 -4.68503445e-01 -3.52914989e-01
3.74389976e-01 1.01346099e+00 -9.20138180e-01 -9.67264056e-01
-9.85300913e-02 -2.16476113e-01 -7.54766345e-01 5.04919551e-02
-7.90698707e-01 -4.54892218e-01 -4.68534231e-01 1.03667545e+00
4.45716053e-01 -7.22927749e-01 7.10604429e-01 -5.79622328e-01
2.49648556e-01 5.97951829e-01 -1.20591259e+00 1.55937254e+00
-5.38638294e-01 3.44058663e-01 -4.45160806e-01 -1.03020871e+00
6.13206744e-01 3.00734162e-01 1.78746192e-03 -7.40050793e-01
-2.70552695e-01 1.25389501e-01 -2.88664401e-02 -8.70281875e-01
2.38154113e-01 -4.27504838e-01 -1.27532214e-01 7.17315450e-02
3.40096295e-01 -8.09820369e-02 5.67018807e-01 1.92697793e-01
1.27119386e+00 1.97364345e-01 6.86652660e-01 -4.36364889e-01
5.84438324e-01 4.45816368e-01 4.20922250e-01 8.59107256e-01
3.38490874e-01 -4.74668831e-01 5.87522566e-01 -4.43755120e-01
-6.41252577e-01 -1.11469269e+00 -8.41954887e-01 1.23841870e+00
3.21825147e-01 -8.55175853e-01 -2.41966426e-01 -6.71898246e-01
3.55516642e-01 1.19056809e+00 -2.10592553e-01 -1.01950333e-01
-2.17655599e-01 -7.95769453e-01 1.17136288e+00 7.67419517e-01
4.88160491e-01 -1.25938189e+00 -2.43056580e-01 2.50576615e-01
-4.02130216e-01 -1.59830308e+00 8.32565486e-01 1.71873018e-01
-9.10526574e-01 -1.24040484e+00 3.75345677e-01 -6.42685592e-01
4.63118583e-01 -3.76343995e-01 1.58142090e+00 -4.49479930e-03
-1.85966849e-01 3.97220433e-01 -3.85017246e-01 -6.57171607e-01
-3.47667009e-01 2.39100158e-01 2.19430596e-01 -8.25052798e-01
1.05284083e+00 -6.54467344e-01 9.96209085e-02 5.72456956e-01
-1.06336987e+00 -8.74879509e-02 4.21474367e-01 3.66569340e-01
2.09467962e-01 8.83094788e-01 6.07047379e-01 -1.05649674e+00
4.97411817e-01 -7.47223020e-01 -7.29287982e-01 7.63307810e-01
-7.20109761e-01 3.55945110e-01 5.28095901e-01 -1.75860777e-01
-1.17151737e+00 -6.11909926e-01 9.18983147e-02 6.77407309e-02
-2.57860035e-01 1.22076511e+00 -5.17312109e-01 6.50520772e-02
8.44794571e-01 -7.83906430e-02 -4.65084672e-01 -3.48850787e-01
5.51372826e-01 7.40858972e-01 4.31794792e-01 -1.67856622e+00
7.22378314e-01 4.41559583e-01 -2.04456404e-01 -7.19780445e-01
-1.30606067e+00 -2.83447087e-01 -6.21716261e-01 3.45117092e-01
3.65467310e-01 -1.09362125e+00 -1.01658309e+00 -3.52201611e-01
-9.70259011e-01 -5.21756589e-01 -2.38490775e-01 2.43506685e-01
-4.94086564e-01 2.14139283e-01 -5.54246426e-01 -8.13361943e-01
-3.67961265e-02 -5.68556547e-01 7.98850656e-01 6.24096803e-02
-4.96489882e-01 -1.10479736e+00 6.64342865e-02 4.01518703e-01
1.93810284e-01 -1.92586720e-01 1.63446236e+00 -8.44833374e-01
-7.06308007e-01 -2.25715026e-01 -5.06707251e-01 4.35954690e-01
1.37240902e-01 -2.67760068e-01 -9.34947312e-01 -3.73456292e-02
-6.86425209e-01 -1.01539493e+00 2.12936953e-01 -2.32623234e-01
1.32856548e+00 -6.77898049e-01 -6.17762148e-01 2.75922716e-01
1.59527063e+00 6.05621934e-02 1.04662919e+00 7.30231285e-01
2.22465068e-01 5.62884927e-01 6.73085213e-01 3.19728494e-01
9.49425936e-01 4.07051891e-01 9.62197408e-02 3.43885779e-01
3.36116791e-01 -3.82135451e-01 1.80880427e-01 5.74318647e-01
-3.09225678e-01 -3.17392409e-01 -1.13347864e+00 5.99064708e-01
-1.81971478e+00 -8.89936507e-01 1.37486547e-01 1.80907774e+00
1.57670152e+00 4.39185739e-01 -2.29169309e-01 -2.53037177e-03
1.47708982e-01 -3.35325688e-01 -5.45250773e-01 -8.42104778e-02
-4.09951746e-01 6.68256879e-01 3.95228118e-01 4.74947959e-01
-8.94061863e-01 1.36686730e+00 6.89164257e+00 7.04102218e-01
-5.57259202e-01 -1.12963490e-01 -2.03115791e-01 1.32897764e-01
-4.36067522e-01 3.76723647e-01 -1.15416896e+00 1.20473631e-01
1.20238400e+00 -4.45535511e-01 5.63868523e-01 1.07220960e+00
-4.62545812e-01 -2.72322521e-02 -1.55598569e+00 4.39054430e-01
-1.90583065e-01 -1.43337607e+00 5.39787233e-01 9.29808244e-03
3.69235933e-01 4.28249948e-02 -6.49927557e-01 1.00862229e+00
9.65792418e-01 -1.21208334e+00 6.40944123e-01 7.84738421e-01
7.21453309e-01 -3.49386156e-01 9.73071098e-01 3.44521672e-01
-1.19263649e+00 -2.62168050e-01 -8.25240612e-01 -2.85349250e-01
-4.14233059e-02 4.53905791e-01 -8.59267473e-01 8.89997721e-01
1.00994766e+00 5.72841108e-01 -7.38172710e-01 5.51025867e-01
-4.08509791e-01 5.28428733e-01 -6.70561433e-01 -8.17633644e-02
-9.61356387e-02 2.16241658e-01 -1.93359748e-01 1.21691823e+00
-2.57473923e-02 4.36704218e-01 1.79792374e-01 1.10766065e+00
-1.81976780e-01 -4.90115955e-02 -4.72114295e-01 -3.57378423e-01
8.08868408e-01 9.38949764e-01 -1.13964543e-01 -8.79613996e-01
-7.59373963e-01 4.06916648e-01 6.00753307e-01 4.18464005e-01
-8.95608485e-01 -3.86285692e-01 7.66074836e-01 1.18401237e-01
1.66662440e-01 1.87601596e-01 3.50155801e-01 -1.37228990e+00
1.11083522e-01 -9.07268643e-01 9.03997838e-01 -1.25688243e+00
-1.66100550e+00 1.96001932e-01 6.53747261e-01 -5.74403048e-01
-2.16515839e-01 -1.06164491e+00 -3.44125777e-02 5.10614336e-01
-1.84750342e+00 -1.19259989e+00 -4.05945748e-01 5.95272481e-01
-5.27721904e-02 -1.94958914e-02 1.32443857e+00 4.41500872e-01
-4.11770970e-01 5.67576408e-01 -4.18119580e-01 2.97727406e-01
5.95729172e-01 -1.22375226e+00 3.14372666e-02 1.43081784e-01
-3.77076416e-04 1.20055532e+00 4.21106189e-01 -6.00615680e-01
-1.48696697e+00 -1.08457589e+00 9.90500689e-01 -7.30183601e-01
1.15510857e+00 -1.55708343e-01 -1.08529091e+00 1.45605552e+00
-1.62051827e-01 2.06043541e-01 9.41933990e-01 7.23591328e-01
-6.46557391e-01 -1.63540125e-01 -1.13986874e+00 5.47270119e-01
1.65561545e+00 -7.93741167e-01 -1.10879588e+00 3.52330655e-01
8.48530948e-01 -7.69282179e-03 -1.74684870e+00 9.09013808e-01
8.39988172e-01 -7.07255363e-01 1.32044590e+00 -1.53630555e+00
4.70675111e-01 -5.49792051e-01 -5.83962023e-01 -7.94291079e-01
-3.67555022e-01 2.05522135e-01 -5.43153465e-01 1.16195309e+00
7.23010838e-01 -8.72514844e-01 5.90050936e-01 8.04398715e-01
1.55813932e-01 -5.19990623e-01 -6.97827756e-01 -1.09800231e+00
-5.96106127e-02 -6.95496202e-01 9.76658821e-01 1.37451065e+00
6.06957912e-01 5.92054248e-01 2.61592388e-01 6.57416999e-01
3.81706238e-01 2.14921832e-01 8.99376214e-01 -1.66909397e+00
-1.58315331e-01 -1.73424274e-01 -5.91251373e-01 -8.13953757e-01
5.27621686e-01 -1.17124104e+00 -2.93868005e-01 -1.72721827e+00
2.50677615e-01 -7.63178766e-01 -2.51392871e-01 1.16756940e+00
2.01937705e-01 -1.45668581e-01 -5.13525128e-01 -1.32064208e-01
-9.47471797e-01 3.05641353e-01 7.80842781e-01 -3.66304070e-01
2.43281856e-01 -5.31068921e-01 -8.24881554e-01 9.46315467e-01
3.69945168e-01 -3.05055380e-01 -7.74883747e-01 -1.69407889e-01
7.96470821e-01 -1.15344271e-01 3.72183621e-01 -1.02297723e+00
4.04977113e-01 -4.49342549e-01 2.00478062e-02 -2.73892134e-01
1.66613281e-01 -8.73112023e-01 2.70198017e-01 2.36711930e-02
-3.50961834e-01 -3.88047725e-01 6.24404192e-01 1.75056934e-01
-2.65833497e-01 -5.37158251e-01 3.19711834e-01 -4.71785575e-01
-1.16996264e+00 7.46195018e-02 -1.45828366e-01 4.17572290e-01
5.61661541e-01 -3.59563082e-02 -5.62666595e-01 1.32210804e-02
-7.62302041e-01 3.25073332e-01 2.86804885e-01 3.92193019e-01
3.45016748e-01 -1.24570775e+00 -3.43182653e-01 -1.32945016e-01
8.13236117e-01 3.61469209e-01 -3.08158875e-01 4.87247139e-01
-4.73976046e-01 8.72580945e-01 -1.44037694e-01 -2.53397912e-01
-7.69861817e-01 6.77991331e-01 3.86320829e-01 -1.38034955e-01
-4.69359130e-01 7.59748578e-01 1.10669926e-01 -1.03719127e+00
4.56875831e-01 -7.13500857e-01 -1.06255926e-01 -1.55002519e-01
5.67928016e-01 2.55722940e-01 -4.78858650e-02 9.25888643e-02
-6.25699282e-01 1.36429325e-01 6.27370477e-02 3.85220885e-01
1.23354304e+00 1.45175800e-01 -6.06726646e-01 5.87570071e-01
7.65403569e-01 -1.49754062e-01 -2.61191934e-01 -6.83326125e-01
5.47487557e-01 -2.13950410e-01 -2.41034806e-01 -1.13145614e+00
-3.12242150e-01 3.53029191e-01 1.20843137e-02 3.15402508e-01
7.54382432e-01 5.11740029e-01 4.67394143e-01 1.29649115e+00
1.04890323e+00 -9.85864401e-01 -2.99544722e-01 5.39142609e-01
7.47021019e-01 -1.05817008e+00 2.33853161e-01 -7.80458689e-01
-1.55090153e-01 1.10271692e+00 9.19263124e-01 3.83975893e-01
7.03990400e-01 5.97341239e-01 -3.38724375e-01 -5.32013237e-01
-1.03612161e+00 -4.18024421e-01 1.21016435e-01 5.58044434e-01
4.41733837e-01 7.96083733e-02 2.82831103e-01 9.91878211e-01
-5.15235484e-01 6.70026541e-01 7.76589215e-02 1.06656706e+00
-4.50597942e-01 -1.00175905e+00 -8.41860771e-02 5.63762724e-01
2.14424767e-02 -4.35171932e-01 -3.26069355e-01 1.24050617e+00
3.76286387e-01 7.41562068e-01 -1.32167758e-02 -1.78208098e-01
6.38137758e-01 5.78216553e-01 8.49578857e-01 -7.52260149e-01
1.92319617e-01 -9.26173508e-01 9.53785956e-01 -5.85382760e-01
-5.16548336e-01 -2.08319366e-01 -1.48456740e+00 -5.14148057e-01
-4.83983755e-01 2.49852002e-01 1.46621153e-01 1.25923395e+00
1.23841889e-01 3.78135055e-01 -2.71115154e-01 1.53383210e-01
-3.40786457e-01 -1.04084241e+00 -4.51198161e-01 2.98952788e-01
-3.42208207e-01 -8.62615287e-01 -1.26456276e-01 1.27862498e-01] | [9.974486351013184, 7.868868827819824] |
35f5607b-e3ed-4cd5-8421-4f4b5e6e4ac3 | a-latent-diffusion-model-for-protein | 2305.04120 | null | https://arxiv.org/abs/2305.04120v1 | https://arxiv.org/pdf/2305.04120v1.pdf | A Latent Diffusion Model for Protein Structure Generation | Proteins are complex biomolecules that perform a variety of crucial functions within living organisms. Designing and generating novel proteins can pave the way for many future synthetic biology applications, including drug discovery. However, it remains a challenging computational task due to the large modeling space of protein structures. In this study, we propose a latent diffusion model that can reduce the complexity of protein modeling while flexibly capturing the distribution of natural protein structures in a condensed latent space. Specifically, we propose an equivariant protein autoencoder that embeds proteins into a latent space and then uses an equivariant diffusion model to learn the distribution of the latent protein representations. Experimental results demonstrate that our method can effectively generate novel protein backbone structures with high designability and efficiency. | ['Shuiwang Ji', 'Xiaoning Qian', 'Kanji Uchino', 'Koji Maruhashi', 'Tao Komikado', 'Michael McThrow', 'Wing Yee Au', 'Limei Wang', 'Keqiang Yan', 'Cong Fu'] | 2023-05-06 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 1.76941320e-01 2.18526088e-02 -5.83408661e-02 -2.59329349e-01
-2.73215979e-01 -6.29658997e-01 4.22938496e-01 -7.53432373e-03
-1.06447197e-01 8.88935626e-01 4.92219359e-01 -3.50026876e-01
9.00436193e-02 -9.38193679e-01 -9.61797655e-01 -1.24869013e+00
2.22347975e-01 5.04985154e-01 2.01339424e-02 -1.92428529e-01
-2.49708846e-01 6.17738485e-01 -9.23616469e-01 1.06491588e-01
1.05256653e+00 2.05187276e-01 4.33878303e-01 4.01726633e-01
-1.56743303e-01 5.92641234e-01 -3.94642919e-01 -1.24648429e-01
-1.18714139e-01 -6.72551930e-01 -4.44749564e-01 1.23927012e-01
-3.50939006e-01 -2.80570537e-02 -4.74021971e-01 9.75206733e-01
3.69521260e-01 3.03448826e-01 8.75118375e-01 -5.07884562e-01
-1.19595563e+00 3.81720364e-01 -1.48501799e-01 -1.94581568e-01
9.05583948e-02 4.86063868e-01 9.23474908e-01 -7.74184942e-01
8.35462868e-01 1.33651531e+00 2.72757828e-01 7.48810947e-01
-1.65663981e+00 -4.97896492e-01 1.54347733e-01 -7.65510974e-03
-1.02687693e+00 -1.21253945e-01 8.09726954e-01 -7.84173787e-01
9.70578969e-01 -2.14426458e-01 7.19467282e-01 1.20690608e+00
5.38784027e-01 6.17157817e-01 6.55852377e-01 4.03285399e-02
5.41195631e-01 -2.91539311e-01 3.42119969e-02 7.61057973e-01
5.42872027e-02 -2.35090569e-01 -3.36585492e-01 -4.75429624e-01
7.38548100e-01 5.79987407e-01 -2.21999094e-01 -7.80523598e-01
-1.41853297e+00 1.16897321e+00 3.71707946e-01 2.86923259e-01
-7.31256902e-01 -1.18745640e-02 3.33426371e-02 -1.24668688e-01
2.70963401e-01 4.03430670e-01 -3.88502598e-01 5.13185672e-02
-3.99957925e-01 3.46193880e-01 6.31956935e-01 6.77396715e-01
7.56946862e-01 1.22658856e-01 5.35783581e-02 6.59268320e-01
2.95427114e-01 4.32296902e-01 6.80819631e-01 -8.06524456e-01
-9.78435129e-02 7.51969516e-01 1.77612945e-01 -9.11771536e-01
-2.21898571e-01 -1.51236385e-01 -9.36571300e-01 -6.17060065e-02
1.14144094e-01 -4.53673489e-02 -8.03424656e-01 1.80131197e+00
4.79080826e-01 -2.04147756e-01 4.65535700e-01 7.36044884e-01
3.59535336e-01 1.21565115e+00 4.04666275e-01 -3.92961562e-01
1.09492898e+00 -9.78015900e-01 -7.52268076e-01 6.10941686e-02
4.30901378e-01 -5.83128512e-01 8.97266924e-01 -4.77704182e-02
-8.34486663e-01 -4.28234398e-01 -1.07913351e+00 -2.21155837e-01
-1.63454413e-01 3.28981280e-02 9.39950287e-01 1.18125506e-01
-4.90121990e-01 7.93940783e-01 -1.27645671e+00 -1.89299479e-01
3.44351918e-01 4.74044383e-01 -4.59212840e-01 2.61564683e-02
-1.04063225e+00 5.09280324e-01 6.78772926e-01 -9.67214555e-02
-9.17329609e-01 -4.06499445e-01 -7.74909258e-01 2.23752901e-01
7.46077970e-02 -7.69569099e-01 8.09226632e-01 -5.42314053e-01
-1.79316151e+00 4.24690723e-01 -3.23112607e-01 -4.11198854e-01
1.00170121e-01 1.02859095e-01 -9.45063010e-02 -1.90214306e-01
-7.47871548e-02 6.80347562e-01 6.91885948e-01 -7.19832063e-01
3.92813236e-02 -3.73350292e-01 -1.26928404e-01 -6.00829069e-03
-2.09065482e-01 -4.18738246e-01 -7.15519339e-02 -9.14076746e-01
2.17068136e-01 -1.22819018e+00 -6.49639308e-01 7.20845461e-02
-5.80664650e-02 -3.92641932e-01 6.55402303e-01 -4.64363307e-01
7.92608082e-01 -2.19323230e+00 9.90904450e-01 -1.40637225e-02
6.11028850e-01 1.71502575e-01 -4.61895838e-02 5.65675080e-01
-1.01515457e-01 -2.14140788e-01 -4.89920735e-01 1.62686363e-01
9.30260983e-04 2.82989204e-01 -3.73857439e-01 4.16644722e-01
8.26843753e-02 1.14642727e+00 -8.17228138e-01 1.13890748e-02
7.39063993e-02 8.22940230e-01 -7.20054448e-01 4.48035568e-01
-7.83389449e-01 5.76301038e-01 -8.38276625e-01 3.57710004e-01
4.86838669e-01 -7.48484135e-01 4.98031825e-01 -2.67631859e-01
-2.85905190e-02 -2.64741238e-02 -4.69420969e-01 1.72693980e+00
1.93815976e-01 2.68397510e-01 -3.03243548e-01 -1.01737559e+00
1.15408754e+00 9.58193615e-02 6.41469061e-01 -9.01713893e-02
-2.10946258e-02 -2.16082986e-02 2.02719837e-01 -2.80575991e-01
6.99439049e-02 -5.10906994e-01 4.63233702e-02 4.98352528e-01
4.66695847e-03 2.24675506e-01 -5.16027100e-02 1.15863696e-01
9.33788598e-01 2.51466334e-01 1.62082389e-01 -1.96179941e-01
4.50722724e-01 2.72946935e-02 9.01977360e-01 1.59241959e-01
-1.69200540e-01 1.01453803e-01 5.52483022e-01 -8.47383618e-01
-1.45033371e+00 -1.30867779e+00 2.50110894e-01 8.72904778e-01
-6.46165386e-02 -3.37017715e-01 -9.73787904e-01 -4.88261312e-01
8.99620652e-02 2.57750243e-01 -4.54158276e-01 -7.03080773e-01
-6.07933700e-01 -1.01275039e+00 2.69459635e-01 2.85795867e-01
4.86568827e-03 -1.27313769e+00 -5.83501942e-02 6.24272227e-01
-1.28448695e-01 -7.15938687e-01 -6.36063337e-01 1.85726136e-01
-8.35464656e-01 -8.60585928e-01 -8.25727701e-01 -1.09192419e+00
7.43592918e-01 2.71765620e-01 4.96388823e-01 -3.82206589e-01
-4.14735734e-01 -2.36956924e-01 -3.71232033e-02 -1.74793169e-01
-7.14454055e-01 1.22828379e-01 3.87083262e-01 9.57155228e-02
7.42044926e-01 -7.01979280e-01 -7.56083190e-01 6.77311122e-02
-1.20448375e+00 3.93555164e-01 5.58880150e-01 1.00660741e+00
9.59261417e-01 -6.22191932e-03 5.29853284e-01 -8.27800751e-01
7.26517677e-01 -5.59411347e-01 -6.83902860e-01 3.36155444e-01
-3.57343048e-01 7.47370899e-01 6.46919727e-01 -6.91448510e-01
-9.73132372e-01 4.41276252e-01 -2.27991596e-01 -1.34760931e-01
2.05030721e-02 4.96910512e-01 -4.08361763e-01 1.99206382e-01
4.55495477e-01 5.81317186e-01 4.77060974e-01 -5.64907849e-01
4.98127192e-01 4.33638662e-01 2.45248258e-01 -7.25236952e-01
5.96406043e-01 4.41244215e-01 -5.14607504e-02 -7.71221995e-01
-5.67286432e-01 3.95303480e-02 -6.48422241e-01 2.93857306e-01
1.21620917e+00 -8.88543487e-01 -9.46777165e-01 4.23075169e-01
-1.09326398e+00 -1.87761843e-01 -1.37903318e-01 5.30591905e-01
-6.69063270e-01 4.83889878e-01 -8.30567539e-01 -1.28871381e-01
-3.21961015e-01 -1.59447730e+00 8.71817410e-01 1.92591205e-01
-1.75094992e-01 -9.75126922e-01 5.24616957e-01 2.38123402e-01
2.38613486e-01 4.50738072e-01 1.42080343e+00 -2.04681039e-01
-9.54674363e-01 8.44901204e-02 8.06446448e-02 1.23360090e-01
6.32865310e-01 -1.28471419e-01 -4.27312225e-01 -4.16372359e-01
-2.88983937e-02 -2.16867939e-01 8.83242130e-01 3.13049376e-01
9.25617933e-01 -4.87591088e-01 -5.39122701e-01 6.11912906e-01
1.05135179e+00 4.52615023e-01 4.23230201e-01 8.73565860e-03
9.05361474e-01 7.13521481e-01 2.13156492e-01 3.99755776e-01
1.16875529e-01 5.68980932e-01 -4.12510745e-02 -7.13948831e-02
3.73169452e-01 -5.22085607e-01 4.51313108e-01 1.16615033e+00
7.80942142e-02 -1.57893956e-01 -8.40337336e-01 2.44145066e-01
-2.07417083e+00 -9.72564280e-01 3.45850199e-01 1.62319386e+00
1.17021000e+00 -1.10774636e-01 -1.73093397e-02 -4.43467617e-01
6.93701267e-01 8.15609321e-02 -1.16721952e+00 -1.08763114e-01
-2.51472235e-01 1.27302423e-01 2.45035380e-01 4.65536773e-01
-8.36386502e-01 1.25956845e+00 6.69598103e+00 5.30481040e-01
-1.12191010e+00 -8.30400437e-02 3.81650150e-01 1.39253139e-01
-4.69484478e-01 2.72482149e-02 -8.02336216e-01 5.07749736e-01
8.27571452e-01 -3.47801208e-01 4.52214539e-01 8.96493733e-01
2.04184890e-01 6.94877803e-01 -8.50978434e-01 7.70525575e-01
-2.24110246e-01 -1.74131775e+00 3.92710328e-01 2.69479841e-01
8.05336595e-01 -1.33960024e-01 9.79205370e-02 2.25017127e-02
5.36259651e-01 -9.76589739e-01 1.12819143e-01 8.43258440e-01
4.27513748e-01 -8.08209717e-01 1.66972384e-01 5.44715881e-01
-7.63561487e-01 1.78876504e-01 -6.86470211e-01 8.05328321e-03
1.55736536e-01 5.46985507e-01 -6.62485361e-01 -3.06081593e-01
3.05985540e-01 9.06953037e-01 -8.19106549e-02 2.49474883e-01
-1.58696100e-01 4.06631380e-01 -4.81466651e-02 -2.21517354e-01
1.44938782e-01 -7.45852292e-01 3.31172019e-01 7.81584799e-01
1.73924714e-01 3.40523750e-01 4.59676683e-01 1.24344957e+00
-2.63606369e-01 1.70286030e-01 -6.12245381e-01 -7.60690272e-01
1.61854878e-01 8.69790792e-01 -6.17304325e-01 -1.92200631e-01
-2.61017501e-01 1.23789418e+00 4.94603395e-01 5.88485181e-01
-8.68061185e-01 -3.85708719e-01 1.04787326e+00 -3.52697708e-02
3.13754737e-01 -6.70652211e-01 4.59813923e-01 -1.48678577e+00
-9.44810212e-02 -9.33456302e-01 -1.01479337e-01 -4.61913139e-01
-1.22960746e+00 2.88511932e-01 -5.59733093e-01 -7.51767516e-01
-7.28511736e-02 -6.03627086e-01 -7.22435564e-02 8.53716493e-01
-1.03672528e+00 -1.04130387e+00 8.40966925e-02 3.33331078e-01
5.48461199e-01 -4.91048962e-01 1.09936452e+00 -4.67140786e-02
-6.78731561e-01 2.60033876e-01 9.05214489e-01 -4.64860611e-02
6.05284929e-01 -1.13604856e+00 7.56762862e-01 4.55672950e-01
-6.26899898e-02 1.30089676e+00 7.42633224e-01 -8.67636621e-01
-1.50771582e+00 -1.25563955e+00 6.47518158e-01 -3.13922197e-01
6.02239847e-01 -5.74514091e-01 -1.25195336e+00 6.65233433e-01
-1.08348355e-01 -2.08254486e-01 1.00851023e+00 -2.17792764e-01
-3.26439768e-01 1.95416912e-01 -8.78491759e-01 7.71692514e-01
8.16437364e-01 -7.91732788e-01 -4.77254301e-01 5.95881402e-01
1.35327625e+00 -1.63339347e-01 -1.24618196e+00 3.56675863e-01
6.77778721e-01 -3.51523966e-01 1.13149977e+00 -9.47570384e-01
3.99219036e-01 -4.92599994e-01 -1.97213236e-02 -1.10442352e+00
-8.07080984e-01 -6.31638527e-01 -4.13832963e-01 8.23030055e-01
2.99285322e-01 -4.39345479e-01 1.08117580e+00 6.73404276e-01
1.38233095e-01 -6.64975524e-01 -4.60755885e-01 -4.10454452e-01
2.86592841e-01 4.43526089e-01 6.46108568e-01 9.73448277e-01
-9.93563011e-02 5.95124602e-01 -4.38731670e-01 -8.02124143e-02
5.73828697e-01 2.99645334e-01 6.79080784e-01 -1.08072162e+00
-6.49652123e-01 -2.20214322e-01 -4.15784806e-01 -1.34045923e+00
4.80020821e-01 -8.31223547e-01 -1.66249037e-01 -1.31832314e+00
4.46398705e-01 -1.46718830e-01 -3.95720303e-01 3.61948520e-01
-2.24672273e-01 -2.47546628e-01 -1.47126868e-01 4.62842107e-01
-4.78903681e-01 1.08576012e+00 1.18223214e+00 -4.08583730e-01
-3.00025105e-01 -2.29623228e-01 -6.58012867e-01 5.41830420e-01
8.83902788e-01 -4.98566389e-01 -4.43660527e-01 -1.90053880e-01
1.89770177e-01 5.03571481e-02 -8.00748020e-02 -7.22804844e-01
-1.18091609e-02 -6.30187392e-01 4.19419259e-01 -3.55154604e-01
3.18721294e-01 -5.70422173e-01 4.42333728e-01 7.15041339e-01
-4.87181067e-01 -4.39401111e-03 -7.51967654e-02 9.91766036e-01
-5.29088005e-02 2.55859852e-01 8.75929058e-01 -1.25077680e-01
-3.48767847e-01 6.00356340e-01 -7.30011165e-01 -4.50376153e-01
1.18434560e+00 2.28982300e-01 -7.00066388e-02 6.40851036e-02
-1.06418598e+00 -3.32262143e-02 9.03413296e-01 5.79892397e-01
6.70400023e-01 -1.35930216e+00 -3.50668252e-01 4.19083774e-01
1.55734420e-01 -1.11038759e-01 3.56650829e-01 1.47221595e-01
-8.68427575e-01 5.01348734e-01 -3.72124314e-01 -5.59039891e-01
-1.10762584e+00 9.08256650e-01 2.76564658e-01 -1.37729213e-01
-7.06244111e-01 5.29507220e-01 7.39248455e-01 -6.55191600e-01
1.93540081e-02 1.00848945e-02 -4.40760031e-02 -4.13317293e-01
5.16493261e-01 -4.41328660e-02 -4.77467060e-01 -6.58327758e-01
-6.53731124e-03 4.95448649e-01 -4.50013340e-01 3.52768719e-01
1.62047231e+00 1.40222222e-01 -4.24043119e-01 2.69960105e-01
1.20781231e+00 -1.84887424e-01 -1.63123059e+00 -3.47542465e-01
3.08275409e-02 -3.18223890e-03 -2.84174770e-01 -3.44323307e-01
-6.11885309e-01 8.82166803e-01 6.44941568e-01 -3.63369137e-01
6.11647010e-01 7.19884783e-02 1.09226918e+00 8.62164855e-01
2.98524559e-01 -8.21249187e-01 2.25125983e-01 2.87169427e-01
5.17420769e-01 -1.07316899e+00 -9.91090387e-02 -9.47976410e-02
-4.55670953e-01 1.00824511e+00 4.22048420e-01 -5.81332482e-02
5.88795662e-01 5.24637252e-02 -1.98827848e-01 -2.72123247e-01
-8.33195686e-01 2.63334930e-01 -4.67195623e-02 5.95861614e-01
6.60948038e-01 2.98792243e-01 -3.13710272e-01 6.65116668e-01
1.94465026e-01 -1.40531406e-01 2.00999781e-01 9.30521965e-01
-7.61650264e-01 -1.64666200e+00 -6.93544513e-03 -1.97220743e-02
-4.71184492e-01 2.23521851e-02 -5.33079624e-01 7.99669251e-02
-9.60956514e-02 4.25115287e-01 -3.60649973e-01 -2.35615015e-01
1.24063805e-01 1.94485977e-01 3.15231234e-01 -5.93525946e-01
-7.96263441e-02 2.50755966e-01 -6.44163370e-01 -2.34094471e-01
-4.02211070e-01 -4.99652773e-01 -1.46798754e+00 -3.81845981e-01
-8.23903307e-02 5.77063262e-01 5.36184490e-01 7.83067942e-01
8.86286557e-01 6.08297706e-01 4.81211752e-01 -4.69153702e-01
-5.14295340e-01 -6.22541249e-01 -4.60725635e-01 5.59548974e-01
3.26972783e-01 -5.76479435e-01 1.71265855e-01 5.31141818e-01] | [4.6735734939575195, 5.695454120635986] |
9aa14daf-3747-40cb-bb34-f6b2b96c4204 | optimization-of-iot-enabled-physical-location | 2204.04664 | null | https://arxiv.org/abs/2204.04664v1 | https://arxiv.org/pdf/2204.04664v1.pdf | Optimization of IoT-Enabled Physical Location Monitoring Using DT and VAR | This study shows an enhancement of IoT that gets sensor data and performs real-time face recognition to screen physical areas to find strange situations and send an alarm mail to the client to make remedial moves to avoid any potential misfortune in the environment. Sensor data is pushed onto the local system and GoDaddy Cloud whenever the camera detects a person to optimize the physical location monitoring system by reducing the bandwidth requirement and storage cost onto the cloud using edge computation. The study reveals that decision tree (DT) and random forest give reasonably similar macro average f1-scores to predict a person using sensor data. Experimental results show that DT is the most reliable predictive model for the cloud datasets of three different physical locations to predict a person using timestamp with an accuracy of 83.99%, 88.92%, and 80.97%. This study also explains multivariate time series prediction using vector auto regression that gives reasonably good root mean squared error to predict temperature, humidity, light-dependent resistor, and gas time series. | ['Manoj Himmatrao Devare', 'Ajitkumar Sureshrao Shitole'] | 2022-04-10 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 1.48675805e-02 -3.99691463e-01 8.81315768e-03 -5.83058596e-01
2.90953308e-01 -2.13389561e-01 7.68761337e-02 -3.87744568e-02
-5.59934042e-02 7.55339622e-01 -1.96567163e-01 -3.08880270e-01
-3.54601026e-01 -1.17676687e+00 5.54874539e-02 -7.85832286e-01
3.40820625e-02 1.08896323e-01 -6.09267913e-02 1.72644779e-01
1.67493477e-01 8.77326310e-01 -1.88360286e+00 1.05996467e-01
5.51985443e-01 1.48658502e+00 -1.09718472e-01 7.63701200e-01
2.39948198e-01 5.26298225e-01 -9.30265903e-01 -2.25315511e-01
2.67307013e-01 1.41550884e-01 -1.38110772e-01 -1.69380531e-01
-1.07251987e-01 -1.97671771e-01 6.60597235e-02 4.89626616e-01
7.74816155e-01 6.03650101e-02 5.37515879e-01 -1.86078596e+00
-5.12118340e-01 -1.25341117e-01 -6.53232098e-01 4.88500118e-01
4.70847487e-01 -1.29948342e-02 -1.40610859e-01 -4.70847011e-01
-1.33788688e-02 7.68211544e-01 1.02674818e+00 3.75159234e-01
-5.46430230e-01 -1.06212342e+00 -3.13679367e-01 6.00465715e-01
-1.84713352e+00 -3.62462252e-01 6.48857176e-01 2.93913372e-02
1.51023066e+00 8.03577185e-01 8.87682021e-01 5.53031385e-01
3.69622558e-01 -2.80418426e-01 1.25302088e+00 -4.08881277e-01
3.28589201e-01 4.57943857e-01 -3.02871373e-02 6.54816568e-01
4.66761470e-01 2.22150888e-02 -7.56687641e-01 -2.47629613e-01
3.07903826e-01 6.78700566e-01 1.79771692e-01 7.12508678e-01
-3.89963269e-01 3.02552998e-01 2.16143548e-01 3.21064711e-01
-8.59130323e-01 -3.13064545e-01 -1.05451964e-01 1.16746582e-01
3.50212693e-01 -1.69689402e-01 -7.63232946e-01 -4.07690644e-01
-8.34445119e-01 -3.95193875e-01 7.71840274e-01 6.85211360e-01
3.31575722e-01 1.67157903e-01 1.76423207e-01 2.78720796e-01
3.12022299e-01 1.09875250e+00 3.84988904e-01 -7.20577896e-01
1.71569251e-02 8.84737730e-01 1.81875944e-01 -1.27036333e+00
-5.76182961e-01 -2.10586175e-01 -7.87710130e-01 5.20898327e-02
-1.19456992e-01 -5.72613835e-01 -5.15932560e-01 8.04895163e-01
6.35101080e-01 7.87613869e-01 -7.36806905e-05 6.40558004e-01
7.23896682e-01 6.35749042e-01 2.04960302e-01 -5.87444365e-01
1.45946169e+00 -3.82141948e-01 -7.93736041e-01 1.90143455e-02
1.38760448e-01 -5.17776549e-01 5.94781220e-01 3.72307986e-01
-5.68423092e-01 -3.36828649e-01 -7.74361551e-01 6.20669544e-01
-8.19468200e-01 1.73206210e-01 8.00051153e-01 1.42294860e+00
-9.86368656e-01 4.84014243e-01 -8.96404803e-01 -8.52861464e-01
2.40090281e-01 6.80378318e-01 9.27515104e-02 6.25460446e-02
-6.97528005e-01 7.61394799e-01 -1.66230142e-01 2.55065244e-02
-1.10897101e-01 -6.49408758e-01 -3.44290793e-01 -1.62625030e-01
-3.31719965e-01 -5.26171446e-01 5.45526922e-01 -6.72401488e-01
-1.43159497e+00 5.45933366e-01 -5.90024173e-01 -1.57519713e-01
6.45967647e-02 2.31622398e-01 -1.33579385e+00 7.39221200e-02
-1.66600242e-01 3.19862604e-01 7.97939837e-01 -4.37753707e-01
-7.01422870e-01 -9.13507342e-01 -6.92887187e-01 2.72197882e-03
-6.20672405e-01 2.96860963e-01 2.24342003e-01 -1.60358518e-01
-2.44715326e-02 -7.83153176e-01 1.03514388e-01 -3.82954776e-01
-2.43706703e-02 -4.05325681e-01 1.38716531e+00 -7.87585855e-01
1.29472709e+00 -1.82697260e+00 -9.46063757e-01 2.87748069e-01
-2.38001987e-01 1.85671151e-01 3.78588438e-01 1.39493525e-01
-1.05948314e-01 1.71698570e-01 3.89413923e-01 1.43540502e-01
-3.92792761e-01 1.40277013e-01 1.55798241e-01 4.14236188e-01
-5.67319915e-02 6.56599462e-01 -4.55486059e-01 -3.90757531e-01
4.51052278e-01 9.22630966e-01 8.66305009e-02 1.35150164e-01
4.44151789e-01 2.24150643e-01 -4.95159119e-01 1.15707755e+00
7.45069265e-01 -2.00198561e-01 -5.08104227e-02 4.10170071e-02
-1.34118363e-01 3.44904214e-02 -1.21294880e+00 7.68302321e-01
-3.71841252e-01 6.28121197e-01 -1.61508709e-01 -7.35479295e-01
1.10589588e+00 5.54034352e-01 9.12324488e-01 -7.43199289e-01
2.80678630e-01 -2.83516020e-01 -6.66554451e-01 -1.07758808e+00
1.95479736e-01 1.60742745e-01 1.95735246e-01 4.23194200e-01
-8.14484894e-01 5.63327432e-01 -6.47449553e-01 -2.75105059e-01
1.19576812e+00 -2.25476995e-01 1.85665920e-01 -4.99134064e-02
2.59276122e-01 -1.40001342e-01 4.47462201e-01 2.43827075e-01
-6.23213887e-01 -1.50141507e-01 -4.55795616e-01 -7.82864869e-01
-3.80552381e-01 -8.31175566e-01 -1.35357678e-01 9.40344870e-01
4.00185436e-02 -2.20894724e-01 -5.67440093e-01 -3.61764580e-01
1.31471902e-01 7.50378251e-01 -2.62351543e-01 -8.31142515e-02
-2.30332855e-02 -9.67209041e-01 2.48489603e-01 5.39336622e-01
8.48115802e-01 -8.40892136e-01 -1.24605632e+00 -2.16011316e-01
-1.82513773e-01 -9.10122275e-01 2.28051648e-01 -9.27941948e-02
-9.43899870e-01 -9.46314216e-01 6.24431781e-02 -4.35323000e-01
7.15633571e-01 5.23351729e-01 8.88945639e-01 2.98966885e-01
-5.56923330e-01 5.09075165e-01 -9.94373858e-03 -1.00035036e+00
6.02412999e-01 -4.24203753e-01 4.34047550e-01 -1.14755621e-02
1.33782840e+00 -6.44116879e-01 -8.35313678e-01 4.59592044e-01
-1.84381679e-01 -3.55879605e-01 -3.90399098e-02 -7.62665048e-02
4.30801511e-01 9.10823762e-01 5.64086854e-01 -1.49115518e-01
4.99837399e-01 -7.70589352e-01 -6.38311803e-01 4.04923230e-01
-1.25864780e+00 -5.31782091e-01 2.75360674e-01 -3.47162694e-01
-8.65054190e-01 2.20158964e-01 5.49114823e-01 -2.83240259e-01
-4.93964523e-01 -2.19656780e-01 3.21762636e-03 -1.58310235e-01
4.93734330e-01 4.27523144e-02 -4.55487192e-01 -3.30601037e-01
-4.26830053e-01 1.31312561e+00 2.41242394e-01 1.93289325e-01
6.43091083e-01 6.26306593e-01 2.61967003e-01 -9.92102742e-01
-1.95518225e-01 -4.85389471e-01 -6.32564902e-01 -7.69520640e-01
9.27037239e-01 -8.57035756e-01 -1.49755740e+00 3.89154166e-01
-6.75669432e-01 7.12324604e-02 1.95315868e-01 3.18440616e-01
3.43095005e-01 -3.18802267e-01 -9.52263996e-02 -1.62646091e+00
-6.03217185e-01 -3.34785700e-01 1.00668430e+00 9.06847537e-01
-2.70933479e-01 -7.71832347e-01 -4.67806995e-01 6.29725635e-01
6.72283173e-01 5.69685221e-01 1.48604602e-01 -2.22060814e-01
-3.94739479e-01 -6.02978289e-01 -1.53044090e-01 -3.57039958e-01
4.57515717e-01 3.00380141e-01 -1.29316616e+00 -8.19114074e-02
5.73551357e-02 5.04205167e-01 1.48857832e-02 4.67361361e-01
1.21938717e+00 -5.23214579e-01 -7.39581704e-01 6.60575569e-01
1.64079463e+00 6.07888997e-01 6.55759156e-01 1.32323623e-01
3.13880235e-01 5.38780093e-01 4.19232577e-01 8.21288705e-01
6.57799006e-01 4.34134454e-01 3.58087689e-01 2.88010478e-01
2.16068640e-01 -1.55663908e-01 4.16872352e-01 5.12689769e-01
-3.25554371e-01 -1.21601976e-01 -8.11628759e-01 3.48174900e-01
-1.54091012e+00 -1.08780503e+00 -3.84969234e-01 2.28223872e+00
1.38229534e-01 -3.53172898e-01 3.98323357e-01 5.49701333e-01
7.58541048e-01 -4.53151405e-01 -5.27016342e-01 -4.54149127e-01
1.65825978e-01 2.25360081e-01 8.85046661e-01 2.32977435e-01
-7.62099326e-01 4.90986019e-01 6.47645378e+00 1.30742535e-01
-1.41128206e+00 2.61603504e-01 8.77565742e-01 -5.38227916e-01
3.81469160e-01 -3.81684542e-01 -8.83170903e-01 9.45556164e-01
1.66834569e+00 7.34505197e-03 8.38770688e-01 9.80524302e-01
6.53234422e-01 -5.11740744e-01 -4.96263623e-01 1.56774366e+00
3.98197845e-02 -9.06965792e-01 -7.07473159e-01 1.48878366e-01
6.38661683e-01 -3.50655802e-02 -8.78385082e-02 -8.55095312e-02
1.74508095e-01 -1.12553239e+00 1.44632142e-02 1.01510191e+00
7.28438675e-01 -8.67537498e-01 6.85541630e-01 4.14302915e-01
-1.44895780e+00 -3.40396285e-01 -3.31229866e-01 -7.27616787e-01
-1.63564920e-01 6.28872097e-01 -1.11037505e+00 -6.69177845e-02
1.40016890e+00 2.32761234e-01 -6.45881414e-01 7.48130620e-01
2.53566522e-02 7.35983908e-01 -8.61323774e-01 -4.62703526e-01
-6.16190493e-01 -6.85561374e-02 2.14116678e-01 7.55615771e-01
6.80107892e-01 7.64066041e-01 -1.72937915e-01 1.30866259e-01
3.72989416e-01 -6.60997108e-02 -8.68295729e-01 3.96550417e-01
9.25469220e-01 1.26550019e+00 -8.66388559e-01 -1.22169450e-01
-2.54854172e-01 9.65175390e-01 -3.28677595e-01 3.07700932e-01
-8.64987850e-01 -4.44798112e-01 6.01461351e-01 2.78829992e-01
1.58608228e-01 -1.05866179e-01 -8.93159926e-01 -6.64162576e-01
1.21017084e-01 -9.81996208e-02 6.61649168e-01 -1.32417238e+00
-9.14788961e-01 4.67391640e-01 -2.10222051e-01 -9.42079425e-01
-1.29090652e-01 -4.33885455e-01 -7.37234294e-01 8.76078129e-01
-1.32828712e+00 -8.09648395e-01 -8.17807138e-01 1.07508624e+00
1.23526193e-02 -2.23181084e-01 1.12477481e+00 2.79390454e-01
-9.31427777e-01 6.38448656e-01 -1.10623680e-01 -1.23154104e-01
2.43362963e-01 -7.98007131e-01 -1.86152920e-01 8.78247559e-01
-3.92478168e-01 4.10905808e-01 4.71046984e-01 -9.37817574e-01
-1.58902931e+00 -1.18945706e+00 1.38822603e+00 -6.78317666e-01
5.60785830e-02 -7.81139284e-02 -3.98834378e-01 3.77303332e-01
7.19888136e-02 1.88603267e-01 1.03629327e+00 -8.21272731e-02
7.16459602e-02 -8.24653924e-01 -2.05740428e+00 3.06481481e-01
7.86163688e-01 -3.39907914e-01 1.52301282e-01 5.43419123e-01
3.61166477e-01 2.65769273e-01 -1.07045031e+00 1.07731573e-01
7.01120436e-01 -9.52277839e-01 7.47630179e-01 -3.13582361e-01
-4.17899609e-01 -3.69351655e-01 -2.77516425e-01 -7.23088145e-01
-3.00317287e-01 -8.09946716e-01 -3.14084262e-01 1.27584028e+00
3.21492732e-01 -9.81432319e-01 1.07683873e+00 1.46136189e+00
6.29521668e-01 -5.21369755e-01 -9.27327037e-01 -5.46498656e-01
-1.00474501e+00 -7.33600259e-01 1.33562374e+00 1.01530254e+00
-1.90071613e-02 1.63478628e-01 -2.41216570e-01 6.69477701e-01
6.42470062e-01 -6.19937778e-02 5.58346272e-01 -1.55202794e+00
3.04238617e-01 1.73795044e-01 -6.06661916e-01 3.69830541e-02
-3.89296502e-01 -3.39483082e-01 -6.59224689e-01 -1.43760896e+00
-1.68182820e-01 -4.49989170e-01 -4.63384092e-01 7.83696473e-01
1.36628062e-01 4.87592071e-01 -1.90583661e-01 -2.11424027e-02
-2.76207536e-01 -4.50432710e-02 4.84379083e-01 1.64977431e-01
-4.23093081e-01 5.10770082e-01 -5.79797089e-01 6.35926127e-01
1.07299006e+00 -5.37389755e-01 -4.98455495e-01 -3.50109696e-01
2.63351828e-01 2.56520748e-01 4.13433701e-01 -1.36699033e+00
4.67618197e-01 -5.66969156e-01 1.18373561e+00 -5.29647887e-01
4.33814108e-01 -1.64387751e+00 6.66184604e-01 5.49128413e-01
4.70615536e-01 6.64282203e-01 1.04258873e-01 3.97936255e-01
6.86100304e-01 2.24256143e-01 4.90810871e-01 3.64855677e-01
-5.27764142e-01 2.52994120e-01 -3.29429626e-01 -9.93269801e-01
1.51646698e+00 -9.19285715e-01 -3.72414947e-01 -5.14827251e-01
-5.98142385e-01 -1.05858020e-01 3.79873604e-01 4.61910695e-01
6.32409930e-01 -1.26338184e+00 -2.60409772e-01 6.50379837e-01
-2.48701766e-01 -6.87825382e-01 1.09587736e-01 6.61189973e-01
-3.83769542e-01 5.31976879e-01 -2.48068050e-01 -4.86051798e-01
-1.85484314e+00 4.37977254e-01 4.20558661e-01 3.31421375e-01
-3.19366336e-01 7.55960822e-01 -1.12796175e+00 1.19310059e-02
3.15220326e-01 -2.37447783e-01 -2.91925400e-01 -2.29467988e-01
8.51397514e-01 1.10158634e+00 3.91766787e-01 -6.37995541e-01
-9.39320683e-01 6.59306228e-01 6.80204153e-01 1.17489405e-01
1.49430943e+00 -4.17276829e-01 -2.07389414e-01 1.17047794e-01
8.98313761e-01 -1.21248991e-03 -9.80303168e-01 1.83478504e-01
-6.31156787e-02 -6.25616968e-01 1.74286559e-01 -1.26082778e+00
-1.29654860e+00 4.57992703e-01 1.48510730e+00 5.37136912e-01
1.90052044e+00 -4.63445842e-01 6.95304096e-01 2.45702863e-01
5.82844973e-01 -1.12870622e+00 -4.67715025e-01 -3.23860794e-02
2.23871693e-01 -1.22900105e+00 2.38861606e-01 -2.93941021e-01
-5.83938539e-01 8.77610505e-01 5.39244652e-01 2.59421915e-01
1.23311424e+00 5.94433963e-01 7.20839109e-03 -1.63156152e-01
-8.69357109e-01 1.77985728e-01 -1.11249082e-01 1.27487588e+00
3.81005742e-02 6.76798642e-01 1.56969890e-01 6.55950069e-01
-5.70443153e-01 3.21254015e-01 -1.56956562e-03 8.10530424e-01
-5.38244605e-01 -5.31087160e-01 -8.30337763e-01 7.34259784e-01
-5.86257756e-01 2.42711231e-01 -4.80213076e-01 2.17273518e-01
4.76662308e-01 1.87210310e+00 3.96496534e-01 -8.02352011e-01
3.75196636e-02 4.39567178e-01 1.09839462e-01 1.16516434e-01
-6.51722193e-01 -3.60979259e-01 -1.12374514e-01 -5.55957735e-01
-4.36698824e-01 -7.00989544e-01 -1.22064126e+00 -8.19157541e-01
-1.87681928e-01 1.73702672e-01 1.39019179e+00 8.89527440e-01
8.93153608e-01 4.06791687e-01 9.07122970e-01 -4.15071011e-01
4.06328797e-01 -9.43953514e-01 -4.67813075e-01 -2.16073692e-02
1.34896591e-01 -4.14742053e-01 -6.32865429e-02 2.53600627e-01] | [13.346487045288086, 1.4653000831604004] |
c88ed29f-5f2b-437c-a575-76c7d62aa46d | 190600544 | 1906.00544 | null | https://arxiv.org/abs/1906.00544v1 | https://arxiv.org/pdf/1906.00544v1.pdf | 3D Magic Mirror: Automatic Video to 3D Caricature Translation | Caricature is an abstraction of a real person which distorts or exaggerates certain features, but still retains a likeness. While most existing works focus on 3D caricature reconstruction from 2D caricatures or translating 2D photos to 2D caricatures, this paper presents a real-time and automatic algorithm for creating expressive 3D caricatures with caricature style texture map from 2D photos or videos. To solve this challenging ill-posed reconstruction problem and cross-domain translation problem, we first reconstruct the 3D face shape for each frame, and then translate 3D face shape from normal style to caricature style by a novel identity and expression preserving VAE-CycleGAN. Based on a labeling formulation, the caricature texture map is constructed from a set of multi-view caricature images generated by CariGANs. The effectiveness and efficiency of our method are demonstrated by comparison with baseline implementations. The perceptual study shows that the 3D caricatures generated by our method meet people's expectations of 3D caricature style. | ['Juyong Zhang', 'Yudong Guo', 'Luo Jiang', 'Lin Cai'] | 2019-06-03 | null | null | null | null | ['caricature'] | ['computer-vision'] | [ 3.56311172e-01 3.15378010e-01 1.82112023e-01 -4.16522115e-01
-4.54209089e-01 -9.06320274e-01 7.43795633e-01 -1.21720350e+00
4.33343589e-01 5.10030210e-01 4.23767060e-01 2.26767287e-01
5.35658896e-01 -5.89339375e-01 -7.76046813e-01 -5.46240330e-01
5.35505950e-01 4.71571982e-01 -5.67560077e-01 -4.06156778e-01
-2.73234732e-02 1.10068810e+00 -1.50148177e+00 3.80818039e-01
3.09686273e-01 7.71631062e-01 -2.97998376e-02 6.57907844e-01
-2.04589188e-01 4.49704826e-01 -5.57074666e-01 -8.85240436e-01
6.10914886e-01 -6.86936796e-01 -5.05212605e-01 7.28408813e-01
8.95674050e-01 -5.64995706e-01 -6.59767613e-02 1.17683983e+00
3.88464630e-01 -3.85180354e-01 7.09469020e-01 -1.32201374e+00
-1.17942119e+00 9.78494249e-03 -7.83604980e-01 -8.20133567e-01
6.84949815e-01 -2.44637534e-01 3.37253600e-01 -1.03791749e+00
9.88812923e-01 2.04909563e+00 7.27088213e-01 9.96573865e-01
-1.46985471e+00 -8.64940822e-01 -2.45902792e-01 -5.60215056e-01
-1.75089335e+00 -7.57461548e-01 1.23057163e+00 -2.96546251e-01
2.65712798e-01 5.05176485e-01 7.13518023e-01 1.31154871e+00
1.73497304e-01 6.80280447e-01 1.50573933e+00 -4.97726530e-01
-2.88184255e-01 4.56683546e-01 -6.56569183e-01 8.05269539e-01
-2.16575697e-01 3.49318385e-01 6.35472387e-02 -1.99270532e-01
1.49222827e+00 -2.51828600e-03 -1.82794467e-01 -4.49869126e-01
-1.08460033e+00 6.77594006e-01 -1.31450057e-01 1.29794627e-01
-8.82118046e-02 2.09000617e-01 -2.65526548e-02 4.43567902e-01
6.64983273e-01 1.64074332e-01 -1.23163648e-01 1.04566328e-01
-6.56548440e-01 4.29185957e-01 6.04168475e-01 1.36226392e+00
6.56520545e-01 3.67065847e-01 8.44586268e-02 9.79911089e-01
1.67411730e-01 1.07968211e+00 3.02295778e-02 -1.30251336e+00
-2.62083530e-01 3.02626014e-01 8.21823478e-02 -1.06675053e+00
2.78250337e-01 2.98090279e-01 -8.02304208e-01 7.76257336e-01
1.71678543e-01 -2.49090064e-02 -6.14977181e-01 1.66448414e+00
5.02030373e-01 -7.63755590e-02 9.12373289e-02 1.07927012e+00
9.15770710e-01 6.73150659e-01 -4.67849463e-01 -1.10353574e-01
1.40619731e+00 -6.70811236e-01 -1.23082387e+00 2.14438483e-01
1.69201925e-01 -1.29427898e+00 1.15731537e+00 3.73792619e-01
-1.47716951e+00 -6.67452931e-01 -7.81994760e-01 -4.21336085e-01
9.76698697e-02 3.63608509e-01 2.99255282e-01 1.10802519e+00
-1.28718567e+00 1.78842202e-01 -4.36935760e-02 -2.57191151e-01
4.77828830e-01 3.25841099e-01 -1.01599526e+00 -1.27267465e-01
-6.44078195e-01 9.25709248e-01 -5.42889759e-02 -1.12941474e-01
-8.11376512e-01 -5.85575163e-01 -9.75946605e-01 -4.76815552e-01
5.61564714e-02 -5.25520384e-01 1.02175701e+00 -1.63686717e+00
-2.06682563e+00 1.61492729e+00 -2.58733273e-01 1.57327056e-01
5.99356055e-01 -1.35923810e-02 -6.50631964e-01 5.23064360e-02
-2.30915412e-01 9.74632323e-01 1.44062614e+00 -1.65107536e+00
-2.68481284e-01 -3.86420935e-01 -4.83388826e-02 2.49664128e-01
1.53952599e-01 1.91950649e-01 -6.90750420e-01 -1.08642054e+00
-1.53923973e-01 -1.03872967e+00 5.22878885e-01 6.36893511e-01
-3.23300183e-01 1.90201044e-01 1.59370530e+00 -7.69015968e-01
5.23045182e-01 -2.39069438e+00 3.28512132e-01 7.45029300e-02
1.43241420e-01 5.73702119e-02 -2.83947408e-01 1.64443195e-01
-1.79823130e-01 9.39825326e-02 -1.15039244e-01 -6.97304308e-01
-2.97770798e-02 3.73748869e-01 -5.57554305e-01 6.71093702e-01
3.16465884e-01 9.66754138e-01 -5.33969760e-01 -5.95007956e-01
2.43450046e-01 1.00283754e+00 -7.83673763e-01 3.10022861e-01
7.93387145e-02 7.26396501e-01 -1.27109572e-01 9.00278628e-01
1.28808534e+00 4.86157715e-01 8.33701417e-02 -3.05289030e-01
-1.95387751e-02 -6.26112461e-01 -8.78056586e-01 1.82742810e+00
-4.96499836e-01 6.35947347e-01 5.54138601e-01 -4.60027039e-01
1.37708771e+00 4.22337145e-01 2.27931991e-01 -5.19632697e-01
3.52343321e-01 2.48586744e-01 -6.43312216e-01 -5.10595679e-01
4.45274562e-01 -7.81400502e-01 -3.10263067e-01 3.82454604e-01
-4.45778295e-02 -1.12559092e+00 -6.73125803e-01 -1.87206790e-01
1.46323338e-01 6.69520378e-01 1.63662419e-01 -3.38335395e-01
8.01480591e-01 -2.40421951e-01 2.73317188e-01 -6.40046671e-02
-2.27026399e-02 1.04135776e+00 4.27763432e-01 -6.93677127e-01
-1.49745846e+00 -1.14766026e+00 3.64357568e-02 3.81613910e-01
5.66243902e-02 9.73536596e-02 -1.33727026e+00 -5.17689943e-01
-6.01844341e-02 7.23727047e-01 -9.37693238e-01 -2.83576269e-02
-7.01048553e-01 1.46838740e-01 8.34763587e-01 -6.96824342e-02
7.66692519e-01 -7.71935105e-01 -1.70808762e-01 -3.33123028e-01
-2.17301458e-01 -1.17486131e+00 -1.27263880e+00 -9.50872123e-01
-5.18327475e-01 -8.03762794e-01 -8.86173844e-01 -1.01659811e+00
8.41244340e-01 4.91901577e-01 9.88592207e-01 -7.47907460e-02
-1.72939435e-01 5.86169064e-01 -2.36189887e-01 -3.90052319e-01
-8.35668683e-01 -7.89783001e-01 3.35541397e-01 5.81657708e-01
2.00322926e-01 -5.90908825e-01 -2.61957377e-01 3.99261326e-01
-1.00027943e+00 4.17569727e-01 3.57961297e-01 7.53049552e-01
7.11165130e-01 -1.87023655e-01 2.64421195e-01 -9.48841214e-01
4.96938616e-01 1.96784943e-01 -5.24692357e-01 1.61468893e-01
-6.22242838e-02 -5.68912625e-02 5.74252367e-01 -7.41084933e-01
-1.43470991e+00 2.50789076e-01 -1.37996867e-01 -1.13870966e+00
-2.33007684e-01 -6.18128181e-01 -5.92300057e-01 -2.51608074e-01
4.17198956e-01 2.32040927e-01 8.08233142e-01 -4.38316137e-01
7.53876328e-01 7.87150979e-01 1.01413727e+00 -7.21621573e-01
9.84410346e-01 1.06411707e+00 -1.40032738e-01 -9.09652710e-01
-3.46671283e-01 3.72801751e-01 -7.06492126e-01 -4.72672164e-01
9.57287192e-01 -1.02498055e+00 -9.33429360e-01 5.82322121e-01
-1.33781552e+00 -2.33429298e-02 -6.04242086e-01 4.51720841e-02
-8.53752553e-01 2.93543041e-01 -3.42638999e-01 -6.98759437e-01
-2.54638791e-01 -1.14506519e+00 1.66711366e+00 -1.77855492e-01
-3.80495667e-01 -8.86291802e-01 4.94987778e-02 6.04282737e-01
2.78506756e-01 7.54322886e-01 8.42459798e-01 3.25813144e-01
-3.27196777e-01 -2.27755427e-01 -8.36719722e-02 5.68519235e-01
3.99880350e-01 4.87936810e-02 -1.21014380e+00 -2.25278616e-01
1.89693838e-01 -1.79941058e-01 1.46417767e-01 1.82773590e-01
8.13780725e-01 -4.62717503e-01 -5.86843453e-02 8.69891763e-01
1.09747803e+00 2.13398531e-01 9.72478151e-01 -5.33359945e-01
5.81879497e-01 8.35707247e-01 4.16621476e-01 2.37186581e-01
2.25158095e-01 8.53073716e-01 4.03405100e-01 -2.63822138e-01
-6.44852817e-01 -8.34644854e-01 6.02920413e-01 5.80476284e-01
-1.60354197e-01 9.25268680e-02 -1.57601058e-01 3.13700795e-01
-1.08438075e+00 -9.85188842e-01 6.63138777e-02 1.85660303e+00
6.81820393e-01 -5.91651738e-01 5.42228110e-03 -7.17286533e-03
8.77173424e-01 -1.44787982e-01 -3.48560184e-01 -7.64437258e-01
-5.32013834e-01 1.75687537e-01 1.60195664e-01 6.63006842e-01
-6.57815874e-01 1.28013134e+00 6.67599249e+00 9.40340638e-01
-1.30619979e+00 1.22629419e-01 6.54220521e-01 -1.54117877e-02
-7.42767036e-01 -7.32196271e-02 -4.09137309e-01 1.07822031e-01
1.87752113e-01 7.91546926e-02 8.11409950e-01 6.35370851e-01
2.52295792e-01 3.56812775e-01 -1.04031599e+00 1.59591067e+00
5.26969790e-01 -1.41146708e+00 2.53100872e-01 2.05757007e-01
1.07053113e+00 -7.85052776e-01 5.23781836e-01 1.43362498e-02
3.09790283e-01 -1.21451545e+00 1.25842309e+00 5.98311901e-01
1.80630839e+00 -8.08640659e-01 2.17933550e-01 1.37504116e-01
-8.00155282e-01 3.61735880e-01 -2.13791490e-01 8.69044065e-02
8.22429135e-02 1.15200996e-01 -5.76001585e-01 2.18712106e-01
6.03660166e-01 4.89303738e-01 8.74336436e-03 -5.04082777e-02
3.29764076e-02 -5.62053770e-02 -1.07196032e-03 2.22670346e-01
-1.24759205e-01 -6.82622373e-01 5.66097677e-01 8.85547042e-01
7.95256972e-01 4.48358089e-01 -3.71528059e-01 1.21598935e+00
-3.37277293e-01 -3.75197157e-02 -1.28857863e+00 -2.08011065e-02
3.91471356e-01 1.13085842e+00 -1.27666518e-01 -2.92456329e-01
-3.25383961e-01 1.45840740e+00 -9.34327543e-02 1.80797264e-01
-7.71666467e-01 -1.84059247e-01 9.63014126e-01 2.62473285e-01
1.13758728e-01 -3.07592144e-03 -5.21433055e-02 -1.28018332e+00
-3.41020286e-01 -1.04714048e+00 -2.62861967e-01 -1.34732258e+00
-1.08523345e+00 8.41771424e-01 7.63774570e-03 -1.35499322e+00
-3.36068302e-01 -4.76049811e-01 -3.94923061e-01 7.33600974e-01
-1.03045237e+00 -1.84224439e+00 -1.87609792e-01 1.04741848e+00
4.68322903e-01 -4.95264679e-01 1.08691037e+00 1.05295442e-01
-2.49267798e-02 7.03359008e-01 -1.56578094e-01 -6.41460866e-02
8.70890975e-01 -6.13050699e-01 4.92710322e-01 4.57455814e-01
-1.52738718e-02 2.79157788e-01 7.97636569e-01 -3.93688291e-01
-1.87360704e+00 -1.10542870e+00 7.72365093e-01 -7.09259868e-01
-8.37436467e-02 -6.79798901e-01 -4.81816113e-01 9.30680156e-01
3.53031427e-01 1.38633028e-01 5.38611472e-01 -7.57400990e-01
-5.37039340e-01 -4.62899655e-02 -1.88419926e+00 8.64710033e-01
1.33623624e+00 -5.73294818e-01 -4.89594817e-01 -6.71152994e-02
4.53464299e-01 -5.39632738e-01 -8.83462667e-01 -5.32149673e-02
8.53435338e-01 -1.01440084e+00 9.56339836e-01 -1.73833013e-01
4.15427059e-01 -4.05220598e-01 -4.77299213e-01 -1.16332150e+00
-1.47851214e-01 -1.22161889e+00 3.03813070e-01 1.24796796e+00
-2.01699004e-01 -3.89963597e-01 5.72284579e-01 3.29403758e-01
-6.29560500e-02 -7.24593252e-02 -8.90154362e-01 -5.35789430e-01
2.61621892e-01 -2.21748427e-01 1.15568638e+00 1.24041378e+00
-3.89293045e-01 2.28378549e-01 -1.06920111e+00 -1.56733274e-01
6.76064909e-01 3.46963525e-01 1.39399731e+00 -9.18353319e-01
2.20235027e-02 -1.59561470e-01 -2.63723046e-01 -9.77821350e-01
5.22129416e-01 -9.29716647e-01 -3.97922665e-01 -6.14718199e-01
-5.68097793e-02 -2.85515636e-01 8.13594520e-01 8.50041136e-02
5.02855718e-01 7.61417329e-01 4.29485261e-01 1.22960545e-01
2.66447544e-01 5.62998354e-01 2.01786113e+00 -3.66790779e-02
-5.30545004e-02 -3.59724075e-01 -9.03721571e-01 8.18180442e-01
3.63083988e-01 -1.34089068e-01 -4.73662138e-01 -5.98810077e-01
-2.31889471e-01 4.28958654e-01 5.06324470e-01 -4.21386480e-01
-4.28139418e-01 -3.09144288e-01 4.53109175e-01 -4.45941985e-01
9.08222020e-01 -1.08273721e+00 9.46523428e-01 2.58915961e-01
-1.04798123e-01 3.48365903e-02 1.23777688e-01 2.82073140e-01
-2.57095635e-01 1.85426846e-01 1.09239280e+00 -2.33334422e-01
-4.24810231e-01 4.45874304e-01 -2.14415357e-01 -2.99864948e-01
1.20153296e+00 -5.54900944e-01 4.61601615e-02 -7.11147368e-01
-5.64581931e-01 -5.40856123e-01 1.05317581e+00 7.19609201e-01
1.01804519e+00 -2.08539128e+00 -9.64346826e-01 1.06100726e+00
-2.87030749e-02 -3.01495373e-01 3.31300646e-01 2.92402536e-01
-1.05463982e+00 1.88881099e-01 -6.51205719e-01 -4.88680750e-01
-1.53699005e+00 6.07865751e-01 5.33987403e-01 5.26638865e-01
-7.53158092e-01 6.60324633e-01 7.06550181e-01 -7.53309369e-01
-3.73027742e-01 -6.99013146e-03 6.03517480e-02 -3.90626967e-01
5.23863316e-01 -9.15265158e-02 -4.65232700e-01 -1.31305182e+00
-9.98253599e-02 1.13136888e+00 2.38447741e-01 -3.90047669e-01
1.15163350e+00 -3.28386128e-01 -2.60203660e-01 6.85190558e-02
1.58466017e+00 4.47852522e-01 -1.29126894e+00 -1.11580223e-01
-9.14485157e-01 -1.10238624e+00 -1.40344843e-01 -4.24150229e-01
-1.27739537e+00 8.06443393e-01 6.00976646e-01 -4.05761451e-01
1.32109857e+00 6.27837703e-02 8.42105448e-01 -7.28840157e-02
4.56781477e-01 -8.00921679e-01 3.49492915e-02 1.37591898e-01
1.59954524e+00 -8.70535970e-01 -3.32324982e-01 -5.75513184e-01
-9.37021077e-01 1.12737107e+00 1.90050945e-01 -4.16051149e-01
6.25916958e-01 4.12739992e-01 3.32328111e-01 -3.92748863e-02
-5.07346630e-01 4.13639277e-01 1.54837474e-01 9.49893177e-01
1.15181871e-01 2.21465886e-01 5.96590415e-02 3.15729052e-01
-6.08112454e-01 2.89909337e-02 6.09005332e-01 4.71572787e-01
3.20995860e-02 -1.20322847e+00 -9.08897400e-01 -4.17447388e-01
-5.04207551e-01 3.30852032e-01 -7.31994331e-01 9.63398457e-01
3.81401330e-01 7.59586930e-01 4.33968216e-01 -4.52235073e-01
5.26889265e-01 -1.27211306e-02 1.10393775e+00 -1.61446065e-01
-3.45195800e-01 5.18243730e-01 9.41037294e-03 -2.95761049e-01
-5.55926323e-01 -5.83930612e-01 -8.61567497e-01 -1.03918993e+00
7.29391500e-02 -2.43084684e-01 5.63366294e-01 4.31226492e-01
4.01665717e-01 -1.30291492e-01 9.44928527e-01 -1.02226949e+00
-5.72728142e-02 -6.63254738e-01 -8.40008616e-01 8.43809903e-01
3.84089500e-01 -5.35119355e-01 -1.06441289e-01 6.25144660e-01] | [12.689620018005371, -0.2617032527923584] |
8b0abb56-c0c0-40ca-a1d9-57df6a6d8a48 | towards-label-free-scene-understanding-by | 2306.03899 | null | https://arxiv.org/abs/2306.03899v1 | https://arxiv.org/pdf/2306.03899v1.pdf | Towards Label-free Scene Understanding by Vision Foundation Models | Vision foundation models such as Contrastive Vision-Language Pre-training (CLIP) and Segment Anything (SAM) have demonstrated impressive zero-shot performance on image classification and segmentation tasks. However, the incorporation of CLIP and SAM for label-free scene understanding has yet to be explored. In this paper, we investigate the potential of vision foundation models in enabling networks to comprehend 2D and 3D worlds without labelled data. The primary challenge lies in effectively supervising networks under extremely noisy pseudo labels, which are generated by CLIP and further exacerbated during the propagation from the 2D to the 3D domain. To tackle these challenges, we propose a novel Cross-modality Noisy Supervision (CNS) method that leverages the strengths of CLIP and SAM to supervise 2D and 3D networks simultaneously. In particular, we introduce a prediction consistency regularization to co-train 2D and 3D networks, then further impose the networks' latent space consistency using the SAM's robust feature representation. Experiments conducted on diverse indoor and outdoor datasets demonstrate the superior performance of our method in understanding 2D and 3D open environments. Our 2D and 3D network achieves label-free semantic segmentation with 28.4% and 33.5% mIoU on ScanNet, improving 4.7% and 7.9%, respectively. And for nuScenes dataset, our performance is 26.8% with an improvement of 6%. Code will be released (https://github.com/runnanchen/Label-Free-Scene-Understanding). | ['Wenping Wang', 'Tongliang Liu', 'Yuexin Ma', 'Xinge Zhu', 'Nenglun Chen', 'Lingdong Kong', 'Youquan Liu', 'Runnan Chen'] | 2023-06-06 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [ 3.94021988e-01 2.69117713e-01 -1.44392177e-01 -6.94026411e-01
-6.65302157e-01 -5.71013749e-01 6.16246343e-01 -3.38403970e-01
-4.66412842e-01 3.28253448e-01 -9.18061808e-02 -3.58170152e-01
2.11677983e-01 -5.44997692e-01 -8.90533447e-01 -4.05230880e-01
2.21461102e-01 4.20485079e-01 2.73801029e-01 4.65563275e-02
-2.67361403e-01 2.79203653e-01 -1.50821555e+00 4.54006493e-02
9.06734705e-01 9.34995711e-01 5.01729846e-01 5.25392592e-01
-2.25197256e-01 6.92869544e-01 -2.41075650e-01 -1.10354744e-01
5.19598186e-01 -1.12177782e-01 -8.59411955e-01 4.13700312e-01
7.41882980e-01 -5.66801012e-01 -2.48627603e-01 1.10433269e+00
2.88175166e-01 1.43909290e-01 4.49015290e-01 -1.23678827e+00
-7.41419554e-01 2.94800967e-01 -5.82460463e-01 -6.05698600e-02
1.36869073e-01 3.54381084e-01 8.26096773e-01 -8.04534793e-01
7.58578718e-01 1.30144894e+00 6.83752954e-01 7.83671796e-01
-1.31697130e+00 -6.78705871e-01 4.83749241e-01 -7.03076497e-02
-1.20465875e+00 -4.97861832e-01 7.05973744e-01 -4.92377460e-01
9.26410377e-01 -1.05950935e-02 6.19344115e-01 1.38826799e+00
-2.97897279e-01 9.91253138e-01 1.36781228e+00 -2.42260054e-01
8.55021998e-02 1.50907561e-01 3.06015164e-01 7.77567506e-01
2.35479660e-02 1.99990928e-01 -4.41855460e-01 4.85867292e-01
8.23935747e-01 -8.86442214e-02 -9.53945220e-02 -2.63917029e-01
-1.07945442e+00 6.18866980e-01 7.82608211e-01 -2.98218019e-02
-1.32714137e-01 3.15287977e-01 1.55395851e-01 -4.22811136e-02
8.23846400e-01 2.44623169e-01 -4.83679563e-01 1.26900569e-01
-9.98933613e-01 3.04612983e-02 6.38113022e-01 1.25087118e+00
7.97669351e-01 -6.58417866e-02 2.97066271e-02 9.60160375e-01
6.27501369e-01 7.03454375e-01 -8.78244489e-02 -1.41970694e+00
2.98215747e-01 4.38196659e-01 -8.33324417e-02 -5.94887555e-01
-4.97411788e-01 -6.10067010e-01 -7.93482423e-01 1.48034871e-01
3.75043690e-01 -2.79601365e-02 -1.66525543e+00 1.93018472e+00
4.48184311e-01 5.12039721e-01 8.64962414e-02 1.18369031e+00
1.04499304e+00 4.70776200e-01 1.98431760e-01 2.79231966e-01
1.16009390e+00 -1.32750392e+00 -3.95058513e-01 -7.55727291e-01
5.46828926e-01 -5.57511568e-01 1.08434558e+00 9.79623199e-03
-1.04898071e+00 -7.62918532e-01 -9.91198123e-01 -4.31605577e-01
-2.43678197e-01 -1.41734472e-02 6.64879084e-01 3.82484257e-01
-1.19582188e+00 2.94263393e-01 -1.15693915e+00 -4.80254829e-01
9.83765602e-01 1.67854875e-01 -4.09366667e-01 -5.26363134e-01
-9.25619125e-01 7.68718183e-01 4.91400957e-01 6.34044707e-02
-1.33184838e+00 -8.54449630e-01 -9.93976474e-01 -3.49868983e-01
4.90888745e-01 -9.77545083e-01 1.21923876e+00 -7.99986839e-01
-1.30108023e+00 1.16241717e+00 -2.50140965e-01 -3.70901227e-01
6.36177480e-01 -3.44006687e-01 -6.40234817e-03 3.03252101e-01
5.41345417e-01 1.39818788e+00 5.17398715e-01 -1.64993811e+00
-4.92072582e-01 -4.72786278e-01 3.19021583e-01 4.30213004e-01
1.79281682e-01 -3.35278541e-01 -8.89119029e-01 -3.11803788e-01
5.25426984e-01 -1.06132889e+00 -3.16535890e-01 3.11075866e-01
-7.21082330e-01 -1.49295896e-01 6.84209764e-01 -5.54120660e-01
3.30290824e-01 -2.13289118e+00 -4.42320332e-02 -1.55769605e-02
3.71952772e-01 2.93676525e-01 -4.16216016e-01 -1.46562442e-01
2.12401122e-01 3.30749363e-01 -4.93157059e-01 -9.47089255e-01
3.85347828e-02 6.83815360e-01 -6.55770525e-02 3.66675317e-01
5.41770518e-01 1.05772877e+00 -9.66727078e-01 -3.63577724e-01
4.91053730e-01 6.26335979e-01 -4.29806501e-01 1.48598716e-01
-4.69992459e-01 8.42141688e-01 -4.42078888e-01 6.96514368e-01
8.27274203e-01 -4.88234878e-01 -1.17218502e-01 -2.42176086e-01
-1.17460430e-01 3.23406249e-01 -9.75013077e-01 2.17624736e+00
-4.76321995e-01 5.36155045e-01 2.40633503e-01 -8.54025960e-01
7.65530229e-01 4.59403247e-02 1.97263092e-01 -1.00192785e+00
1.50472417e-01 1.44660935e-01 -4.19474363e-01 -5.28446198e-01
1.67487979e-01 -8.22775811e-02 7.44682774e-02 2.75621206e-01
2.83625573e-01 -4.02374834e-01 1.75262883e-01 3.39238465e-01
8.93264949e-01 4.21842247e-01 -3.61435413e-01 -1.63927525e-02
1.64985687e-01 1.86079934e-01 6.10724986e-01 1.02656925e+00
-4.73521829e-01 7.33446360e-01 2.46630460e-01 -8.56956691e-02
-9.98149872e-01 -1.25533783e+00 -1.71232879e-01 9.02916968e-01
4.91631866e-01 -1.19285546e-02 -7.11977661e-01 -7.85881460e-01
-6.36851788e-02 8.25694978e-01 -4.40558702e-01 -3.37308981e-02
-2.68162876e-01 -3.68870944e-01 6.80376589e-01 6.13967001e-01
1.07311499e+00 -5.94035864e-01 -3.16795290e-01 -9.91069824e-02
-2.76577085e-01 -1.63741934e+00 -1.07619993e-01 3.19617540e-01
-6.36872590e-01 -9.52462733e-01 -5.23928761e-01 -8.19219887e-01
7.24547446e-01 6.80143833e-01 1.04340529e+00 1.45788621e-02
-1.26508966e-01 5.50175846e-01 -2.41253257e-01 -3.64639938e-01
-3.40400457e-01 1.46532550e-01 1.65698212e-02 -3.04154813e-01
2.54351526e-01 -6.12519205e-01 -5.24123073e-01 4.16992813e-01
-8.63366604e-01 5.85957289e-01 4.88065541e-01 6.58532381e-01
8.38009417e-01 -3.08814168e-01 3.85026008e-01 -9.46443200e-01
6.48212340e-03 -5.09016275e-01 -4.44129169e-01 7.47546032e-02
-5.23003399e-01 -1.70829043e-01 2.86239773e-01 -1.37975633e-01
-1.29061651e+00 2.31931984e-01 -3.29567939e-01 -6.12471879e-01
-7.86479056e-01 3.22823942e-01 -1.74026057e-01 -7.90787712e-02
5.30183733e-01 2.06139516e-02 4.56262231e-02 -5.13139069e-01
7.22080767e-01 5.93173623e-01 7.05935538e-01 -5.74419498e-01
8.94814730e-01 8.58510792e-01 -1.89172536e-01 -7.51099288e-01
-1.59609175e+00 -7.65562236e-01 -7.70097256e-01 -2.65408814e-01
1.30715978e+00 -1.41096282e+00 -1.82733104e-01 6.45644546e-01
-1.09748757e+00 -7.53272891e-01 -2.89777607e-01 3.83415550e-01
-4.93666351e-01 3.12209755e-01 -4.82520133e-01 -5.41783631e-01
-6.23115003e-02 -1.23508215e+00 1.23037493e+00 3.79280031e-01
3.35663632e-02 -1.06182575e+00 -2.85879672e-01 1.00757289e+00
2.02010036e-01 2.57547885e-01 6.05025589e-01 -5.32041728e-01
-7.57516801e-01 1.98094442e-01 -8.56501698e-01 7.17520058e-01
-1.39200211e-01 -4.03732687e-01 -1.42740643e+00 -9.07810107e-02
-2.01579947e-02 -6.81277931e-01 1.14832783e+00 4.27322179e-01
1.04751241e+00 2.30883554e-01 -2.33902082e-01 9.48411107e-01
1.32290351e+00 -2.23306552e-01 2.29429349e-01 2.19288960e-01
1.06318629e+00 7.06443548e-01 4.84629333e-01 -1.78209618e-02
7.54900217e-01 3.81521344e-01 7.05992222e-01 -2.06649676e-01
-6.88826740e-01 -4.09591585e-01 1.51069865e-01 6.53879225e-01
1.88261241e-01 -2.19140902e-01 -1.10050273e+00 6.62810087e-01
-1.81655824e+00 -5.09462416e-01 -3.61793250e-01 1.59301937e+00
7.74854183e-01 3.58234197e-01 -2.98513800e-01 -3.11699688e-01
4.70378757e-01 2.58704543e-01 -8.79720867e-01 3.01019475e-03
-2.57092237e-01 -3.57026681e-02 6.50865495e-01 6.05313659e-01
-1.26687968e+00 1.33625329e+00 4.93002558e+00 6.94770455e-01
-8.88536811e-01 3.77331644e-01 7.37713575e-01 -1.28347725e-01
-3.25650394e-01 8.26449543e-02 -8.44340324e-01 2.19506964e-01
7.13778377e-01 5.43900192e-01 2.70882726e-01 6.74411476e-01
3.87426674e-01 -2.69545257e-01 -1.08348668e+00 9.48622584e-01
3.94782275e-02 -1.21339095e+00 -3.58165540e-02 -3.41158584e-02
1.12071145e+00 7.19829798e-01 5.39202131e-02 3.24633777e-01
3.58655512e-01 -1.04299021e+00 9.42886770e-01 4.81262296e-01
9.43996072e-01 -3.47664416e-01 6.35210931e-01 6.13880157e-01
-8.30619097e-01 2.15564653e-01 -2.61297613e-01 -7.27171600e-02
4.06633854e-01 7.59824395e-01 -8.12050104e-01 5.15704989e-01
8.18264902e-01 1.02573502e+00 -5.78687072e-01 7.71316707e-01
-5.68706930e-01 7.92380691e-01 -6.25511706e-01 4.85487908e-01
5.90575874e-01 -1.72799900e-01 5.85632205e-01 1.10091209e+00
-3.05124151e-04 -2.16150600e-02 4.65351760e-01 1.34092951e+00
-1.83837846e-01 -4.82910633e-01 -5.79080164e-01 1.90205082e-01
4.23234969e-01 1.10795391e+00 -7.83517361e-01 -2.36555696e-01
-3.82420301e-01 1.07292724e+00 2.01423675e-01 5.88722169e-01
-9.50619280e-01 7.51190335e-02 5.99953711e-01 -1.45074517e-01
1.96154818e-01 -5.27269304e-01 -6.36279464e-01 -1.13948071e+00
9.09729768e-03 -3.89672726e-01 -4.08330113e-02 -1.02005470e+00
-1.40514290e+00 3.23966146e-01 4.44338247e-02 -7.01084971e-01
1.67513847e-01 -6.13464236e-01 -3.77084851e-01 8.20815384e-01
-1.92120981e+00 -1.60771561e+00 -5.74907243e-01 5.63839614e-01
6.59761548e-01 3.02604169e-01 6.00760818e-01 3.06085110e-01
-6.39454663e-01 3.55625153e-01 1.10168522e-02 1.36696890e-01
5.42966545e-01 -1.32971501e+00 5.81899047e-01 8.81944418e-01
2.54700571e-01 3.46776932e-01 3.79083782e-01 -6.40515447e-01
-1.01723170e+00 -1.52509964e+00 6.39160275e-01 -6.29765987e-01
4.35107201e-01 -5.21174967e-01 -8.21152389e-01 7.13399291e-01
-9.81831849e-02 1.76086292e-01 5.49940467e-01 8.43215212e-02
-6.76023126e-01 7.97426626e-02 -1.09624851e+00 5.92142701e-01
1.57812047e+00 -7.47330964e-01 -3.84681851e-01 4.56871867e-01
1.15777493e+00 -4.58358288e-01 -7.21507847e-01 4.50678736e-01
2.41402477e-01 -9.79668558e-01 1.12518132e+00 -3.19123209e-01
4.16655362e-01 -3.59439433e-01 -3.89701515e-01 -1.00233734e+00
-9.78570879e-02 -2.30150849e-01 7.37082213e-02 1.22265100e+00
4.00564402e-01 -4.19814765e-01 7.62387872e-01 6.34272158e-01
-5.67851067e-01 -6.81643128e-01 -8.44329834e-01 -7.32356668e-01
-1.31832128e-02 -1.11143017e+00 7.75236115e-02 9.34305668e-01
-8.27501833e-01 5.01339316e-01 -8.02905709e-02 4.45494354e-01
8.73212576e-01 -1.31504864e-01 6.61418140e-01 -1.11653936e+00
-1.16880342e-01 -2.29605943e-01 -2.55329788e-01 -1.46677136e+00
4.27738547e-01 -1.27142417e+00 1.64933369e-01 -1.91653895e+00
2.02767566e-01 -6.57520175e-01 -1.34902820e-01 6.73250794e-01
1.10061085e-02 5.98642945e-01 3.99461389e-01 3.03062141e-01
-1.01008189e+00 5.93448400e-01 1.44631410e+00 -2.89873570e-01
-7.14855045e-02 -9.27318186e-02 -6.65895522e-01 9.19823885e-01
8.96396339e-01 -3.49016726e-01 -6.21149540e-01 -1.01304960e+00
-2.07423978e-02 -4.92592186e-01 9.86495376e-01 -9.72103655e-01
1.69499919e-01 5.63493334e-02 3.12276483e-01 -6.22711241e-01
4.65704560e-01 -6.83569908e-01 -9.74625275e-02 3.40345427e-02
-3.04246247e-01 -6.64457023e-01 2.94709563e-01 7.65992522e-01
-2.18211897e-02 -1.00206263e-01 8.01876307e-01 -2.73355275e-01
-1.11709762e+00 3.32642764e-01 -1.14449416e-03 3.25879961e-01
8.87377918e-01 -3.17592889e-01 -5.31866968e-01 -1.77899569e-01
-9.23926294e-01 6.46410525e-01 5.41573226e-01 4.91537154e-01
7.27187395e-01 -9.90024030e-01 -6.02975667e-01 2.62735337e-01
1.78420156e-01 8.04447234e-01 5.30318797e-01 8.72016847e-01
-4.90647227e-01 3.43947113e-01 -4.11887243e-02 -1.12791646e+00
-9.43320811e-01 -3.61902043e-02 4.75201845e-01 -2.06373911e-03
-6.13693357e-01 1.26868272e+00 2.70375013e-01 -8.74686658e-01
3.27780038e-01 -3.83112371e-01 1.80100113e-01 -1.94990277e-01
8.91099945e-02 1.59341618e-01 -1.62752926e-01 -6.08192623e-01
-3.15207005e-01 5.86590052e-01 -2.03841671e-01 -1.24269992e-01
1.33436227e+00 -5.25881827e-01 -4.33569215e-02 5.38681209e-01
1.36165142e+00 -6.28491640e-01 -1.89864779e+00 -4.13845956e-01
-1.75181895e-01 -1.81020841e-01 3.01355749e-01 -1.08023822e+00
-1.02781320e+00 1.01674449e+00 7.62805521e-01 -3.36574018e-02
9.98739958e-01 4.26555455e-01 9.55146730e-01 2.66090900e-01
2.26421595e-01 -1.00803375e+00 6.02416694e-02 9.24903989e-01
4.58427638e-01 -1.69621563e+00 -2.82422900e-01 -4.68713254e-01
-5.37433684e-01 5.46937764e-01 8.06267679e-01 -6.74684495e-02
5.67124844e-01 1.00575266e-02 3.53662044e-01 -3.77684563e-01
-5.28440833e-01 -5.97529113e-01 2.58524060e-01 8.05867732e-01
7.54341707e-02 1.07882947e-01 4.56502914e-01 4.24203873e-01
-3.93280312e-02 -1.29700705e-01 2.86361843e-01 7.63282716e-01
-3.28712314e-01 -7.94967175e-01 -3.93382609e-02 2.38961190e-01
-1.21816806e-01 -6.88641146e-02 -2.82963991e-01 7.08851755e-01
6.30551457e-01 1.13409662e+00 8.20377655e-03 -3.09711158e-01
2.82975554e-01 8.68525729e-02 4.03683245e-01 -9.21921432e-01
-1.60984844e-01 1.62636772e-01 1.40724406e-01 -7.60542154e-01
-8.34565401e-01 -3.96857053e-01 -1.39700198e+00 -2.14367150e-03
-3.06255996e-01 -4.10222918e-01 9.68060136e-01 1.09798670e+00
4.93105948e-01 6.78329170e-01 2.89728135e-01 -9.38345969e-01
-3.41592193e-01 -7.06116974e-01 -3.77172083e-01 4.34275448e-01
4.22637314e-01 -7.21606851e-01 -4.11046565e-01 1.04809582e-01] | [9.643895149230957, 0.705441415309906] |
e284fd59-6e36-4f94-8a6c-e7221c8be763 | defect-transformer-an-efficient-hybrid | 2207.08319 | null | https://arxiv.org/abs/2207.08319v1 | https://arxiv.org/pdf/2207.08319v1.pdf | Defect Transformer: An Efficient Hybrid Transformer Architecture for Surface Defect Detection | Surface defect detection is an extremely crucial step to ensure the quality of industrial products. Nowadays, convolutional neural networks (CNNs) based on encoder-decoder architecture have achieved tremendous success in various defect detection tasks. However, due to the intrinsic locality of convolution, they commonly exhibit a limitation in explicitly modeling long-range interactions, critical for pixel-wise defect detection in complex cases, e.g., cluttered background and illegible pseudo-defects. Recent transformers are especially skilled at learning global image dependencies but with limited local structural information necessary for detailed defect location. To overcome the above limitations, we propose an efficient hybrid transformer architecture, termed Defect Transformer (DefT), for surface defect detection, which incorporates CNN and transformer into a unified model to capture local and non-local relationships collaboratively. Specifically, in the encoder module, a convolutional stem block is firstly adopted to retain more detailed spatial information. Then, the patch aggregation blocks are used to generate multi-scale representation with four hierarchies, each of them is followed by a series of DefT blocks, which respectively include a locally position-aware block for local position encoding, a lightweight multi-pooling self-attention to model multi-scale global contextual relationships with good computational efficiency, and a convolutional feed-forward network for feature transformation and further location information learning. Finally, a simple but effective decoder module is proposed to gradually recover spatial details from the skip connections in the encoder. Extensive experiments on three datasets demonstrate the superiority and efficiency of our method compared with other CNN- and transformer-based networks. | ['Zhengsheng Wang', 'Jinjin Wang', 'Fuju Yan', 'Guili Xu', 'Junpu Wang'] | 2022-07-17 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 2.32736930e-01 -2.59334117e-01 2.35825807e-01 -2.45973751e-01
-5.89241624e-01 1.97713211e-01 1.65245086e-01 1.30925581e-01
3.09968852e-02 2.16017216e-01 4.88677435e-02 -2.28645802e-02
-1.49770640e-02 -1.16984153e+00 -8.18039715e-01 -9.06950057e-01
1.72527313e-01 4.91140075e-02 7.57439852e-01 -2.23385647e-01
3.29987586e-01 2.95662135e-01 -1.51696002e+00 5.89953363e-01
1.05425501e+00 1.46858001e+00 6.87497795e-01 1.60885304e-01
-2.47744516e-01 8.18458498e-01 -3.41466904e-01 9.85949580e-03
-1.77915841e-01 -3.15023631e-01 -4.66135919e-01 3.53939503e-01
3.94758768e-02 -3.19097102e-01 -4.55496162e-01 1.00772715e+00
6.01268589e-01 -2.53735483e-01 4.23282534e-01 -6.24843538e-01
-8.25226009e-01 1.55868694e-01 -7.57516503e-01 1.24001905e-01
2.38668934e-01 2.91358471e-01 1.05421281e+00 -1.11347699e+00
3.02170008e-01 1.29298103e+00 6.98040783e-01 1.12168603e-01
-9.24289882e-01 -5.15869379e-01 3.61350119e-01 2.79294133e-01
-1.28459346e+00 -6.51379079e-02 1.31659627e+00 -2.19932437e-01
1.17619014e+00 -3.63159887e-02 6.74828172e-01 7.72910893e-01
5.76437593e-01 7.34579742e-01 6.22274816e-01 -1.36749923e-01
-8.21441934e-02 -3.75904411e-01 -1.48832902e-01 9.27848220e-01
1.79531068e-01 -2.71517001e-02 -5.64099252e-01 1.76841199e-01
1.28877699e+00 4.56740588e-01 -2.34769449e-01 -3.36353630e-01
-1.10427403e+00 7.05177069e-01 9.88226950e-01 5.52946210e-01
-4.58322525e-01 1.20395988e-01 5.50245583e-01 3.66251469e-01
6.18545771e-01 1.16173863e-01 -5.02902389e-01 3.45677882e-01
-5.47918558e-01 5.43382904e-03 2.42499813e-01 7.73734987e-01
9.98784065e-01 -9.01027918e-02 -1.65800452e-01 1.19592142e+00
3.75912011e-01 1.61626607e-01 5.64448118e-01 -2.89492637e-01
6.48852885e-01 1.25809228e+00 -1.71126455e-01 -1.26139057e+00
-3.39113533e-01 -7.35366642e-01 -1.22422945e+00 9.87093970e-02
-2.45625526e-01 3.45897526e-01 -1.02553451e+00 1.15548110e+00
1.93828061e-01 1.82910666e-01 -2.59915084e-01 8.04428995e-01
8.83867681e-01 6.74097061e-01 -4.03352678e-02 -8.07066411e-02
1.38698792e+00 -1.06797147e+00 -5.13727248e-01 -3.21697235e-01
7.04744458e-01 -8.38327467e-01 1.07863724e+00 3.13104242e-01
-1.06056213e+00 -7.98681259e-01 -1.19296575e+00 -3.64233881e-01
-2.51691520e-01 4.09903318e-01 5.78941464e-01 1.95424724e-02
-7.55866528e-01 5.61179340e-01 -9.32933152e-01 -9.92721468e-02
8.12784433e-01 4.71485436e-01 -2.56296486e-01 -5.09863317e-01
-9.97767150e-01 3.70783240e-01 1.11868739e-01 5.46126425e-01
-9.61187720e-01 -5.50108373e-01 -1.07501614e+00 1.36655241e-01
2.58463323e-01 -6.58246160e-01 7.69025564e-01 -9.59364474e-01
-1.42795897e+00 6.22907579e-01 -1.30890101e-01 7.23956618e-03
2.49620140e-01 -1.64992660e-01 -3.30771953e-01 3.20712365e-02
3.16426605e-01 3.06936741e-01 8.42610180e-01 -1.16173363e+00
-7.15510547e-01 -4.57439512e-01 8.12184662e-02 2.62419522e-01
-4.14562464e-01 -4.55131941e-02 -6.57697797e-01 -8.44837308e-01
7.26804197e-01 -3.43458265e-01 -3.40124667e-01 1.01797603e-01
-4.97231990e-01 -2.91124105e-01 1.09146976e+00 -7.25244761e-01
1.32443655e+00 -2.30590034e+00 1.89944476e-01 1.34726584e-01
3.17248225e-01 1.62269995e-01 -2.65546858e-01 5.03026903e-01
-2.44244739e-01 -1.35453761e-01 -4.91911381e-01 -3.98789465e-01
-3.05851310e-01 1.50850192e-01 1.23507820e-01 3.31264496e-01
5.94178736e-01 9.71908271e-01 -7.64623106e-01 -4.26218778e-01
4.48189616e-01 4.79503572e-01 -6.01747811e-01 3.59742463e-01
-2.55356163e-01 4.91136640e-01 -7.82405734e-01 9.86026347e-01
9.59471464e-01 -3.59257072e-01 -2.21943736e-01 -5.49144924e-01
-2.86697179e-01 3.40745956e-01 -1.00037539e+00 1.94256949e+00
-7.26066887e-01 7.91710839e-02 2.67661065e-01 -1.32328522e+00
1.07964385e+00 1.80796757e-01 5.23199320e-01 -1.03105295e+00
1.01424575e-01 4.09550250e-01 -1.54450342e-01 -6.43352032e-01
-3.45610119e-02 -1.97740853e-01 1.22385450e-01 1.14342496e-01
-1.37147263e-01 -4.68715392e-02 -2.53232569e-01 -1.27209425e-01
1.35997093e+00 3.06322388e-02 -1.32138744e-01 -1.69252664e-01
8.52539837e-01 -3.31635535e-01 7.53045619e-01 9.82186273e-02
1.31386444e-01 9.82128263e-01 4.72709447e-01 -7.43255496e-01
-7.89727449e-01 -7.50469208e-01 1.11854738e-02 5.05979180e-01
6.27906680e-01 -4.48199451e-01 -4.27131861e-01 -7.06397057e-01
2.70168353e-02 -6.25942573e-02 -6.21669352e-01 -5.37697017e-01
-7.90080547e-01 -7.90179729e-01 -5.80881664e-04 7.16484189e-01
1.16190851e+00 -1.23445213e+00 -3.04888517e-01 4.72259521e-01
-5.64487763e-02 -7.59527683e-01 -4.48991388e-01 1.85253605e-01
-9.18059051e-01 -1.09810638e+00 -6.86047137e-01 -1.25644660e+00
8.36947620e-01 5.10231376e-01 8.53499413e-01 6.28033340e-01
-5.94897628e-01 -2.00927034e-01 -3.77543449e-01 7.27872923e-02
1.73476055e-01 -2.57725324e-02 -5.56126535e-01 2.43624613e-01
1.64356679e-01 -9.12531912e-01 -1.04033887e+00 6.39850378e-01
-1.04739535e+00 2.05844238e-01 1.09709287e+00 1.28171849e+00
7.87358284e-01 6.24384642e-01 4.25047249e-01 -8.45319331e-01
4.30674046e-01 -4.02698189e-01 -2.11591080e-01 4.05393355e-02
-2.52923280e-01 -1.88741580e-01 6.45129502e-01 -1.45525828e-01
-1.24630606e+00 -6.01773374e-02 -5.25831759e-01 -4.58399355e-01
1.79925319e-02 6.41361237e-01 -7.09120989e-01 -3.54052991e-01
2.09782094e-01 4.61266100e-01 -3.01922131e-02 -6.98684573e-01
-2.11224392e-01 6.05783522e-01 1.19048439e-01 -5.78281760e-01
7.06568241e-01 4.43805993e-01 -1.42043620e-01 -6.76545203e-01
-6.81100726e-01 -3.45394492e-01 -7.48167694e-01 8.50486085e-02
8.30075562e-01 -1.05436254e+00 -5.29631674e-01 1.14311945e+00
-1.22216868e+00 -2.07808226e-01 -1.85490340e-01 8.21378380e-02
-3.91953498e-01 3.86125922e-01 -9.37742293e-01 -2.92091846e-01
-2.41332591e-01 -1.25142968e+00 1.39622152e+00 1.74107790e-01
4.92320925e-01 -9.63157833e-01 -4.44935918e-01 3.60606164e-01
4.72337961e-01 1.92585647e-01 1.17561865e+00 2.86565293e-02
-9.86437500e-01 -1.17826827e-01 -6.02104068e-01 6.97085261e-01
5.70615351e-01 -3.72277588e-01 -6.90843880e-01 -3.15021068e-01
7.31706098e-02 -2.23116145e-01 1.15621650e+00 3.01740021e-01
1.46768582e+00 1.16762873e-02 -6.30468905e-01 5.11847496e-01
1.59203947e+00 2.10883245e-01 8.87671649e-01 2.50337690e-01
1.14428675e+00 3.20698708e-01 7.16135442e-01 3.95950049e-01
3.47539067e-01 4.30614859e-01 6.40387297e-01 -4.50164855e-01
-3.67142200e-01 -3.38534176e-01 1.85353771e-01 1.02585733e+00
-3.38505283e-02 -5.65614998e-02 -6.43285513e-01 6.09562755e-01
-1.73010170e+00 -4.63451058e-01 -5.03998250e-03 1.90920234e+00
4.87533480e-01 3.78579527e-01 -3.76255810e-01 3.61719400e-01
5.78126431e-01 3.17650259e-01 -6.86222970e-01 1.13736443e-01
-2.21530944e-01 2.49034151e-01 2.73152232e-01 1.54040858e-01
-1.05121243e+00 9.48947430e-01 4.91194010e+00 1.12357807e+00
-1.11647725e+00 1.33840859e-01 6.42956614e-01 2.76173532e-01
-4.41921920e-01 -9.47124138e-02 -2.87152916e-01 4.62226331e-01
1.30730391e-01 5.71011245e-01 1.45261073e-02 5.70021331e-01
1.47408113e-01 -5.17616086e-02 -7.04313397e-01 1.04548681e+00
-2.72821002e-02 -1.32760620e+00 2.68905789e-01 9.31029115e-03
6.23002410e-01 -1.13246918e-01 8.97846371e-02 7.69454762e-02
-2.09030211e-02 -8.84131730e-01 6.42664552e-01 3.51642668e-01
8.54799509e-01 -8.78880739e-01 9.17897046e-01 1.20238744e-01
-1.60182488e+00 -3.78391325e-01 -7.09603071e-01 -1.07851125e-01
-3.27622369e-02 1.00658751e+00 1.61691271e-02 8.76601696e-01
9.50753629e-01 1.48044896e+00 -3.83285373e-01 9.05524254e-01
-2.05659971e-01 3.17615837e-01 -1.41289875e-01 3.19587797e-01
3.05210352e-01 -9.14252326e-02 2.00378314e-01 8.01169693e-01
5.24481237e-01 1.25147521e-01 1.61529303e-01 7.84313321e-01
5.89191215e-03 3.94877158e-02 -4.44903255e-01 2.77406335e-01
1.71589360e-01 1.19700575e+00 -7.06591368e-01 3.07138152e-02
-8.35795403e-01 1.26823640e+00 4.21540201e-01 2.50068754e-01
-5.73920548e-01 -5.95520675e-01 7.69644439e-01 3.81155491e-01
6.49200618e-01 -1.88776165e-01 -3.46157998e-01 -1.03714085e+00
4.09359217e-01 -7.36438155e-01 2.40911677e-01 -7.04030275e-01
-1.32477701e+00 5.51333725e-01 -6.11749828e-01 -1.41939080e+00
6.55349314e-01 -7.75814056e-01 -1.01293492e+00 8.96111488e-01
-1.77014828e+00 -1.75176632e+00 -4.54694301e-01 5.80625117e-01
6.66988134e-01 1.29146576e-02 3.75428170e-01 6.32604659e-01
-9.72539663e-01 3.37997288e-01 -7.85642862e-02 1.90136004e-02
4.61812526e-01 -9.12536442e-01 3.04009318e-01 6.94735885e-01
-3.15712810e-01 4.88125771e-01 8.18248615e-02 -8.52510512e-01
-1.38828456e+00 -1.36037588e+00 6.26476228e-01 -4.90096807e-02
4.15650547e-01 -3.76056731e-01 -1.08805978e+00 5.64684033e-01
-1.40769854e-01 3.38572979e-01 1.08792558e-01 6.38969466e-02
-1.64521098e-01 -3.55015367e-01 -9.26439643e-01 2.34737784e-01
1.27453911e+00 -6.20753407e-01 -2.85890490e-01 2.83773184e-01
8.23369920e-01 -6.12917960e-01 -8.94877195e-01 6.24450982e-01
2.26546258e-01 -1.23786187e+00 1.07378733e+00 1.23437539e-01
7.75474072e-01 -6.18993521e-01 1.53327966e-02 -1.12485695e+00
-6.67182803e-01 -2.14280963e-01 2.61542559e-01 1.30711567e+00
1.09639287e-01 -7.50863075e-01 8.24512184e-01 1.81973986e-02
-8.39157581e-01 -1.28804529e+00 -9.98347938e-01 -3.89451534e-01
-2.02050284e-01 -3.61593366e-01 6.99348807e-01 5.66349149e-01
-3.17998588e-01 2.94614077e-01 -2.97177732e-01 4.13713694e-01
2.81343609e-01 2.12189704e-01 3.24980378e-01 -1.19434142e+00
-1.07046394e-02 -3.41078788e-01 -6.69286370e-01 -1.46915901e+00
-1.38921231e-01 -5.87687254e-01 1.48458511e-01 -1.74057794e+00
1.92345783e-01 -7.42169023e-01 -6.43768430e-01 4.38219160e-01
-1.67124376e-01 2.96836823e-01 -5.00415206e-01 3.15760940e-01
-6.01984441e-01 1.00526524e+00 1.78020442e+00 -3.77498120e-01
-6.25211969e-02 -5.03487773e-02 -6.67147934e-01 5.69404304e-01
3.91089290e-01 -1.93392113e-01 -3.35407883e-01 -7.94293106e-01
1.62715748e-01 -1.16874233e-01 6.51228964e-01 -1.23251212e+00
2.77492225e-01 1.59191445e-01 5.62403262e-01 -6.73597753e-01
2.15724334e-01 -8.89702618e-01 -8.46243501e-02 5.22360384e-01
1.75867319e-01 -1.76495567e-01 5.35604395e-02 6.90584719e-01
-8.99307251e-01 1.10924378e-01 6.86156809e-01 -2.63268262e-01
-7.34135687e-01 8.99399579e-01 2.56896503e-02 -4.50240701e-01
9.66639996e-01 -3.38568330e-01 -3.52074057e-02 1.28523663e-01
-5.06640196e-01 2.83935070e-01 5.63477993e-01 4.39196855e-01
1.01990974e+00 -1.49085391e+00 -4.70426619e-01 7.48677373e-01
3.22027951e-01 7.20752001e-01 9.47795987e-01 8.41980457e-01
-6.68740511e-01 2.73289204e-01 -1.74595669e-01 -6.56318188e-01
-9.08460081e-01 4.15265977e-01 5.14382958e-01 -4.91629243e-01
-9.10420716e-01 1.15163279e+00 6.96388066e-01 -1.15681596e-01
-1.36741146e-01 -6.99797511e-01 -5.48122115e-02 -3.01871896e-01
3.44732910e-01 4.73017506e-02 3.39993566e-01 -6.25480771e-01
-2.08177939e-01 1.10306287e+00 -2.37910494e-01 6.19686782e-01
1.49118829e+00 -2.88533032e-01 -5.44175565e-01 1.96737453e-01
1.18154049e+00 -1.96364701e-01 -1.45767975e+00 -4.86770928e-01
-2.88000554e-01 -4.47160661e-01 2.43402034e-01 -4.96179909e-01
-1.57037592e+00 1.17563653e+00 5.01681626e-01 5.62222935e-02
1.48137558e+00 5.29896431e-02 1.23397160e+00 6.58710152e-02
3.65755349e-01 -8.50057542e-01 5.81888437e-01 3.78878534e-01
1.03077340e+00 -1.28592241e+00 -2.33505979e-01 -8.45046043e-01
-2.28377908e-01 1.08077979e+00 1.00630701e+00 -3.46637785e-01
9.95227158e-01 1.73891544e-01 -1.51378065e-01 -6.03739202e-01
-5.37725210e-01 -9.39731374e-02 7.13117197e-02 5.12044072e-01
4.20327127e-01 -2.83288181e-01 -4.05261889e-02 7.85787821e-01
4.29689735e-01 -9.79861990e-02 -2.05113143e-01 1.04857934e+00
-5.09151161e-01 -1.19361985e+00 7.49043841e-03 6.55619919e-01
-1.64734334e-01 -1.13482393e-01 9.29169133e-02 5.05712092e-01
7.31209397e-01 7.13357270e-01 2.21821621e-01 -6.39137387e-01
6.16551518e-01 -5.02372861e-01 2.71403134e-01 -8.73550236e-01
-5.93880653e-01 2.49944791e-01 -2.05288649e-01 -7.85915732e-01
-2.79967517e-01 -5.54034531e-01 -1.29438210e+00 8.88534039e-02
-3.63627762e-01 -1.34005815e-01 1.74296126e-01 9.93264318e-01
4.37259376e-01 1.09101558e+00 8.53649914e-01 -8.55264425e-01
6.20787926e-02 -1.09401524e+00 -8.75791907e-01 4.18709397e-01
3.44918668e-01 -8.20818305e-01 -2.13491708e-01 -1.43611431e-01] | [7.541024684906006, 1.524172067642212] |
65d6b9b1-cad9-4396-867d-cb16c4f307d8 | adaptive-intrusion-detection-in-the | null | null | https://ieeexplore.ieee.org/document/9296578 | https://ieeexplore.ieee.org/document/9296578 | Adaptive Intrusion Detection in the Networking of Large-Scale LANs with Segmented Federated Learning | Predominant network intrusion detection systems (NIDS) aim to identify malicious traffic patterns based on a handcrafted dataset of rules. Recently, the application of machine learning in NIDS helps alleviate the enormous effort of human observation. Federated learning (FL) is a collaborative learning scheme concerning distributed data. Instead of sharing raw data, it allows a participant to share only a trained local model. Despite the success of existing FL solutions, in NIDS, a network's traffic data distribution does not always fit into the single global model of FL; some networks have similarities with each other but other networks do not. We propose Segmented-Federated Learning (Segmented-FL), where by employing periodic local model evaluation and network segmentation, we aim to bring similar network environments to the same group. A comparison between FL and our method was conducted against a range of metrics including the weighted precision, recall, and F1 score, using a collected dataset from 20 massively distributed networks within 60 days. By studying the optimized hyperparameters of Segmented-FL and employing three evaluation methods, it shows that Segmented-FL has better performance in all three types of intrusion detection tasks, achieving validation weighted F1 scores of 0.964, 0.803, and 0.912 with Method A, Method B, and Method C respectively. For each method, this scheme shows a gain of 0.1%, 4.0% and 1.1% in performance compared with FL. | ['Hideya Ochiai.', 'Hiroshi Esaki', 'Yuwei Sun'] | 2020-12-16 | null | null | null | ieee-open-journal-of-the-communications | ['network-intrusion-detection'] | ['miscellaneous'] | [-1.36720061e-01 -4.02410448e-01 -2.74975151e-01 -3.60121310e-01
-8.50671902e-02 -5.68390846e-01 6.55693114e-01 1.35531977e-01
-3.26198846e-01 7.92023242e-01 -5.49147427e-01 -5.43057740e-01
-6.52291238e-01 -8.45492661e-01 -1.81193516e-01 -4.45605814e-01
-3.30516040e-01 6.39078975e-01 7.13123024e-01 -1.61901154e-02
3.72591794e-01 9.71950769e-01 -1.15468812e+00 3.21592540e-01
4.65883255e-01 1.17272019e+00 -4.26194072e-01 5.30570626e-01
-5.92294276e-01 1.04892612e+00 -1.00243592e+00 -5.86119294e-01
6.50429904e-01 -2.65198022e-01 -7.04977453e-01 -8.48202854e-02
1.33726105e-01 -1.59869250e-02 -2.86034614e-01 8.99933159e-01
1.65940776e-01 1.69015676e-01 3.77902567e-01 -1.85760975e+00
-1.54501513e-01 4.52735871e-01 -6.20796800e-01 4.22312975e-01
-7.29822321e-03 2.38826796e-01 6.66612327e-01 -1.93032026e-01
6.32912874e-01 1.15325308e+00 6.34445667e-01 5.04856765e-01
-1.22913301e+00 -1.01074672e+00 6.55255765e-02 2.37210974e-01
-1.26460290e+00 -4.33139175e-01 6.31680787e-01 -1.83304533e-01
8.25210631e-01 3.65438610e-01 -2.22656457e-03 9.06702757e-01
2.89520532e-01 2.24321380e-01 1.01260531e+00 -3.75521541e-01
5.29220045e-01 4.44355369e-01 3.45150173e-01 5.93056083e-01
5.05896688e-01 2.07584053e-01 -2.05407903e-01 -5.71479917e-01
3.29705298e-01 4.67357606e-01 2.29745984e-01 -3.57938528e-01
-6.40380263e-01 7.97809720e-01 2.19162375e-01 5.98996699e-01
-4.18160468e-01 -4.61051404e-01 6.39083266e-01 5.47347367e-01
3.32470119e-01 2.57593125e-01 -6.96197391e-01 -5.61760403e-02
-7.53019094e-01 -1.84652377e-02 1.24887824e+00 4.28562284e-01
9.60070074e-01 2.23623440e-01 2.79538482e-02 6.79301679e-01
2.92859644e-01 3.20161253e-01 4.82243091e-01 -7.81509221e-01
3.07594776e-01 1.07758653e+00 -1.27102211e-01 -1.24242663e+00
-3.07101607e-01 -2.95345992e-01 -7.33445883e-01 2.49351218e-01
4.42313403e-01 -4.66143936e-01 -5.57328463e-01 1.68227684e+00
3.85524869e-01 5.04953921e-01 -1.73506681e-02 2.14695215e-01
9.91182923e-02 4.00973916e-01 1.02425061e-01 -3.33125144e-01
5.68120658e-01 -9.13753390e-01 -4.39741552e-01 7.82332793e-02
4.28001076e-01 -6.50358081e-01 4.01081473e-01 4.96343940e-01
-4.78294104e-01 -3.99167657e-01 -8.06165993e-01 1.06233215e+00
-6.57745123e-01 -6.37527108e-01 6.01457238e-01 1.18330014e+00
-8.51509571e-01 6.88293636e-01 -4.91320252e-01 -5.37093401e-01
6.09528720e-01 4.77009773e-01 -3.23173046e-01 -4.23222035e-02
-8.56679618e-01 6.14756167e-01 2.56853670e-01 -5.08056462e-01
-9.07871008e-01 -7.47831285e-01 -1.59062341e-01 -3.19674723e-02
4.55016315e-01 -2.02297047e-01 9.42455351e-01 -7.36254990e-01
-1.26680088e+00 3.13732475e-01 -4.48891940e-03 -7.48626053e-01
3.11976552e-01 1.26373306e-01 -1.06239927e+00 2.06011206e-01
-4.18851748e-02 4.45593260e-02 7.97985971e-01 -1.33474934e+00
-8.60290229e-01 -4.85835642e-01 -1.20836645e-02 -5.40232778e-01
-4.64024872e-01 4.64037925e-01 -3.34139019e-02 -2.26471618e-01
-2.61516184e-01 -5.30456841e-01 -4.20623988e-01 -1.75122291e-01
-3.41001034e-01 -1.97977930e-01 1.63338721e+00 -1.82195425e-01
1.40823007e+00 -1.89947486e+00 -5.93550682e-01 8.87849867e-01
4.14966643e-01 9.61764038e-01 -5.71700782e-02 4.86059725e-01
-5.49719520e-02 4.55856889e-01 -5.89922778e-02 -1.53308913e-01
1.85495075e-02 3.50955904e-01 -1.10928752e-01 2.42296115e-01
-1.61794811e-01 4.48396742e-01 -5.57188809e-01 -4.26299989e-01
3.93437743e-01 1.96252301e-01 -4.99199718e-01 3.66656899e-01
-1.29805759e-01 3.11020553e-01 -3.94137710e-01 7.43722498e-01
5.82627892e-01 -2.50902683e-01 3.52352500e-01 1.71457008e-02
2.82741059e-02 -2.06175879e-01 -1.44553995e+00 8.70442092e-01
-3.91675442e-01 2.50886649e-01 1.01292916e-01 -1.05354488e+00
1.37518346e+00 3.65127146e-01 8.89184594e-01 -8.93550277e-01
1.99621424e-01 1.75870925e-01 2.28910759e-01 -4.01768893e-01
-2.28271589e-01 1.00084610e-01 2.12581992e-01 1.16193378e+00
2.18980417e-01 7.06511199e-01 2.77092308e-01 3.15386802e-01
1.71214747e+00 -7.92439044e-01 2.83208668e-01 5.01160249e-02
8.21816981e-01 -2.49322832e-01 5.62237263e-01 1.13254952e+00
-7.15644956e-01 -2.03844786e-01 5.51159561e-01 -7.47665644e-01
-8.23321223e-01 -1.09353364e+00 1.02682486e-02 1.06894183e+00
2.36677919e-02 -3.29125822e-01 -4.46308315e-01 -1.18148875e+00
1.26882434e-01 4.69932228e-01 -2.63795704e-01 -2.99574316e-01
-5.17836511e-01 -6.98917747e-01 9.00714993e-01 -5.30729890e-02
9.73492026e-01 -1.17831731e+00 -2.80449599e-01 1.43805951e-01
3.14511478e-01 -1.04369891e+00 -8.83763134e-02 4.52729017e-02
-6.39357507e-01 -1.61029255e+00 1.71117708e-01 -4.75577205e-01
4.73706901e-01 3.91291201e-01 8.35221291e-01 3.34333330e-01
-4.39027518e-01 3.29514831e-01 -4.70025867e-01 -3.04123491e-01
-4.71722215e-01 1.07830670e-02 2.58741885e-01 6.65881455e-01
5.96168876e-01 -8.24328125e-01 -3.04228902e-01 7.34678149e-01
-8.39368522e-01 -7.99331725e-01 5.45204520e-01 4.42716181e-01
-7.13245720e-02 4.95099097e-01 8.87028158e-01 -1.28346038e+00
7.14831233e-01 -9.45063233e-01 -3.45531046e-01 3.61144543e-01
-9.64368820e-01 -3.17572773e-01 8.98745358e-01 -5.28666258e-01
-1.01126754e+00 -3.36599737e-01 2.41473421e-01 -5.86724818e-01
-5.79991460e-01 -3.78040783e-02 -1.96739942e-01 -4.29933220e-01
8.96867216e-01 -8.36393684e-02 2.61435986e-01 -4.88473594e-01
2.28298660e-02 1.04026103e+00 1.70585334e-01 -6.04767025e-01
9.15047050e-01 4.01957721e-01 -9.91612524e-02 -6.67953551e-01
-5.70943296e-01 -5.86438239e-01 -4.65611935e-01 -2.78817356e-01
2.06364676e-01 -7.46319965e-02 -9.84784663e-01 4.30602133e-01
-9.18550551e-01 -5.22680320e-02 -2.80222744e-01 2.48945296e-01
4.57121152e-03 3.84259552e-01 -4.05112743e-01 -8.87649357e-01
-5.03028035e-01 -8.05496573e-01 2.56426223e-02 2.46960476e-01
-2.49376982e-01 -1.14862597e+00 2.05572680e-01 4.25027043e-01
1.11454797e+00 4.02603418e-01 1.00813580e+00 -1.65118372e+00
-2.14947537e-01 -5.64880967e-01 -3.06352168e-01 5.12492299e-01
6.03788137e-01 1.68913960e-01 -7.55586922e-01 -3.19263637e-01
8.18562210e-02 -1.05169997e-01 2.68661261e-01 -2.85131782e-01
1.14250362e+00 -5.52359402e-01 -2.81996101e-01 3.24710280e-01
1.52984226e+00 9.22988653e-01 4.13641810e-01 3.61732394e-01
4.54847246e-01 6.00357294e-01 1.05672807e-01 5.28305590e-01
-5.04372194e-02 3.28775704e-01 5.48958540e-01 2.32542247e-01
2.76848767e-02 -9.79573429e-02 8.42899457e-02 3.25095475e-01
2.87751436e-01 -2.03988135e-01 -1.02549744e+00 2.99536943e-01
-1.72246730e+00 -1.21398199e+00 7.19467551e-02 2.25720096e+00
3.57477039e-01 4.93782371e-01 4.66714680e-01 4.02688473e-01
1.08961487e+00 9.03773978e-02 -7.67499030e-01 -7.17443705e-01
6.06447458e-02 6.07697487e-01 4.36967343e-01 2.85346657e-01
-9.93269920e-01 7.23436236e-01 5.99192524e+00 9.14736569e-01
-1.18474758e+00 1.27420887e-01 5.36664069e-01 5.16960286e-02
2.71188349e-01 2.02096850e-01 -6.66233182e-01 6.95927143e-01
1.53843880e+00 -4.41794872e-01 6.15980148e-01 8.03024352e-01
1.62514895e-01 1.42722577e-01 -7.41940141e-01 8.48504663e-01
-8.14694911e-02 -1.39931929e+00 2.40755022e-01 2.33328819e-01
7.75555789e-01 1.83937982e-01 -2.95558393e-01 4.01262939e-01
5.23303330e-01 -9.97057259e-01 1.77381013e-03 5.29283226e-01
1.13565698e-01 -1.04883695e+00 7.83395112e-01 4.80273426e-01
-1.01685405e+00 -5.12688100e-01 -1.61257237e-01 3.04428637e-02
-2.47242481e-01 5.07433832e-01 -6.75971508e-01 5.66910028e-01
5.94867051e-01 3.29969674e-01 -6.68505728e-01 1.13102496e+00
2.91785032e-01 7.83928335e-01 -3.64550263e-01 5.87284453e-02
1.77226678e-01 -3.32233198e-02 5.73713601e-01 9.65055823e-01
-2.20769018e-01 -2.69784033e-01 4.02611852e-01 5.95932305e-01
-1.15015097e-01 2.23844588e-01 -6.74149513e-01 1.78334694e-02
8.87531400e-01 1.44368696e+00 -5.58928251e-01 -1.47851571e-01
-3.79479855e-01 3.85447294e-01 1.71162963e-01 1.85336158e-01
-7.10371017e-01 -6.91062152e-01 8.04957151e-01 1.63249388e-01
1.25910729e-01 -2.38804799e-02 -2.16801137e-01 -6.56285167e-01
-2.77464688e-01 -1.14564097e+00 6.32462680e-01 -7.64059275e-03
-1.63876271e+00 9.04906809e-01 -1.13819711e-01 -9.23597217e-01
-1.26380861e-01 -4.49873745e-01 -1.09470487e+00 5.47893524e-01
-9.67913091e-01 -1.03930402e+00 -1.89890131e-01 1.19894016e+00
1.76555976e-01 -9.20361221e-01 8.86441290e-01 5.62109590e-01
-1.04685104e+00 7.54259229e-01 9.47090015e-02 1.98281318e-01
7.09898949e-01 -7.66691923e-01 9.36314613e-02 8.35896194e-01
2.48893693e-01 7.08335161e-01 2.15810865e-01 -6.64084136e-01
-9.88232970e-01 -1.34126878e+00 7.65798986e-01 -1.37379885e-01
8.23813975e-01 -1.02124857e-02 -8.84174824e-01 5.96639395e-01
-2.50482708e-02 3.19889784e-01 9.61109698e-01 1.24639645e-01
-6.53356910e-01 -5.70613980e-01 -1.97847283e+00 3.50775689e-01
8.49726021e-01 -3.25671971e-01 -1.80424616e-01 3.44504088e-01
5.10184109e-01 3.96721065e-01 -8.99651408e-01 2.35839337e-01
2.76495129e-01 -1.30917966e+00 7.14713693e-01 -9.69380915e-01
-3.41348559e-01 -2.38419458e-01 -3.98622751e-01 -8.40303540e-01
-1.72257632e-01 -7.42122948e-01 -4.69336450e-01 1.43800187e+00
3.13902080e-01 -1.19990766e+00 9.74027097e-01 5.41311562e-01
4.03051555e-01 -7.29461551e-01 -9.05681551e-01 -1.01053548e+00
-1.70895070e-01 -5.26726007e-01 1.00392580e+00 1.01731837e+00
-2.29817986e-01 7.73421898e-02 -2.12991297e-01 1.20655365e-01
1.02614498e+00 -4.50099744e-02 9.32540655e-01 -1.63112843e+00
-2.15714529e-01 -3.80019128e-01 -5.99853992e-01 -1.45957172e-01
2.53003776e-01 -6.99492395e-01 -9.64214265e-01 -8.53397667e-01
-1.11692362e-01 -8.29060078e-01 -6.87419951e-01 9.48364973e-01
4.94854659e-01 2.70745337e-01 3.03474903e-01 5.22988856e-01
-8.38631868e-01 -1.11463197e-01 5.58055222e-01 -6.75207675e-02
-3.49488050e-01 3.31410795e-01 -7.29024172e-01 6.48868680e-01
1.52206683e+00 -6.36398077e-01 -4.72833186e-01 5.38473688e-02
-4.48874950e-01 -7.35100657e-02 2.68250585e-01 -1.24162138e+00
6.92766786e-01 -4.03154075e-01 3.58173937e-01 -1.81893960e-01
-2.94396374e-02 -9.66134727e-01 3.00804377e-01 5.64715326e-01
-2.67568119e-02 6.14124388e-02 6.04987368e-02 5.25378227e-01
-1.02822877e-01 -1.38271019e-01 9.44198251e-01 -1.52139008e-01
-6.59969151e-01 4.07176614e-01 -1.99557111e-01 5.09890914e-02
1.39501297e+00 -3.05560142e-01 -5.25744081e-01 -1.28769964e-01
-6.79095507e-01 1.28786951e-01 1.47746071e-01 4.90472853e-01
4.44667310e-01 -1.00842738e+00 -3.24192494e-01 5.44661701e-01
-2.23835543e-01 -5.25477827e-01 1.43578738e-01 5.93799293e-01
-3.68906528e-01 5.44919431e-01 -4.90297407e-01 -3.83974999e-01
-1.16811073e+00 4.29142743e-01 5.18954933e-01 -5.13793945e-01
-3.30177158e-01 2.53901541e-01 -5.56724787e-01 -8.63562167e-01
3.82543743e-01 6.18704200e-01 -8.60374272e-02 -1.80333793e-01
7.09856451e-01 9.11637127e-01 3.43146861e-01 -4.33240175e-01
-5.87786019e-01 1.16774432e-01 -4.34670597e-01 2.30503738e-01
1.28199363e+00 8.65005478e-02 -3.21020514e-01 1.35278881e-01
1.20111823e+00 -8.93047079e-03 -6.90603316e-01 -4.84075278e-01
5.00208199e-01 -6.76433027e-01 -2.38963008e-01 -1.02202368e+00
-1.47501969e+00 1.35463730e-01 6.20195091e-01 5.46110213e-01
9.77080762e-01 -2.90131450e-01 7.66378880e-01 4.96848822e-01
7.22310305e-01 -6.55371249e-01 1.50727436e-01 5.69642365e-01
1.64853051e-01 -1.09335065e+00 -2.61278123e-01 -4.65666093e-02
-4.45399851e-01 1.01435781e+00 9.12435710e-01 -2.56037861e-01
9.60379779e-01 2.20500186e-01 7.08800834e-03 -2.19955117e-01
-1.01185298e+00 9.34770852e-02 -3.51740688e-01 8.22814047e-01
-2.04018891e-01 -4.79186811e-02 -1.10561460e-01 4.19212610e-01
1.40355244e-01 -2.01752275e-01 2.50317365e-01 1.06752110e+00
-6.37292087e-01 -1.50788152e+00 -3.96252990e-01 7.27628946e-01
-4.49439764e-01 5.34479618e-01 -6.53540850e-01 7.86532819e-01
3.15870196e-01 1.50412619e+00 2.02413127e-02 -7.92670786e-01
3.49055320e-01 2.52785176e-01 -1.66543543e-01 -4.24530864e-01
-9.68750596e-01 -4.73721802e-01 -1.68194517e-01 -7.10871518e-01
-2.64954984e-01 -1.67010486e-01 -9.92474258e-01 -9.71068501e-01
-1.59059078e-01 5.95357478e-01 7.27567434e-01 8.72653306e-01
5.05052567e-01 3.41070920e-01 1.40726519e+00 -3.09211284e-01
-5.85357249e-01 -6.42634153e-01 -6.98734403e-01 3.97664338e-01
-3.46341692e-02 -5.19466102e-01 -7.68813312e-01 -5.27370930e-01] | [5.288809776306152, 7.187138080596924] |
dc6579a6-c36d-4a83-9743-400eab36dea6 | contrimix-unsupervised-disentanglement-of | 2306.04527 | null | https://arxiv.org/abs/2306.04527v2 | https://arxiv.org/pdf/2306.04527v2.pdf | ContriMix: Unsupervised disentanglement of content and attribute for domain generalization in microscopy image analysis | Domain generalization is critical for real-world applications of machine learning models to microscopy images, including histopathology and fluorescence imaging. Artifacts in histopathology arise through a complex combination of factors relating to tissue collection and laboratory processing, as well as factors intrinsic to patient samples. In fluorescence imaging, these artifacts stem from variations across experimental batches. The complexity and subtlety of these artifacts make the enumeration of data domains intractable. Therefore, augmentation-based methods of domain generalization that require domain identifiers and manual fine-tuning are inadequate in this setting. To overcome this challenge, we introduce ContriMix, a domain generalization technique that learns to generate synthetic images by disentangling and permuting the biological content ("content") and technical variations ("attributes") in microscopy images. ContriMix does not rely on domain identifiers or handcrafted augmentations and makes no assumptions about the input characteristics of images. We assess the performance of ContriMix on two pathology datasets (Camelyon17-WILDS and a prostate cell classification dataset) and one fluorescence microscopy dataset (RxRx1-WILDS). ContriMix outperforms current state-of-the-art methods in all datasets, motivating its usage for microscopy image analysis in real-world settings where domain information is hard to come by. | ['Amaro Taylor-Weiner', 'Justin Lee', 'John Abel', 'Anand Sampat', 'Michael Griffin', 'Sai Chowdary Gullapally', 'Chintan Shah', 'Shima Nofallah', 'Aaditya Prakash', 'Jin Li', 'Dinkar Juyal', 'Tan H. Nguyen'] | 2023-06-07 | null | null | null | null | ['domain-generalization', 'disentanglement'] | ['methodology', 'methodology'] | [ 7.99748003e-01 -2.35270262e-01 3.46632265e-02 -6.37283385e-01
-7.93557346e-01 -8.58066261e-01 6.95384920e-01 2.93656915e-01
-7.51392007e-01 1.10939145e+00 -1.55803099e-01 -2.90245593e-01
-1.31536245e-01 -4.26186740e-01 -8.49245071e-01 -1.02009141e+00
8.23729858e-02 7.03253865e-01 -1.22138672e-01 1.48841143e-01
8.95657167e-02 7.49882400e-01 -1.09616315e+00 5.00155985e-01
6.33237958e-01 6.54443860e-01 1.32466763e-01 7.84248829e-01
-4.86402772e-02 3.11945677e-01 -7.23811686e-01 -4.42976207e-02
2.66949445e-01 -4.69457090e-01 -8.74124229e-01 2.74094284e-01
4.43377465e-01 -2.66499311e-01 -6.26435205e-02 9.98592317e-01
5.24134576e-01 -2.09299281e-01 9.64345396e-01 -1.35356581e+00
-5.16067326e-01 1.51037678e-01 -6.02072716e-01 3.92531276e-01
-1.48432881e-01 6.29175842e-01 4.99707669e-01 -4.37215626e-01
1.19738948e+00 9.99292433e-01 6.66815579e-01 9.35776353e-01
-2.14468145e+00 -6.93922341e-01 -1.99372172e-01 -1.34069666e-01
-1.05978942e+00 -4.07259852e-01 2.69728780e-01 -7.95032680e-01
4.09844249e-01 1.62989825e-01 1.41760990e-01 1.56556296e+00
2.29106531e-01 4.19402927e-01 1.64545178e+00 -1.82588041e-01
5.01945376e-01 2.88300633e-01 -2.33469605e-02 3.05504113e-01
2.37231955e-01 1.08758815e-01 -3.32787782e-01 -4.69839156e-01
9.44085836e-01 7.05882758e-02 -3.76160949e-01 -5.53744733e-01
-1.49053943e+00 6.13808453e-01 1.38577059e-01 1.41589537e-01
-1.56057954e-01 -5.06331921e-02 5.54172218e-01 4.48931962e-01
5.99074364e-02 8.24333847e-01 -6.50354922e-01 2.30264932e-01
-8.20820570e-01 2.42438152e-01 3.80245149e-01 9.10944283e-01
6.96614742e-01 -3.26283276e-01 -6.00581840e-02 7.31828988e-01
-2.92355001e-01 3.15652668e-01 7.70271122e-01 -9.76744533e-01
-1.76959280e-02 6.64692521e-01 7.65611231e-02 -3.95474583e-01
-6.05311632e-01 -3.58658582e-01 -9.38138008e-01 3.36051524e-01
9.64810789e-01 -1.17965758e-01 -1.33729863e+00 1.81449270e+00
2.19621330e-01 1.08742006e-01 3.11594419e-02 7.16492772e-01
7.30453312e-01 -1.57496370e-02 2.43579566e-01 -6.52686134e-02
1.41956699e+00 -2.75070310e-01 -3.18309397e-01 -1.42647669e-01
7.55338788e-01 -4.31236893e-01 1.22817254e+00 4.11767423e-01
-6.10185802e-01 -1.02451451e-01 -8.44078720e-01 -3.49315293e-02
-5.51214039e-01 1.10809445e-01 6.70661926e-01 4.45160240e-01
-7.67990053e-01 6.06339812e-01 -9.68397319e-01 -6.25402391e-01
9.48633134e-01 6.79300725e-01 -9.18797731e-01 -2.85816938e-01
-6.11642540e-01 5.80972910e-01 1.91556305e-01 -2.20068529e-01
-8.86160374e-01 -1.38798428e+00 -6.38591349e-01 -1.97735608e-01
8.33070204e-02 -8.74379277e-01 1.14209855e+00 -8.22366953e-01
-1.03339529e+00 1.27629435e+00 -8.86049867e-03 -3.54527980e-01
6.84275985e-01 4.92363155e-01 -2.24861652e-01 1.60835415e-01
6.60024807e-02 9.14635718e-01 6.35145187e-01 -1.20159805e+00
-3.31147581e-01 -5.31171858e-01 -2.85033464e-01 -3.11515719e-01
-3.19198295e-02 -2.25861534e-01 1.61447786e-02 -6.07838869e-01
-2.29771912e-01 -1.11915755e+00 -4.73376393e-01 4.99420524e-01
-3.43248457e-01 4.03769791e-01 9.18863058e-01 -3.32578093e-01
3.93741250e-01 -2.30938888e+00 -5.87259158e-02 1.91754460e-01
2.91165918e-01 3.73927534e-01 -4.57039714e-01 2.47577444e-01
-2.42771953e-01 2.37927124e-01 -2.71393418e-01 -2.21594974e-01
-1.94726527e-01 2.65892774e-01 -7.73581937e-02 7.29915082e-01
4.90546554e-01 6.93664908e-01 -1.03291333e+00 -5.56177616e-01
1.46855507e-02 4.36433226e-01 -5.15182674e-01 -1.05103645e-02
-4.08762217e-01 9.60472584e-01 -2.34677136e-01 6.75854981e-01
7.24760532e-01 -5.46427548e-01 3.46889377e-01 -2.86421776e-01
3.01390946e-01 1.42794056e-02 -8.16045582e-01 1.72706556e+00
-2.60986090e-01 4.66104925e-01 2.13878348e-01 -8.57208371e-01
7.14203656e-01 1.05127782e-01 4.53096062e-01 -4.79303956e-01
1.97944596e-01 2.53514826e-01 2.00485438e-01 -3.74361843e-01
-1.21348917e-01 -6.04899764e-01 1.54346555e-01 5.00007629e-01
3.01271051e-01 -1.25461087e-01 2.03748867e-01 5.57726286e-02
1.59015226e+00 -1.21701963e-01 2.31169611e-01 -4.13510531e-01
3.09104741e-01 2.30944335e-01 8.17183912e-01 6.75984919e-01
-4.37599093e-01 8.07782590e-01 8.10777426e-01 -4.98336494e-01
-1.22370172e+00 -1.25897217e+00 -4.62105006e-01 6.12008631e-01
-2.63219357e-01 1.57330632e-01 -5.81005991e-01 -9.73777115e-01
2.45656997e-01 2.50474960e-01 -1.15948713e+00 -1.77411452e-01
-2.40418300e-01 -1.11196280e+00 6.54012263e-01 5.23234546e-01
1.29149482e-01 -8.70664954e-01 -5.58882475e-01 6.04064837e-02
4.95297313e-02 -1.37781394e+00 -3.81832868e-01 3.80310118e-01
-8.66004944e-01 -1.28538167e+00 -6.07856214e-01 -4.97839451e-01
1.11204278e+00 1.31950984e-02 1.12867236e+00 -1.23000070e-01
-9.57453251e-01 1.95892304e-01 -1.31037623e-01 -6.13580167e-01
-7.19684601e-01 7.79462326e-03 -4.60709445e-02 6.38785437e-02
4.49288607e-01 -6.83512211e-01 -7.88769722e-01 3.10838848e-01
-1.37210989e+00 -1.35877833e-01 9.87927973e-01 1.42142248e+00
9.62461948e-01 -3.84978652e-02 5.70353627e-01 -1.50062776e+00
4.38938916e-01 -3.98386121e-01 -5.54850042e-01 9.91080701e-02
-2.39238203e-01 7.44578242e-02 7.41862893e-01 -7.91653097e-01
-7.42937922e-01 3.90152216e-01 3.35627139e-01 -6.44023836e-01
-4.93313104e-01 2.47611180e-01 -8.02929103e-02 -1.93078637e-01
9.62483108e-01 2.54968941e-01 5.96518874e-01 -2.53885806e-01
-1.17006101e-01 4.77794468e-01 8.46866667e-01 -5.83611965e-01
6.30539715e-01 8.50295961e-01 4.38458681e-01 -8.78928244e-01
-4.81269151e-01 -3.11708659e-01 -5.79930365e-01 3.07979167e-01
6.45504296e-01 -7.25937605e-01 -6.48823798e-01 5.17239928e-01
-7.37933993e-01 -5.33296943e-01 -3.11404377e-01 3.11777592e-01
-6.27319574e-01 1.71031147e-01 -5.54952860e-01 -2.55632013e-01
-7.82729015e-02 -1.27650380e+00 1.04266918e+00 1.07916065e-01
-4.50697362e-01 -1.08083165e+00 -1.21050395e-01 2.82743841e-01
3.68178904e-01 7.47338653e-01 1.29301023e+00 -8.33291411e-01
-5.06533027e-01 -9.26305726e-02 -3.50733817e-01 3.18542570e-01
3.88245434e-01 5.68794757e-02 -1.03631139e+00 -3.99792492e-01
-1.88037038e-01 -5.91693997e-01 6.42822266e-01 3.12328577e-01
1.41557491e+00 -1.71414107e-01 -4.89850670e-01 7.25345492e-01
1.45494401e+00 -1.02705946e-02 7.07093060e-01 4.43749249e-01
2.54363388e-01 7.22006559e-01 3.57390463e-01 3.29515785e-01
1.83757879e-02 4.09531802e-01 2.84891069e-01 -5.12061656e-01
-1.17001526e-01 -9.47385468e-03 -3.06039482e-01 -4.12764907e-01
3.46077830e-01 6.44225106e-02 -9.44212079e-01 7.85915911e-01
-1.34022248e+00 -7.80621052e-01 4.67805825e-02 2.15118814e+00
1.17184186e+00 -9.29132551e-02 1.94815621e-01 -5.13704345e-02
6.42742574e-01 -5.98326564e-01 -1.00055289e+00 -3.06940991e-02
-2.27651969e-01 2.31225476e-01 5.22556365e-01 2.56359531e-03
-9.17457521e-01 4.83840734e-01 6.24469757e+00 4.89392191e-01
-1.43875110e+00 -2.48130068e-01 1.04013026e+00 -1.90495908e-01
-1.88754722e-01 -2.20310569e-01 -5.90670764e-01 5.78500152e-01
7.72125065e-01 8.36492181e-02 2.54435062e-01 4.89290148e-01
1.38721928e-01 -5.24492674e-02 -1.72852063e+00 8.60575914e-01
-2.86247283e-01 -1.52972198e+00 -5.53057976e-02 3.71710300e-01
6.60306692e-01 4.81079519e-02 3.86793792e-01 1.07765282e-02
3.95802557e-01 -1.30248773e+00 2.21203477e-03 2.56029636e-01
1.09089530e+00 -3.54567796e-01 8.51092041e-01 1.21022493e-01
-1.97009832e-01 6.75964803e-02 -2.29622662e-01 4.66241151e-01
-3.08536768e-01 7.17426181e-01 -1.43653595e+00 1.22022614e-01
6.18686974e-01 3.67269605e-01 -5.75455606e-01 8.37150574e-01
3.17580551e-01 4.69026893e-01 -3.07767838e-01 3.06784511e-01
-3.88600561e-03 1.03518561e-01 2.29749873e-01 1.38715744e+00
7.73468465e-02 -5.92050292e-02 -2.54101902e-01 1.06704807e+00
-3.13259691e-01 -2.57470280e-01 -4.73407596e-01 -3.92577291e-01
6.45628631e-01 1.41871727e+00 -8.22615564e-01 -4.80480529e-02
-1.78698853e-01 7.11937666e-01 1.71502516e-01 3.48404229e-01
-4.84687388e-01 -2.14609355e-01 1.10718620e+00 3.94177616e-01
2.13499561e-01 9.51415673e-02 -3.51482213e-01 -1.00195169e+00
-1.50326505e-01 -1.25154924e+00 4.92274940e-01 -4.67530906e-01
-1.65951121e+00 4.02714223e-01 -4.23642062e-02 -1.31261170e+00
-9.23823342e-02 -8.95281136e-01 -3.43694836e-01 8.69743407e-01
-1.64117908e+00 -1.25780809e+00 -3.59539092e-01 6.16981924e-01
5.67487776e-02 -1.17863543e-01 1.05220127e+00 1.08969569e-01
-3.96811157e-01 7.59383857e-01 3.31385046e-01 1.82576522e-01
1.17457831e+00 -1.28566873e+00 1.14926718e-01 3.77871215e-01
-3.52031708e-01 8.72207880e-01 8.77683580e-01 -2.70019770e-01
-1.29072011e+00 -1.32534349e+00 5.46674669e-01 -6.02937996e-01
7.75958359e-01 -5.02154946e-01 -9.83587563e-01 6.97674692e-01
-3.34826082e-01 5.48172653e-01 1.21639383e+00 -8.79338533e-02
-7.02168941e-01 -2.03245744e-01 -1.82769632e+00 6.37516260e-01
6.80585146e-01 -3.89142007e-01 -6.83409125e-02 4.88986045e-01
2.74852097e-01 -3.88927609e-01 -1.09484780e+00 2.38249555e-01
5.26302636e-01 -7.53699124e-01 8.40358913e-01 -8.09885263e-01
5.38419127e-01 -3.74855369e-01 -1.50876781e-02 -1.39198351e+00
-5.89559041e-02 -4.87129301e-01 4.87373680e-01 1.19930542e+00
4.23372626e-01 -7.27354050e-01 1.05527055e+00 7.08785951e-01
1.68529823e-01 -5.90327799e-01 -8.81675899e-01 -9.21879411e-01
4.60191846e-01 2.76099235e-01 6.96524382e-01 1.16391778e+00
-3.99819948e-03 1.84779227e-01 1.64803520e-01 7.13392124e-02
6.96519434e-01 -4.45973910e-02 9.86714482e-01 -1.33984208e+00
-3.99243742e-01 -3.06117386e-01 -8.16706002e-01 -2.69054353e-01
2.04537854e-01 -6.71463907e-01 -5.14035113e-02 -9.69553351e-01
5.12595892e-01 -6.70503795e-01 -2.73963839e-01 6.07953310e-01
-8.32238495e-02 3.29488873e-01 -1.42758608e-01 1.96609855e-01
-1.83560133e-01 5.41863143e-02 1.32529902e+00 -3.00210088e-01
-7.00101107e-02 -2.46127397e-01 -9.72604752e-01 4.70433444e-01
7.48933256e-01 -5.36802530e-01 -5.04835844e-01 -3.43246222e-01
-1.66273221e-01 -2.14456823e-02 6.41807973e-01 -9.59506929e-01
2.86876615e-02 -3.79240751e-01 7.01848924e-01 -1.00219175e-01
2.45958954e-01 -8.13057780e-01 3.44744951e-01 5.63610792e-01
-3.95715117e-01 -1.08061127e-01 3.84053379e-01 6.21160209e-01
3.19500417e-02 -7.18231946e-02 1.21540225e+00 -3.50427181e-01
-2.85567343e-01 2.78857738e-01 -3.50649625e-01 1.57242551e-01
1.03450155e+00 -3.33787709e-01 -7.40149856e-01 -5.67941442e-02
-6.85723305e-01 1.77517548e-01 9.29152906e-01 3.54545638e-02
3.09546441e-01 -8.70112360e-01 -7.92449594e-01 3.75904113e-01
4.35072154e-01 2.91261196e-01 4.74492848e-01 8.58553946e-01
-4.96576786e-01 2.46984035e-01 -5.26583850e-01 -8.39177489e-01
-1.31635869e+00 5.03782153e-01 3.45678568e-01 -4.65576470e-01
-2.56950289e-01 6.99588895e-01 5.73984981e-01 -6.36715114e-01
-6.78382963e-02 -3.93554837e-01 2.93747306e-01 -3.65145177e-01
5.70142090e-01 -2.35099569e-02 3.26170832e-01 -1.25745788e-01
-4.10216510e-01 8.37238654e-02 -6.48584306e-01 1.22126400e-01
1.37745166e+00 2.69392282e-01 1.00969128e-01 9.13771093e-02
1.17213237e+00 -4.50996846e-01 -1.25011778e+00 -1.38294354e-01
-1.17919862e-01 -4.27649796e-01 -2.34198213e-01 -1.27442753e+00
-8.08767438e-01 8.36159110e-01 5.76874614e-01 -3.10665637e-01
1.26931870e+00 -1.36542335e-01 4.30282354e-01 3.09266210e-01
3.15674782e-01 -8.36733103e-01 5.39538041e-02 1.37541801e-01
6.75252497e-01 -1.29519594e+00 1.54114040e-02 -3.74359310e-01
-5.32612681e-01 7.73441732e-01 6.56003296e-01 1.61965668e-01
2.86067247e-01 5.87740779e-01 4.50367332e-01 2.18462050e-02
-8.82645547e-01 2.22316772e-01 -2.74307132e-01 1.28475320e+00
2.65547782e-01 -2.11995885e-01 -1.02084219e-01 4.82446969e-01
-1.53724495e-02 3.46825391e-01 8.31830919e-01 1.04734421e+00
1.93731591e-01 -1.39866459e+00 -3.14172924e-01 6.61398828e-01
-6.52946234e-01 1.84753776e-01 -5.67113280e-01 9.95401382e-01
1.41276658e-01 5.38817227e-01 1.19167566e-01 -9.27831605e-02
1.73942000e-01 -7.54291713e-02 6.19599223e-01 -6.89598978e-01
-4.46232885e-01 -9.66241136e-02 -1.88275963e-01 -2.85727739e-01
-3.36976856e-01 -9.74946022e-01 -1.31342304e+00 -2.27719188e-01
-6.08642679e-03 -1.16806254e-01 7.04512417e-01 7.83766389e-01
6.98149025e-01 5.43385088e-01 3.61736149e-01 -6.53801024e-01
-7.09183633e-01 -8.38972092e-01 -7.26165533e-01 8.40846956e-01
5.38429201e-01 -6.52783811e-01 -3.05772811e-01 5.41958153e-01] | [15.039206504821777, -2.9207327365875244] |
bd1b83c5-feda-49c7-bee1-e55a5d0950fa | blind-image-deblurring-using-class-adapted | 1709.01710 | null | http://arxiv.org/abs/1709.01710v1 | http://arxiv.org/pdf/1709.01710v1.pdf | Blind image deblurring using class-adapted image priors | Blind image deblurring (BID) is an ill-posed inverse problem, usually
addressed by imposing prior knowledge on the (unknown) image and on the
blurring filter. Most of the work on BID has focused on natural images, using
image priors based on statistical properties of generic natural images.
However, in many applications, it is known that the image being recovered
belongs to some specific class (e.g., text, face, fingerprints), and exploiting
this knowledge allows obtaining more accurate priors. In this work, we propose
a method where a Gaussian mixture model (GMM) is used to learn a class-adapted
prior, by training on a dataset of clean images of that class. Experiments show
the competitiveness of the proposed method in terms of restoration quality when
dealing with images containing text, faces, or fingerprints. Additionally,
experiments show that the proposed method is able to handle text images at high
noise levels, outperforming state-of-the-art methods specifically designed for
BID of text images. | ['Mário A. T. Figueiredo', 'Marina Ljubenović'] | 2017-09-06 | null | null | null | null | ['blind-image-deblurring'] | ['computer-vision'] | [ 5.90347648e-01 -2.42749974e-01 1.78881213e-01 -2.44005816e-03
-4.53236163e-01 -4.23880786e-01 7.83529937e-01 -2.32356504e-01
-3.91974032e-01 7.76961923e-01 2.34541908e-01 1.77605689e-01
-2.68505484e-01 -4.84492093e-01 -7.12204456e-01 -9.85021114e-01
2.71681190e-01 2.66485006e-01 -1.95777752e-02 1.04858078e-01
4.19598669e-01 5.47032535e-01 -1.77866888e+00 -2.11413633e-02
9.67976451e-01 8.28139186e-01 6.24957561e-01 4.70828861e-01
3.37851457e-02 7.75463104e-01 -6.83212459e-01 -3.04220438e-01
2.00026095e-01 -5.05189419e-01 -6.02106512e-01 5.08502960e-01
5.28967381e-01 -3.10420632e-01 -2.80004889e-01 1.40007091e+00
2.00450182e-01 2.36765876e-01 8.70704710e-01 -6.97103262e-01
-6.60115480e-01 1.06684051e-01 -4.94679272e-01 5.06765321e-02
2.74101138e-01 -1.29720598e-01 4.32651430e-01 -1.05333066e+00
5.47747672e-01 9.76241171e-01 4.19796973e-01 3.77359509e-01
-1.46413910e+00 -4.33744013e-01 -1.78263579e-02 4.07222271e-01
-1.48625815e+00 -7.52410650e-01 8.20291698e-01 -5.91111124e-01
8.70117918e-02 2.33684242e-01 1.16695739e-01 1.01723623e+00
1.43381387e-01 6.37285590e-01 1.47332537e+00 -6.17962956e-01
4.49082851e-01 4.37615424e-01 9.12975296e-02 2.32030436e-01
4.37356830e-01 -9.76474211e-02 -3.93215865e-01 -1.27332255e-01
6.86472178e-01 -1.01412781e-01 -9.24223304e-01 -6.04457557e-01
-8.40082943e-01 5.24197102e-01 1.77273437e-01 6.60137832e-01
-7.45203674e-01 -2.91004002e-01 -9.49537531e-02 -1.01945482e-01
5.43590546e-01 1.06878638e-01 5.70050953e-03 7.83698037e-02
-1.34663641e+00 5.86301759e-02 8.73900712e-01 6.46022320e-01
8.54046404e-01 -1.16628781e-01 -1.96140558e-01 9.87242222e-01
2.68463612e-01 7.00308025e-01 5.51750660e-01 -6.38738990e-01
1.36703461e-01 1.84972391e-01 5.13455451e-01 -1.06971669e+00
9.25800949e-02 -3.99687827e-01 -9.96866167e-01 2.23094866e-01
6.09740674e-01 1.70756102e-01 -9.30556357e-01 1.73025393e+00
1.37004733e-01 4.38446641e-01 2.90695578e-01 1.09489369e+00
3.59017849e-01 5.42375386e-01 -4.00296092e-01 -4.59177494e-01
1.36014867e+00 -5.86780190e-01 -9.63995218e-01 -4.32140201e-01
-3.10157806e-01 -1.07217562e+00 7.16148853e-01 8.39794099e-01
-7.09160030e-01 -6.52777493e-01 -8.82866204e-01 2.40450427e-01
-9.97728258e-02 5.58083951e-01 -1.12938099e-01 1.00418878e+00
-9.09468114e-01 3.77730757e-01 -6.03883743e-01 -5.63542902e-01
1.44409999e-01 1.09232903e-01 -5.04439771e-01 -6.51203513e-01
-8.84605706e-01 9.39231813e-01 2.45279685e-01 2.32560962e-01
-9.88803208e-01 -2.21868664e-01 -7.28939712e-01 3.82336713e-02
3.19426447e-01 -5.35723627e-01 8.59009564e-01 -1.09117317e+00
-1.64662170e+00 6.68672740e-01 -3.42704922e-01 -4.93435204e-01
6.27979934e-01 -5.88804066e-01 -4.36253190e-01 5.13690472e-01
-1.19328946e-01 1.07596023e-02 1.80068982e+00 -1.64908886e+00
-1.18102796e-01 -5.18878877e-01 -1.04541272e-01 -3.82691734e-02
-4.47396994e-01 -6.63379654e-02 -5.94259501e-01 -7.75019050e-01
9.99403745e-02 -8.49908113e-01 1.21694230e-01 -8.98027718e-02
-3.22175294e-01 3.90061289e-01 1.00656235e+00 -1.16663182e+00
9.94744122e-01 -2.24322367e+00 3.66189539e-01 2.47888863e-01
-1.10167861e-01 5.03318906e-01 2.76913848e-02 4.78439897e-01
-6.53975606e-02 -4.22340155e-01 -6.14676058e-01 -4.03986543e-01
-2.54024774e-01 2.68755436e-01 -1.17915295e-01 8.26952100e-01
-6.82774633e-02 6.19325079e-02 -6.18982613e-01 -1.75210357e-01
5.00237346e-01 8.87404025e-01 -1.51624724e-01 4.03281748e-01
-6.21423079e-03 8.30360115e-01 -2.28539467e-01 1.32574752e-01
1.20057905e+00 -5.07548079e-02 1.73547372e-01 -3.63739014e-01
-2.95209233e-02 -4.63907421e-01 -1.53157759e+00 1.40669096e+00
-6.33614063e-01 5.47919452e-01 5.99827409e-01 -1.13778901e+00
1.03300321e+00 3.31531197e-01 2.51630217e-01 -1.64084852e-01
3.71816345e-02 2.62512356e-01 -2.42597565e-01 -7.44376361e-01
4.74647224e-01 -1.11244179e-01 5.31921446e-01 3.54767680e-01
2.44329441e-02 -1.44417629e-01 2.40416214e-01 -9.27166492e-02
7.24974751e-01 3.56333964e-02 1.94683865e-01 -3.03733855e-01
1.11687684e+00 -4.92590576e-01 1.14312656e-01 8.05735886e-01
2.55765885e-01 8.80341768e-01 3.45143229e-02 4.15823497e-02
-8.78548503e-01 -6.79166913e-01 -3.84542555e-01 3.98255318e-01
2.71279514e-01 1.05556585e-01 -1.21170866e+00 -2.63295650e-01
-2.33583793e-01 8.28205764e-01 -4.10769820e-01 -6.70204544e-03
-3.18824470e-01 -8.66430998e-01 1.56295881e-01 -1.47272497e-01
8.04454148e-01 -7.12579489e-01 -4.95754331e-01 1.85688734e-01
-3.22143108e-01 -1.36377132e+00 -3.99965882e-01 -3.13183099e-01
-8.60178471e-01 -1.31777120e+00 -1.35938513e+00 -6.11529529e-01
9.18103993e-01 4.95153755e-01 6.80886149e-01 -2.13298723e-01
-1.52637646e-01 6.70011282e-01 -4.69736338e-01 1.59843832e-01
-5.94467402e-01 -5.10370016e-01 9.42067802e-03 1.11344337e+00
1.29095437e-02 -4.03015822e-01 -4.46661323e-01 5.27545273e-01
-1.32521915e+00 -1.46435872e-01 8.31467927e-01 9.95330572e-01
1.37992039e-01 6.01393700e-01 7.63260499e-02 -7.78789580e-01
6.19848788e-01 -2.75079310e-01 -7.29423821e-01 2.86468476e-01
-5.44054508e-01 1.47038400e-01 6.15080535e-01 -5.03741205e-01
-1.51989853e+00 2.37455234e-01 1.26445815e-01 -4.77776170e-01
-5.87751806e-01 3.97185862e-01 -3.24688554e-01 -3.17847818e-01
6.01162255e-01 6.57922208e-01 2.71832854e-01 -9.27680433e-01
2.56155640e-01 9.49630260e-01 8.49097133e-01 -5.05504429e-01
8.55728984e-01 5.84045887e-01 -5.89078814e-02 -1.25821841e+00
-4.94554698e-01 -5.33759296e-01 -5.32770395e-01 -1.87853351e-01
6.27743423e-01 -7.49464154e-01 -2.93832958e-01 9.81073260e-01
-1.04428494e+00 2.43365578e-03 1.41681686e-01 8.22848439e-01
-4.62962091e-01 8.81844699e-01 -4.03043032e-01 -1.06016374e+00
-1.14177890e-01 -1.22166049e+00 9.21605885e-01 2.74718255e-01
2.84823954e-01 -9.56521809e-01 -2.05344126e-01 5.35282850e-01
6.27258837e-01 -1.16967544e-01 5.50741553e-01 -4.10463721e-01
-5.48021257e-01 -4.30264652e-01 -2.44874150e-01 9.08760250e-01
3.86318803e-01 -4.47649390e-01 -1.17860186e+00 -4.84415591e-01
5.69308341e-01 2.00761870e-01 8.23637724e-01 5.23369789e-01
1.00766468e+00 -3.91839653e-01 -1.62730411e-01 2.66156167e-01
1.53160691e+00 -1.17489584e-01 8.28882039e-01 1.08301796e-01
3.12687367e-01 6.01630986e-01 5.77730060e-01 4.39050585e-01
3.49388346e-02 9.50584888e-01 5.22403359e-01 -2.23410800e-02
-2.44368613e-01 2.07777664e-01 2.67586917e-01 1.26396298e-01
-9.90898311e-02 -3.69291574e-01 -5.92936575e-01 6.03570998e-01
-1.78985429e+00 -8.20169926e-01 -1.28694966e-01 2.61086679e+00
6.23885751e-01 -1.22108288e-01 -2.75559068e-01 3.67771894e-01
1.23235214e+00 7.78203681e-02 -2.81988263e-01 1.52183577e-01
-1.38742566e-01 1.95250139e-01 5.36738575e-01 5.54512084e-01
-1.16214573e+00 5.81743479e-01 5.40907574e+00 8.80312741e-01
-1.09399474e+00 6.35562539e-02 3.04912716e-01 4.93619651e-01
1.02380179e-01 -2.60647549e-03 -4.55470800e-01 6.59293354e-01
5.93641043e-01 -5.85152768e-02 7.63305366e-01 4.61288035e-01
3.35665315e-01 -5.15554845e-01 -7.78748333e-01 1.11383414e+00
6.03842795e-01 -6.62265241e-01 -2.10238457e-01 9.51566473e-02
6.91409826e-01 -5.22130787e-01 2.04183057e-01 -2.76872814e-01
-1.70424268e-01 -8.56878400e-01 5.70600867e-01 8.98506343e-01
5.41544259e-01 -4.57862467e-01 1.05705976e+00 5.91573894e-01
-6.00847781e-01 7.09251687e-03 -2.96376348e-01 1.08996488e-01
5.86511903e-02 1.09708405e+00 -6.86337173e-01 8.25277269e-01
6.04906440e-01 6.82606101e-01 -4.93206471e-01 1.26642096e+00
-3.91914785e-01 5.79090893e-01 -2.49296620e-01 3.81835401e-01
-1.28145084e-01 -4.40726608e-01 8.27940226e-01 1.11265767e+00
6.38138711e-01 -1.71947833e-02 -2.03551859e-01 8.52635324e-01
1.58259273e-01 1.54353708e-01 -4.47903961e-01 2.52532754e-02
1.13064393e-01 1.19870305e+00 -6.65228248e-01 -3.54431599e-01
-3.76388103e-01 1.35181260e+00 -3.91306162e-01 6.43056750e-01
-4.56794292e-01 -2.06721529e-01 3.52335453e-01 -8.85733403e-03
6.96335375e-01 -1.79408282e-01 6.55971244e-02 -1.27950358e+00
1.36924818e-01 -9.65603709e-01 4.60791029e-02 -9.37671721e-01
-1.22120762e+00 5.34353137e-01 1.37917446e-02 -1.38453281e+00
-1.46781102e-01 -6.74065411e-01 -3.26095849e-01 1.11998856e+00
-1.54644465e+00 -9.66752350e-01 -5.62848032e-01 6.59052074e-01
4.92156744e-01 -1.06694631e-01 6.27487838e-01 2.74190187e-01
-3.20683658e-01 1.38159648e-01 6.63596332e-01 -1.24030620e-01
8.85861814e-01 -1.01574099e+00 -9.64239687e-02 1.27060807e+00
9.57816243e-02 6.74465656e-01 1.18202782e+00 -4.34394449e-01
-1.42392325e+00 -8.28428388e-01 4.23133552e-01 8.97012204e-02
3.22534561e-01 -8.31707716e-02 -1.09031677e+00 2.75559098e-01
2.17730805e-01 1.13580875e-01 3.23519349e-01 -3.27675343e-01
-1.59340099e-01 -1.19202547e-01 -1.40913868e+00 1.03217237e-01
4.01637286e-01 -5.37916422e-01 -5.66059768e-01 2.76218563e-01
-1.11002155e-01 -3.87635708e-01 -7.29537129e-01 2.36136228e-01
3.54111999e-01 -1.19656861e+00 1.03886187e+00 1.54405102e-01
1.76135883e-01 -6.17152154e-01 -1.01423152e-01 -1.52578640e+00
-6.84045404e-02 -3.53710860e-01 -1.94241583e-01 1.31153595e+00
-2.49010578e-01 -6.15297198e-01 5.72161853e-01 3.68882805e-01
1.65499762e-01 1.25150338e-01 -7.99579680e-01 -8.36173117e-01
-4.16091025e-01 -7.30542541e-02 3.57984930e-01 7.41471887e-01
-4.58919853e-01 5.76793589e-02 -8.64583492e-01 5.74534535e-01
8.79514337e-01 1.28029034e-01 8.81996632e-01 -1.27755916e+00
-4.92309242e-01 -9.38286483e-02 -5.34741998e-01 -1.09461403e+00
1.87940598e-01 -4.02479917e-01 2.91272372e-01 -1.41346955e+00
1.06953688e-01 -7.40344524e-02 -3.42242718e-02 1.43211577e-02
-2.66580671e-01 3.96135807e-01 1.54015318e-01 3.69444072e-01
-9.54196751e-02 4.46061105e-01 1.06138706e+00 -2.56690621e-01
4.92209047e-02 3.63440633e-01 -4.53055918e-01 5.44004202e-01
5.19536197e-01 -3.54575425e-01 -2.00761303e-01 -2.05916986e-01
-3.63520086e-01 5.56123257e-02 6.00043595e-01 -1.19196534e+00
3.26020569e-01 1.10824846e-01 3.09457600e-01 -3.27686697e-01
6.22285008e-01 -1.11705661e+00 5.37529171e-01 3.42571110e-01
-2.65066605e-02 -7.03892052e-01 1.20936610e-01 8.12153637e-01
-4.79638755e-01 -8.03875446e-01 1.05937123e+00 -4.83774357e-02
-6.18726015e-01 -2.71780550e-01 -4.35775846e-01 -3.48339915e-01
7.42720902e-01 -1.90145224e-01 -2.07563281e-01 -6.87742591e-01
-6.94899142e-01 -3.95190716e-01 6.89576030e-01 2.35831946e-01
5.70487320e-01 -8.18711519e-01 -7.69945621e-01 3.45916450e-01
6.66556060e-02 -2.77551651e-01 4.15081769e-01 9.79087830e-01
-3.32826436e-01 2.05185533e-01 -2.24474177e-01 -5.31586468e-01
-1.34718156e+00 7.94425130e-01 2.37398431e-01 -3.32429893e-02
-4.76497471e-01 5.18550813e-01 2.22611770e-01 4.26387452e-02
1.09725975e-01 4.69655618e-02 -4.67829794e-01 -6.58668429e-02
6.97607160e-01 4.51076120e-01 3.47787887e-01 -1.07148325e+00
-1.51336133e-01 8.99415851e-01 1.32876515e-01 -2.43307278e-01
1.19790220e+00 -3.36275071e-01 -3.67923826e-01 -2.34688334e-02
1.03338468e+00 3.43561292e-01 -1.15202463e+00 -5.23101747e-01
5.25423151e-04 -8.68802011e-01 4.00546610e-01 -5.76456547e-01
-1.01630986e+00 7.45684743e-01 9.20268774e-01 2.13376418e-01
1.42857301e+00 -2.92094439e-01 3.55379969e-01 2.03024700e-01
4.04693991e-01 -8.28180254e-01 1.24276364e-02 1.03592575e-01
9.99584556e-01 -1.06561053e+00 3.39938998e-02 -4.81561631e-01
-3.57481092e-01 1.36087632e+00 3.81241948e-03 -6.25316203e-02
6.06289983e-01 -1.21214435e-01 -7.77782947e-02 1.17068768e-01
5.03107198e-02 -2.74521470e-01 3.80726784e-01 5.72828472e-01
1.48197487e-01 -2.44169250e-01 -4.10215646e-01 2.62166560e-01
1.62953958e-01 1.91257477e-01 7.41390765e-01 7.63647139e-01
-4.41721588e-01 -1.11636019e+00 -1.28791225e+00 1.95213899e-01
-5.56277931e-01 -9.61939692e-02 -1.45879939e-01 4.37193632e-01
7.15225562e-02 1.34280944e+00 -2.66793132e-01 6.80218190e-02
1.37973174e-01 8.78306665e-03 6.11038268e-01 -4.28804129e-01
-1.53975394e-02 2.67589331e-01 -2.58648515e-01 -2.03262255e-01
-8.45472455e-01 -7.71623969e-01 -5.16608953e-01 8.70490521e-02
-5.51597118e-01 2.73342103e-01 9.64872539e-01 9.03522730e-01
7.36816376e-02 2.00767577e-01 6.74951732e-01 -1.13459325e+00
-4.20810223e-01 -1.11598027e+00 -9.43320096e-01 5.54411471e-01
5.89208901e-01 -8.18048239e-01 -6.82048261e-01 5.37274003e-01] | [11.60057258605957, -2.563585042953491] |
15509130-bef8-4e3d-9ea2-692ccd20f828 | correcting-for-selection-bias-and-missing | 2303.16800 | null | https://arxiv.org/abs/2303.16800v2 | https://arxiv.org/pdf/2303.16800v2.pdf | Correcting for Selection Bias and Missing Response in Regression using Privileged Information | When estimating a regression model, we might have data where some labels are missing, or our data might be biased by a selection mechanism. When the response or selection mechanism is ignorable (i.e., independent of the response variable given the features) one can use off-the-shelf regression methods; in the nonignorable case one typically has to adjust for bias. We observe that privileged information (i.e. information that is only available during training) might render a nonignorable selection mechanism ignorable, and we refer to this scenario as Privilegedly Missing at Random (PMAR). We propose a novel imputation-based regression method, named repeated regression, that is suitable for PMAR. We also consider an importance weighted regression method, and a doubly robust combination of the two. The proposed methods are easy to implement with most popular out-of-the-box regression algorithms. We empirically assess the performance of the proposed methods with extensive simulated experiments and on a synthetically augmented real-world dataset. We conclude that repeated regression can appropriately correct for bias, and can have considerable advantage over weighted regression, especially when extrapolating to regions of the feature space where response is never observed. | ['Onno Zoeter', 'Joris M. Mooij', 'Mathijs de Jong', 'Noud de Kroon', 'Philip Boeken'] | 2023-03-29 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [ 5.90252697e-01 -3.01883910e-02 -5.48110306e-01 -7.54315376e-01
-8.90186250e-01 -3.40560079e-01 4.24989879e-01 -4.96192910e-02
-7.14434147e-01 1.50488961e+00 -1.26298845e-01 -4.87966359e-01
-2.88042814e-01 -9.98915493e-01 -1.04573619e+00 -9.01697755e-01
1.28732607e-01 4.85256314e-01 -3.03304851e-01 -5.30847274e-02
3.52663606e-01 2.99650133e-01 -1.52841794e+00 -1.50366902e-01
1.10032213e+00 7.89632618e-01 -2.61103928e-01 2.72818625e-01
1.72259212e-01 9.97309864e-01 -5.51985621e-01 -3.06297064e-01
3.49513561e-01 -1.82290912e-01 -4.48467314e-01 4.57452872e-04
1.57849580e-01 -1.78650558e-01 1.09796494e-01 8.21211576e-01
2.86625236e-01 1.51370496e-01 8.64253461e-01 -1.47564232e+00
-5.00337899e-01 5.69512427e-01 -1.03985155e+00 -3.26603025e-01
3.41475397e-01 -1.93739206e-01 6.29213691e-01 -9.93940413e-01
3.54754359e-01 1.00322223e+00 4.70591933e-01 5.74358165e-01
-1.65247047e+00 -1.04114640e+00 2.21840754e-01 -1.45021468e-01
-1.25241637e+00 -6.24508023e-01 7.96666026e-01 -5.14353693e-01
8.93078670e-02 5.49213111e-01 -4.44341898e-02 1.38205838e+00
3.01687151e-01 5.17888308e-01 1.79109216e+00 -5.26511729e-01
3.90087694e-01 4.71350670e-01 2.96995282e-01 1.94202185e-01
6.28328502e-01 5.03711343e-01 -5.80117881e-01 -8.52820933e-01
4.77273732e-01 3.82889241e-01 -3.19796383e-01 -6.58834577e-01
-1.25048649e+00 1.00996137e+00 4.92275171e-02 -4.37760413e-01
-5.76770663e-01 9.29969549e-02 1.26548419e-02 6.52894914e-01
5.20933211e-01 1.31879300e-01 -7.05312252e-01 3.46127480e-01
-9.84460652e-01 4.46251065e-01 7.23777831e-01 7.33079076e-01
8.78627419e-01 -4.01945747e-02 -1.76860645e-01 9.33479965e-01
1.54243320e-01 5.72182000e-01 3.52086931e-01 -7.38688588e-01
5.78649938e-01 3.71490926e-01 8.50097477e-01 -5.64409614e-01
-3.39663237e-01 -2.04551980e-01 -1.12290084e+00 3.60437483e-01
7.37718582e-01 -4.32157904e-01 -7.40422964e-01 2.01686311e+00
4.25191402e-01 1.09393358e-01 1.49408191e-01 8.00931036e-01
3.40506822e-01 3.62858564e-01 1.87404141e-01 -6.65450871e-01
9.98854995e-01 -3.37454289e-01 -6.82658136e-01 -1.72265396e-01
3.20061088e-01 -6.41198277e-01 8.63017142e-01 6.68495357e-01
-8.76930535e-01 -2.76557297e-01 -9.60021555e-01 3.68150800e-01
-6.40265346e-02 5.13255149e-02 6.68100953e-01 6.72310948e-01
-4.62876409e-01 5.27175665e-01 -6.05068564e-01 -1.24406265e-02
2.10570946e-01 6.90175831e-01 -6.41977310e-01 -2.44207785e-01
-1.34589863e+00 8.76237392e-01 -4.74820752e-03 2.44001940e-01
-7.03112245e-01 -6.30652368e-01 -6.80974603e-01 -2.45766118e-01
8.40875685e-01 -6.35918379e-01 9.40055966e-01 -1.25645828e+00
-1.08862007e+00 4.79937136e-01 -4.85181391e-01 -3.14811915e-01
1.01201022e+00 -1.89938694e-01 -4.98219848e-01 -5.13273299e-01
-6.32343143e-02 1.34497225e-01 1.36903119e+00 -1.31405485e+00
-5.39233863e-01 -5.27278543e-01 -1.28485069e-01 -5.55398986e-02
1.67662166e-02 1.22717887e-01 2.67116666e-01 -7.83364892e-01
1.76732838e-01 -1.09860098e+00 -4.03297424e-01 -1.78155959e-01
-5.33334076e-01 5.53776063e-02 4.58330542e-01 -4.36712831e-01
1.13754368e+00 -1.77612114e+00 -1.94057897e-01 5.62872589e-01
-9.44003314e-02 -2.52337724e-01 1.32075191e-01 3.55696440e-01
-4.33927745e-01 1.71171069e-01 -5.28294623e-01 6.94140280e-03
-2.58108318e-01 1.27683282e-01 -4.89282966e-01 8.77249181e-01
7.63587803e-02 3.70505810e-01 -6.09022617e-01 -3.56873691e-01
1.18374981e-01 3.34647238e-01 -3.41140747e-01 2.11928174e-01
1.50157690e-01 6.15504622e-01 -6.27314746e-01 6.43144011e-01
9.66872752e-01 -9.20215994e-02 1.57241672e-01 2.17144057e-01
-1.77114695e-01 1.90606136e-02 -1.58794641e+00 8.23349297e-01
-7.07499266e-01 8.54830593e-02 -1.34426385e-01 -1.02613795e+00
1.16827691e+00 2.96413153e-01 3.01324517e-01 -1.72180757e-01
-7.91388154e-02 2.97979653e-01 -1.32866859e-01 -1.60026982e-01
4.10145015e-01 -2.14215234e-01 -2.65094906e-01 5.55889487e-01
-3.15582842e-01 3.67008030e-01 -3.78807127e-01 -1.69389576e-01
1.15615726e+00 2.05659598e-01 9.19291258e-01 -3.09697390e-01
5.20463467e-01 -2.22542182e-01 8.94923508e-01 1.13909054e+00
2.20109165e-01 5.39629579e-01 5.99363625e-01 -2.98162013e-01
-7.56496787e-01 -7.31115401e-01 -5.28564274e-01 1.06924236e+00
-1.55665681e-01 3.88939530e-01 -2.57957667e-01 -8.63883555e-01
3.27355474e-01 8.10216427e-01 -8.25276971e-01 -5.36628142e-02
-3.83810103e-01 -1.25234103e+00 -3.72109488e-02 3.23177993e-01
2.00166464e-01 -7.45906830e-01 -2.73114294e-01 3.02099794e-01
-9.24070701e-02 -5.26360154e-01 4.12691645e-02 4.08734024e-01
-9.38820660e-01 -1.13054395e+00 -5.35543323e-01 -1.22208953e-01
1.06868482e+00 2.36959442e-01 9.63594437e-01 1.90908089e-01
4.16072369e-01 2.31424044e-03 -3.02566886e-01 -4.23206896e-01
-3.55260074e-01 -1.55671746e-01 2.17065275e-01 3.83887470e-01
6.66700482e-01 -2.97832727e-01 -6.60400152e-01 6.94639504e-01
-7.34028816e-01 -2.59396344e-01 6.31384969e-01 1.30637562e+00
6.79919064e-01 2.50514634e-02 9.84822452e-01 -1.86687660e+00
4.25580800e-01 -8.89209390e-01 -9.03149366e-01 3.07011157e-01
-9.08430636e-01 2.27173448e-01 7.81011403e-01 -7.37306178e-01
-1.20055652e+00 2.91360170e-01 1.26303509e-01 -9.07177329e-02
-1.74436241e-01 5.11106193e-01 -3.66215855e-01 -2.55556554e-01
6.47764385e-01 -1.36179850e-01 -2.41039574e-01 -5.46455562e-01
-9.32994857e-02 9.96633053e-01 1.51886582e-01 -7.78056979e-01
8.13144803e-01 4.24590290e-01 3.41997921e-01 -3.98943901e-01
-6.26107574e-01 -1.26394629e-01 -4.52653855e-01 9.86671001e-02
1.22497484e-01 -9.84490454e-01 -8.38360190e-01 3.82378876e-01
-6.43470168e-01 -6.48347959e-02 -2.00905129e-01 7.96081305e-01
-4.69573677e-01 8.51747543e-02 2.98283901e-02 -1.19064510e+00
9.14303362e-02 -9.79998827e-01 5.49981952e-01 -7.00824261e-02
-3.52331758e-01 -7.45633304e-01 5.73884136e-05 2.36657426e-01
1.66196331e-01 5.95266223e-01 8.37578297e-01 -9.80085433e-01
-4.36413199e-01 -5.96869946e-01 8.56701657e-02 1.88518122e-01
1.45462947e-02 -3.73293348e-02 -1.10644174e+00 -3.51518124e-01
1.96380809e-01 -3.07442367e-01 9.98589933e-01 5.46475351e-01
1.44025576e+00 -5.76036930e-01 -1.81981042e-01 2.96474695e-01
1.40153944e+00 1.10084051e-02 5.98489940e-01 3.35625947e-01
3.86165500e-01 7.72120893e-01 1.16864288e+00 6.70462072e-01
2.23028064e-01 7.39829123e-01 3.55050772e-01 -1.75317690e-01
6.33881986e-01 -3.60301346e-01 2.52484292e-01 3.40643942e-01
-1.77561909e-01 -2.02637836e-01 -4.57800031e-01 4.03097123e-01
-2.04706025e+00 -8.60242128e-01 -3.99504066e-01 3.04233694e+00
9.84934509e-01 6.23434484e-02 1.31693832e-03 3.57518852e-01
9.12248373e-01 -1.23181045e-01 -6.91896737e-01 -5.14677346e-01
-6.21123686e-02 1.63081989e-01 1.02083004e+00 5.82147121e-01
-1.01091373e+00 2.68263906e-01 6.08767128e+00 7.83098757e-01
-7.04318404e-01 2.13670462e-01 7.80677319e-01 -5.06397747e-02
-4.46712613e-01 2.75005013e-01 -8.14973235e-01 6.57580614e-01
8.48297775e-01 -5.56938127e-02 4.15016890e-01 9.72110033e-01
3.04289699e-01 -4.37389731e-01 -1.38279688e+00 5.18623769e-01
-1.30415946e-01 -6.97809756e-01 -3.06438684e-01 2.01563075e-01
7.98623264e-01 -4.18363392e-01 9.27567631e-02 4.29438204e-01
5.35829842e-01 -1.07313275e+00 3.33998650e-01 7.57983983e-01
8.98774147e-01 -8.56877863e-01 1.03929210e+00 6.80167794e-01
-4.44206178e-01 -2.74892282e-02 -5.72396815e-01 -1.80584654e-01
-2.50114918e-01 9.96539831e-01 -4.13455606e-01 4.95305359e-01
2.66364902e-01 2.47115403e-01 -3.48474532e-01 9.58636880e-01
-2.95398980e-01 7.36495018e-01 -3.23918909e-01 9.42732319e-02
-4.05891925e-01 -1.78644985e-01 3.52093726e-01 6.60675168e-01
5.55911303e-01 1.42338946e-02 -5.62503263e-02 5.64825654e-01
-1.94271579e-01 1.43914536e-01 -6.35317028e-01 7.42716074e-01
6.96824133e-01 1.04505956e+00 -2.01375484e-01 -9.50265303e-02
-6.91444814e-01 4.24410254e-01 1.27917230e-01 5.95991731e-01
-7.64869869e-01 -1.70748308e-01 3.96498859e-01 1.04643069e-01
1.65129408e-01 3.51393402e-01 -3.17602247e-01 -1.31914687e+00
2.09630523e-02 -9.66586113e-01 5.32811403e-01 -5.66879332e-01
-1.65118158e+00 3.12213987e-01 2.79849082e-01 -1.36516619e+00
-4.98677194e-01 -3.64748627e-01 -2.31649294e-01 1.19440019e+00
-1.69142509e+00 -8.40098560e-01 -5.87170012e-02 7.87002981e-01
1.67724386e-01 -9.13585424e-02 8.05416286e-01 5.19166589e-02
-4.48468715e-01 6.88584566e-01 2.68888354e-01 -3.69866043e-01
1.11901057e+00 -1.12296104e+00 -2.39408705e-02 6.17073596e-01
-2.45800465e-01 8.33326340e-01 9.18619514e-01 -6.86213195e-01
-1.25923920e+00 -1.22829258e+00 8.95727336e-01 -5.12914717e-01
4.08448428e-01 -4.91258204e-01 -1.11355352e+00 7.39891231e-01
-4.25735414e-01 2.12947488e-01 9.06228125e-01 4.48254347e-01
-3.05592209e-01 -3.99576664e-01 -1.62572837e+00 4.37375695e-01
5.50689638e-01 -2.41973121e-02 -6.05405271e-01 2.06505284e-01
2.87875324e-01 -2.46749222e-01 -7.75663733e-01 6.09797835e-01
7.34450102e-01 -6.96094692e-01 7.78152764e-01 -6.14082932e-01
2.24424019e-01 -3.19751322e-01 -2.39281833e-01 -1.29256451e+00
-1.89586729e-01 -5.18076897e-01 -1.34896889e-01 1.34580374e+00
5.10388374e-01 -1.02350509e+00 7.57125735e-01 1.08222461e+00
7.16215193e-01 -5.49597740e-01 -1.08383930e+00 -6.62427604e-01
7.24897310e-02 -4.25104231e-01 1.03464150e+00 1.09720659e+00
-2.78297216e-01 6.36391267e-02 -1.07707345e+00 1.55472487e-01
8.34562719e-01 1.76266924e-01 9.06116962e-01 -1.48953998e+00
-3.97872686e-01 3.76996547e-01 -6.17187209e-02 -5.80561936e-01
3.93491626e-01 -3.81120265e-01 -1.07651874e-01 -7.58255243e-01
3.72334629e-01 -7.73655713e-01 -5.89026332e-01 6.30434394e-01
-2.86113918e-01 1.98382601e-01 -2.26107270e-01 2.98124671e-01
1.00984611e-01 2.80519754e-01 8.74518156e-01 1.19550884e-01
-2.76803941e-01 8.32716227e-01 -1.02202225e+00 9.00428176e-01
7.75372088e-01 -1.06716812e+00 -2.30678186e-01 1.69134244e-01
4.37287331e-01 4.55201983e-01 5.26439071e-01 -3.94960284e-01
8.18351563e-03 -7.16093838e-01 6.46192431e-01 -4.62382585e-01
1.81821048e-01 -1.29842365e+00 4.47779804e-01 1.08207315e-01
-6.90567791e-01 -3.73689592e-01 -3.44590396e-01 6.50615990e-01
3.97824012e-02 -4.82899994e-01 6.85139298e-01 2.99884230e-02
-1.00056268e-01 8.32283869e-02 -1.88645422e-01 -1.45768046e-01
9.64302540e-01 -1.57018647e-01 -4.30550963e-01 -4.60337490e-01
-5.51842988e-01 2.65881240e-01 5.18205047e-01 1.20334737e-01
5.88873088e-01 -1.22024846e+00 -9.15877163e-01 4.00164455e-01
1.96521387e-01 -4.08035040e-01 1.35955259e-01 8.04837108e-01
1.90226257e-01 1.77447096e-01 2.76271012e-02 -1.63306236e-01
-1.16889656e+00 7.20062733e-01 -1.00813195e-01 -2.15164095e-01
-1.28082350e-01 2.58191824e-01 2.50622988e-01 -5.81482708e-01
4.65629920e-02 -2.39657126e-02 -2.33727276e-01 -4.34050448e-02
5.04920542e-01 5.87151587e-01 1.47857919e-01 -5.23455679e-01
-3.62788528e-01 4.26597327e-01 -1.56002060e-01 -1.23991244e-01
1.19695127e+00 -1.41821012e-01 -1.79221585e-01 7.46942520e-01
7.45817721e-01 4.90899771e-01 -1.16515076e+00 -2.75245309e-01
-2.56104976e-01 -9.86386240e-01 1.30220443e-01 -9.67892706e-01
-8.08503330e-01 5.69295168e-01 3.80317718e-01 2.00683251e-01
1.26792479e+00 -5.35514116e-01 8.81750975e-03 3.84499341e-01
7.16258824e-01 -9.14196610e-01 -8.08000207e-01 -1.45915121e-01
7.73714244e-01 -1.67900193e+00 5.36449909e-01 -3.96678537e-01
-9.16989326e-01 8.08828652e-01 5.47941089e-01 -2.45217443e-01
5.51049650e-01 1.37861103e-01 -1.82206646e-01 4.19862509e-01
-1.08165920e+00 3.03584605e-01 1.53323561e-01 5.60112119e-01
4.10051674e-01 1.76262319e-01 -7.22332954e-01 7.97996342e-01
9.47236419e-02 3.01303118e-01 9.51076567e-01 8.39783370e-01
-1.24767117e-01 -1.29373801e+00 -8.71501863e-01 9.16051269e-01
-7.70305336e-01 7.43437707e-02 -1.88569933e-01 1.05840886e+00
-4.76225428e-02 1.25043547e+00 -2.17082441e-01 -1.38609940e-02
4.17822957e-01 6.33631200e-02 2.12220848e-01 -3.12944978e-01
-4.81817186e-01 7.27107078e-02 1.29416943e-01 -4.43498194e-01
-5.34470141e-01 -8.39777112e-01 -6.23874605e-01 -5.55847347e-01
-6.65232062e-01 1.99538261e-01 6.48597479e-01 7.88122773e-01
-2.98578292e-02 2.11475551e-01 1.14824402e+00 -2.09548652e-01
-9.30012703e-01 -8.74207258e-01 -7.80085921e-01 1.16660081e-01
6.67832494e-01 -9.50445652e-01 -7.13377178e-01 -2.47352362e-01] | [8.744263648986816, 4.6465983390808105] |
4bad6632-cb69-4105-a362-a771f131ffb8 | long-scale-error-control-in-low-light-image | 2206.01334 | null | https://arxiv.org/abs/2206.01334v1 | https://arxiv.org/pdf/2206.01334v1.pdf | Long Scale Error Control in Low Light Image and Video Enhancement Using Equivariance | Image frames obtained in darkness are special. Just multiplying by a constant doesn't restore the image. Shot noise, quantization effects and camera non-linearities mean that colors and relative light levels are estimated poorly. Current methods learn a mapping using real dark-bright image pairs. These are very hard to capture. A recent paper has shown that simulated data pairs produce real improvements in restoration, likely because huge volumes of simulated data are easy to obtain. In this paper, we show that respecting equivariance -- the color of a restored pixel should be the same, however the image is cropped -- produces real improvements over the state of the art for restoration. We show that a scale selection mechanism can be used to improve reconstructions. Finally, we show that our approach produces improvements on video restoration as well. Our methods are evaluated both quantitatively and qualitatively. | ['David Forsyth', 'Sara Aghajanzadeh'] | 2022-06-02 | null | null | null | null | ['video-enhancement', 'video-restoration'] | ['computer-vision', 'computer-vision'] | [ 3.71124536e-01 -3.55891079e-01 3.29866827e-01 -3.16798598e-01
-7.37113535e-01 -5.40586114e-01 3.57300341e-01 -3.25785100e-01
-3.34015936e-01 9.82337296e-01 2.84222186e-01 -2.44010482e-02
4.09834951e-01 -6.04072213e-01 -8.72648239e-01 -8.41338336e-01
-1.12183601e-01 4.35619196e-03 3.62762690e-01 -3.79145682e-01
4.56520647e-01 5.13581395e-01 -1.53116405e+00 3.04845393e-01
6.54738784e-01 5.32181203e-01 3.12906414e-01 1.14430296e+00
1.20491683e-01 8.32388461e-01 -5.99922121e-01 -2.20910519e-01
9.73473430e-01 -9.35386956e-01 -5.31866372e-01 5.34493983e-01
1.04588604e+00 -8.70633304e-01 -6.15770102e-01 1.36314273e+00
4.25625265e-01 1.03087783e-01 4.07711923e-01 -1.04602182e+00
-7.07795560e-01 -7.89045244e-02 -8.23267937e-01 2.18663260e-01
5.30073345e-01 4.70563442e-01 5.91355264e-01 -7.72009075e-01
8.13050032e-01 1.33141518e+00 6.28859103e-01 4.96098429e-01
-1.69449615e+00 -4.15538341e-01 -3.34776968e-01 1.68722108e-01
-1.20432842e+00 -8.00185978e-01 5.09419918e-01 -1.07612893e-01
7.68112242e-01 3.61234903e-01 6.80281997e-01 6.24513805e-01
3.95047337e-01 1.80175841e-01 1.55289555e+00 -7.65365660e-01
2.35387012e-01 6.18708916e-02 -5.52469373e-01 6.26444697e-01
3.39213938e-01 5.38026810e-01 -6.39367163e-01 2.99624145e-01
1.25583351e+00 -1.06823534e-01 -7.09981143e-01 -4.17429894e-01
-1.47284234e+00 3.11073273e-01 5.25407612e-01 8.11622217e-02
-1.81301415e-01 4.21213388e-01 3.52874486e-04 6.33581519e-01
3.44574541e-01 4.35350120e-01 -7.89475664e-02 1.85064860e-02
-1.04713905e+00 -3.74204740e-02 8.41639996e-01 8.80132377e-01
9.68548954e-01 2.58301705e-01 2.47906312e-01 6.97643578e-01
7.55241290e-02 7.87023425e-01 5.63194752e-02 -1.76244521e+00
2.49810576e-01 -8.35211799e-02 5.44444144e-01 -9.44498658e-01
-1.17258362e-01 7.32959509e-02 -9.43959534e-01 1.17904723e+00
7.39841998e-01 -1.32059693e-01 -1.07735968e+00 1.47703838e+00
-9.81193557e-02 3.01057398e-01 1.78142458e-01 1.15725982e+00
2.65261024e-01 7.80119538e-01 -3.81060451e-01 -5.41740179e-01
9.89276826e-01 -6.34224176e-01 -1.03448725e+00 -3.67258042e-01
5.09968847e-02 -1.25274503e+00 1.22243726e+00 6.99560225e-01
-1.51645887e+00 -5.55796683e-01 -1.30216634e+00 -2.63028026e-01
5.10809645e-02 -2.72192806e-01 2.67489672e-01 6.36112154e-01
-1.52953768e+00 9.68594372e-01 -6.65867746e-01 -5.11040390e-01
2.50497490e-01 1.18243672e-01 -4.32322919e-01 -4.91787255e-01
-7.63970017e-01 1.09553063e+00 1.56790763e-02 -2.62372911e-01
-7.78311789e-01 -4.92983580e-01 -6.57874346e-01 -9.11542699e-02
5.61404303e-02 -6.48234546e-01 1.00720036e+00 -1.15124893e+00
-1.56012428e+00 9.85682547e-01 -2.43268043e-01 -3.77267182e-01
6.65536344e-01 -3.00927106e-02 -4.34435785e-01 4.62279916e-01
-1.05075836e-01 7.90676594e-01 7.79789329e-01 -1.79087472e+00
-6.46837771e-01 -8.73826295e-02 -2.73418315e-02 2.73938209e-01
-9.68111306e-02 -8.72569904e-02 -4.96052951e-01 -5.33407271e-01
3.52731466e-01 -6.75886571e-01 -2.36525252e-01 5.78752875e-01
-1.45797670e-01 7.89691746e-01 7.36762404e-01 -8.29469502e-01
6.95435166e-01 -2.11698079e+00 -1.34657249e-01 -1.71690877e-03
2.83696968e-02 -6.77341074e-02 -2.87949234e-01 4.17877108e-01
-9.49386805e-02 -6.34883344e-02 -3.14850599e-01 -2.87327737e-01
-2.17209339e-01 5.28593361e-01 -3.82341743e-01 9.38502431e-01
-7.70843923e-02 4.50743049e-01 -8.34789336e-01 -5.17374277e-01
7.95559704e-01 8.08576167e-01 -5.84425509e-01 4.24225852e-02
1.23189434e-01 3.04373831e-01 5.83745778e-01 4.46642905e-01
1.06112790e+00 -4.71818745e-02 1.28290191e-01 -4.49162990e-01
-1.03451081e-01 -1.69752464e-01 -1.36431980e+00 1.64097202e+00
-3.55760336e-01 1.13265717e+00 2.55708516e-01 -5.28598487e-01
6.82243526e-01 1.64229229e-01 3.53193969e-01 -1.03225005e+00
-1.31532088e-01 2.28735968e-01 -1.09993242e-01 -4.32753772e-01
6.08417690e-01 -5.54523826e-01 6.34211957e-01 2.37879068e-01
-2.62841642e-01 -8.16066742e-01 3.61173958e-01 2.72649020e-01
9.86547232e-01 2.10262865e-01 1.65724561e-01 -3.98821592e-01
5.53217670e-03 7.88723826e-02 5.80370367e-01 6.67633414e-01
-3.30508143e-01 1.21278346e+00 2.11001083e-01 -3.64605933e-01
-1.70873809e+00 -1.25721681e+00 -3.90907191e-02 5.56862593e-01
7.80877769e-01 -8.25925320e-02 -8.02251697e-01 2.13147141e-02
-3.36956948e-01 6.91604316e-01 -4.12636101e-01 2.13622779e-01
-5.84035039e-01 -7.12799489e-01 8.23976919e-02 2.67789453e-01
5.51493645e-01 -6.57611609e-01 -7.79181600e-01 1.08665228e-01
-2.13306665e-01 -1.27908731e+00 -4.25758660e-01 2.76570708e-01
-1.05147016e+00 -9.78285611e-01 -7.57750332e-01 -8.68737280e-01
8.50382626e-01 8.81617606e-01 1.32398868e+00 2.66995460e-01
-4.48369354e-01 2.07707301e-01 -1.70647442e-01 1.50359035e-01
-5.15025795e-01 -9.73046720e-01 8.07815120e-02 -2.38317266e-01
-1.72436595e-01 -7.09005237e-01 -1.05795014e+00 4.13465977e-01
-1.05763304e+00 3.28180611e-01 3.31125736e-01 8.28996837e-01
6.17684603e-01 1.64126739e-01 -5.39078452e-02 -6.14645064e-01
2.30449617e-01 3.15397531e-01 -6.73734069e-01 1.34814009e-01
-7.48947561e-01 1.06989354e-01 6.33795440e-01 -2.53259927e-01
-1.15972292e+00 3.44211608e-01 1.99812382e-01 -2.65575945e-01
-1.99163884e-01 -4.89478022e-01 4.49449308e-02 -4.37783897e-01
1.00928068e+00 4.81431782e-02 1.73554420e-01 -3.60454440e-01
3.66043299e-01 3.74807835e-01 1.11134911e+00 -1.52570963e-01
9.59836006e-01 1.12203705e+00 1.64214984e-01 -8.59879315e-01
-3.38453472e-01 -3.89663905e-01 -5.29163718e-01 -2.19365537e-01
8.07756364e-01 -9.74991441e-01 -4.45481837e-01 3.52144152e-01
-1.07162917e+00 -6.48781717e-01 -5.86976171e-01 4.62641776e-01
-9.35679257e-01 5.66981912e-01 -8.45774353e-01 -6.78817272e-01
2.17585564e-01 -1.06208324e+00 8.98959816e-01 4.34808135e-01
2.60950655e-01 -9.37336445e-01 1.08438849e-01 1.41162932e-01
5.43105304e-01 4.09018109e-03 4.58788961e-01 4.58060950e-01
-8.78112853e-01 1.37337312e-01 -5.02542436e-01 5.34701943e-01
2.83061296e-01 2.10578293e-01 -1.04985666e+00 -4.21362162e-01
-2.97334604e-03 -4.61922958e-02 8.40644717e-01 5.47558784e-01
7.64961362e-01 -1.92613959e-01 2.08478179e-02 8.78818333e-01
1.94915140e+00 5.57217188e-02 1.49014854e+00 4.39539790e-01
4.34744686e-01 3.46950978e-01 4.52593565e-01 3.06464016e-01
-1.30983349e-03 6.50229871e-01 5.54282188e-01 -6.75655067e-01
-8.84999275e-01 2.58510429e-02 4.25800622e-01 6.11985981e-01
-2.13866308e-01 -3.24339181e-01 -6.15609467e-01 4.31743324e-01
-1.30710924e+00 -1.24552929e+00 -4.18189645e-01 2.38428903e+00
9.69786048e-01 -5.94202653e-02 -2.93860286e-01 2.90165454e-01
6.62034631e-01 -4.44821492e-02 -2.69361049e-01 -1.56930000e-01
-7.22096801e-01 1.04088746e-01 1.06472409e+00 9.98710692e-01
-7.34697938e-01 8.19471061e-01 7.66870451e+00 3.21940184e-01
-9.94990408e-01 -9.05113202e-03 7.03341484e-01 -1.85828418e-01
-2.88227469e-01 3.39305550e-01 -9.35743824e-02 4.37545478e-01
7.43122280e-01 -2.80302600e-03 7.44670033e-01 1.95261449e-01
4.88314956e-01 -6.60677850e-01 -1.00194097e+00 1.29539442e+00
3.52803051e-01 -1.24981415e+00 -1.59387589e-01 1.93531707e-01
1.19195759e+00 -8.85882322e-03 6.41567037e-02 -4.57400680e-01
5.26411295e-01 -8.47566247e-01 5.90407133e-01 8.00805688e-01
8.17520916e-01 -5.65610349e-01 4.63518590e-01 -5.29182069e-02
-7.25277722e-01 1.37991875e-01 -6.95704281e-01 -3.98102067e-02
4.18357968e-01 5.70921123e-01 -5.20765722e-01 2.46134341e-01
8.34876597e-01 6.49001122e-01 -6.75608158e-01 1.42217159e+00
-1.49789512e-01 2.61780411e-01 -2.29414940e-01 5.23751676e-01
-2.15679020e-01 -4.85209882e-01 3.30458015e-01 1.18185866e+00
4.50475335e-01 2.50318497e-01 1.95082668e-02 4.43239748e-01
4.19771783e-02 -1.29845098e-01 -5.70529759e-01 4.75665718e-01
7.89112747e-02 9.46063995e-01 -9.07782793e-01 -5.87620199e-01
-5.19892037e-01 1.55653441e+00 -3.24099720e-01 7.83377528e-01
-4.10631359e-01 -3.65103126e-01 6.75157845e-01 3.09817344e-01
1.07183546e-01 -2.66002685e-01 -4.81342256e-01 -1.26640141e+00
-2.62397379e-01 -1.00103915e+00 4.28169593e-03 -1.56342125e+00
-1.17776370e+00 3.27581167e-01 -2.57013321e-01 -1.64302659e+00
-2.01281607e-01 -5.38595557e-01 -3.59705806e-01 7.45128691e-01
-1.51813090e+00 -5.67012072e-01 -6.31957114e-01 5.65476418e-01
5.09835064e-01 1.18153512e-01 6.37587368e-01 4.09846544e-01
2.30720192e-02 1.16988383e-01 5.66460729e-01 -8.24173763e-02
1.18557525e+00 -1.35560715e+00 4.03234988e-01 1.32737446e+00
3.38586926e-01 1.93163887e-01 1.31090069e+00 -3.79654467e-01
-1.45619488e+00 -4.64269668e-01 4.49376076e-01 -1.66773185e-01
5.29759884e-01 4.30441424e-02 -1.01269758e+00 4.66304749e-01
6.30868018e-01 3.83954555e-01 1.06189147e-01 -5.75315356e-01
-3.95331770e-01 -3.62457514e-01 -1.32553363e+00 6.22047663e-01
9.32724357e-01 -5.66868007e-01 -2.49743506e-01 4.82761174e-01
3.31254542e-01 -3.70084882e-01 -5.35048544e-01 -1.00748837e-01
5.34803331e-01 -1.53156531e+00 1.29553783e+00 -2.35281482e-01
5.43837070e-01 -6.32550657e-01 -5.37601113e-01 -1.55103791e+00
-1.19810127e-01 -6.02734089e-01 1.70274124e-01 8.83572280e-01
7.23451748e-02 -4.82445985e-01 5.87234199e-01 5.49219251e-01
1.49581924e-01 1.41291216e-01 -8.83717239e-01 -9.29137766e-01
-1.27573937e-01 -2.10572481e-02 2.14045733e-01 8.66330564e-01
-1.46348521e-01 1.11787483e-01 -7.54323483e-01 1.44875035e-01
1.15896571e+00 4.83491080e-04 8.53366673e-01 -9.00080025e-01
-2.67746210e-01 -1.83205232e-01 -8.07397306e-01 -9.62592244e-01
-3.83335739e-01 -2.60544509e-01 3.18244249e-01 -1.70750451e+00
2.42862716e-01 -2.27771580e-01 -5.13573252e-02 1.29979745e-01
-4.44692336e-02 9.79901314e-01 3.58768344e-01 3.35961521e-01
-4.12253588e-01 5.30854091e-02 1.51743460e+00 -5.93261980e-02
-3.53627950e-02 -5.64842820e-01 -4.63224262e-01 6.16024017e-01
8.64839494e-01 -2.73401022e-01 -3.32169980e-01 -6.04895353e-01
1.63069502e-01 2.08925590e-01 5.80203176e-01 -1.25780678e+00
2.44143814e-01 -1.10677883e-01 8.95776093e-01 -2.45882347e-01
7.22233474e-01 -9.47262108e-01 2.85876602e-01 4.50718999e-01
-1.31121635e-01 1.11672603e-01 6.65631071e-02 6.91156626e-01
-1.11997657e-01 -1.36874527e-01 1.25389218e+00 -2.60717809e-01
-8.08172524e-01 -1.41444832e-01 -1.60572693e-01 -1.38547802e-02
7.96197593e-01 -4.95475471e-01 -5.86606383e-01 -8.13563466e-01
-4.78987068e-01 -2.83529013e-01 1.25543332e+00 -5.30343428e-02
6.62319064e-01 -1.26957786e+00 -8.79685640e-01 2.22778022e-01
-2.01763958e-01 -5.13920784e-01 3.33672881e-01 7.98768580e-01
-1.26316237e+00 -4.09147352e-01 -5.06162524e-01 -7.50707150e-01
-1.48360074e+00 5.63378274e-01 5.20156622e-01 4.24190283e-01
-1.05004835e+00 6.00209713e-01 1.44446880e-01 2.29260191e-01
1.95238233e-01 -2.37829193e-01 5.17355740e-01 -2.96294510e-01
8.80108476e-01 3.36355299e-01 -9.17916223e-02 -5.86344719e-01
-1.18790001e-01 8.01605940e-01 2.40457639e-01 -6.57544672e-01
1.43197918e+00 -6.08916163e-01 -1.19836405e-01 3.08358818e-01
1.20830321e+00 2.20717505e-01 -1.84124374e+00 -6.01279847e-02
-5.05153716e-01 -1.27217698e+00 3.23438466e-01 -8.68355095e-01
-1.33897018e+00 9.09192324e-01 1.25698662e+00 2.54688829e-01
1.52680004e+00 -3.12947631e-01 5.86937547e-01 3.58066678e-01
3.24795872e-01 -1.15654576e+00 1.16567373e-01 -9.35498700e-02
7.06468403e-01 -1.55738974e+00 4.49851602e-01 -2.96916366e-01
-5.17017663e-01 1.17902279e+00 2.78246194e-01 -2.54282385e-01
3.08110178e-01 6.79600358e-01 5.00126183e-01 1.30781054e-01
-6.25923097e-01 -2.19837651e-01 -2.10328192e-01 8.34287703e-01
2.57602811e-01 -1.75557137e-01 3.68940048e-02 -8.00241292e-01
-6.12572879e-02 4.57198024e-02 1.12769091e+00 8.25819254e-01
-8.32712054e-01 -1.00874650e+00 -9.79406536e-01 1.07113190e-01
-1.97429150e-01 -1.61169648e-01 -1.59522057e-01 7.14541376e-01
-1.41804293e-01 1.21556723e+00 1.86351717e-01 -2.03001082e-01
1.50152102e-01 -3.35364580e-01 9.87134576e-01 -2.01070935e-01
-1.20837368e-01 2.98922241e-01 -1.69153437e-01 -8.00116658e-01
-6.46901786e-01 -4.00678545e-01 -1.31593609e+00 -7.99385488e-01
-6.94193393e-02 -2.36550599e-01 8.04812431e-01 3.75909865e-01
-1.80524848e-02 3.74382675e-01 7.03564167e-01 -1.07909942e+00
-1.94525272e-01 -5.82973480e-01 -7.93303728e-01 6.60547376e-01
7.39985347e-01 -2.53475606e-01 -8.34829807e-01 8.25346708e-01] | [10.834465026855469, -2.3189361095428467] |
2a699390-bcbb-4677-a19c-e3e1bc489c69 | image-segmentation-based-unsupervised | 2212.10124 | null | https://arxiv.org/abs/2212.10124v1 | https://arxiv.org/pdf/2212.10124v1.pdf | Image Segmentation-based Unsupervised Multiple Objects Discovery | Unsupervised object discovery aims to localize objects in images, while removing the dependence on annotations required by most deep learning-based methods. To address this problem, we propose a fully unsupervised, bottom-up approach, for multiple objects discovery. The proposed approach is a two-stage framework. First, instances of object parts are segmented by using the intra-image similarity between self-supervised local features. The second step merges and filters the object parts to form complete object instances. The latter is performed by two CNN models that capture semantic information on objects from the entire dataset. We demonstrate that the pseudo-labels generated by our method provide a better precision-recall trade-off than existing single and multiple objects discovery methods. In particular, we provide state-of-the-art results for both unsupervised class-agnostic object detection and unsupervised image segmentation. | ['Quoc-Cuong Pham', 'Florian Chabot', 'Hejer Ammar', 'Sandra Kara'] | 2022-12-20 | null | null | null | null | ['object-discovery', 'class-agnostic-object-detection'] | ['computer-vision', 'computer-vision'] | [ 4.83986467e-01 2.58818507e-01 -2.87499100e-01 -6.02554083e-01
-8.95393312e-01 -6.37400687e-01 4.52935427e-01 5.84455788e-01
-4.91052538e-01 2.54387081e-01 -2.89829284e-01 3.39274436e-01
-1.70671687e-01 -8.27005565e-01 -8.14576328e-01 -6.34895563e-01
1.54211059e-01 7.58229017e-01 8.15659523e-01 4.11703408e-01
2.38116995e-01 7.11316466e-01 -2.00736642e+00 5.09967387e-01
7.06507921e-01 1.29733157e+00 3.20376784e-01 4.51910287e-01
-4.98952061e-01 5.77816963e-01 -4.03809577e-01 -2.10247442e-01
3.15712452e-01 -3.31989378e-02 -1.08976197e+00 7.51899779e-01
6.26992524e-01 -3.21506530e-01 1.90163210e-01 1.26422155e+00
4.19207215e-02 -7.40997940e-02 5.99221945e-01 -1.02217650e+00
-3.53557110e-01 6.25901937e-01 -7.15639889e-01 4.50347178e-02
-2.45708406e-01 3.98478024e-02 1.11813688e+00 -1.10200977e+00
5.14827549e-01 1.17589390e+00 4.71952975e-01 4.00039136e-01
-1.38030672e+00 -5.54200411e-01 2.93941885e-01 -1.25563741e-01
-1.49430525e+00 -2.53330261e-01 6.30898535e-01 -6.27842546e-01
5.61887980e-01 9.41947326e-02 4.33939427e-01 3.93785387e-01
-4.43765640e-01 1.09971130e+00 9.02900755e-01 -4.69426215e-01
2.90774822e-01 3.61669809e-01 5.07827103e-01 7.55144358e-01
5.40345728e-01 -1.74324766e-01 -3.81800681e-01 2.96573173e-02
7.12932408e-01 2.92029560e-01 3.87021422e-01 -6.80911958e-01
-9.55462635e-01 6.28449738e-01 5.68826437e-01 4.81206357e-01
-5.98024309e-01 1.32474989e-01 2.30546296e-01 -3.17242235e-01
4.40740526e-01 3.45984399e-01 -6.71448112e-01 4.17297393e-01
-1.23638725e+00 1.90132990e-01 5.23010075e-01 1.12019229e+00
1.20588946e+00 -4.91217107e-01 -3.61896485e-01 7.89846957e-01
4.96126056e-01 9.14628431e-03 3.13710392e-01 -8.03087354e-01
2.68955845e-02 1.15608847e+00 -8.71940516e-03 -7.01809704e-01
-4.57589924e-01 -6.74295068e-01 -3.43585193e-01 7.86284059e-02
4.24924344e-01 2.93303639e-01 -1.29267859e+00 1.32881737e+00
7.36149788e-01 1.77431665e-02 -1.08119845e-01 7.77110338e-01
1.08451235e+00 7.74437115e-02 3.77621025e-01 1.81302056e-02
1.70561337e+00 -1.24202633e+00 -4.99515235e-01 -3.54498714e-01
3.50703090e-01 -6.09403670e-01 6.72184050e-01 1.03910968e-01
-9.98326719e-01 -7.09697604e-01 -8.17069948e-01 -2.40663961e-02
-5.24517655e-01 4.84586030e-01 6.12874746e-01 4.71464723e-01
-6.92206740e-01 4.61963147e-01 -8.10913444e-01 -3.44237208e-01
1.07634699e+00 5.76396942e-01 -2.38055959e-01 5.04936371e-03
-5.40380359e-01 3.90700459e-01 9.33154166e-01 9.04997736e-02
-1.38356078e+00 -6.40893877e-01 -7.52671778e-01 2.24983767e-01
6.90381408e-01 -3.87202531e-01 1.36746287e+00 -1.29041648e+00
-1.12944388e+00 1.29207087e+00 -2.47155353e-01 -5.67609370e-01
3.62029910e-01 -4.08340275e-01 -1.37515934e-04 3.87407809e-01
4.17684466e-01 9.81140196e-01 9.28891182e-01 -1.80633831e+00
-1.11232257e+00 -4.19636965e-01 7.21898749e-02 -9.20641720e-02
-3.63148689e-01 1.52441710e-01 -9.13246453e-01 -5.25470555e-01
3.91557753e-01 -6.54584587e-01 -3.90337139e-01 1.07704841e-01
-8.67510259e-01 -3.54113340e-01 9.61265504e-01 -1.58984452e-01
8.13749909e-01 -2.01908493e+00 -1.15886241e-01 3.05988818e-01
4.70345080e-01 2.83281595e-01 -1.83816552e-01 -4.53463979e-02
2.72727367e-02 1.54812440e-01 -4.77435738e-01 -5.78034937e-01
-1.48047373e-01 1.46558210e-01 -4.26867418e-03 4.06427294e-01
6.23213410e-01 9.46049392e-01 -9.21199799e-01 -8.41271102e-01
2.16260925e-01 1.60325974e-01 -2.77231723e-01 1.88710958e-01
-4.33535159e-01 4.65382010e-01 -4.39066261e-01 9.88576770e-01
7.51007020e-01 -5.47839642e-01 2.03035921e-01 -2.96714991e-01
-1.02500759e-01 1.29424959e-01 -1.25430131e+00 1.61575401e+00
-2.66445931e-02 4.01916504e-01 5.28260544e-02 -1.16563606e+00
7.12669194e-01 8.16475376e-02 7.20112205e-01 -4.06797260e-01
2.04541340e-01 3.64374161e-01 -1.45054802e-01 -5.32513022e-01
3.86589974e-01 2.11878438e-02 1.37539223e-01 2.44302943e-01
4.21323001e-01 8.44292268e-02 5.25447667e-01 2.05275506e-01
7.55257249e-01 1.86756745e-01 3.25234085e-01 -3.78697187e-01
4.38711703e-01 2.24797860e-01 5.94666481e-01 1.06129682e+00
-1.46554038e-01 7.77348399e-01 2.13676229e-01 -3.63725185e-01
-7.04499543e-01 -1.05748284e+00 -2.42050841e-01 1.17125344e+00
4.01603043e-01 -2.20993444e-01 -1.05660915e+00 -1.14129496e+00
1.54315010e-01 2.42812127e-01 -7.42869854e-01 1.58444762e-01
-4.60300297e-01 -7.58991003e-01 2.83665836e-01 6.97323382e-01
5.09230614e-01 -1.12154078e+00 -7.03965485e-01 1.78288981e-01
-4.79439311e-02 -1.31731462e+00 -1.75358266e-01 4.37613517e-01
-9.92798924e-01 -1.24866080e+00 -5.10660410e-01 -1.03440571e+00
1.20315945e+00 2.94181406e-01 1.13230622e+00 2.93281615e-01
-6.97176456e-01 2.61631012e-01 -3.12338680e-01 -3.90574485e-01
-1.27104267e-01 1.08921662e-01 -1.24684274e-01 5.71978986e-01
4.45058823e-01 -2.74349093e-01 -5.69984853e-01 2.94586897e-01
-1.03397942e+00 -2.21154261e-02 1.02156794e+00 4.07242894e-01
1.15853655e+00 2.36215159e-01 3.88477445e-01 -1.10250700e+00
-1.14261985e-01 -2.69566149e-01 -6.58634782e-01 2.00672209e-01
-3.68973702e-01 -1.94951799e-02 2.47134045e-01 -5.13193786e-01
-9.82866406e-01 8.22848499e-01 2.96679437e-01 -3.30749869e-01
-7.76587844e-01 1.94005221e-01 -3.12776893e-01 -9.36152115e-02
4.85688150e-01 2.45182410e-01 -3.29520017e-01 -8.37543666e-01
5.05917668e-01 6.55491054e-01 6.92392528e-01 -4.93907094e-01
8.80159259e-01 8.43358099e-01 -2.45631263e-01 -6.28383338e-01
-1.38846338e+00 -1.03919494e+00 -1.25772953e+00 -1.71186373e-01
1.11131048e+00 -9.96154964e-01 -3.90604794e-01 4.56345677e-01
-1.19172585e+00 -2.03586429e-01 -5.76593339e-01 2.51666367e-01
-3.81335616e-01 2.22048715e-01 -3.64462107e-01 -7.37193704e-01
-2.02925786e-01 -1.05489743e+00 1.41468000e+00 5.19206107e-01
-2.07942668e-02 -5.00117362e-01 -2.62380362e-01 5.22999465e-01
-9.00513399e-03 8.14670250e-02 6.18425429e-01 -1.14951169e+00
-9.94065464e-01 -3.05863529e-01 -6.29801691e-01 2.46683493e-01
3.04581255e-01 -1.12295695e-01 -1.17099607e+00 -4.87295687e-02
-3.61797720e-01 -2.95955688e-01 1.15274239e+00 3.23523432e-01
1.45194244e+00 -2.94936985e-01 -6.39294624e-01 3.40665489e-01
1.58501649e+00 -4.43212986e-02 2.47390613e-01 2.96274364e-01
9.36168671e-01 8.64782512e-01 6.84005499e-01 3.71759951e-01
2.60664642e-01 4.39986289e-01 4.97000545e-01 -3.50625813e-01
-2.67474085e-01 -1.26606664e-02 -2.98491597e-01 2.32591689e-01
1.56089798e-01 -6.68952689e-02 -9.60828006e-01 1.04988766e+00
-1.99780321e+00 -5.93804419e-01 -3.33618909e-01 2.01762819e+00
8.38341415e-01 4.74726260e-01 2.11009428e-01 4.53899689e-02
9.22682762e-01 -2.48416305e-01 -6.15875304e-01 3.22116315e-01
4.47278842e-04 2.30656356e-01 6.25110269e-01 9.99115855e-02
-1.56147003e+00 1.15991926e+00 6.47662973e+00 8.42335403e-01
-6.82104468e-01 1.77481920e-01 5.81691206e-01 3.10023353e-02
1.56421602e-01 7.51266181e-02 -1.02400613e+00 1.57899216e-01
3.74353141e-01 4.21325803e-01 -1.70701876e-01 1.09953737e+00
3.08278631e-02 -2.83595502e-01 -1.17423308e+00 6.66194439e-01
-6.77764229e-03 -1.33212006e+00 1.12395301e-01 -9.62688178e-02
9.26345408e-01 -6.22090697e-02 -2.64062941e-01 -5.69371786e-03
2.00364634e-01 -6.95540369e-01 1.04019856e+00 3.49902272e-01
4.18149978e-01 -6.09565854e-01 6.82257533e-01 3.35155547e-01
-1.41989362e+00 -2.78431475e-01 1.60897914e-02 2.66089916e-01
-2.88611986e-02 8.78154397e-01 -7.42213488e-01 3.92314970e-01
9.21386123e-01 5.21088898e-01 -7.16089845e-01 1.40294158e+00
-2.68610895e-01 5.89379609e-01 -3.39168727e-01 1.81006610e-01
2.98694640e-01 2.18165398e-01 4.29410428e-01 1.36683917e+00
-4.24016774e-01 -4.88425903e-02 5.45917153e-01 1.12123692e+00
-2.59648770e-01 1.12369485e-01 -9.04936492e-02 -1.08853392e-01
3.34953815e-01 1.68772364e+00 -1.59036112e+00 -7.50009298e-01
-2.56144196e-01 8.61988366e-01 4.19647247e-01 2.22310871e-01
-6.23977721e-01 -4.85914886e-01 3.54454130e-01 1.61497697e-01
7.77170658e-01 -2.08636865e-01 -4.88784879e-01 -8.52882802e-01
1.12343930e-01 -5.37667453e-01 3.19358945e-01 -3.30930173e-01
-1.26750076e+00 2.72876948e-01 7.21894056e-02 -1.05864477e+00
1.37396321e-01 -6.27306819e-01 -4.03743088e-01 3.41909409e-01
-1.60537839e+00 -1.44649971e+00 -4.97330159e-01 2.21722946e-01
7.85727084e-01 1.35880366e-01 5.64616859e-01 4.18666005e-01
-5.61659455e-01 3.32883030e-01 -1.10693704e-02 4.62592512e-01
2.51780868e-01 -1.31160891e+00 1.72793329e-01 9.05561924e-01
4.50508505e-01 6.45188451e-01 3.56291533e-01 -7.39232600e-01
-8.22374821e-01 -1.39728856e+00 6.09397113e-01 -4.31061536e-01
3.10232639e-01 -4.86308187e-01 -9.37235892e-01 4.67463970e-01
-2.26585612e-01 4.26950008e-01 5.78370988e-01 -1.45552251e-02
-5.42309403e-01 -6.81623295e-02 -1.16136801e+00 1.62855431e-01
9.19242382e-01 -3.11382204e-01 -4.38911647e-01 4.33741599e-01
7.97383785e-01 -1.88114971e-01 -6.13004982e-01 6.12263799e-01
4.52704012e-01 -7.39262342e-01 1.06603169e+00 -6.12399280e-01
3.34888488e-01 -6.39808178e-01 -3.04054655e-02 -5.56074202e-01
-2.62364805e-01 -2.97749311e-01 -2.33136207e-01 1.35925221e+00
4.65028375e-01 -2.27525681e-01 9.04170156e-01 3.55838865e-01
-1.01738244e-01 -6.42354608e-01 -5.72553158e-01 -8.38926017e-01
-3.64482313e-01 -3.74023527e-01 4.54900861e-01 7.61498451e-01
-6.09002650e-01 1.10745929e-01 1.79842800e-01 4.70154881e-01
1.05793381e+00 5.39707720e-01 6.21727049e-01 -1.57425332e+00
-1.73932597e-01 -4.83946830e-01 -4.83002692e-01 -9.53245103e-01
5.93127385e-02 -8.61254930e-01 5.29599190e-01 -1.73883915e+00
6.73496962e-01 -6.05722964e-01 -5.21396101e-01 8.72909129e-01
-2.08784267e-01 7.08464265e-01 2.41697263e-02 5.00705302e-01
-1.15290213e+00 1.69917434e-01 8.99507225e-01 -2.93099314e-01
-3.22163105e-01 -3.70997004e-02 -6.55599833e-01 1.02671218e+00
6.83749616e-01 -8.61530781e-01 -8.45137145e-03 -4.68947262e-01
-3.67848635e-01 -6.43944204e-01 5.81540346e-01 -1.12704098e+00
2.92228878e-01 -1.35461852e-01 5.01970708e-01 -8.81657302e-01
1.34849563e-01 -7.73488343e-01 -2.46721417e-01 4.78091359e-01
-4.27328825e-01 -6.65215850e-01 1.59408152e-01 6.93240941e-01
-2.05145925e-01 -4.06561881e-01 1.04617333e+00 -3.45519215e-01
-8.44608426e-01 4.70340699e-01 -3.24943453e-01 -8.83665234e-02
1.13456869e+00 -2.54164815e-01 5.73281534e-02 2.75014549e-01
-9.46709275e-01 2.34750807e-01 1.22023016e-01 4.68364686e-01
4.83088613e-01 -1.09280789e+00 -4.22895968e-01 7.64887258e-02
5.04121363e-01 5.68848014e-01 5.94961829e-02 7.74139762e-01
-2.66435444e-01 2.14639992e-01 6.72861189e-02 -9.27611411e-01
-1.35641074e+00 6.43633723e-01 2.71036118e-01 -1.25509515e-01
-5.43539524e-01 1.04988790e+00 5.39189994e-01 -4.63090181e-01
3.15752536e-01 -2.49866292e-01 -2.85254151e-01 2.14076549e-01
4.31317568e-01 1.46578074e-01 8.32136199e-02 -7.40272999e-01
-5.59282422e-01 6.68299139e-01 -3.12848359e-01 9.38582513e-03
1.34086764e+00 -1.07129775e-01 -2.40105793e-01 2.85443127e-01
1.14381969e+00 -1.76990375e-01 -1.20466495e+00 -4.95193750e-01
5.07471204e-01 -4.23179507e-01 5.56090735e-02 -8.17222714e-01
-1.13170767e+00 6.63969278e-01 8.04918230e-01 3.34955484e-01
1.02863193e+00 5.20248652e-01 6.28788650e-01 3.72940391e-01
8.63728225e-02 -1.27218807e+00 3.36169064e-01 2.46938273e-01
3.61811519e-01 -1.58250546e+00 1.82425734e-02 -8.81165266e-01
-2.14591354e-01 1.03265262e+00 8.26077104e-01 -4.01955135e-02
5.67679167e-01 1.77540198e-01 -1.09976441e-01 -5.36097884e-01
-1.80484027e-01 -9.24635947e-01 6.02376759e-01 3.88115227e-01
1.74663231e-01 -1.87300555e-02 -7.55198672e-02 7.38930821e-01
4.27949339e-01 -5.40428311e-02 4.41810954e-03 1.15537512e+00
-8.14078689e-01 -1.03272426e+00 -3.36916983e-01 5.14158189e-01
-6.41466379e-01 4.09566909e-02 -6.26904309e-01 6.67632997e-01
6.29156888e-01 8.61268580e-01 3.10683131e-01 -1.05044976e-01
1.78731620e-01 -1.46090701e-01 4.24975842e-01 -1.07901359e+00
-6.19941831e-01 2.28445485e-01 -1.82583943e-01 -4.99135256e-01
-7.70454109e-01 -4.85296190e-01 -1.58071816e+00 6.62990451e-01
-7.61238337e-01 -5.96479066e-02 7.29603171e-01 1.16940749e+00
4.36316878e-01 6.57461882e-01 4.84330684e-01 -9.62066293e-01
-1.06249914e-01 -7.66366363e-01 -3.62182856e-01 5.46559572e-01
3.39234143e-01 -7.64642835e-01 -1.34845287e-01 3.57381314e-01] | [9.440848350524902, 0.6542792916297913] |
002ca917-378b-4729-9fcb-16dbd19c21af | triplet-loss-less-center-loss-sampling | 2302.04108 | null | https://arxiv.org/abs/2302.04108v1 | https://arxiv.org/pdf/2302.04108v1.pdf | Triplet Loss-less Center Loss Sampling Strategies in Facial Expression Recognition Scenarios | Facial expressions convey massive information and play a crucial role in emotional expression. Deep neural network (DNN) accompanied by deep metric learning (DML) techniques boost the discriminative ability of the model in facial expression recognition (FER) applications. DNN, equipped with only classification loss functions such as Cross-Entropy cannot compact intra-class feature variation or separate inter-class feature distance as well as when it gets fortified by a DML supporting loss item. The triplet center loss (TCL) function is applied on all dimensions of the sample's embedding in the embedding space. In our work, we developed three strategies: fully-synthesized, semi-synthesized, and prediction-based negative sample selection strategies. To achieve better results, we introduce a selective attention module that provides a combination of pixel-wise and element-wise attention coefficients using high-semantic deep features of input samples. We evaluated the proposed method on the RAF-DB, a highly imbalanced dataset. The experimental results reveal significant improvements in comparison to the baseline for all three negative sample selection strategies. | ['Fatemeh Afghah', 'Adham Atyabi', 'Fatemeh Lotfi', 'Hossein Rajoli'] | 2023-02-08 | null | null | null | null | ['metric-learning', 'facial-expression-recognition', 'metric-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 2.80647218e-01 -1.86993912e-01 -1.93480134e-01 -8.31333816e-01
-6.22221291e-01 2.42645033e-02 4.30925727e-01 -3.78686190e-01
-5.72001576e-01 7.32346356e-01 1.37316570e-01 2.94757038e-01
-3.77959548e-03 -6.30410433e-01 -4.24420029e-01 -9.13444340e-01
-4.23965417e-02 6.58939704e-02 -3.59736174e-01 -3.46148938e-01
1.40511334e-01 8.06980193e-01 -1.67469954e+00 6.11418188e-01
6.47570372e-01 1.81196988e+00 -2.32268244e-01 1.31975174e-01
-2.84549266e-01 9.39921677e-01 -4.60430115e-01 -6.64755940e-01
2.92174906e-01 -4.31090027e-01 -4.70185429e-01 -1.11901485e-04
2.75577098e-01 -2.46898130e-01 -1.06225535e-01 1.13076115e+00
9.68922317e-01 5.19011170e-02 7.22133458e-01 -1.59881139e+00
-4.87295687e-01 -3.16580869e-02 -9.27050471e-01 2.11240232e-01
9.55796316e-02 8.13254416e-02 9.16713178e-01 -1.45538759e+00
5.62291026e-01 1.35150468e+00 5.47630489e-01 8.13943267e-01
-1.00240111e+00 -9.21453357e-01 5.03003486e-02 4.31468248e-01
-1.33884168e+00 -6.53290808e-01 1.24408472e+00 -3.40917617e-01
7.24473536e-01 3.49216759e-01 5.81869841e-01 1.12592828e+00
6.70781434e-02 1.04752302e+00 9.05065835e-01 -3.31484586e-01
1.71018437e-01 3.02191526e-01 -1.34245872e-01 6.55325532e-01
-3.97793591e-01 -8.52452815e-02 -6.18669093e-01 -1.13024175e-01
3.47327143e-01 3.82919274e-02 -2.21862167e-01 -3.50129157e-01
-6.05642915e-01 9.11151171e-01 4.72325593e-01 2.25336581e-01
-5.39390087e-01 4.23901267e-02 8.08173299e-01 3.58996838e-01
8.23968768e-01 2.24974558e-01 -4.49022770e-01 -1.03552885e-01
-7.18331814e-01 1.27629787e-01 2.37823665e-01 4.98319715e-01
7.59016871e-01 1.41987503e-01 -4.52059239e-01 1.18608701e+00
6.22856990e-02 1.76743567e-01 6.54111087e-01 -7.41793692e-01
5.41799366e-01 7.11259305e-01 -1.24766544e-01 -1.49127591e+00
-3.83140445e-01 -4.38994378e-01 -1.02621698e+00 3.58598590e-01
-5.22079282e-02 -2.00709656e-01 -5.46220720e-01 2.08556342e+00
5.68606913e-01 -5.63492253e-02 -2.15596687e-02 1.17564261e+00
7.68019140e-01 5.74885786e-01 1.71538755e-01 -2.97060192e-01
1.12554407e+00 -9.20713484e-01 -9.54238594e-01 1.61105484e-01
8.04092050e-01 -5.35823822e-01 1.14335990e+00 4.52214092e-01
-9.32269275e-01 -5.24741232e-01 -1.01811469e+00 -1.36633649e-01
-3.71087939e-01 4.80595499e-01 5.61479747e-01 5.66014469e-01
-8.68403614e-01 4.40296501e-01 -4.47494090e-01 1.59897879e-01
1.04891753e+00 4.51195300e-01 -7.05782652e-01 -9.08128452e-03
-1.30510533e+00 6.25097990e-01 -7.97497928e-02 6.03640676e-01
-6.53615177e-01 -7.97466874e-01 -9.48111594e-01 3.03435996e-02
-8.94994587e-02 -2.26426363e-01 6.63134396e-01 -1.89625263e+00
-1.69020212e+00 1.05477822e+00 1.57707892e-02 -1.38184413e-01
6.12511933e-01 -1.82613894e-01 -4.63708282e-01 2.81164378e-01
-1.64354652e-01 7.36552835e-01 8.65774035e-01 -8.57812643e-01
-2.54872650e-01 -7.27537334e-01 -2.05399126e-01 2.43094653e-01
-7.85972714e-01 3.48194778e-01 5.70701770e-02 -6.35139942e-01
-1.33793458e-01 -6.01109982e-01 3.88197936e-02 6.99263036e-01
-3.15353096e-01 -3.50895405e-01 1.08369946e+00 -7.08375454e-01
1.21782625e+00 -2.37006044e+00 1.55842274e-01 2.42068201e-01
-1.77624915e-02 4.84578550e-01 -3.75130653e-01 6.72257394e-02
-4.81470853e-01 -1.30789742e-01 -1.95632830e-01 -5.90255976e-01
1.56712979e-01 -9.63354632e-02 -1.11784205e-01 6.46076143e-01
8.58666897e-01 6.91431284e-01 -6.36383891e-01 -4.96078730e-01
7.53552094e-02 7.23248839e-01 -6.75854564e-01 1.84193596e-01
4.65130843e-02 1.26201078e-01 -2.88944304e-01 7.04097033e-01
9.74196494e-01 8.82929936e-02 -2.25704566e-01 -4.42152649e-01
3.74689579e-01 -1.64418980e-01 -9.57484305e-01 1.65890419e+00
-5.25976598e-01 5.57148516e-01 2.02007949e-01 -1.22448134e+00
1.36770678e+00 1.45811096e-01 4.58071798e-01 -9.72724974e-01
3.14340979e-01 2.33311981e-01 -1.24371335e-01 -5.96771479e-01
2.96586424e-01 -4.47408646e-01 1.76865548e-01 1.23272471e-01
1.83975995e-01 5.25546014e-01 -2.39206418e-01 -1.57630399e-01
7.11723626e-01 6.00864217e-02 -2.90998947e-02 -2.57669985e-01
8.30536902e-01 -6.99395120e-01 7.80322969e-01 -1.98507726e-01
-5.53816676e-01 5.74772596e-01 9.21085358e-01 -5.41101873e-01
-8.54572833e-01 -6.24091685e-01 -2.67033875e-01 1.18914294e+00
-4.99171764e-02 -1.04305983e-01 -6.62586272e-01 -8.23432088e-01
-4.85555194e-02 5.16353369e-01 -9.67517674e-01 -7.00702846e-01
-3.03691804e-01 -9.90386963e-01 3.84146750e-01 5.77516913e-01
5.76871872e-01 -1.09830070e+00 -3.94981921e-01 -7.65730143e-02
-6.04219027e-02 -7.44192481e-01 -3.43734801e-01 1.74127132e-01
-4.35331613e-01 -7.90341854e-01 -7.88960040e-01 -7.35674977e-01
7.11756349e-01 -2.50404596e-01 8.41836035e-01 -1.97230846e-01
-4.62101787e-01 -7.66743645e-02 -3.35451841e-01 -3.08494568e-01
1.51269987e-01 -1.99235752e-01 -7.47061595e-02 7.95471132e-01
7.79747427e-01 -5.38953900e-01 -7.30071723e-01 2.93252647e-01
-6.56391740e-01 -2.25786433e-01 3.60919893e-01 1.21942568e+00
5.61660409e-01 -3.33532870e-01 8.07766914e-01 -4.78893518e-01
4.80027437e-01 -4.97197777e-01 -2.10761353e-01 1.82367340e-01
-1.95593134e-01 -1.18666962e-01 6.41306877e-01 -4.15804297e-01
-1.02401114e+00 -9.84352827e-02 -3.34114879e-01 -7.48555601e-01
1.43656477e-01 1.16579436e-01 -5.58317065e-01 -7.55541250e-02
5.54456651e-01 8.28521177e-02 3.12567711e-01 -2.42542759e-01
-6.96204528e-02 7.74834335e-01 -4.04347405e-02 -5.19017458e-01
3.57423089e-02 5.14795780e-01 1.42182946e-01 -7.20813692e-01
-6.35849118e-01 -2.64405049e-02 -3.15800846e-01 -3.33626479e-01
6.95344448e-01 -9.02680099e-01 -7.17548430e-01 6.21702373e-01
-1.04314196e+00 -4.35669571e-02 -3.19641471e-01 3.96866828e-01
-4.72792119e-01 1.21912487e-01 -4.82984215e-01 -1.05048108e+00
-5.80875754e-01 -1.13577318e+00 1.25782323e+00 1.29456609e-01
-6.86509814e-03 -6.24946713e-01 -4.14784625e-02 2.41884068e-01
4.59731996e-01 6.29004657e-01 6.01475835e-01 -6.27927184e-01
5.24970666e-02 -3.03508312e-01 -4.33600277e-01 9.84661639e-01
-5.39167076e-02 5.46189323e-02 -1.21883059e+00 -1.92231506e-01
-1.10419914e-02 -8.67339194e-01 9.41537321e-01 1.20976716e-01
1.72674692e+00 -3.10773849e-01 4.90220524e-02 8.08808565e-01
1.24650526e+00 4.48714606e-02 9.37439382e-01 2.00070694e-01
6.24366343e-01 9.19798970e-01 6.94507420e-01 8.34942102e-01
1.27704620e-01 7.65742660e-01 5.50420642e-01 -2.72879601e-01
1.20472327e-01 1.85294468e-02 4.23655480e-01 4.67368782e-01
7.53993019e-02 -8.35828558e-02 -3.49232793e-01 4.07048792e-01
-1.45183814e+00 -1.05197203e+00 2.93136984e-01 1.88254166e+00
9.50775146e-01 -1.46938890e-01 1.03981689e-01 4.13418353e-01
7.24212646e-01 3.78026068e-01 -6.38133764e-01 -7.67382026e-01
-5.02409756e-01 3.56892586e-01 -1.15523420e-01 2.17811331e-01
-1.21799827e+00 6.13858998e-01 4.61490870e+00 1.36465585e+00
-1.61991858e+00 8.97192210e-03 1.32355273e+00 -5.31032205e-01
-2.47327730e-01 -6.92470014e-01 -5.67265749e-01 6.70295060e-01
6.87670052e-01 2.17646405e-01 -6.21951465e-03 9.87605989e-01
1.29546747e-01 2.58968115e-01 -8.49131286e-01 1.49195230e+00
2.34783098e-01 -1.08896554e+00 9.71038565e-02 -1.37582093e-01
4.23907340e-01 -3.51961344e-01 4.20412898e-01 3.40878725e-01
-5.47514319e-01 -1.07665610e+00 6.25145912e-01 5.24027884e-01
1.01888466e+00 -1.16210914e+00 8.54006588e-01 -1.54872611e-01
-8.37540329e-01 -3.02006364e-01 -6.19514346e-01 8.98917112e-03
-7.40078092e-02 6.24811351e-01 -3.36666644e-01 3.74379992e-01
7.31435359e-01 8.78776729e-01 -2.88800985e-01 5.06455064e-01
8.38221386e-02 3.08688492e-01 -1.76817074e-01 -1.66721791e-01
1.53206915e-01 -4.68929738e-01 3.53955716e-01 1.07774830e+00
2.74806082e-01 -5.73720708e-02 -4.00223821e-01 7.47842550e-01
-2.70553470e-01 4.72315937e-01 -3.95033240e-01 -7.57884514e-03
-2.28879657e-02 1.48202074e+00 -1.09002508e-01 -6.26719669e-02
-1.84751391e-01 1.17776978e+00 5.13911963e-01 2.59019971e-01
-8.62531066e-01 -6.45104647e-01 1.07360840e+00 7.65258223e-02
3.72688174e-01 4.57795322e-01 -1.08677782e-01 -9.66932833e-01
4.14478540e-01 -7.48063266e-01 3.25500697e-01 -7.54783690e-01
-1.39246190e+00 8.20072114e-01 -5.67908943e-01 -1.21875906e+00
-7.91583061e-02 -7.23641038e-01 -3.90234262e-01 9.19584334e-01
-1.69325578e+00 -1.09812391e+00 -3.07706058e-01 7.27548420e-01
1.56130552e-01 -3.52819473e-01 8.36584747e-01 9.03112948e-01
-7.64204443e-01 1.23473382e+00 -6.08303174e-02 3.35449100e-01
5.90887964e-01 -7.72961557e-01 -4.69909847e-01 3.85358721e-01
-2.85916388e-01 2.21030980e-01 4.05647665e-01 9.71819758e-02
-1.04881871e+00 -1.19041955e+00 8.35528016e-01 -2.44017337e-02
3.59095097e-01 -5.54429948e-01 -7.53828943e-01 2.62269884e-01
-1.92172989e-01 6.03630304e-01 1.03217876e+00 -2.43764475e-01
-5.14373124e-01 -7.86573410e-01 -1.55763125e+00 2.42756158e-01
9.78003561e-01 -5.43277264e-01 1.27278902e-02 1.84365883e-01
4.69547212e-01 -7.59264827e-02 -7.09494948e-01 7.20163882e-01
7.74992704e-01 -1.25694323e+00 7.15702951e-01 -9.34433758e-01
8.23714435e-01 -4.38061729e-02 -4.92608219e-01 -9.71270680e-01
-1.11894742e-01 -3.40606302e-01 5.83308861e-02 1.27477932e+00
2.62952447e-01 -4.47894394e-01 9.18033659e-01 5.42964935e-01
-9.54637229e-02 -1.49947250e+00 -1.16714549e+00 -4.12524372e-01
5.88389998e-03 -2.82352090e-01 6.69098735e-01 1.05845213e+00
-1.69343367e-01 1.44979864e-01 -5.32979608e-01 -4.17750418e-01
3.02975327e-01 -6.29022047e-02 5.80073595e-01 -9.89909947e-01
5.34804761e-02 -6.14903450e-01 -9.63765562e-01 -6.99292958e-01
5.78604162e-01 -7.81219184e-01 -3.82187665e-01 -8.54395747e-01
1.84606805e-01 -2.88687050e-01 -6.65728271e-01 3.99666727e-01
2.80790925e-02 4.62607652e-01 -6.72790268e-03 -2.64787167e-01
-6.08600140e-01 1.33633661e+00 1.20409024e+00 -1.19162209e-01
6.10133410e-02 -2.81569600e-01 -6.01534128e-01 5.08211195e-01
6.88360751e-01 -2.26106808e-01 -3.17140043e-01 -1.99762747e-01
2.66581953e-01 -1.33735210e-01 3.19453984e-01 -7.47473419e-01
-2.51201838e-01 6.79972023e-02 7.49997437e-01 -4.36118484e-01
6.40319407e-01 -8.44950259e-01 -1.87650532e-01 2.00636402e-01
-5.76117456e-01 -1.01702794e-01 1.81988031e-01 2.98701286e-01
-6.33513153e-01 2.13827550e-01 1.21225107e+00 3.51581156e-01
-6.14530742e-01 5.86241663e-01 2.18833819e-01 -1.87234432e-02
1.11557770e+00 -3.73406440e-01 -3.32931168e-02 -2.24676281e-01
-5.78560948e-01 9.36578512e-02 1.20780535e-01 3.95258814e-01
7.62946427e-01 -1.87059939e+00 -7.73791552e-01 5.01270711e-01
3.83167565e-01 -2.30496004e-01 5.44175982e-01 1.04972589e+00
-3.68931472e-01 -2.86833756e-02 -5.76577485e-01 -3.94084841e-01
-1.25774372e+00 4.14838850e-01 6.23981118e-01 -8.56200978e-02
-6.08842000e-02 1.31557035e+00 3.55948865e-01 -6.01584911e-01
2.22538456e-01 8.45912471e-02 -3.25707287e-01 3.62505913e-01
8.13267648e-01 3.93614292e-01 1.14734039e-01 -9.36325490e-01
-5.43658733e-01 5.53997040e-01 4.35605459e-03 9.73062292e-02
1.42911720e+00 1.07414067e-01 -3.36582243e-01 3.05689722e-01
1.96375024e+00 -3.78620446e-01 -1.25675213e+00 -1.97284073e-01
-2.11116910e-01 -5.50722599e-01 1.27037510e-01 -7.13982999e-01
-1.54929054e+00 1.26519096e+00 9.00547922e-01 -3.58077705e-01
1.67267823e+00 -2.35503197e-01 7.58625686e-01 4.24902625e-02
1.10241643e-03 -1.37698317e+00 3.95035535e-01 1.26021460e-01
1.22759295e+00 -1.22307920e+00 -3.14173847e-01 -8.66184533e-02
-7.42087364e-01 1.09817564e+00 7.66307771e-01 -2.41431758e-01
8.54638517e-01 8.55205655e-02 -6.66022748e-02 -2.09673405e-01
-8.60876620e-01 1.89669594e-01 2.44678497e-01 4.52593356e-01
4.05969799e-01 -7.25124255e-02 -4.14940923e-01 9.86357391e-01
-2.97018532e-02 1.32967532e-01 -1.58403009e-01 6.28393292e-01
-2.05843866e-01 -7.52255440e-01 -4.79592988e-03 4.72818851e-01
-5.24454892e-01 7.42488429e-02 -3.26105595e-01 5.65182328e-01
2.60261595e-01 5.46267390e-01 3.03985536e-01 -5.88133037e-01
3.18619341e-01 3.80297713e-02 4.95676339e-01 -4.29799631e-02
-4.83324409e-01 -1.49704188e-01 6.28387406e-02 -8.65722597e-01
-5.07776618e-01 -5.60284734e-01 -1.00652122e+00 -1.36466533e-01
-3.39103192e-01 -4.58920635e-02 6.28106415e-01 5.91319084e-01
6.03942573e-01 2.03203350e-01 1.20311296e+00 -6.87777042e-01
-6.63890302e-01 -9.88071442e-01 -7.88693368e-01 7.40105569e-01
3.66083413e-01 -7.90699244e-01 -6.91704810e-01 -2.76258618e-01] | [13.578397750854492, 1.6433298587799072] |
7b260f8c-2d07-4009-82b1-da78f9931db9 | robust-and-efficient-memory-network-for-video | 2304.11840 | null | https://arxiv.org/abs/2304.11840v1 | https://arxiv.org/pdf/2304.11840v1.pdf | Robust and Efficient Memory Network for Video Object Segmentation | This paper proposes a Robust and Efficient Memory Network, referred to as REMN, for studying semi-supervised video object segmentation (VOS). Memory-based methods have recently achieved outstanding VOS performance by performing non-local pixel-wise matching between the query and memory. However, these methods have two limitations. 1) Non-local matching could cause distractor objects in the background to be incorrectly segmented. 2) Memory features with high temporal redundancy consume significant computing resources. For limitation 1, we introduce a local attention mechanism that tackles the background distraction by enhancing the features of foreground objects with the previous mask. For limitation 2, we first adaptively decide whether to update the memory features depending on the variation of foreground objects to reduce temporal redundancy. Second, we employ a dynamic memory bank, which uses a lightweight and differentiable soft modulation gate to decide how many memory features need to be removed in the temporal dimension. Experiments demonstrate that our REMN achieves state-of-the-art results on DAVIS 2017, with a $\mathcal{J\&F}$ score of 86.3% and on YouTube-VOS 2018, with a $\mathcal{G}$ over mean of 85.5%. Furthermore, our network shows a high inference speed of 25+ FPS and uses relatively few computing resources. | ['Enhua Wu', 'Zhi-Xin Yang', 'Dingwei Zhang', 'Yadang Chen'] | 2023-04-24 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'video-object-segmentation', 'video-semantic-segmentation'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.61078060e-01 -2.96614528e-01 -2.06922442e-01 -2.59480953e-01
-5.83150923e-01 -7.32406080e-02 -5.15196808e-02 -1.29740775e-01
-7.77036190e-01 5.91520607e-01 -4.17684764e-01 -1.38696671e-01
2.47891366e-01 -7.09174931e-01 -9.86245573e-01 -6.82490408e-01
4.08944637e-02 2.82346457e-02 9.73740041e-01 3.27504218e-01
2.02320620e-01 3.72575969e-01 -1.76341224e+00 3.17228347e-01
8.34471643e-01 1.58067405e+00 5.13489783e-01 5.81965506e-01
-3.11633974e-01 7.04442441e-01 -7.09499538e-01 -2.89236363e-02
3.32224816e-01 -3.75689715e-01 -8.05399418e-01 -1.94033480e-03
7.79561102e-01 -5.22569120e-01 -6.88977361e-01 1.14685750e+00
5.32571614e-01 3.97401631e-01 2.85651267e-01 -1.14183104e+00
-4.05581176e-01 1.27069652e-01 -7.86160231e-01 8.80528390e-01
-1.00652091e-01 5.58086038e-01 5.03492773e-01 -8.70828867e-01
5.99250495e-01 1.09022963e+00 2.72851557e-01 7.21391976e-01
-9.03860271e-01 -8.19499969e-01 5.34096003e-01 4.55577731e-01
-1.45713913e+00 -5.01341224e-01 5.38182437e-01 -3.52131218e-01
9.00157332e-01 2.27860898e-01 6.13482177e-01 6.13159001e-01
-2.16119792e-02 1.11015844e+00 9.66888309e-01 -8.43995437e-02
2.72630334e-01 -1.61307976e-02 2.20182940e-01 9.26327586e-01
-3.65138403e-03 -2.29562759e-01 -7.88567722e-01 2.93301255e-01
9.54684973e-01 3.38965538e-03 -2.34339997e-01 6.88582584e-02
-1.01920116e+00 5.13350904e-01 3.73422682e-01 1.23149484e-01
-2.78987348e-01 4.26502973e-01 4.19196308e-01 8.11109692e-03
3.62992734e-01 -5.65487072e-02 -4.34854686e-01 -2.92720616e-01
-1.36657655e+00 8.41003470e-03 3.27359647e-01 1.14365840e+00
6.75070345e-01 2.59611279e-01 -4.75527853e-01 6.51220322e-01
2.27007955e-01 6.27824068e-01 3.63155574e-01 -1.12251246e+00
5.33683658e-01 5.47142982e-01 -1.63823254e-02 -1.10480118e+00
-2.47181118e-01 -1.09670535e-01 -8.48463178e-01 9.03788954e-02
3.78684670e-01 2.25344915e-02 -1.20734644e+00 1.59543276e+00
3.45453620e-01 5.91858208e-01 -3.51245373e-01 1.18478405e+00
1.17199624e+00 7.45293915e-01 3.29820484e-01 -3.93597901e-01
1.22728121e+00 -1.22945583e+00 -8.00773144e-01 -4.08749580e-01
1.41326562e-01 -4.63915408e-01 1.16063309e+00 2.05164596e-01
-1.53462887e+00 -9.40834045e-01 -1.14440727e+00 -9.71501842e-02
-2.77981579e-01 6.78277016e-02 4.36553687e-01 4.78789419e-01
-1.09176743e+00 7.59077847e-01 -1.14593124e+00 4.36128303e-02
9.17372227e-01 5.59047937e-01 2.81433403e-01 1.60913616e-01
-1.11323977e+00 4.14108664e-01 2.36086756e-01 1.01596318e-01
-9.11067247e-01 -6.26188636e-01 -6.51294291e-01 1.59237266e-01
7.12950528e-01 -4.81084138e-01 8.84171426e-01 -1.09491551e+00
-1.31125689e+00 8.38638186e-01 -4.64021713e-01 -4.35876638e-01
4.45907652e-01 -2.65754282e-01 -5.07835805e-01 4.72840309e-01
4.10603248e-02 1.15242970e+00 1.15550447e+00 -8.40210557e-01
-8.05605769e-01 -3.07027310e-01 -1.10366113e-01 6.41114116e-02
-4.54630136e-01 4.49280702e-02 -1.19951618e+00 -7.65896559e-01
3.35581332e-01 -8.15791309e-01 -1.35537639e-01 4.23306018e-01
-8.41987804e-02 -1.33187771e-01 1.00967574e+00 -5.67820489e-01
1.60652804e+00 -2.33383536e+00 -7.33541101e-02 2.25746389e-02
2.81391114e-01 6.68551624e-01 -9.42563731e-03 -5.62086582e-01
4.58800405e-01 1.46533564e-01 -1.85430154e-01 -3.50944787e-01
-2.19065025e-01 8.33444744e-02 -1.89473271e-01 4.40168291e-01
3.11416507e-01 1.01155722e+00 -7.22994030e-01 -9.23183382e-01
2.20867872e-01 4.21773493e-01 -4.57198262e-01 1.32474288e-01
-2.69584745e-01 1.63015917e-01 -3.05265695e-01 8.41029286e-01
7.32116759e-01 -4.85125721e-01 -1.29065529e-01 -2.56249517e-01
-1.15326770e-01 2.58989602e-01 -1.50090885e+00 1.54951096e+00
1.37630686e-01 7.62054861e-01 2.47557625e-01 -7.69195616e-01
6.38689220e-01 -3.87038179e-02 4.27779287e-01 -9.31068242e-01
3.75386715e-01 9.03763622e-02 -2.88018107e-01 -5.07282078e-01
5.69500327e-01 5.44995010e-01 3.46206963e-01 1.50659651e-01
2.12859251e-02 3.45172822e-01 3.83907825e-01 1.00838006e-01
9.15377796e-01 2.75938213e-01 -2.30818808e-01 -2.45646328e-01
5.34784377e-01 -1.68571740e-01 9.15993571e-01 8.88818860e-01
-7.48558342e-01 4.92219061e-01 2.92026490e-01 -3.56771737e-01
-5.20760953e-01 -1.07483327e+00 -3.74553464e-02 1.21230340e+00
7.28483379e-01 -2.06588149e-01 -8.85170698e-01 -5.57379663e-01
-1.57172441e-01 3.14432055e-01 -3.36528420e-01 -2.38144204e-01
-7.55764902e-01 -5.90366900e-01 5.27235925e-01 9.54989374e-01
1.06421328e+00 -1.24215436e+00 -1.13879704e+00 1.67610407e-01
-2.39183560e-01 -1.26503778e+00 -7.98021197e-01 6.78511560e-02
-1.14774394e+00 -8.73788357e-01 -7.14900076e-01 -8.64614367e-01
6.54397547e-01 4.82150525e-01 9.82778251e-01 3.28832000e-01
-5.03122270e-01 3.46746624e-01 1.85545385e-02 -1.84760913e-01
1.48641586e-01 4.22133459e-03 -5.38604222e-02 7.80738667e-02
4.75928158e-01 -2.16934025e-01 -9.84284937e-01 4.20186132e-01
-9.12643015e-01 1.03567019e-01 4.16762233e-01 6.52917743e-01
1.13141727e+00 -5.47491908e-02 5.29134095e-01 -5.60132384e-01
-4.46730144e-02 -2.03110963e-01 -6.70019209e-01 1.96256533e-01
-4.78996336e-01 -1.89302519e-01 2.40363881e-01 -7.79458523e-01
-7.79879153e-01 1.53717279e-01 1.75588742e-01 -7.59794533e-01
-1.18724450e-01 -1.53829396e-01 -2.03623712e-01 -1.35113209e-01
1.62698731e-01 4.06832397e-01 -1.44445285e-01 -3.28408569e-01
5.09904660e-02 4.61231142e-01 6.98503673e-01 -3.72150481e-01
4.72740382e-01 6.24590218e-01 -1.06278665e-01 -6.73518002e-01
-6.24397993e-01 -4.94392872e-01 -5.84628880e-01 -3.74457031e-01
1.14621198e+00 -9.82143223e-01 -9.59333301e-01 6.21863127e-01
-9.98734951e-01 -5.46779811e-01 -1.73367247e-01 2.49677971e-01
-3.84149343e-01 2.92545825e-01 -8.42876315e-01 -8.16587508e-01
-5.49346983e-01 -1.21806324e+00 8.13927412e-01 5.88624358e-01
-6.18022382e-02 -5.05669296e-01 -6.99904442e-01 4.66571122e-01
5.17014205e-01 6.03948571e-02 5.09035945e-01 -2.06456125e-01
-1.12631106e+00 2.94320375e-01 -5.69015622e-01 3.33752036e-01
-1.37128994e-01 -3.73199694e-02 -9.22816336e-01 -3.26451838e-01
-8.60757306e-02 -6.50018826e-02 1.22957587e+00 5.32608092e-01
1.63064814e+00 -8.53337869e-02 -3.88248414e-01 7.13103354e-01
1.04246807e+00 4.73481327e-01 6.97559655e-01 2.01458842e-01
7.89172113e-01 1.75509021e-01 8.35214734e-01 3.44916999e-01
2.59874076e-01 5.81856191e-01 4.13721353e-01 -8.39625001e-02
-4.41776007e-01 8.07193667e-02 3.73167962e-01 5.79712331e-01
4.44259495e-02 -2.13975251e-01 -6.29289389e-01 6.12769306e-01
-1.76853526e+00 -8.18939984e-01 -1.57228559e-01 2.25719762e+00
8.92524004e-01 5.40742457e-01 1.06679566e-01 -4.25940827e-02
8.45761180e-01 3.96799892e-01 -9.02508020e-01 -1.19211316e-01
-2.82899857e-01 3.23872268e-01 5.98342717e-01 2.54931778e-01
-1.32022858e+00 1.16498423e+00 5.57044983e+00 8.60441685e-01
-1.22746217e+00 2.25092918e-01 8.63140583e-01 -6.94782794e-01
1.91762611e-01 -3.36141914e-01 -9.53920841e-01 9.86473739e-01
8.13476205e-01 3.13771844e-01 3.83675665e-01 7.35114038e-01
1.36091337e-01 -5.33008933e-01 -8.91649187e-01 1.29594469e+00
7.87340328e-02 -1.45756567e+00 -1.08693518e-01 -3.08634818e-01
7.33573914e-01 -8.59223828e-02 2.56743342e-01 1.79323629e-01
-4.03878212e-01 -8.28887403e-01 8.83987844e-01 4.86064494e-01
8.56945455e-01 -6.82434499e-01 3.83441299e-01 5.03452383e-02
-1.40202129e+00 -1.31952325e-02 -3.09088916e-01 1.07724778e-01
2.31386214e-01 4.52333093e-01 -1.14349611e-01 -4.76224767e-03
1.23639309e+00 4.19420749e-01 -5.52399516e-01 8.46815109e-01
2.02504527e-02 6.60108685e-01 -4.06918347e-01 9.11106076e-03
2.06086323e-01 1.20991513e-01 4.41517562e-01 1.38217676e+00
8.88095349e-02 3.28559041e-01 2.46004879e-01 8.98591697e-01
-1.57343149e-01 -4.97121401e-02 1.06667325e-01 2.49656569e-02
4.77752298e-01 9.39605296e-01 -1.07704234e+00 -7.21282959e-01
-2.89378047e-01 1.28141510e+00 3.65190841e-02 3.34768414e-01
-1.26739228e+00 -3.05006385e-01 5.85336149e-01 2.13822380e-01
6.75732911e-01 -2.87456036e-01 -4.19807792e-01 -9.39478874e-01
2.03013167e-01 -4.59311843e-01 4.65641886e-01 -6.08686328e-01
-6.70664251e-01 4.83927757e-01 -2.07167402e-01 -1.04695868e+00
3.15764725e-01 -3.79386008e-01 -3.65245759e-01 5.92017591e-01
-1.50034451e+00 -6.79269433e-01 -5.80079556e-01 6.60363793e-01
7.70862877e-01 -5.98253682e-02 2.60028958e-01 7.60239542e-01
-7.97631443e-01 7.17043281e-01 -3.82151634e-01 2.08790585e-01
6.73323393e-01 -8.78383338e-01 3.23950738e-01 9.01701987e-01
-2.18901485e-02 4.01725739e-01 3.46474111e-01 -7.75621474e-01
-1.41057408e+00 -1.18916631e+00 6.60476983e-01 -1.26724124e-01
2.14040741e-01 -1.76426262e-01 -1.20649517e+00 3.14892590e-01
-9.22600403e-02 6.68165982e-01 3.47217441e-01 -5.70129573e-01
-1.11547463e-01 -1.62090376e-01 -1.26223862e+00 5.97260833e-01
1.34804356e+00 -4.38733846e-01 -1.63948461e-01 4.82339412e-02
7.02430964e-01 -6.61598325e-01 -6.62660778e-01 5.08704603e-01
4.01848197e-01 -7.82250464e-01 8.63125324e-01 -5.43163359e-01
1.68849096e-01 -6.67642057e-01 -1.82754323e-01 -4.04096216e-01
-2.27297828e-01 -6.65135443e-01 -6.99665904e-01 1.19938838e+00
5.89569472e-02 -2.65585631e-01 8.86705041e-01 7.90939987e-01
-1.68553088e-02 -1.01868522e+00 -1.09457731e+00 -7.60983169e-01
-3.00302744e-01 -3.62784803e-01 2.86733329e-01 4.87700492e-01
-4.26317185e-01 3.90223972e-02 -2.86415011e-01 1.99008018e-01
3.83196294e-01 1.03731997e-01 4.50255990e-01 -8.12846065e-01
-2.64775753e-01 -5.16389966e-01 -4.03662980e-01 -1.63600671e+00
1.87939871e-03 -4.18985218e-01 1.20674163e-01 -1.17929626e+00
2.19148159e-01 -4.15122062e-01 -5.58973312e-01 4.30473596e-01
-3.85537803e-01 5.35350680e-01 4.46873903e-01 1.20556451e-01
-1.31045842e+00 3.81917596e-01 1.17746377e+00 -3.15935522e-01
-5.01447499e-01 -5.47778569e-02 -2.44386703e-01 7.93972850e-01
6.64514422e-01 -3.50357205e-01 -1.34229511e-01 -5.48570156e-01
-1.74310714e-01 -8.10735449e-02 4.97249544e-01 -1.13418078e+00
5.19777477e-01 -2.22113624e-01 5.83511055e-01 -9.05163825e-01
4.54356939e-01 -6.44650817e-01 -1.47978589e-01 5.67379951e-01
-1.51854351e-01 7.54813701e-02 5.06588459e-01 5.62762856e-01
-1.60169363e-01 -1.32910460e-01 1.04270041e+00 -5.89505211e-02
-1.19387209e+00 5.12712777e-01 -4.05291438e-01 3.53520423e-01
1.17239714e+00 -5.01127601e-01 -2.15181306e-01 -7.85760507e-02
-6.96717024e-01 4.20366555e-01 2.49134362e-01 4.33641136e-01
7.63232589e-01 -1.14895558e+00 -2.84604639e-01 2.43791610e-01
-3.16661179e-01 1.48007736e-01 6.71331942e-01 9.14978981e-01
-3.95555556e-01 2.39148945e-01 -9.38124657e-02 -7.38959789e-01
-1.45498657e+00 6.20143890e-01 2.40774155e-01 1.58472195e-01
-5.58385253e-01 1.10342789e+00 8.08582678e-02 4.70925838e-01
6.27284944e-01 -3.67172003e-01 6.52265102e-02 1.11265428e-01
7.99872935e-01 7.37756789e-01 -1.66953921e-01 -6.63986385e-01
-5.46380877e-01 5.11689365e-01 1.87474452e-02 1.54810563e-01
9.63938653e-01 -2.68315732e-01 1.03955090e-01 3.71192515e-01
1.06693912e+00 -3.84207278e-01 -1.84157419e+00 -3.74681681e-01
-1.07558250e-01 -6.38081014e-01 1.45732358e-01 -6.14106476e-01
-1.59927404e+00 9.90778446e-01 1.02939653e+00 -1.95771202e-01
1.35065389e+00 -8.52041319e-02 1.19899273e+00 9.16136429e-02
1.02251880e-01 -1.58450222e+00 3.09955806e-01 5.38181007e-01
4.15997893e-01 -1.39657116e+00 -2.97000855e-02 -4.02265191e-01
-5.58759272e-01 8.86705220e-01 1.05948579e+00 -1.59828924e-02
6.22888207e-01 3.82774442e-01 -5.93402199e-02 -7.29023069e-02
-4.76129025e-01 -2.82588989e-01 4.00330514e-01 2.87289917e-01
2.26926319e-02 -1.64640337e-01 -2.06794202e-01 5.65132737e-01
3.91982019e-01 5.85058592e-02 6.06149323e-02 9.87369597e-01
-6.15823925e-01 -4.29565609e-01 -1.79719225e-01 4.98199105e-01
-6.21138215e-01 -8.57734904e-02 -2.24683876e-03 4.76299286e-01
3.74657363e-01 8.84884715e-01 4.23330307e-01 -2.90336877e-01
2.10574716e-01 1.19098619e-01 4.29270208e-01 -2.21366704e-01
-5.39694548e-01 2.04244405e-01 -3.40093791e-01 -9.49102938e-01
-5.28702199e-01 -4.69621420e-01 -1.69932401e+00 -3.30920607e-01
-2.20260441e-01 -3.68302792e-01 4.50695544e-01 8.97728741e-01
6.50319278e-01 8.31318736e-01 2.02447891e-01 -9.19739127e-01
-2.65645146e-01 -5.27107239e-01 -3.68198693e-01 3.33971173e-01
2.41691649e-01 -7.32793689e-01 -4.84262146e-02 2.17199966e-01] | [9.19897747039795, -0.13850954174995422] |
af68eeb6-e404-4c56-b55d-d0400f14647a | unsupervised-learning-of-object-keypoints-for | 1906.11883 | null | https://arxiv.org/abs/1906.11883v2 | https://arxiv.org/pdf/1906.11883v2.pdf | Unsupervised Learning of Object Keypoints for Perception and Control | The study of object representations in computer vision has primarily focused on developing representations that are useful for image classification, object detection, or semantic segmentation as downstream tasks. In this work we aim to learn object representations that are useful for control and reinforcement learning (RL). To this end, we introduce Transporter, a neural network architecture for discovering concise geometric object representations in terms of keypoints or image-space coordinates. Our method learns from raw video frames in a fully unsupervised manner, by transporting learnt image features between video frames using a keypoint bottleneck. The discovered keypoints track objects and object parts across long time-horizons more accurately than recent similar methods. Furthermore, consistent long-term tracking enables two notable results in control domains -- (1) using the keypoint co-ordinates and corresponding image features as inputs enables highly sample-efficient reinforcement learning; (2) learning to explore by controlling keypoint locations drastically reduces the search space, enabling deep exploration (leading to states unreachable through random action exploration) without any extrinsic rewards. | ['Volodymyr Mnih', 'Tejas Kulkarni', 'Sebastian Borgeaud', 'Catalin Ionescu', 'Andrew Zisserman', 'Malcolm Reynolds', 'Ankush Gupta'] | 2019-06-19 | unsupervised-learning-of-object-keypoints-for-1 | http://papers.nips.cc/paper/9256-unsupervised-learning-of-object-keypoints-for-perception-and-control | http://papers.nips.cc/paper/9256-unsupervised-learning-of-object-keypoints-for-perception-and-control.pdf | neurips-2019-12 | ['3d-human-action-recognition'] | ['computer-vision'] | [ 1.66014016e-01 1.34552583e-01 -6.04429662e-01 -1.45859241e-01
-6.33371532e-01 -5.90612054e-01 7.59390295e-01 2.93246299e-01
-5.73234856e-01 6.50752008e-01 -8.31317753e-02 1.27803281e-01
-3.96954060e-01 -6.21043205e-01 -1.28053403e+00 -7.50514269e-01
-4.97005433e-01 2.78194547e-01 3.17701787e-01 7.32376501e-02
5.76795280e-01 8.93812418e-01 -1.65048611e+00 -7.43059069e-02
3.82717669e-01 1.14465451e+00 6.15381062e-01 8.76431406e-01
8.90652388e-02 9.29674983e-01 -3.38606477e-01 3.26220691e-01
4.97871786e-01 -3.15020889e-01 -9.20571446e-01 2.76498199e-01
5.24408698e-01 -4.75144267e-01 -4.62114781e-01 1.01068103e+00
-2.00716406e-02 5.55741787e-01 3.98771614e-01 -1.26423073e+00
-8.40744436e-01 2.03376621e-01 -3.97936791e-01 2.93298900e-01
1.96288191e-02 6.27973557e-01 9.14862335e-01 -4.46501732e-01
8.90163481e-01 1.36862719e+00 3.76637995e-01 7.29859114e-01
-1.32436800e+00 -2.52204925e-01 6.28529966e-01 3.21350813e-01
-1.02141023e+00 -2.88764626e-01 7.26532280e-01 -4.39464420e-01
1.21731293e+00 1.55053914e-01 1.02540696e+00 9.67363894e-01
3.26802522e-01 1.05904675e+00 6.21262968e-01 -2.78084993e-01
5.76942265e-01 -2.95732319e-01 -1.14005752e-01 9.27127421e-01
1.12167206e-02 5.14084280e-01 -4.58882928e-01 1.54385686e-01
1.30944300e+00 4.36785996e-01 -1.74305573e-01 -8.84536743e-01
-1.57497489e+00 9.57231641e-01 9.12753224e-01 1.54252574e-01
-5.54653704e-01 1.07022679e+00 1.93086252e-01 3.91085953e-01
1.18776523e-02 9.98476148e-01 -7.03347445e-01 -1.10488042e-01
-4.38161373e-01 5.89481473e-01 2.28677899e-01 1.07675791e+00
8.81017268e-01 2.38364443e-01 -3.77413452e-01 3.52241129e-01
1.36022851e-01 4.83881146e-01 4.82617885e-01 -1.46360934e+00
2.44313955e-01 4.84836370e-01 2.52571046e-01 -9.03189480e-01
-4.66509432e-01 -1.87427893e-01 -5.27814865e-01 6.21108353e-01
3.04261565e-01 1.29208155e-02 -1.14119327e+00 1.79839611e+00
3.24717939e-01 3.62542570e-01 6.42896965e-02 1.01204443e+00
1.46500334e-01 6.89898610e-01 2.30791867e-01 -2.08395213e-01
9.40261424e-01 -9.71192837e-01 -4.24097002e-01 -2.35280722e-01
5.98767400e-01 -1.24955356e-01 9.15834129e-01 3.23449165e-01
-1.21890795e+00 -7.82939613e-01 -7.70651221e-01 4.37940331e-03
-4.85475868e-01 -7.85010010e-02 7.39345729e-01 -3.91084217e-02
-1.15998125e+00 1.04324317e+00 -1.15243471e+00 -1.90593272e-01
8.97065818e-01 6.04826868e-01 -3.08041543e-01 6.29095733e-02
-6.62733972e-01 7.35798419e-01 5.65245092e-01 1.95873622e-02
-1.35485983e+00 -8.17452431e-01 -1.12814867e+00 3.48701030e-02
6.66206896e-01 -5.24034321e-01 1.42302382e+00 -1.18258762e+00
-1.62029576e+00 4.86955136e-01 9.66736302e-02 -8.31399858e-01
2.53515214e-01 -6.10098720e-01 2.14721501e-01 4.90600944e-01
2.01925412e-01 1.46895564e+00 1.25299966e+00 -1.03460169e+00
-9.07621861e-01 -4.52561408e-01 1.94448993e-01 1.93509385e-01
-2.38517355e-02 -3.89177382e-01 -4.19419378e-01 -6.87230527e-01
1.54310152e-01 -9.87654924e-01 -6.66923642e-01 5.25254488e-01
-3.01625818e-01 -3.90987366e-01 9.55546737e-01 -1.98315963e-01
4.79548305e-01 -1.82129836e+00 5.27236700e-01 -6.23777211e-02
2.09188238e-01 1.33811831e-01 -3.44322205e-01 8.98474231e-02
1.57466829e-02 1.67498887e-02 -2.13566814e-02 5.02446257e-02
-1.69054478e-01 4.21887189e-01 -4.90365624e-01 4.53747839e-01
8.07778180e-01 1.52214992e+00 -1.23678315e+00 -2.58594245e-01
5.68723977e-01 2.84295976e-01 -5.82029939e-01 1.06482536e-01
-7.96348512e-01 6.21528089e-01 -9.45175827e-01 6.73564970e-01
5.20997606e-02 -2.40630656e-01 -2.17077181e-01 -6.55707344e-02
-2.40544528e-01 -6.82729557e-02 -9.66737568e-01 2.03154302e+00
-2.26884633e-01 7.22622275e-01 -1.93121165e-01 -1.19744277e+00
8.45267355e-01 1.92331672e-02 7.51407266e-01 -9.09047186e-01
-1.18567824e-01 -1.36804700e-01 -4.73173320e-01 -5.36405981e-01
4.54810381e-01 3.48639876e-01 -2.93512028e-02 1.51984006e-01
1.63418993e-01 -2.49116763e-01 2.15879068e-01 7.51231983e-02
9.66214776e-01 3.89214307e-01 1.30917624e-01 -1.80297092e-01
2.06662819e-01 3.96787912e-01 4.99196202e-01 9.26534355e-01
-1.95409328e-01 4.12687689e-01 4.85485882e-01 -7.91403890e-01
-1.05642045e+00 -1.08917379e+00 1.75756305e-01 9.39695835e-01
3.16824913e-01 -3.01812351e-01 -4.22986716e-01 -6.21751904e-01
2.37190858e-01 5.39304078e-01 -8.13522816e-01 -4.11617815e-01
-8.59727323e-01 1.41503394e-01 -2.26392504e-03 8.71432900e-01
2.81717956e-01 -1.56856358e+00 -1.48468280e+00 3.46517146e-01
4.37502325e-01 -7.56272078e-01 -4.46528286e-01 5.16139805e-01
-1.16449440e+00 -1.28141463e+00 -9.45863903e-01 -7.06328869e-01
7.09612966e-01 2.60412067e-01 8.63659024e-01 6.13233857e-02
-8.16728055e-01 8.69395018e-01 -2.76911885e-01 -1.36768460e-01
-3.74794602e-02 -1.37263924e-01 1.21909335e-01 -3.20393413e-01
1.09658623e-02 -3.23296428e-01 -7.31446326e-01 1.32081965e-02
-7.57356465e-01 6.03590198e-02 5.94357252e-01 7.62630939e-01
9.82584953e-01 -2.39259407e-01 5.75755119e-01 -4.40956354e-01
2.51130104e-01 -2.87683636e-01 -1.15513742e+00 1.43905818e-01
-4.76392478e-01 3.49066794e-01 4.55831558e-01 -5.89454532e-01
-5.39418399e-01 3.36336941e-01 4.14896280e-01 -9.85830426e-01
-2.16915131e-01 9.86602828e-02 4.45143640e-01 -3.70541103e-02
4.23050255e-01 3.78275484e-01 1.54861107e-01 -3.55571657e-01
6.15641832e-01 -9.68533978e-02 5.41629016e-01 -7.20814228e-01
4.93318886e-01 4.87285972e-01 2.08129451e-01 -6.29324794e-01
-7.42166936e-01 -3.58428746e-01 -8.50688577e-01 -2.01312736e-01
1.18776488e+00 -7.50826120e-01 -1.13813245e+00 4.89598401e-02
-9.90479231e-01 -6.41169965e-01 -8.47086370e-01 5.91343641e-01
-1.24762702e+00 -1.22447304e-01 -4.62054670e-01 -8.54900777e-01
4.20055352e-03 -1.33070183e+00 1.24796677e+00 5.06883979e-01
-6.94128945e-02 -7.03109801e-01 -9.94492439e-04 -2.53573656e-01
9.15834084e-02 3.74222308e-01 7.79414952e-01 -1.14626184e-01
-1.25377285e+00 2.12195262e-01 -7.11301640e-02 2.10743286e-02
1.88276917e-02 -9.61429998e-02 -6.49843037e-01 -4.12926793e-01
-4.57924455e-01 -6.63145483e-01 9.48636293e-01 8.17402363e-01
1.69505632e+00 -3.82477969e-01 -5.07770240e-01 7.34513700e-01
1.29928708e+00 4.71423656e-01 2.87122637e-01 6.32197142e-01
5.77096939e-01 5.85037053e-01 6.75812483e-01 4.34269249e-01
-5.93695836e-03 4.91762072e-01 8.66232276e-01 6.78554699e-02
6.43384606e-02 -3.43693405e-01 1.64578661e-01 1.16224535e-01
1.27492204e-01 9.64625627e-02 -6.77896559e-01 7.55042195e-01
-2.17051530e+00 -9.99751806e-01 4.16875988e-01 2.04262280e+00
5.18536747e-01 6.50864020e-02 4.66593429e-02 -2.53029108e-01
5.40522099e-01 5.81495725e-02 -1.19981062e+00 -1.36037707e-01
1.92035839e-01 8.05071443e-02 5.80488861e-01 1.99691847e-01
-1.21827555e+00 1.03361523e+00 6.26248598e+00 3.36826473e-01
-1.06844091e+00 -3.51570606e-01 7.56695330e-01 -3.92879963e-01
6.43963441e-02 -1.13029949e-01 -7.10394025e-01 8.83235484e-02
5.76910436e-01 1.08189493e-01 5.58947742e-01 1.19667470e+00
1.23664901e-01 -5.67657612e-02 -1.40503716e+00 1.06266570e+00
-2.78130382e-01 -1.96825552e+00 1.87169477e-01 9.23364796e-03
8.08571935e-01 8.94737318e-02 3.18693072e-01 3.99367273e-01
3.82358938e-01 -1.04716504e+00 8.90878737e-01 6.79096222e-01
5.66358328e-01 -8.08807135e-01 -7.94357210e-02 2.18525931e-01
-1.03862977e+00 -5.79917669e-01 -5.07331610e-01 9.70548466e-02
6.33663088e-02 -1.08114891e-01 -5.94511569e-01 1.61755368e-01
8.68410170e-01 1.18411708e+00 -3.90000969e-01 1.09262180e+00
-5.08551933e-02 2.04448640e-01 -1.62462905e-01 -1.13813177e-01
8.66147637e-01 -1.07994601e-01 5.32700479e-01 7.56537080e-01
2.09699363e-01 1.39974833e-01 5.55398345e-01 1.15857399e+00
8.72133821e-02 -4.18713748e-01 -7.38616943e-01 -2.12090299e-01
2.29258910e-01 1.07353222e+00 -9.88008499e-01 -3.63149315e-01
2.60924833e-04 8.21426451e-01 4.92683083e-01 5.11295617e-01
-5.78854263e-01 -2.28498578e-01 9.08308685e-01 -8.40149075e-02
7.23540664e-01 -4.83786821e-01 1.70440927e-01 -8.84992540e-01
-2.04805702e-01 -4.68999237e-01 1.98973954e-01 -8.87330472e-01
-8.24585438e-01 3.22504401e-01 1.58314496e-01 -1.05537808e+00
-6.03765965e-01 -7.13445246e-01 -4.35328037e-01 4.25619572e-01
-1.69721067e+00 -9.00403082e-01 -5.26560321e-02 7.80820370e-01
1.03861165e+00 -1.82789505e-01 6.09920979e-01 -3.50437552e-01
-3.07091564e-01 1.71134800e-01 1.14812173e-01 9.81355608e-02
1.57152623e-01 -1.38315880e+00 4.95363981e-01 3.57502520e-01
5.22518694e-01 4.86606121e-01 4.03607994e-01 -5.43284118e-01
-1.73194230e+00 -1.21951306e+00 6.90892860e-02 -4.60005492e-01
4.84698623e-01 -1.23695187e-01 -7.55938768e-01 7.92983055e-01
-1.79452613e-01 4.84667748e-01 -2.09169373e-01 -7.95636103e-02
-2.19104275e-01 -9.39607099e-02 -8.68010879e-01 7.35908389e-01
1.02287805e+00 -2.55525827e-01 -4.26873058e-01 4.87078190e-01
1.01182365e+00 -3.85901779e-01 -5.56446373e-01 1.59204036e-01
4.68055785e-01 -5.84177911e-01 1.21860540e+00 -1.12926519e+00
2.47135490e-01 -2.88119763e-01 -4.15970050e-02 -1.14989090e+00
-3.75229627e-01 -8.19319367e-01 -4.98755723e-01 5.05671620e-01
6.83056861e-02 -2.56455541e-01 9.29328382e-01 4.05010790e-01
-1.92580387e-01 -9.22753692e-01 -9.39338446e-01 -7.82938898e-01
-5.34122288e-02 -2.92863935e-01 3.71169835e-01 5.05084217e-01
-3.46348017e-01 -8.55100453e-02 -1.34322435e-01 1.40912339e-01
6.19568169e-01 3.98279756e-01 4.86266077e-01 -1.06513464e+00
-1.62375495e-01 -6.02557778e-01 -6.41391277e-01 -1.45709920e+00
2.20051289e-01 -6.35601044e-01 2.83946693e-01 -1.26613522e+00
6.54027797e-03 -6.35527253e-01 -4.07678574e-01 6.50419116e-01
5.54020964e-02 -3.50716636e-02 5.25091827e-01 1.91483244e-01
-9.50368166e-01 6.72227323e-01 1.61737382e+00 -3.42059672e-01
-3.37875932e-01 2.38261763e-02 -3.83912683e-01 5.88643730e-01
7.96032667e-01 -3.96010280e-01 -5.39465427e-01 -6.50830209e-01
-2.14980133e-02 2.76250541e-01 9.05599713e-01 -9.76094246e-01
1.96567267e-01 -5.22456944e-01 7.43703783e-01 -6.88613653e-01
5.91594219e-01 -8.71453047e-01 -3.68651092e-01 7.83198357e-01
-8.40544760e-01 2.74564594e-01 2.29956597e-01 1.09924364e+00
6.22448623e-02 -2.74354875e-01 6.12951756e-01 -4.24103588e-01
-1.21479487e+00 5.85969508e-01 -4.46390003e-01 -2.73047574e-02
1.48569119e+00 -3.53678584e-01 3.61622199e-02 -1.28406525e-01
-7.83055186e-01 4.54084963e-01 2.16021463e-01 7.11743057e-01
8.58791411e-01 -1.24441719e+00 -2.39922106e-01 3.75604928e-01
4.51976098e-02 3.74035239e-01 9.25623104e-02 3.93234193e-01
-4.59312409e-01 5.92971444e-01 -5.87374747e-01 -1.11898685e+00
-9.01353598e-01 8.92773807e-01 3.59242052e-01 7.75341094e-02
-1.07230854e+00 8.95011425e-01 3.45863312e-01 -2.88788736e-01
4.64797616e-01 -7.67657459e-01 -1.24162704e-01 -2.88457185e-01
6.08069837e-01 2.77225643e-01 -4.32943255e-01 -2.11747453e-01
-2.67765801e-02 7.64176369e-01 -3.23480606e-01 3.98717858e-02
1.53135896e+00 -9.64123160e-02 2.22439915e-01 5.59409797e-01
1.23035073e+00 -9.01134014e-01 -2.23101926e+00 -1.83489963e-01
2.68782914e-01 -5.41019976e-01 1.90697372e-01 -5.01837790e-01
-1.03755724e+00 8.51300001e-01 6.83530331e-01 2.62571853e-02
7.53407836e-01 2.18200460e-01 6.25386775e-01 9.86522079e-01
4.21682179e-01 -1.15120649e+00 6.80716872e-01 4.91987407e-01
9.07130241e-01 -1.29678929e+00 -2.46655345e-01 2.07821816e-01
-6.76664948e-01 1.35456419e+00 8.22409689e-01 -4.89192903e-01
3.85360420e-01 -1.03008360e-01 -2.31908575e-01 -3.40620726e-01
-9.75230992e-01 -2.88090289e-01 2.32410327e-01 8.80515754e-01
-2.02063724e-01 -3.24377835e-01 5.58213592e-01 -2.14424748e-02
3.70264709e-01 -1.40622094e-01 2.35468790e-01 1.15413332e+00
-8.49893630e-01 -6.92814112e-01 -1.42025337e-01 4.06247824e-01
-1.18527852e-01 2.47040987e-01 -8.02131742e-03 9.50600147e-01
-7.79273286e-02 4.34121579e-01 4.84207362e-01 1.96094751e-01
1.24662787e-01 -3.73169452e-01 7.22090721e-01 -7.08674550e-01
-3.30355734e-01 -2.89593618e-02 -6.05408788e-01 -1.10006022e+00
-3.44967872e-01 -6.81883872e-01 -1.50185561e+00 2.06594452e-01
-2.12091923e-01 6.77371724e-03 6.87947571e-01 8.37057829e-01
5.04792392e-01 7.34300137e-01 4.96238351e-01 -1.32951951e+00
-8.12436223e-01 -4.10220504e-01 -3.77979487e-01 2.66946703e-01
8.82408381e-01 -8.12457502e-01 1.94373906e-01 2.86158293e-01] | [4.698873043060303, 0.7201289534568787] |
a8799b5f-abb7-468d-8ae3-f43baab94b21 | zero-shot-slot-and-intent-detection-in-low | 2304.13292 | null | https://arxiv.org/abs/2304.13292v1 | https://arxiv.org/pdf/2304.13292v1.pdf | Zero-Shot Slot and Intent Detection in Low-Resource Languages | Intent detection and slot filling are critical tasks in spoken and natural language understanding for task-oriented dialog systems. In this work we describe our participation in the slot and intent detection for low-resource language varieties (SID4LR; Aepli et al. (2023)). We investigate the slot and intent detection (SID) tasks using a wide range of models and settings. Given the recent success of multitask-prompted finetuning of large language models, we also test the generalization capability of the recent encoder-decoder model mT0 (Muennighoff et al., 2022) on new tasks (i.e., SID) in languages they have never intentionally seen. We show that our best model outperforms the baseline by a large margin (up to +30 F1 points) in both SID tasks | ['Muhammad Abdul-Mageed', 'Alcides Alcoba Inciarte', 'El Moatez Billah Nagoudi', 'Gagan Bhatia', 'Sang Yun Kwon'] | 2023-04-26 | null | null | null | null | ['intent-detection', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [ 7.88650885e-02 5.77579618e-01 -2.24269480e-01 -4.95565414e-01
-8.22900653e-01 -6.28130436e-01 1.01937664e+00 -2.00257868e-01
-7.39952028e-01 9.05161202e-01 7.62561083e-01 -5.77113390e-01
3.14334780e-01 3.28246877e-02 -1.11512914e-01 3.39816436e-02
-6.54001674e-03 1.00108659e+00 2.52329111e-01 -5.81883132e-01
1.53604690e-02 8.62888899e-03 -1.02413297e+00 5.15172243e-01
8.18929970e-01 4.31104034e-01 5.09635091e-01 9.30141628e-01
-3.18134755e-01 9.43705976e-01 -8.00276279e-01 -2.79761195e-01
-1.55303746e-01 -2.96790808e-01 -1.27358210e+00 1.57217115e-01
2.67895103e-01 -5.34807324e-01 -3.61161232e-01 5.37706733e-01
5.46964109e-01 2.65030742e-01 5.35911143e-01 -1.15100241e+00
-4.16464299e-01 8.36481571e-01 -1.21552683e-01 2.72225618e-01
6.48130953e-01 1.96520075e-01 1.16330051e+00 -1.16068292e+00
7.60013223e-01 1.55988908e+00 4.47952360e-01 8.91309261e-01
-1.23929310e+00 -3.92995149e-01 3.98524106e-01 -8.31821039e-02
-1.34212303e+00 -1.00050759e+00 4.14912283e-01 -2.19787821e-01
1.76718938e+00 2.95571178e-01 -8.01568925e-02 1.40450919e+00
-8.86806920e-02 1.35150206e+00 1.03291476e+00 -7.51083434e-01
2.12216422e-01 4.32467043e-01 3.17128509e-01 4.95155692e-01
-2.35993832e-01 -8.36481005e-02 -8.32463026e-01 -2.18033478e-01
4.21220332e-01 -6.91332519e-01 8.42851214e-03 1.68444604e-01
-1.25263762e+00 1.08041525e+00 -2.14222133e-01 1.67423457e-01
-1.88444123e-01 -2.24176452e-01 8.20372999e-01 5.77192724e-01
7.07550168e-01 6.43647313e-01 -1.06370866e+00 -4.63129848e-01
-5.03974617e-01 3.75667632e-01 1.46105218e+00 1.15741086e+00
3.12865317e-01 1.43667191e-01 -5.16051769e-01 1.40808654e+00
4.01800781e-01 4.03232038e-01 7.86055267e-01 -7.35076427e-01
7.36404061e-01 1.45199984e-01 2.15831578e-01 -1.68207899e-01
-7.04074502e-01 2.04009674e-02 -5.07218242e-01 -3.33561838e-01
5.06344259e-01 -4.84396785e-01 -8.69668484e-01 1.98891592e+00
1.79255247e-01 -4.45356011e-01 5.87749839e-01 5.17487824e-01
9.49709952e-01 8.41214418e-01 6.41446352e-01 -4.03603226e-01
1.62495553e+00 -1.05064666e+00 -9.53295469e-01 -9.36054051e-01
1.01519227e+00 -8.91762972e-01 1.31664598e+00 2.53443718e-01
-8.54622781e-01 -5.17134726e-01 -7.10829675e-01 -2.63219237e-01
-1.89748690e-01 1.95994690e-01 7.69979179e-01 7.67535985e-01
-1.29796576e+00 -1.24381267e-01 -5.85071027e-01 -8.47604752e-01
-1.68245792e-01 2.30104536e-01 -2.70270109e-01 1.29672498e-01
-1.52005947e+00 1.11136150e+00 3.15471232e-01 -3.31927359e-01
-7.05242336e-01 -4.06419396e-01 -1.11357319e+00 -2.47712493e-01
6.20434165e-01 -3.98343891e-01 1.90513432e+00 -4.38494623e-01
-1.68406498e+00 1.17827857e+00 -5.28227329e-01 -7.89739311e-01
3.35495442e-01 -6.13448083e-01 -4.10201430e-01 -2.48914018e-01
2.94736534e-01 1.05090225e+00 5.33164620e-01 -7.42341757e-01
-6.98435664e-01 -1.59617782e-01 2.88108587e-02 6.88951850e-01
-1.21706620e-01 4.63507146e-01 -1.21971518e-01 -6.01208031e-01
-1.38135955e-01 -9.13833737e-01 -1.54580995e-01 -4.42583531e-01
-5.35461307e-01 -6.50042772e-01 7.30035126e-01 -9.33440268e-01
1.12813604e+00 -2.02172780e+00 -9.48572978e-02 -4.95680034e-01
1.00533620e-01 4.60534006e-01 -2.99787700e-01 7.16798782e-01
1.46655470e-01 1.24570549e-01 -1.57254234e-01 -7.33225703e-01
3.84563446e-01 3.85137916e-01 -5.36001205e-01 4.52087596e-02
2.46547148e-01 1.12737978e+00 -3.75836879e-01 -4.03952658e-01
3.75112206e-01 -2.80140359e-02 -5.47757983e-01 5.12496114e-01
-7.33100414e-01 3.37294370e-01 -1.51295871e-01 4.60590184e-01
1.80479094e-01 -9.03004557e-02 2.52537906e-01 2.42504895e-01
-3.43707144e-01 1.17589819e+00 -9.76416349e-01 1.78300166e+00
-7.60470986e-01 6.62574589e-01 9.35592651e-02 -4.60682720e-01
8.31114888e-01 8.08431327e-01 6.58708736e-02 -6.82682931e-01
-1.55323312e-01 2.56879449e-01 2.22888201e-01 -2.37912625e-01
8.78446102e-01 -2.29532883e-01 -6.77560985e-01 7.00155616e-01
2.89135069e-01 -4.46914248e-02 1.13421462e-01 3.01982194e-01
9.38824177e-01 -2.87620962e-01 9.04903769e-01 -5.37506819e-01
5.36398411e-01 1.44348860e-01 1.46949813e-01 1.07531679e+00
-5.58395863e-01 3.67130131e-01 5.05637646e-01 -2.60965586e-01
-1.01941156e+00 -8.56065273e-01 -2.03738719e-01 1.92396295e+00
-2.52128184e-01 -2.89123327e-01 -4.34853882e-01 -7.41578162e-01
-1.32260099e-01 1.22469676e+00 -3.10606390e-01 -3.62267159e-02
-7.03088522e-01 -4.86012042e-01 9.56404567e-01 4.01203692e-01
5.33467054e-01 -1.44260383e+00 -2.56738037e-01 5.02448559e-01
-2.12385476e-01 -1.52609503e+00 -5.37511170e-01 4.34870034e-01
-3.56311530e-01 -5.51882863e-01 -6.39172554e-01 -9.71461535e-01
-1.87048689e-01 -5.95118888e-02 1.36498988e+00 -2.09293142e-01
-8.62403139e-02 2.97686040e-01 -4.03361857e-01 -3.28732669e-01
-8.52586627e-01 4.02652651e-01 2.99899787e-01 -5.37391722e-01
6.08788431e-01 7.50421807e-02 1.45733535e-01 3.35702509e-01
-5.76912761e-01 2.16144428e-01 2.97024131e-01 1.08690858e+00
-1.51292995e-01 -5.67896366e-01 8.15846205e-01 -1.02051103e+00
8.71193111e-01 -3.14853847e-01 -5.24228692e-01 2.22297281e-01
-4.42504615e-01 2.41032869e-01 3.49668413e-01 -3.20130616e-01
-1.48220754e+00 -1.92903671e-02 -7.44979739e-01 2.73817509e-01
-6.16208017e-01 2.67869800e-01 -4.30903494e-01 2.35302240e-01
5.38019061e-01 2.57632196e-01 -1.72650456e-01 -6.32790744e-01
5.86198032e-01 1.00242686e+00 2.74984896e-01 -3.69451076e-01
2.87730247e-01 -1.20609283e-01 -7.28677750e-01 -1.22325897e+00
-9.23303962e-01 -7.38240898e-01 -6.85765088e-01 1.42713636e-01
9.37399447e-01 -1.17857957e+00 -5.45439482e-01 5.07844031e-01
-1.54040599e+00 -7.24599481e-01 -1.79080173e-01 3.38701159e-01
-6.15230203e-01 3.13016713e-01 -8.68458331e-01 -1.01361883e+00
-3.76363397e-01 -9.84305620e-01 1.27367461e+00 -1.96582489e-02
-9.15713549e-01 -1.29380739e+00 2.58308619e-01 3.23917359e-01
4.40963745e-01 -6.41056716e-01 8.51214468e-01 -1.43471098e+00
-1.67462468e-01 3.29727411e-01 -2.04363868e-01 5.08546829e-02
-6.81076869e-02 -7.46597588e-01 -1.18960202e+00 -1.76078200e-01
1.12577654e-01 -8.78872752e-01 7.80550361e-01 3.00608754e-01
3.46457303e-01 -5.02962947e-01 -1.48478001e-01 9.61243734e-02
8.06743801e-01 4.19283360e-01 4.47066575e-01 2.70110071e-01
5.19004405e-01 7.51052678e-01 7.93422401e-01 3.84112179e-01
7.44687021e-01 9.44543660e-01 -2.40709469e-01 -6.98943660e-02
-2.63557136e-01 -4.21182185e-01 5.84311962e-01 7.43446827e-01
8.01924825e-01 -7.69051850e-01 -1.08424091e+00 6.05680168e-01
-1.79759145e+00 -4.87489194e-01 2.38372367e-02 1.86578166e+00
1.17997336e+00 2.01448843e-01 1.48973256e-01 -4.37049508e-01
5.46010673e-01 5.76049864e-01 -3.90335619e-01 -8.42497885e-01
-7.19359368e-02 5.97741269e-02 2.47840285e-01 1.03051627e+00
-1.39759552e+00 1.89161372e+00 6.70906496e+00 7.71658421e-01
-5.33681035e-01 2.13143080e-01 6.41859233e-01 4.24364239e-01
-1.70813113e-01 -9.72720534e-02 -1.36872077e+00 6.31021559e-02
1.27060866e+00 -1.90986753e-01 3.11807781e-01 9.29272592e-01
-1.19885787e-01 -3.45106840e-01 -1.17661905e+00 7.32250035e-01
3.04989964e-01 -8.45864952e-01 -2.38994777e-01 -1.99496612e-01
4.40204203e-01 2.23223463e-01 -2.81839818e-02 8.93140137e-01
7.09567845e-01 -1.02907729e+00 3.69654328e-01 -1.03624806e-01
9.15461659e-01 -3.85666251e-01 6.76739037e-01 6.10197902e-01
-9.65806663e-01 1.55371085e-01 -3.48594666e-01 -2.64542907e-01
6.47033572e-01 1.97573185e-01 -1.50323737e+00 -5.36765978e-02
1.67284906e-01 4.83111829e-01 -3.38204890e-01 4.08419102e-01
-2.71856427e-01 7.49364853e-01 -3.30376446e-01 -2.62292325e-01
2.78203428e-01 3.13636601e-01 8.40295613e-01 1.44727683e+00
-2.94935703e-01 3.23204324e-02 5.06269276e-01 4.96394664e-01
-2.01087445e-02 1.19368009e-01 -6.68671012e-01 -2.22514018e-01
4.54412013e-01 8.94072771e-01 -5.30986845e-01 -3.85861933e-01
-4.06220526e-01 1.14734674e+00 3.67302448e-01 1.87388927e-01
-3.40006977e-01 -1.23982346e-02 9.45162654e-01 -3.88779968e-01
1.65412292e-01 -5.09305418e-01 -3.63046408e-01 -1.10632086e+00
-2.69953787e-01 -1.06499970e+00 4.68809128e-01 -4.97669458e-01
-1.31016195e+00 8.63783360e-01 1.49872601e-01 -5.97780049e-01
-1.08157563e+00 -7.38263488e-01 -3.27816099e-01 1.01989329e+00
-1.29231262e+00 -1.16659033e+00 3.58036339e-01 3.38790089e-01
1.39930367e+00 -4.79932338e-01 1.17313075e+00 8.45323578e-02
-4.00799155e-01 6.11784756e-01 -2.59233594e-01 1.86456591e-01
8.86195421e-01 -1.24635422e+00 1.23130000e+00 6.66800559e-01
2.95139730e-01 4.14325029e-01 8.85279596e-01 -8.29952598e-01
-1.10187745e+00 -8.60991240e-01 1.57735956e+00 -6.60908401e-01
7.69102633e-01 -9.61413443e-01 -9.56187844e-01 1.06709504e+00
4.79274422e-01 -7.65987217e-01 6.50218487e-01 5.42352140e-01
-2.57456779e-01 5.31931758e-01 -9.79935408e-01 7.96210766e-01
1.10490227e+00 -7.88482666e-01 -9.39983189e-01 5.43355823e-01
1.23464239e+00 -5.14137626e-01 -4.41219091e-01 3.40220600e-01
2.86812097e-01 -7.17216015e-01 9.22048688e-01 -7.91246593e-01
2.37469561e-02 3.01191747e-01 -2.89308429e-01 -1.25288141e+00
-5.99402264e-02 -6.12146497e-01 9.65776294e-02 1.26013696e+00
5.49936473e-01 -6.22185409e-01 4.20645386e-01 7.83525944e-01
-3.09044331e-01 -1.57448560e-01 -1.14749765e+00 -6.60496235e-01
1.68697864e-01 -7.00972855e-01 2.20307350e-01 6.16137207e-01
4.01871353e-01 1.00900364e+00 -6.80182755e-01 -3.68569754e-02
8.05894658e-02 -1.36880144e-01 7.59566963e-01 -1.04268587e+00
-3.03646505e-01 -1.01918519e-01 6.23939559e-03 -1.63108277e+00
6.06894255e-01 -5.71057379e-01 4.43678498e-01 -1.28428209e+00
1.16677396e-02 -4.66726124e-01 2.17264533e-01 6.94013059e-01
-1.99821517e-01 -7.66083375e-02 2.16425508e-01 2.28372574e-01
-1.00946558e+00 7.69220352e-01 6.27154469e-01 1.82921529e-01
-4.48872805e-01 8.70978236e-02 -6.06041729e-01 6.34311795e-01
8.15997183e-01 -2.97587186e-01 -4.25819218e-01 -3.45856011e-01
-2.46376678e-01 3.31968367e-01 3.17888930e-02 -5.47916293e-01
1.75610632e-01 -6.57225633e-03 -8.98388848e-02 -6.24428928e-01
5.50437033e-01 -2.67291158e-01 -4.74871397e-01 4.51276153e-01
-8.48584473e-01 4.78619114e-02 6.30374491e-01 1.37150437e-01
-1.21203765e-01 -2.69770086e-01 4.72985893e-01 -2.45373502e-01
-1.17081952e+00 -2.65588015e-02 -9.14589643e-01 6.25628412e-01
6.17166758e-01 1.99977279e-01 -5.41522503e-01 -7.42327988e-01
-6.62469983e-01 2.67486304e-01 7.86006972e-02 8.14287484e-01
3.69552970e-01 -8.04715633e-01 -7.69705713e-01 4.87465143e-01
2.97103435e-01 -4.02783990e-01 1.44266952e-02 6.07189715e-01
-1.34439632e-01 1.07150626e+00 1.47330146e-02 -4.53558326e-01
-1.21079850e+00 1.64151177e-01 1.44696310e-01 -4.85692263e-01
-4.55956131e-01 1.15204525e+00 4.28389788e-01 -8.99856091e-01
4.19873863e-01 -4.21714149e-02 -3.17531854e-01 -3.71110328e-02
5.75302243e-01 3.83313112e-02 -4.61413115e-02 -6.38130367e-01
-5.31004727e-01 -3.96775872e-01 -5.65360010e-01 -7.17282534e-01
8.11412096e-01 -4.73214298e-01 1.45785898e-01 7.74848580e-01
1.03046858e+00 -2.76737437e-02 -1.01577199e+00 -4.79102999e-01
5.27516782e-01 -9.47262719e-02 -3.22356701e-01 -1.10020173e+00
-4.21025939e-02 7.17969894e-01 2.99583018e-01 3.52424175e-01
5.58428168e-01 3.48549396e-01 8.16050291e-01 7.36653626e-01
3.62666756e-01 -1.10429883e+00 -4.65817638e-02 1.45937514e+00
9.12442684e-01 -1.67368066e+00 -3.26708347e-01 -3.59614134e-01
-1.26222312e+00 7.20257461e-01 7.42979705e-01 4.38485831e-01
3.49126011e-01 4.38641816e-01 4.45247471e-01 -6.07317388e-02
-1.27358603e+00 -5.06998897e-01 2.34223187e-01 7.19263375e-01
8.00553977e-01 2.30661035e-01 -3.88795406e-01 4.94610339e-01
-3.76610994e-01 -4.13062185e-01 4.07007664e-01 7.52794504e-01
-6.81148231e-01 -1.16751921e+00 -2.19340980e-01 4.37238455e-01
-1.54334500e-01 -5.96235991e-01 -5.24666727e-01 9.39588428e-01
-5.56681871e-01 1.20087934e+00 1.20887626e-02 -1.96064353e-01
2.56371707e-01 6.66700125e-01 -1.72637537e-01 -1.18298292e+00
-4.97275770e-01 1.36129186e-01 8.93488586e-01 -4.55917507e-01
3.47513817e-02 -8.48914027e-01 -1.13275766e+00 1.04262248e-01
-1.56514704e-01 9.99372900e-02 3.92289698e-01 1.29277623e+00
6.10893555e-02 2.51946628e-01 3.49353939e-01 -4.58757758e-01
-7.50690937e-01 -1.43187594e+00 -5.67444682e-01 -4.52412702e-02
3.84845734e-01 -5.27346313e-01 -4.11949575e-01 -6.18667267e-02] | [12.562447547912598, 7.889482021331787] |
a5987816-5f51-4073-b322-4b60244ba797 | learning-human-objectives-by-evaluating | 1912.05652 | null | https://arxiv.org/abs/1912.05652v2 | https://arxiv.org/pdf/1912.05652v2.pdf | Learning Human Objectives by Evaluating Hypothetical Behavior | We seek to align agent behavior with a user's objectives in a reinforcement learning setting with unknown dynamics, an unknown reward function, and unknown unsafe states. The user knows the rewards and unsafe states, but querying the user is expensive. To address this challenge, we propose an algorithm that safely and interactively learns a model of the user's reward function. We start with a generative model of initial states and a forward dynamics model trained on off-policy data. Our method uses these models to synthesize hypothetical behaviors, asks the user to label the behaviors with rewards, and trains a neural network to predict the rewards. The key idea is to actively synthesize the hypothetical behaviors from scratch by maximizing tractable proxies for the value of information, without interacting with the environment. We call this method reward query synthesis via trajectory optimization (ReQueST). We evaluate ReQueST with simulated users on a state-based 2D navigation task and the image-based Car Racing video game. The results show that ReQueST significantly outperforms prior methods in learning reward models that transfer to new environments with different initial state distributions. Moreover, ReQueST safely trains the reward model to detect unsafe states, and corrects reward hacking before deploying the agent. | ['Siddharth Reddy', 'Jan Leike', 'Shane Legg', 'Sergey Levine', 'Anca D. Dragan'] | 2019-12-05 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/664-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/664-Paper.pdf | icml-2020-1 | ['carracing-v0'] | ['playing-games'] | [-1.31496400e-01 3.76604050e-01 -3.19669515e-01 -2.79199362e-01
-8.37562621e-01 -7.66456783e-01 5.87857902e-01 -2.19995916e-01
-7.32363522e-01 8.20142329e-01 2.98171416e-02 -4.40248668e-01
1.45503223e-01 -7.06791759e-01 -8.60687256e-01 -5.15599370e-01
-4.01727289e-01 7.61684239e-01 3.42565030e-02 -4.79076087e-01
2.30576217e-01 2.70970136e-01 -1.58912051e+00 -2.58070350e-01
8.99024606e-01 6.82979465e-01 4.14433151e-01 1.05666375e+00
3.39012504e-01 1.14373648e+00 -5.03575981e-01 1.95078820e-01
4.20241743e-01 -3.57889473e-01 -7.51702845e-01 2.32187994e-02
-2.42304459e-01 -1.11450589e+00 -5.75418890e-01 9.78045404e-01
1.74362212e-01 5.71131527e-01 4.97121811e-01 -1.66505933e+00
-4.54959452e-01 4.15675759e-01 2.00081453e-01 -1.46776572e-01
4.62542266e-01 9.88028288e-01 7.69298136e-01 2.55967351e-03
5.08701622e-01 1.22115767e+00 -6.30814629e-03 1.06723714e+00
-1.20518720e+00 -4.45145637e-01 4.46242422e-01 -3.16177122e-02
-9.73936915e-01 -2.86291242e-01 3.69971931e-01 -4.29418981e-01
8.91002238e-01 1.46486843e-02 8.09521973e-01 1.37567282e+00
-3.66571732e-02 9.04176652e-01 8.86299372e-01 1.50825068e-01
7.83926547e-01 1.62134722e-01 -2.02544048e-01 9.09749329e-01
-3.30729246e-01 9.37938929e-01 -1.72924846e-01 -4.12562698e-01
9.17167127e-01 6.84703737e-02 -6.00157604e-02 -7.01324761e-01
-1.22649384e+00 6.63677752e-01 2.12104633e-01 -3.62401336e-01
-7.23771870e-01 5.84719062e-01 -8.22787061e-02 4.94969696e-01
-2.68367618e-01 7.20635533e-01 -4.66151595e-01 -6.93400919e-01
-1.71264589e-01 8.47261548e-01 8.86452436e-01 1.14899957e+00
7.36226380e-01 3.58376622e-01 -3.63542050e-01 3.06147873e-01
2.62178421e-01 9.58463371e-01 2.83992052e-01 -1.68563247e+00
3.18756133e-01 2.82123357e-01 1.15218198e+00 -3.34888548e-01
-2.46737525e-01 -1.10233150e-01 4.84248810e-02 6.23900890e-01
4.85221356e-01 -6.66818678e-01 -1.03889370e+00 2.28244066e+00
4.02484328e-01 3.56265724e-01 4.77439761e-01 1.08029068e+00
3.53429057e-02 7.54296958e-01 -1.10861892e-02 -9.13772825e-03
6.83533967e-01 -8.12201321e-01 -4.40157682e-01 -4.20196265e-01
6.71728134e-01 1.67666905e-04 1.37087619e+00 3.52254122e-01
-1.22801208e+00 -3.22129220e-01 -6.85575783e-01 5.11045218e-01
4.14665751e-02 -4.37647104e-04 4.25781429e-01 4.34660822e-01
-1.11201775e+00 8.24410737e-01 -1.21049917e+00 -1.91454083e-01
5.94924875e-02 6.59124017e-01 2.05084428e-01 1.76870733e-01
-9.89198029e-01 9.42963362e-01 3.80269200e-01 -3.69339794e-01
-2.13049722e+00 -2.54962981e-01 -7.87634671e-01 1.75520450e-01
9.96567488e-01 -5.89769244e-01 1.96240067e+00 -6.87063277e-01
-2.24889922e+00 1.20006301e-01 2.75751412e-01 -4.72177476e-01
2.81207323e-01 -1.80689782e-01 -4.07883376e-02 -1.57337680e-01
1.18624963e-01 8.05390060e-01 9.39110458e-01 -1.29973435e+00
-9.17069852e-01 -1.55203640e-01 5.38110495e-01 6.73197865e-01
1.15053251e-01 -5.79291761e-01 -2.43716583e-01 4.45128530e-02
-4.33957845e-01 -1.36419594e+00 -5.93029082e-01 -3.74002784e-01
-2.22806379e-01 -5.98730966e-02 6.08941615e-01 -2.25521117e-01
7.21515775e-01 -2.05449581e+00 2.27709144e-01 5.02694368e-01
-6.31280318e-02 -2.14104772e-01 -5.29310524e-01 2.39999279e-01
4.21633929e-01 -2.70931065e-01 -4.88728173e-02 -1.20167129e-01
3.00884992e-01 5.43066204e-01 -5.24215341e-01 1.75825432e-01
-6.74209893e-02 8.43502402e-01 -1.39007556e+00 -4.55721840e-02
1.89573929e-01 -7.14400932e-02 -1.05345476e+00 1.00010860e+00
-8.12251151e-01 8.24603915e-01 -8.25325489e-01 3.60595554e-01
4.02138010e-02 -1.35929987e-01 4.34972465e-01 6.00777149e-01
-1.63152888e-01 3.56010318e-01 -1.14223373e+00 1.56190467e+00
-3.55866730e-01 -1.85141459e-01 1.50026888e-01 -7.34264493e-01
6.91437840e-01 1.42794758e-01 6.18677914e-01 -6.66135907e-01
3.21042389e-01 2.74636247e-03 -8.71357769e-02 -7.45701194e-01
6.00515366e-01 -3.40739600e-02 -4.89949495e-01 7.15444684e-01
1.31060198e-01 -5.43963194e-01 -1.64365977e-01 3.71672958e-01
1.31670916e+00 5.67034662e-01 -6.78569004e-02 2.19895765e-01
-1.03086911e-01 1.60626680e-01 3.11279625e-01 1.20390069e+00
-4.00997341e-01 -9.76521745e-02 7.58386672e-01 -1.87756538e-01
-9.25112784e-01 -1.21179545e+00 7.78725564e-01 1.31270957e+00
4.55947191e-01 -1.06869556e-01 -7.80887008e-01 -8.18633974e-01
7.22604897e-03 1.14022875e+00 -5.56092858e-01 -4.39560056e-01
-4.51699376e-01 -2.21806064e-01 1.83878928e-01 3.16630512e-01
2.25356415e-01 -1.16524255e+00 -1.14232731e+00 1.86281100e-01
-3.03381592e-01 -7.73538828e-01 -7.22871125e-01 2.06953138e-01
-5.35232604e-01 -9.10415113e-01 -4.41150218e-01 -5.31683266e-01
7.11591721e-01 4.38891947e-02 7.62519121e-01 -1.65624619e-01
1.52989730e-01 1.03871667e+00 4.53777723e-02 -1.24703340e-01
-7.62277842e-01 -1.57406196e-01 3.37710738e-01 -2.61449397e-01
-8.97439942e-02 -2.92714447e-01 -5.06810904e-01 3.26407552e-01
-6.40656769e-01 -3.93922701e-02 2.56479651e-01 8.73196006e-01
4.39456969e-01 -1.48525074e-01 4.90455687e-01 -5.86020172e-01
1.04077446e+00 -7.00985193e-01 -1.22399449e+00 9.32313278e-02
-6.52885318e-01 5.49551904e-01 6.74484015e-01 -7.24867284e-01
-1.05408430e+00 4.15470958e-01 9.36690122e-02 -4.64055955e-01
-3.05330396e-01 1.05047300e-01 -3.67380194e-02 3.92052948e-01
8.83084595e-01 1.65822312e-01 2.72804648e-01 8.38896930e-02
6.79972351e-01 3.47462267e-01 6.86050236e-01 -1.12692237e+00
7.45567739e-01 1.51410803e-01 -3.18701774e-01 -3.53035688e-01
-5.24279952e-01 -1.22270584e-01 2.71383417e-03 -4.58035260e-01
7.12513268e-01 -8.19776714e-01 -1.42858791e+00 3.35647404e-01
-8.40128958e-01 -1.27445364e+00 -5.93289316e-01 5.52966058e-01
-1.24718595e+00 -1.54766729e-02 -3.49398583e-01 -1.22439611e+00
3.15527588e-01 -1.50841892e+00 8.93778622e-01 4.16874677e-01
-1.29768178e-01 -5.28446794e-01 2.80278116e-01 -5.02134338e-02
5.06529510e-01 -5.58880679e-02 7.58084655e-01 -4.64496434e-01
-1.07611370e+00 6.41735494e-02 3.14832062e-01 8.86420079e-04
1.75838582e-02 -4.24955368e-01 -6.59983933e-01 -3.07628393e-01
-1.45195678e-01 -8.23133111e-01 1.12454832e-01 3.92382771e-01
1.00120783e+00 -7.50006557e-01 -2.19017133e-01 2.83674836e-01
8.31404626e-01 6.45373642e-01 3.58541429e-01 2.32475922e-01
1.30797967e-01 4.71064061e-01 8.24119866e-01 8.37736905e-01
7.81533897e-01 5.60424089e-01 1.00610232e+00 3.93211573e-01
7.06689894e-01 -7.07108378e-01 9.12506342e-01 -6.29949123e-02
-1.03779629e-01 -1.70019105e-01 -5.28424561e-01 3.33858490e-01
-2.20038056e+00 -1.08136797e+00 6.70420170e-01 2.44268012e+00
8.03755999e-01 2.83859491e-01 5.91152430e-01 -7.09274292e-01
1.97998837e-01 -2.44215339e-01 -1.35766947e+00 -2.89127320e-01
6.61295831e-01 -4.04083841e-02 6.36500239e-01 9.00180042e-01
-7.83351421e-01 1.25055361e+00 6.46660709e+00 1.37921691e-01
-9.15806651e-01 -1.77155659e-01 5.79370558e-01 -3.14736307e-01
-3.66820544e-01 1.83563262e-01 -6.57197535e-01 3.06686729e-01
1.17566144e+00 -1.44307494e-01 1.41908014e+00 1.27255130e+00
4.45411921e-01 -2.14568943e-01 -1.28412652e+00 6.39298081e-01
-2.74164319e-01 -1.01568604e+00 -4.92714554e-01 2.22615823e-02
5.18009186e-01 2.06023619e-01 3.70641321e-01 1.09106481e+00
1.44292569e+00 -8.60453963e-01 8.39345038e-01 6.43616498e-01
5.75579286e-01 -7.71368921e-01 2.35386714e-01 9.40256178e-01
-6.37822628e-01 -4.71399933e-01 7.98879564e-02 -1.62416607e-01
2.02319458e-01 -5.16832292e-01 -1.13234222e+00 -2.83807904e-01
4.67543483e-01 2.89473385e-01 3.55708897e-02 5.62977374e-01
-3.78019542e-01 3.81237209e-01 -5.03133833e-01 -3.76518488e-01
4.56615955e-01 -3.77356559e-01 5.99112213e-01 3.55870008e-01
4.60435033e-01 3.39995503e-01 7.01487720e-01 1.25573182e+00
2.15961114e-01 -3.96432817e-01 -7.85320044e-01 -1.94771618e-01
4.33140010e-01 9.32874978e-01 -2.94057816e-01 -4.36099231e-01
1.21925741e-01 9.82907236e-01 4.17633802e-01 7.33656466e-01
-1.06195343e+00 -1.83038469e-02 9.71616924e-01 -1.35994226e-01
1.38793178e-02 -4.03259963e-01 3.99683058e-01 -1.10160828e+00
-5.21257460e-01 -9.83713210e-01 1.39980957e-01 -8.22672009e-01
-6.90292299e-01 3.59986693e-01 1.34029210e-01 -9.74761963e-01
-9.05665576e-01 -2.54694253e-01 -4.74602729e-01 6.80011511e-01
-1.31418025e+00 -4.75743502e-01 -1.26156554e-01 8.04507375e-01
4.58915502e-01 -1.87185943e-01 7.72931933e-01 -2.65796095e-01
-4.76795733e-01 2.61307508e-01 -1.62050337e-01 -2.71497101e-01
4.29846197e-01 -1.39903426e+00 5.01936197e-01 4.33455378e-01
-3.02157670e-01 4.58932579e-01 8.33823025e-01 -7.89263904e-01
-1.50798094e+00 -9.72620487e-01 -2.87374444e-02 -6.75803840e-01
5.19990504e-01 -1.52641863e-01 -6.88022852e-01 9.44976270e-01
-2.38454193e-02 -1.02478601e-01 2.23403960e-01 -2.90001541e-01
1.02446139e-01 3.71005327e-01 -1.14684474e+00 1.11809647e+00
9.40809548e-01 -2.06680626e-01 -3.58602732e-01 2.78394639e-01
8.91893208e-01 -8.32823396e-01 -3.57151896e-01 -5.70687540e-02
4.61427629e-01 -7.45615840e-01 8.36029053e-01 -1.10059428e+00
-3.10208686e-02 -2.88978070e-01 -2.20202804e-01 -1.71991110e+00
-1.07214667e-01 -1.06678605e+00 -2.34322995e-01 3.25366855e-01
4.62709367e-01 -4.93643880e-01 1.03396511e+00 1.25700140e+00
-1.90395758e-01 -5.13819337e-01 -6.91416204e-01 -7.51943648e-01
-1.45688549e-01 -4.36121613e-01 9.05432820e-01 3.93701136e-01
3.08212340e-01 2.79704243e-01 -7.82823622e-01 4.84544426e-01
5.58579385e-01 -4.04223166e-02 1.21245015e+00 -6.43816054e-01
-6.78096890e-01 -2.12029785e-01 4.36285198e-01 -1.61026669e+00
3.35370243e-01 -5.37851989e-01 6.95655525e-01 -1.23626530e+00
-2.80923881e-02 -6.67643368e-01 -7.49764070e-02 6.03553712e-01
-2.59108897e-02 -8.54878306e-01 2.26249740e-01 6.16043471e-02
-1.00373149e+00 8.86573195e-01 1.41111517e+00 6.98230863e-02
-7.85521865e-01 4.58856404e-01 -5.01498163e-01 6.15430892e-01
9.85110581e-01 -4.13712174e-01 -9.61464763e-01 -2.15504989e-01
2.13756874e-01 7.04908550e-01 3.92157823e-01 -7.56216109e-01
1.34929910e-01 -1.00981367e+00 -9.06518847e-02 -3.50784719e-01
5.75197160e-01 -8.66515279e-01 -1.19310655e-01 7.43571520e-01
-8.01148593e-01 1.05412640e-01 -1.76983401e-02 8.96951318e-01
5.02161801e-01 -1.27222374e-01 6.01857543e-01 -3.53732646e-01
-6.37328207e-01 4.85426009e-01 -9.48282182e-01 1.36094913e-01
1.15311551e+00 2.31923565e-01 -2.39364326e-01 -1.09809005e+00
-1.00239789e+00 8.37691009e-01 4.95198846e-01 4.14205670e-01
8.47538114e-01 -1.19016385e+00 -1.51773885e-01 4.47874755e-01
-3.84334698e-02 -1.16708785e-01 7.66791552e-02 2.66051084e-01
-5.47928549e-02 5.81005923e-02 -2.06451818e-01 -4.20496106e-01
-7.00639188e-01 7.07879603e-01 5.91327131e-01 -2.09697366e-01
-2.82850206e-01 5.02531409e-01 6.21201694e-02 -8.70429337e-01
5.22038758e-01 -5.47885001e-01 -1.75088078e-01 -6.48181200e-01
1.91424936e-01 2.40596175e-01 -6.70745015e-01 -1.28222317e-01
1.48450002e-01 -1.74172357e-01 8.81901830e-02 -6.75201833e-01
9.88626480e-01 -1.34877816e-01 6.44667864e-01 2.70951241e-01
6.18524671e-01 -4.40324098e-01 -2.11569834e+00 -1.95905462e-01
-1.48770675e-01 -5.29294848e-01 -1.74351990e-01 -9.18499351e-01
-5.74101150e-01 4.25984979e-01 6.37123704e-01 1.77130520e-01
6.16394758e-01 -1.23685725e-01 7.91430175e-01 8.86461794e-01
8.26160669e-01 -1.38515091e+00 7.00575054e-01 7.46391177e-01
5.56818068e-01 -1.34190488e+00 -5.85897624e-01 4.31919664e-01
-1.25531125e+00 5.95729828e-01 1.09124637e+00 -1.20109811e-01
4.14679199e-01 3.55290025e-01 2.09447853e-02 -4.59903963e-02
-1.05663466e+00 -3.92758161e-01 -3.00571740e-01 8.44585001e-01
-5.57520807e-01 3.19802731e-01 6.48855150e-01 4.66053426e-01
-2.32826442e-01 2.12392345e-01 7.14227557e-01 1.08892322e+00
-8.91735196e-01 -1.14560223e+00 -3.75972062e-01 4.17163730e-01
1.11970119e-01 3.22952151e-01 -1.05198078e-01 4.86491084e-01
-2.62889802e-01 8.78544748e-01 -8.64465013e-02 -5.28133750e-01
4.44966763e-01 7.67227868e-03 3.46422493e-01 -7.86470711e-01
-3.15522343e-01 6.24896288e-02 3.98291685e-02 -1.08713162e+00
5.72538674e-02 -6.64204776e-01 -1.56357598e+00 -1.16293862e-01
2.97247786e-02 2.51754254e-01 5.26835382e-01 8.28272402e-01
4.61665958e-01 4.43386495e-01 9.25500870e-01 -8.28088224e-01
-1.27597475e+00 -5.12553692e-01 -2.81338364e-01 3.27137023e-01
7.10180461e-01 -7.64640570e-01 -2.53736526e-01 -1.29887953e-01] | [4.1641411781311035, 1.6643182039260864] |
506b8e3c-27cb-4020-8101-51089d1af279 | adaptive-experimentation-at-scale-bayesian | 2303.11582 | null | https://arxiv.org/abs/2303.11582v3 | https://arxiv.org/pdf/2303.11582v3.pdf | Adaptive Experimentation at Scale: A Computational Framework for Flexible Batches | Standard bandit algorithms that assume continual reallocation of measurement effort are challenging to implement due to delayed feedback and infrastructural/organizational difficulties. Motivated by practical instances involving a handful of reallocation epochs in which outcomes are measured in batches, we develop a computation-driven adaptive experimentation framework that can flexibly handle batching. Our main observation is that normal approximations, which are universal in statistical inference, can also guide the design of adaptive algorithms. By deriving a Gaussian sequential experiment, we formulate a dynamic program that can leverage prior information on average rewards. Instead of the typical theory-driven paradigm, we leverage computational tools and empirical benchmarking for algorithm development. In particular, our empirical analysis highlights a simple yet effective algorithm, Residual Horizon Optimization, which iteratively solves a planning problem using stochastic gradient descent. Our approach significantly improves statistical power over standard methods, even when compared to Bayesian bandit algorithms (e.g., Thompson sampling) that require full distributional knowledge of individual rewards. Overall, we expand the scope of adaptive experimentation to settings that are difficult for standard methods, involving a small number of reallocation epochs, low signal-to-noise ratio, and unknown reward distributions. | ['Hongseok Namkoong', 'Ethan Che'] | 2023-03-21 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 1.39747158e-01 9.52330455e-02 -6.23643756e-01 -3.06339592e-01
-1.07488656e+00 -6.63443506e-01 6.38453364e-01 2.40659099e-02
-6.36684120e-01 1.03978395e+00 2.37651974e-01 -8.36513340e-01
-7.02618718e-01 -6.71293736e-01 -8.04929376e-01 -5.39590359e-01
-8.96222070e-02 6.35873258e-01 -1.84140727e-01 1.34901792e-01
4.34383512e-01 3.80303651e-01 -9.96242106e-01 -2.72766173e-01
7.22270370e-01 1.14164698e+00 -7.33548328e-02 6.98170125e-01
8.93354565e-02 9.12329555e-01 -4.78432685e-01 -2.83036560e-01
3.66247028e-01 -3.32368612e-01 -6.09235108e-01 1.75156131e-01
-1.14072271e-01 -7.14455485e-01 -1.18054569e-01 8.58406484e-01
4.71923500e-01 4.05497223e-01 6.01085603e-01 -1.24164188e+00
-3.46466362e-01 1.18126798e+00 -6.49348438e-01 3.66116941e-01
1.13000624e-01 3.24059367e-01 1.17376566e+00 -3.22215050e-01
4.50179279e-02 1.27961814e+00 5.32982051e-01 1.11246713e-01
-1.63154888e+00 -5.49578130e-01 6.12354338e-01 -1.01781106e-02
-9.86891031e-01 -4.73709702e-01 5.11089504e-01 -4.15186137e-01
5.69037974e-01 4.03913230e-01 5.91803789e-01 1.07982779e+00
-1.51903853e-01 1.06340778e+00 1.28267503e+00 -5.91407359e-01
8.08633983e-01 -5.15072048e-02 7.16145858e-02 1.27087489e-01
3.41575265e-01 4.96471524e-01 -2.84759611e-01 -4.15973157e-01
6.36895180e-01 2.19467297e-01 1.05955034e-01 -4.31726813e-01
-1.08951545e+00 1.00845051e+00 -9.85298976e-02 -2.95243591e-01
-8.10663402e-01 5.84684193e-01 3.21928442e-01 2.49184683e-01
4.79497313e-01 4.62710679e-01 -4.66836900e-01 -6.02663159e-01
-1.18787038e+00 4.78287131e-01 9.50993359e-01 9.19522703e-01
4.26832587e-01 6.80537373e-02 -7.89095581e-01 5.09689212e-01
2.58563042e-01 6.27782404e-01 1.24196544e-01 -1.33882928e+00
5.86245477e-01 -1.84381813e-01 1.00275135e+00 -5.78344107e-01
-3.04933786e-01 -6.77870870e-01 -2.71945834e-01 -1.34574473e-01
7.36124396e-01 -7.25428939e-01 -6.13233149e-01 1.52279949e+00
3.75924826e-01 4.81021069e-02 -3.22146595e-01 7.72724688e-01
-4.28036392e-01 2.21394554e-01 -2.14933351e-01 -5.97924948e-01
8.59485447e-01 -7.36148238e-01 -5.57721794e-01 -3.56950462e-01
4.14078802e-01 -5.40956557e-01 1.22920251e+00 4.99664813e-01
-1.18491447e+00 1.65716305e-01 -6.79398775e-01 7.47034252e-01
2.11833030e-01 -4.33814794e-01 1.00555241e+00 1.08123493e+00
-6.69727147e-01 6.03261888e-01 -1.00100124e+00 3.46214287e-02
5.53144693e-01 8.04986954e-02 5.59645534e-01 -1.00359119e-01
-6.65647268e-01 5.65085888e-01 1.75693259e-01 2.48885468e-01
-1.05268371e+00 -7.03583777e-01 -3.51986021e-01 2.55137533e-01
1.02528262e+00 -5.70475996e-01 2.02245378e+00 -8.65257800e-01
-1.77484584e+00 -1.00156128e-01 1.24815680e-01 -6.86803520e-01
7.62370586e-01 -2.73689508e-01 1.46885499e-01 -1.19509608e-01
5.40029705e-02 7.19904751e-02 7.51740277e-01 -8.74849200e-01
-6.25800848e-01 -1.35425538e-01 3.13380420e-01 1.36015639e-01
-2.20314384e-01 2.36539960e-01 3.93955521e-02 -3.79660845e-01
-2.87915528e-01 -9.30057764e-01 -8.35813701e-01 -6.49000347e-01
-4.58143085e-01 1.13995463e-01 -7.00222254e-02 -1.17834106e-01
1.36076665e+00 -1.77826750e+00 -3.73288035e-01 6.39915228e-01
-2.39861533e-01 -4.30846184e-01 3.51716317e-02 6.23189092e-01
2.91323185e-01 1.13630071e-01 1.26960173e-01 -1.47071153e-01
4.95112598e-01 1.66314825e-01 -6.22179866e-01 5.49911201e-01
-2.49600798e-01 7.86600709e-01 -1.02188885e+00 -3.19797754e-01
1.09778859e-01 -4.99552310e-01 -8.12964976e-01 8.73967186e-02
-4.74634409e-01 1.45596430e-01 -5.89064896e-01 6.36870265e-01
2.89643794e-01 -4.20362562e-01 2.27023169e-01 5.19369841e-01
-2.33487263e-01 3.24180931e-01 -1.34182775e+00 1.11308110e+00
-5.90893388e-01 2.77149320e-01 2.27418542e-01 -1.39584172e+00
2.93342978e-01 -6.37314096e-02 4.74665791e-01 -4.38953787e-01
2.69876391e-01 2.56109908e-02 -6.28946722e-02 -2.93188840e-01
4.61204499e-01 -2.31426731e-01 -1.40548140e-01 8.30955327e-01
-3.24749827e-01 -3.30538660e-01 1.74893439e-01 -4.90656756e-02
1.33691275e+00 7.98505638e-03 3.11785996e-01 -1.36723459e-01
-5.57383001e-01 8.10736567e-02 5.19831717e-01 1.62709010e+00
-1.45310968e-01 1.00471340e-01 7.52068937e-01 -2.52586842e-01
-1.03096581e+00 -9.98162627e-01 -1.00704111e-01 1.42835784e+00
-2.32954025e-01 4.82738130e-02 -5.50556004e-01 -4.56309199e-01
2.83041656e-01 1.06003606e+00 -6.29327834e-01 2.06758827e-01
-6.84882998e-02 -1.15347433e+00 1.63556889e-01 5.49938917e-01
-3.10699865e-02 -5.81268191e-01 -8.69065225e-01 5.92299998e-01
1.72153026e-01 -9.04324830e-01 -5.21088183e-01 2.06505522e-01
-9.44510996e-01 -8.31674099e-01 -5.74112594e-01 2.93390632e-01
4.42606807e-01 1.79648012e-01 1.05672312e+00 -5.44153273e-01
-1.10749707e-01 1.04424500e+00 -1.74459219e-01 -8.27861190e-01
1.12017356e-02 -1.54663488e-01 3.80990691e-02 -6.15196712e-02
2.09162474e-01 -5.45342207e-01 -8.47019792e-01 3.01991284e-01
-6.65213585e-01 -2.96162784e-01 7.89550543e-01 1.17700326e+00
4.70450670e-01 -1.09569468e-01 6.85888052e-01 -9.15224254e-01
1.07404411e+00 -7.36952960e-01 -1.12703729e+00 3.15381944e-01
-8.92325640e-01 1.68751493e-01 3.91932219e-01 -7.60214746e-01
-1.04087126e+00 -3.45384985e-01 4.89249408e-01 -2.59822339e-01
6.71157017e-02 8.68301511e-01 4.92583901e-01 2.39178881e-01
8.09831321e-01 -2.75771469e-02 3.39139663e-02 -1.41162768e-01
5.44236064e-01 7.68837214e-01 1.09640926e-01 -1.16205490e+00
4.16961610e-01 3.07831734e-01 -9.30752829e-02 -4.38378632e-01
-1.14937449e+00 -2.80317992e-01 2.34298289e-01 -2.15400428e-01
-4.90497844e-03 -7.20993698e-01 -1.10853684e+00 -5.75117022e-02
-5.48226237e-01 -9.38566446e-01 -6.81949317e-01 9.30937707e-01
-1.13457215e+00 2.09434405e-02 -3.87421668e-01 -1.45149589e+00
2.28396170e-02 -9.24333870e-01 8.36891174e-01 1.14436261e-01
-2.49566704e-01 -8.46530974e-01 1.19256921e-01 3.39894444e-01
5.16120493e-01 7.10338578e-02 3.32233936e-01 -6.61842346e-01
-8.55453730e-01 -2.79365629e-01 1.08324168e-02 7.09209144e-02
-7.38597810e-02 3.95944156e-02 -7.19552875e-01 -4.52552289e-01
-5.99587336e-02 -5.11665404e-01 4.42334890e-01 9.86267805e-01
1.40742850e+00 -6.39846385e-01 -2.07288414e-01 3.50334376e-01
9.82134700e-01 2.18450949e-01 8.75557661e-02 5.73123217e-01
-1.29317597e-01 4.12575155e-01 9.41068172e-01 1.27649629e+00
3.07898194e-01 5.54506958e-01 4.71815348e-01 3.04464608e-01
8.22675824e-01 -2.78530210e-01 2.45578185e-01 7.52987638e-02
3.94424647e-02 5.32988720e-02 -7.80058205e-01 7.50970364e-01
-2.10291100e+00 -1.15747917e+00 4.82253045e-01 2.61980224e+00
9.57072735e-01 4.69387412e-01 6.31031871e-01 -1.50007740e-01
4.84253436e-01 -8.44924748e-02 -8.52377951e-01 -5.84963918e-01
4.21500951e-01 4.79657911e-02 1.12325597e+00 4.24381614e-01
-7.96071589e-01 4.83170211e-01 7.20950270e+00 9.40903485e-01
-8.15676332e-01 8.83292332e-02 8.19598138e-01 -7.82214403e-01
-4.86833543e-01 1.75047725e-01 -7.19348609e-01 6.97078526e-01
1.58407462e+00 -4.92318034e-01 1.00956023e+00 1.08307302e+00
6.61684930e-01 -3.82128865e-01 -1.26583087e+00 6.91969454e-01
-5.58161736e-01 -1.32878160e+00 -7.53235102e-01 2.14175999e-01
8.55324686e-01 2.03390419e-01 2.53043115e-01 4.72335368e-01
1.35723543e+00 -1.00974131e+00 9.31117535e-01 6.88483417e-01
4.40184563e-01 -8.31190705e-01 5.76700509e-01 6.41248167e-01
-3.49217892e-01 -7.51477838e-01 -4.00864124e-01 -5.03540754e-01
1.90637454e-01 1.03491676e+00 -8.84477735e-01 2.12633356e-01
6.26468599e-01 -6.79314360e-02 2.08803862e-01 1.40788317e+00
-1.67322233e-01 1.11639440e+00 -7.90436566e-01 -4.63763863e-01
4.49328005e-01 -2.01202750e-01 2.75755465e-01 1.09499454e+00
4.35303748e-01 2.15544403e-02 4.14717376e-01 8.35352957e-01
1.92826271e-01 7.34599456e-02 -5.59436321e-01 -1.09043650e-01
9.11855221e-01 9.73188877e-01 -6.71583056e-01 -1.88129619e-01
-3.41823876e-01 6.95993528e-02 2.63059348e-01 6.31522477e-01
-8.14715385e-01 -2.96608955e-01 4.17640299e-01 -1.41356692e-01
4.72771466e-01 -3.51158261e-01 -3.87240380e-01 -7.81325996e-01
-6.83366228e-03 -8.76194477e-01 3.59483838e-01 -4.86384898e-01
-1.17790878e+00 -2.81807959e-01 5.38023829e-01 -7.68954635e-01
-8.40447009e-01 -3.35104018e-01 -5.63937545e-01 6.23929918e-01
-1.20323932e+00 -7.26252139e-01 2.39184812e-01 3.32948536e-01
3.92510235e-01 2.34520704e-01 3.88112992e-01 -1.94089919e-01
-6.74071610e-01 6.17568076e-01 6.71160460e-01 -2.39656001e-01
3.88419896e-01 -1.26897132e+00 9.35320742e-03 7.72759020e-01
3.75865377e-03 6.95732236e-01 1.09259319e+00 -4.30929065e-01
-1.39338863e+00 -7.69042015e-01 -1.06946900e-02 -3.20646316e-01
1.18942857e+00 -2.72522479e-01 -8.46899748e-02 9.86836255e-01
-1.00294985e-01 4.42952365e-02 7.17893839e-01 7.33747840e-01
1.05146028e-03 -1.81524530e-01 -9.55816746e-01 8.42055857e-01
9.22563195e-01 -2.65156388e-01 -3.31212580e-01 6.03216290e-01
6.19592845e-01 -5.05365789e-01 -8.34809959e-01 5.10114394e-02
6.40172362e-01 -7.41068602e-01 6.21210754e-01 -8.84859741e-01
-2.23141015e-02 1.48788393e-01 -2.82124162e-01 -1.48052871e+00
-1.31214276e-01 -1.28436112e+00 -3.21365118e-01 1.03660309e+00
4.70856100e-01 -7.48546541e-01 7.53772736e-01 1.15865231e+00
1.16321467e-01 -7.68191814e-01 -8.54093373e-01 -1.05773807e+00
4.56809513e-02 -9.36662316e-01 7.16592908e-01 6.44795477e-01
1.36431709e-01 -1.05687499e-01 -3.81180257e-01 1.91141840e-03
9.65883970e-01 3.78477842e-01 8.37513924e-01 -9.09193337e-01
-8.40485632e-01 -4.99942124e-01 1.20942958e-01 -1.36162913e+00
6.94838166e-02 -1.44870028e-01 2.91347712e-01 -1.09993756e+00
2.74827033e-01 -7.35564888e-01 -5.61366022e-01 4.00354594e-01
-5.31053171e-03 -3.65160406e-01 9.37369931e-03 -1.32608771e-01
-8.64834666e-01 5.73246181e-01 9.10237730e-01 -1.46233067e-02
-4.45463181e-01 6.87640250e-01 -1.14890182e+00 5.49428284e-01
7.98884690e-01 -5.75169802e-01 -5.31118631e-01 -1.64954677e-01
6.08564913e-01 4.90100801e-01 2.86515951e-01 -4.41176921e-01
2.99157232e-01 -1.00546741e+00 4.34699133e-02 -3.32430214e-01
-1.23180393e-02 -6.77316964e-01 6.65419325e-02 1.43296078e-01
-7.11117804e-01 -1.44963473e-01 -1.93499885e-02 8.98521245e-01
1.65987566e-01 -4.09970433e-01 3.41307312e-01 -1.82316989e-01
6.48819283e-02 2.19154522e-01 -5.39770424e-01 3.26857775e-01
9.18220937e-01 -1.22911833e-01 -2.85616606e-01 -8.94363284e-01
-6.40509486e-01 5.88410318e-01 1.75182790e-01 -1.74227640e-01
1.96326599e-01 -9.80691969e-01 -5.90224802e-01 -4.77799982e-01
-1.70486867e-01 -1.02340570e-03 1.31350875e-01 1.23090374e+00
7.69995078e-02 4.77141201e-01 2.59365886e-01 -4.75240737e-01
-5.65199256e-01 6.48317635e-01 2.60488927e-01 -4.04148221e-01
-2.66243070e-01 6.85412467e-01 -9.15586650e-02 -9.10769179e-02
4.21176016e-01 -5.90790331e-01 4.02947068e-01 1.31426066e-01
5.15521586e-01 5.73344469e-01 -4.93236892e-02 3.86460185e-01
3.92860323e-02 -1.94475576e-01 -1.16378799e-01 -5.94681621e-01
1.47817576e+00 -3.12370449e-01 2.14187503e-01 6.53822958e-01
3.32656205e-01 1.97482839e-01 -1.77014399e+00 -4.44272250e-01
3.41061801e-01 -6.99393570e-01 3.29683006e-01 -7.86808312e-01
-5.81217468e-01 4.39443886e-01 2.79678702e-01 6.03323460e-01
8.72221529e-01 -2.31954545e-01 5.52858971e-02 6.76805556e-01
6.27141416e-01 -1.48461580e+00 -1.32875457e-01 2.63962418e-01
4.76061612e-01 -1.04624724e+00 1.31988138e-01 3.92336577e-01
-4.58692133e-01 8.19590926e-01 1.98597923e-01 8.35116506e-02
4.55389649e-01 3.72899324e-01 -3.84442776e-01 1.28189266e-01
-1.01422596e+00 -1.60481215e-01 -1.57799453e-01 3.55111539e-01
9.05041322e-02 3.16775769e-01 -1.83980927e-01 5.89474320e-01
-3.02437514e-01 2.53893405e-01 7.65127778e-01 1.01701999e+00
-5.07305086e-01 -8.41728687e-01 -5.48190892e-01 8.42525661e-01
-9.29317296e-01 -1.09918371e-01 1.70279726e-01 7.78387308e-01
-8.33138883e-01 1.26955581e+00 1.81660771e-01 1.84575155e-01
2.21205637e-01 -1.53310910e-01 4.90374923e-01 -4.99336630e-01
-9.01249871e-02 3.02245378e-01 3.34125727e-01 -8.31216514e-01
-3.10059607e-01 -9.88702297e-01 -6.03888988e-01 -3.94260675e-01
-3.96275759e-01 3.74854714e-01 7.04329550e-01 1.09792650e+00
2.25694135e-01 4.52226877e-01 9.02678013e-01 -9.35774326e-01
-1.57829106e+00 -9.70405698e-01 -7.02197492e-01 -1.11236304e-01
4.07930017e-01 -8.31550598e-01 -4.32084411e-01 -5.57039440e-01] | [4.5119805335998535, 3.2117879390716553] |
c5187280-f5e9-42d3-8d69-6c24906339ee | a-one-class-classifier-based-framework-using | 1612.01349 | null | http://arxiv.org/abs/1612.01349v1 | http://arxiv.org/pdf/1612.01349v1.pdf | A One class Classifier based Framework using SVDD : Application to an Imbalanced Geological Dataset | Evaluation of hydrocarbon reservoir requires classification of petrophysical
properties from available dataset. However, characterization of reservoir
attributes is difficult due to the nonlinear and heterogeneous nature of the
subsurface physical properties. In this context, present study proposes a
generalized one class classification framework based on Support Vector Data
Description (SVDD) to classify a reservoir characteristic water saturation into
two classes (Class high and Class low) from four logs namely gamma ray, neutron
porosity, bulk density, and P sonic using an imbalanced dataset. A comparison
is carried out among proposed framework and different supervised classification
algorithms in terms of g metric means and execution time. Experimental results
show that proposed framework has outperformed other classifiers in terms of
these performance evaluators. It is envisaged that the classification analysis
performed in this study will be useful in further reservoir modeling. | ['Mamata Jenamani', 'Aurobinda Routray', 'Soumi Chaki', 'William K. Mohanty', 'Akhilesh Kumar Verma'] | 2016-12-02 | null | null | null | null | ['one-class-classifier'] | ['methodology'] | [-3.53383794e-02 -3.04768890e-01 1.00887856e-02 -3.94956946e-01
-2.37899736e-01 -5.98791599e-01 9.02895808e-01 9.36295509e-01
-7.87583590e-02 1.04598868e+00 3.93484831e-01 -3.35637003e-01
-4.31346804e-01 -1.25692773e+00 -1.77252337e-01 -9.73159254e-01
-6.64549530e-01 7.16842175e-01 3.33525807e-01 -4.41159129e-01
8.89793575e-01 8.16961944e-01 -1.81366909e+00 1.34668559e-01
7.46061027e-01 9.69999552e-01 2.11520448e-01 4.88945574e-01
2.36879680e-02 6.94207013e-01 -1.99847326e-01 5.29083669e-01
4.47216660e-01 -2.17947498e-01 -5.05697012e-01 -3.49964648e-01
-9.02393740e-03 -3.86110693e-01 2.26528853e-01 9.16671395e-01
4.13472921e-01 2.53617227e-01 1.47534823e+00 -1.10986555e+00
-2.51984090e-01 5.12718260e-01 -3.92629236e-01 6.64138496e-01
2.62905866e-01 3.80309410e-02 9.34925139e-01 -1.05293465e+00
2.25647867e-01 9.92415249e-01 4.60351825e-01 -1.59607038e-01
-9.54546690e-01 -7.65283227e-01 -3.95514876e-01 1.58412650e-01
-1.12964320e+00 -4.22674209e-01 9.41238940e-01 -8.82359922e-01
1.16685271e+00 4.44716841e-01 7.34514892e-01 1.81064069e-01
2.54000306e-01 2.04056591e-01 1.88594675e+00 -3.38643998e-01
5.97035408e-01 3.70936722e-01 5.47586918e-01 3.77623558e-01
8.67724955e-01 3.21126163e-01 -4.98841614e-01 -2.87143767e-01
2.28254199e-01 -3.23342204e-01 -1.82248019e-02 1.35532459e-02
-6.56356573e-01 9.78968501e-01 6.93898648e-02 4.96731222e-01
-5.45673966e-01 -2.07232207e-01 6.37594700e-01 4.11979169e-01
4.42188919e-01 3.11141253e-01 -2.42380261e-01 -1.20244861e-01
-1.18936706e+00 5.41246772e-01 1.00650179e+00 6.60145819e-01
6.71836436e-01 5.47242820e-01 1.53987274e-01 3.31881076e-01
8.78321528e-01 7.20268488e-01 7.36535072e-01 -9.58676189e-02
3.77267629e-01 5.58692932e-01 2.44161114e-01 -9.95015919e-01
-3.55407953e-01 -9.91957262e-02 -6.60385966e-01 5.66993535e-01
1.00232676e-01 2.20948309e-01 -9.57061648e-01 7.11327374e-01
-1.80416405e-02 1.31785735e-01 6.62029803e-01 7.36014962e-01
1.31323957e+00 6.63147330e-01 2.76533514e-01 3.59659083e-02
1.21616805e+00 -1.37348831e-01 -4.67871249e-01 2.95543551e-01
2.36480474e-01 -6.35747850e-01 6.55756593e-01 3.27603191e-01
-7.08346069e-01 -2.22912073e-01 -1.49597597e+00 5.50230682e-01
-6.88737452e-01 -2.58811116e-01 7.45947003e-01 9.83947098e-01
-5.88813901e-01 7.12375998e-01 -1.15513945e+00 -2.50639290e-01
1.22683525e-01 4.32489336e-01 -3.51894230e-01 3.47528130e-01
-1.16925347e+00 8.38145316e-01 7.16613472e-01 2.52696723e-01
-1.06904709e+00 -4.73833025e-01 -7.26128161e-01 -1.09749362e-01
-5.33340991e-01 1.98880658e-01 7.06943989e-01 5.25870249e-02
-1.20171285e+00 4.70087022e-01 1.99941359e-02 -3.49057168e-01
5.39150655e-01 4.39087808e-01 -2.90260702e-01 -2.94531006e-02
-2.24705189e-01 -3.43366116e-01 3.62218708e-01 -1.36749268e+00
-5.86729944e-01 -6.68591321e-01 -4.15161520e-01 1.93085074e-01
-2.00299203e-01 -2.42995515e-01 5.95077038e-01 -1.93020776e-01
6.88144624e-01 -6.04080081e-01 6.46400303e-02 -7.48301566e-01
-3.11881572e-01 -1.56714916e-01 8.95720661e-01 -8.42953503e-01
8.97657633e-01 -1.79713869e+00 -1.70731887e-01 8.62417877e-01
-8.31435621e-02 -7.63151050e-02 5.92440188e-01 7.02189147e-01
5.33385389e-02 1.26808032e-01 -4.06922042e-01 1.00518838e-01
2.35368609e-02 2.44308889e-01 -1.22814626e-01 8.80790234e-01
5.60208075e-02 -8.40813387e-03 -8.37250412e-01 -5.11660874e-01
5.79568446e-01 2.45243117e-01 -4.20327261e-02 2.06568211e-01
1.71246573e-01 4.92794603e-01 -7.10697949e-01 1.29217315e+00
1.14553583e+00 4.27238464e-01 2.19950154e-02 -3.54359418e-01
-4.12843198e-01 -5.72831556e-02 -1.44195724e+00 5.87256193e-01
-3.62059295e-01 4.13727641e-01 -2.06729069e-01 -1.48494387e+00
1.57139504e+00 3.60799164e-01 4.35720831e-01 -5.23300171e-01
1.76786959e-01 7.89985955e-01 2.15407118e-01 -8.11452985e-01
7.59986103e-01 -4.88797933e-01 3.73663664e-01 1.26713604e-01
-2.02104777e-01 -2.20755264e-01 3.58122915e-01 -2.93396175e-01
7.24112511e-01 -1.92556992e-01 2.56093383e-01 -9.49131787e-01
7.73790896e-01 1.61116049e-01 2.44890302e-01 4.71498519e-01
-3.98206890e-01 3.28586139e-02 4.89459813e-01 -1.41311005e-01
-1.40470266e+00 -1.01602507e+00 -8.84323657e-01 5.04551053e-01
1.73719600e-01 2.47771323e-01 1.69314280e-01 1.89117074e-01
5.43559611e-01 3.74299854e-01 -5.87380767e-01 2.86781132e-01
-5.21477818e-01 -1.26610124e+00 5.17125845e-01 5.30284047e-01
4.10511434e-01 -9.47629929e-01 -7.19870269e-01 1.06218860e-01
8.09405923e-01 -5.93837678e-01 8.45149398e-01 4.09538895e-01
-1.39529622e+00 -1.05987167e+00 -3.95813733e-01 -6.57618403e-01
4.27916825e-01 -3.25659394e-01 9.02264714e-01 1.97416723e-01
-1.05533980e-01 -3.57541978e-01 -5.82492769e-01 -5.23125827e-01
-6.25098228e-01 -5.36274016e-02 1.64771676e-01 -4.05660570e-01
6.07100368e-01 -6.07322574e-01 -7.20148325e-01 -9.96771678e-02
-7.19544291e-01 -4.41575378e-01 4.68734354e-01 4.01317537e-01
2.76903868e-01 6.46839261e-01 7.41617441e-01 -9.35917318e-01
7.90035367e-01 -1.25833654e+00 -6.54772699e-01 5.46926111e-02
-9.75604475e-01 1.62476599e-01 4.81750816e-01 2.99732089e-01
-8.98904860e-01 -3.56738389e-01 8.31202120e-02 3.92297864e-01
-1.71625577e-02 9.00479138e-01 1.56586692e-01 -3.27898294e-01
5.53206027e-01 5.11050284e-01 7.16042519e-02 -6.39130950e-01
-5.00652730e-01 1.18066335e+00 4.23512965e-01 -9.07156408e-01
8.42875183e-01 4.24981117e-01 4.65100616e-01 -1.29693544e+00
1.04527153e-01 -6.86700583e-01 -2.55494386e-01 -3.64578784e-01
5.14384210e-01 -1.09344506e+00 -7.54929602e-01 5.21763086e-01
-4.07680601e-01 2.06827000e-01 2.38251492e-01 7.42020965e-01
-1.24859586e-01 2.42176294e-01 -2.81034529e-01 -1.43186235e+00
-6.43242896e-01 -1.24588871e+00 6.22303426e-01 2.40360677e-01
1.27683550e-01 -9.84205723e-01 4.44081165e-02 2.24974483e-01
8.89861763e-01 8.00140202e-01 8.57975900e-01 -9.70609307e-01
-3.71251792e-01 -6.02621138e-01 -2.01665223e-01 3.56364012e-01
1.35654449e-01 2.03561753e-01 -8.79678965e-01 -4.24310088e-01
-7.31222779e-02 -2.29806423e-01 9.87177610e-01 1.94704503e-01
5.55898190e-01 -3.50109884e-03 -3.21891569e-02 2.70734012e-01
2.21081805e+00 4.47377056e-01 4.40656602e-01 7.81569302e-01
2.75087476e-01 3.69637817e-01 6.67285204e-01 8.82355869e-01
-2.69830897e-02 1.48796150e-03 6.24396503e-01 4.14437562e-01
3.22637945e-01 9.45968479e-02 1.78853184e-01 6.48038566e-01
-5.22847116e-01 1.01478897e-01 -1.29489291e+00 6.76454544e-01
-1.35505641e+00 -1.01031733e+00 -3.20533484e-01 2.07443285e+00
5.16297460e-01 1.62057281e-01 1.65610135e-01 8.33348095e-01
4.52227473e-01 1.69963598e-01 -2.43621781e-01 -5.50382853e-01
-1.82331949e-01 4.30324972e-01 1.04978025e+00 5.74556410e-01
-8.48566413e-01 3.11702341e-01 5.54711056e+00 2.87055403e-01
-1.24226415e+00 -1.50651798e-01 3.83874118e-01 3.96683782e-01
-5.35360754e-01 1.66000336e-01 -8.48794818e-01 7.30211318e-01
9.05963004e-01 -3.44257116e-01 7.94996098e-02 6.15363300e-01
4.72065598e-01 -5.12598693e-01 -7.95206249e-01 5.36302626e-01
-1.03159823e-01 -1.25590849e+00 -1.95512339e-01 2.35997308e-02
4.50921059e-01 2.73186177e-01 -1.31009564e-01 1.69937506e-01
6.82915896e-02 -1.16704845e+00 4.49772388e-01 6.29235268e-01
4.96960104e-01 -5.93257427e-01 1.01088905e+00 6.50333762e-02
-1.34497809e+00 -1.92598149e-01 -1.97248533e-01 -3.40810835e-01
-2.07690775e-01 4.16367322e-01 -1.18060064e+00 5.03272057e-01
5.30679882e-01 6.13122642e-01 -2.27767125e-01 1.17737472e+00
4.55660224e-01 7.25243449e-01 -6.84997380e-01 -2.84565032e-01
2.03916743e-01 -5.06530762e-01 3.49995315e-01 1.12122631e+00
4.77336466e-01 5.24615310e-02 3.19767259e-02 4.18661714e-01
4.62522805e-01 5.25219083e-01 -7.28575230e-01 -1.65907234e-01
9.09816980e-01 8.00264835e-01 -9.03952777e-01 -5.50713420e-01
-1.98985606e-01 -1.42722875e-01 -3.43340874e-01 -1.10535137e-01
-2.93136269e-01 -4.52032298e-01 3.20244700e-01 4.53712851e-01
2.91352440e-02 -3.14269960e-01 -5.65991998e-01 -8.47155750e-01
5.03566535e-03 -2.70895422e-01 2.54218400e-01 -7.34684765e-02
-1.08836782e+00 5.13925314e-01 5.14238536e-01 -1.27231145e+00
3.70205492e-02 -6.92408502e-01 -6.99653149e-01 1.22445047e+00
-1.94772792e+00 -9.83912706e-01 -7.73661673e-01 1.79349810e-01
3.19108993e-01 -7.13666439e-01 5.30274689e-01 2.78488249e-01
-2.83083588e-01 1.79525286e-01 7.23650455e-01 -4.08291891e-02
1.37689978e-01 -1.41058969e+00 -3.19303781e-01 5.07843137e-01
-8.90878022e-01 2.89483964e-01 1.37872148e+00 -9.50969934e-01
-1.42753589e+00 -7.20562220e-01 6.72402680e-01 1.47489309e-01
8.11264575e-01 2.18255296e-01 -8.04106355e-01 2.67517179e-01
-1.44758910e-01 2.00139746e-01 7.26633906e-01 -2.47289404e-01
1.83522776e-01 -9.75911692e-03 -1.66920078e+00 -3.27283680e-01
2.35203616e-02 -2.95635372e-01 -5.99134326e-01 4.55394417e-01
-3.78071904e-01 -2.62285352e-01 -1.61103237e+00 5.42857409e-01
7.39998639e-01 -9.85815525e-01 7.25151837e-01 -6.48990452e-01
4.84517694e-01 -4.27175075e-01 -6.73238516e-01 -6.87391400e-01
1.45253539e-01 2.19730213e-01 1.09192722e-01 1.12629807e+00
3.88758063e-01 -5.40217340e-01 1.28210735e+00 6.25441670e-01
-4.77962978e-02 -7.11874902e-01 -9.84788418e-01 -7.87003815e-01
3.29903215e-01 -9.27160755e-02 7.86556244e-01 1.04635608e+00
7.30295554e-02 -1.54963672e-01 -1.85737073e-01 4.22707647e-01
9.29004431e-01 2.91872293e-01 4.28589284e-01 -1.55888176e+00
-4.81621772e-02 -4.59401682e-02 -1.04993081e+00 2.26168498e-01
-5.28614670e-02 -1.09662640e+00 -2.74125934e-01 -1.41073549e+00
2.51087308e-01 -1.08959043e+00 -4.72918957e-01 1.87281847e-01
2.31934905e-01 2.71328986e-01 -1.06146686e-01 6.54408455e-01
5.28077364e-01 5.35126984e-01 7.39841461e-01 -2.19954595e-01
-2.84821868e-01 3.44587974e-02 -7.02545345e-02 4.52354580e-01
1.11889899e+00 -3.30636680e-01 -3.03891033e-01 -1.76998274e-03
2.26017728e-01 2.71156162e-01 4.62781414e-02 -1.25657320e+00
-1.05516046e-01 -4.55440015e-01 3.60125780e-01 -7.21314967e-01
1.65733978e-01 -9.48630452e-01 3.17882776e-01 7.70730734e-01
-4.49700542e-02 -1.14948805e-02 -2.13646628e-02 5.65236151e-01
-3.70743901e-01 -5.79706132e-01 7.56865740e-01 -3.50461394e-01
-9.77992654e-01 -8.67732465e-02 -2.56763190e-01 -4.24372256e-01
1.26962924e+00 -8.60122681e-01 -4.46250170e-01 1.86537027e-01
-8.14779758e-01 1.38797566e-01 3.73217821e-01 -7.53908753e-02
5.47378004e-01 -1.06689847e+00 -1.00552809e+00 1.44156486e-01
1.94382578e-01 -8.38913992e-02 1.01095930e-01 7.05360889e-01
-1.34663606e+00 3.20963532e-01 -7.62973905e-01 -4.96658146e-01
-1.26785469e+00 -2.10868716e-01 3.49755108e-01 -3.47090513e-02
-5.30257165e-01 4.78934437e-01 -8.00545037e-01 -3.23527418e-02
-2.44307831e-01 -2.75191844e-01 -8.89785290e-01 3.74898463e-01
1.69869587e-02 7.04134822e-01 4.42844898e-01 -9.01364267e-01
-5.05128145e-01 4.07983154e-01 2.02111706e-01 -1.65456347e-02
1.77765989e+00 4.81085330e-01 -3.92859191e-01 4.63061303e-01
1.19372594e+00 -2.69364327e-01 -4.95386630e-01 5.41741885e-02
3.09125394e-01 -7.27112770e-01 2.48341978e-01 -4.27708298e-01
-9.47407663e-01 4.07760799e-01 1.12435985e+00 2.33726174e-01
7.97008157e-01 -3.43685210e-01 3.83567095e-01 5.53824365e-01
4.60150577e-02 -1.16076326e+00 -3.43665600e-01 3.32929909e-01
8.08216989e-01 -1.64955175e+00 5.09720206e-01 -1.49274647e-01
-3.46999526e-01 1.36704218e+00 5.00020325e-01 -4.59072262e-01
1.36496854e+00 4.19308692e-01 -2.44112611e-01 -6.55104160e-01
-3.18405569e-01 1.16139010e-01 -3.36997151e-01 4.52760398e-01
8.08243155e-01 2.83949435e-01 -8.50532651e-01 1.89344957e-01
-5.11731565e-01 -4.55221720e-02 7.87745655e-01 1.42818701e+00
-8.98988128e-01 -8.25398922e-01 -5.80272913e-01 1.22442174e+00
-2.24325702e-01 -1.43754082e-02 3.45585495e-01 8.01101327e-01
-8.79343376e-02 8.05678368e-01 1.98639929e-01 -3.11364055e-01
2.33994618e-01 -1.03093453e-01 8.25019032e-02 -4.45525646e-01
-3.87945801e-01 -4.72875893e-01 4.04351026e-01 3.14847261e-01
-6.52040958e-01 -8.46103609e-01 -1.20267272e+00 -3.46397966e-01
-3.46674025e-01 6.37192905e-01 1.25882387e+00 8.24820459e-01
-5.35195172e-01 1.90072030e-01 8.68232250e-01 -9.06219363e-01
-9.92506266e-01 -1.39881504e+00 -1.29457521e+00 4.07503009e-01
4.81752038e-01 -1.14829540e+00 -8.61964822e-01 -1.28065363e-01] | [6.439683437347412, 3.04349946975708] |
fd56a9b9-7cbe-4a61-89e7-6d7e1022fa3a | data-free-point-cloud-network-for-3d-face | 1911.04731 | null | https://arxiv.org/abs/1911.04731v1 | https://arxiv.org/pdf/1911.04731v1.pdf | Data-Free Point Cloud Network for 3D Face Recognition | Point clouds-based Networks have achieved great attention in 3D object classification, segmentation and indoor scene semantic parsing. In terms of face recognition, 3D face recognition method which directly consume point clouds as input is still under study. Two main factors account for this: One is how to get discriminative face representations from 3D point clouds using deep network; the other is the lack of large 3D training dataset. To address these problems, a data-free 3D face recognition method is proposed only using synthesized unreal data from statistical 3D Morphable Model to train a deep point cloud network. To ease the inconsistent distribution between model data and real faces, different point sampling methods are used in train and test phase. In this paper, we propose a curvature-aware point sampling(CPS) strategy replacing the original furthest point sampling(FPS) to hierarchically down-sample feature-sensitive points which are crucial to pass and aggregate features deeply. A PointNet++ like Network is used to extract face features directly from point clouds. The experimental results show that the network trained on generated data generalizes well for real 3D faces. Fine tuning on a small part of FRGCv2.0 and Bosphorus, which include real faces in different poses and expressions, further improves recognition accuracy. | ['Yu', 'Yi', 'Da', 'Ziyu', 'Feipeng', 'Zhang'] | 2019-11-12 | null | null | null | null | ['3d-object-classification'] | ['computer-vision'] | [-1.78756922e-01 -2.33386531e-02 1.91232979e-01 -9.23350096e-01
-3.73476595e-01 -3.80630732e-01 4.62473303e-01 -6.54267907e-01
-5.60392775e-02 1.77855536e-01 -5.97934008e-01 -7.64993131e-02
7.41435140e-02 -9.85278606e-01 -9.30547237e-01 -4.90420461e-01
2.71504316e-02 8.38086188e-01 1.55744433e-01 -2.85312742e-01
2.87625700e-01 1.46816134e+00 -1.99186730e+00 2.43434876e-01
5.83925068e-01 1.28151619e+00 5.15357889e-02 1.56674907e-01
-8.86678696e-01 -1.96939334e-01 -7.10663021e-01 -4.60083902e-01
6.91148400e-01 2.95575634e-02 -3.77892673e-01 2.78617293e-01
7.89707422e-01 -3.75545949e-01 1.65178806e-01 9.19822872e-01
5.30490100e-01 -1.97992355e-01 7.94236004e-01 -1.46850026e+00
-5.08576274e-01 -5.22078127e-02 -7.84015894e-01 -3.33408803e-01
3.52077276e-01 1.43795814e-02 1.82079509e-01 -1.21850824e+00
4.31312263e-01 1.80548215e+00 9.66286719e-01 1.07183897e+00
-8.10464919e-01 -9.68793690e-01 -9.45089757e-03 -1.25963673e-01
-1.59717035e+00 -5.70244908e-01 1.28721249e+00 -3.01558673e-01
9.30945337e-01 2.69282728e-01 6.89457238e-01 1.08700895e+00
-1.39069572e-01 6.49566293e-01 9.11210656e-01 -2.67067403e-01
2.62872905e-01 1.57306597e-01 -2.00379699e-01 7.44757652e-01
-1.80576444e-02 7.83585757e-02 -2.50780970e-01 -7.71958902e-02
1.11834133e+00 3.44963700e-01 4.89440095e-03 -6.23749137e-01
-3.64253491e-01 7.69307077e-01 6.88217044e-01 8.70019943e-02
-1.72307357e-01 -2.71486561e-03 1.20954983e-01 1.68123975e-01
6.24351561e-01 2.38104090e-02 -7.58317351e-01 1.89928353e-01
-1.00572407e+00 3.80927980e-01 6.22273207e-01 1.12427628e+00
1.12095249e+00 1.35110065e-01 3.16572160e-01 9.26014125e-01
5.92175901e-01 7.92769492e-01 2.18402669e-01 -7.63633788e-01
4.29943353e-01 1.17164004e+00 -2.39006802e-01 -1.16992950e+00
-2.81478822e-01 -1.24914952e-01 -6.13953173e-01 5.53300738e-01
2.57350355e-01 2.40068361e-01 -1.37238729e+00 1.30554783e+00
7.50412405e-01 2.89078444e-01 -1.17915936e-01 1.04230738e+00
1.27501237e+00 6.02320433e-01 -2.38098651e-01 2.97495961e-01
1.23072052e+00 -3.67639780e-01 -9.38908234e-02 -7.00254366e-02
4.56019849e-01 -4.85939622e-01 9.54739153e-01 9.23410431e-02
-8.79884243e-01 -7.50113785e-01 -9.07657743e-01 -8.60150233e-02
-6.50405705e-01 2.22772360e-01 5.65163493e-01 1.00341904e+00
-1.07085359e+00 5.55762768e-01 -8.53338718e-01 -3.47454041e-01
1.19264996e+00 8.42672348e-01 -7.36276209e-01 -1.42052785e-01
-7.32221723e-01 5.60190618e-01 1.48686618e-01 3.63107204e-01
-6.98271275e-01 -6.44495666e-01 -9.96097624e-01 -9.78611633e-02
6.66273534e-02 -4.01612967e-01 7.53654718e-01 -1.04556930e+00
-1.78771865e+00 1.23449922e+00 -6.88829198e-02 5.24985790e-02
4.11010802e-01 2.59155799e-02 -9.34065357e-02 1.33861512e-01
6.88199550e-02 9.95773137e-01 1.15737975e+00 -1.44568086e+00
-2.09788755e-01 -1.04096389e+00 -9.77168083e-02 3.36198881e-02
-5.66577546e-05 -3.39115672e-02 -5.35523713e-01 -8.71547535e-02
4.64326233e-01 -7.83338964e-01 3.89195718e-02 3.61957192e-01
-3.47849458e-01 -5.69565475e-01 1.20596135e+00 -2.82970577e-01
6.62451237e-02 -2.18261909e+00 -1.32719263e-01 3.64926100e-01
-1.02353789e-01 4.92146939e-01 -2.68648863e-01 -4.43214625e-02
-2.00863898e-01 1.97064012e-01 -1.36136174e-01 -5.66696763e-01
1.36763193e-02 1.88355073e-01 -1.49370745e-01 4.98058438e-01
8.07655096e-01 8.93687606e-01 -2.99786508e-01 -5.11611164e-01
4.49830383e-01 7.39796460e-01 -7.14023352e-01 1.30902886e-01
-3.53153288e-01 4.25372213e-01 -6.46041930e-01 1.11114895e+00
1.50660062e+00 1.90685049e-01 -4.49034601e-01 -2.10880294e-01
-2.85387188e-02 -9.14407596e-02 -1.24283099e+00 1.97660601e+00
-3.88473272e-01 8.20534155e-02 2.81529516e-01 -8.96479487e-01
1.47015774e+00 9.69934538e-02 4.00647253e-01 -4.54567850e-01
2.90036619e-01 2.57169187e-01 -4.92152542e-01 -4.43616867e-01
-3.80268134e-02 -1.10064358e-01 1.91753671e-01 -1.34564683e-01
2.18969911e-01 -6.21254981e-01 -5.87251186e-01 -4.31009173e-01
5.62076509e-01 4.77983683e-01 -3.05571407e-01 -2.05656469e-01
6.51109040e-01 6.80649420e-03 6.66762650e-01 2.38266066e-01
-8.81241709e-02 8.99490654e-01 4.81713623e-01 -6.97183251e-01
-7.67764509e-01 -9.97659147e-01 -3.40475917e-01 5.68924010e-01
1.91966876e-01 -1.14753649e-01 -8.34616363e-01 -9.90177214e-01
2.52062142e-01 3.76530737e-01 -5.99898040e-01 1.57771960e-01
-6.35613203e-01 -4.33674544e-01 4.52149361e-01 4.03543055e-01
7.02961206e-01 -8.93558443e-01 -5.62494040e-01 -2.42293745e-01
5.60046971e-01 -9.54767942e-01 -2.83928844e-03 2.34483853e-02
-9.31927085e-01 -1.04040515e+00 -6.25079870e-01 -7.60694087e-01
9.28451240e-01 2.84731060e-01 8.62517774e-01 2.28668332e-01
-2.68939227e-01 2.54291743e-01 -2.95409441e-01 -7.46804893e-01
2.98466347e-02 -5.23710717e-03 1.37821421e-01 3.52985971e-02
7.95908451e-01 -6.15445316e-01 -4.14405763e-01 4.80253309e-01
-6.20131195e-01 -1.45523861e-01 4.63999629e-01 6.29312932e-01
5.77133417e-01 -1.33331157e-02 3.54792327e-01 -4.88343447e-01
5.75556830e-02 -2.38590583e-01 -8.35002124e-01 -6.76148161e-02
9.68582407e-02 -1.94952801e-01 5.00882685e-01 -3.21012467e-01
-8.02874327e-01 4.56041306e-01 -5.73650658e-01 -1.01561010e+00
-7.35436201e-01 -2.45416284e-01 -8.14594746e-01 -4.68780577e-01
5.10033190e-01 1.90060079e-01 3.43266129e-01 -5.76520562e-01
1.67964071e-01 7.33508945e-01 2.46425848e-02 -5.91940701e-01
1.00926054e+00 6.73649609e-01 3.34360003e-01 -1.22670019e+00
-3.29157859e-01 -5.51709011e-02 -9.00531650e-01 -1.74681395e-01
9.69582021e-01 -8.47333789e-01 -9.27418232e-01 4.59683299e-01
-1.52551436e+00 -3.52206528e-02 -1.26631230e-01 2.05062166e-01
-3.21918637e-01 3.75035144e-02 -2.25647375e-01 -9.43621635e-01
-1.91858307e-01 -1.38824558e+00 1.76204276e+00 3.79932433e-01
3.02914679e-01 -4.75057036e-01 -3.48677307e-01 2.92541534e-01
1.26622573e-01 4.07358676e-01 6.25308335e-01 -5.58255613e-01
-7.71511972e-01 -3.37470978e-01 -3.55775088e-01 4.77987409e-01
2.12694749e-01 2.08800584e-01 -1.32720661e+00 -4.30041403e-02
2.16896385e-01 -3.67679179e-01 4.92633462e-01 2.05960587e-01
1.48978448e+00 -2.46477015e-02 -4.74490196e-01 1.08613336e+00
1.28651726e+00 2.19593540e-01 5.81898868e-01 -7.74637833e-02
9.49633598e-01 8.39464426e-01 2.52263039e-01 2.73099989e-02
2.38663077e-01 7.06380129e-01 7.14313745e-01 6.20510876e-02
3.08702029e-02 -4.13238853e-01 1.63525045e-01 3.94165754e-01
-7.69803151e-02 2.21180215e-01 -9.53405559e-01 1.10379294e-01
-1.13375890e+00 -7.89008617e-01 2.21149381e-02 1.94161320e+00
2.62427866e-01 2.63940021e-02 -1.58012919e-02 2.47705013e-01
6.69710696e-01 -1.94036528e-01 -5.69367528e-01 -2.26792425e-01
6.80188462e-03 5.66685677e-01 3.07682633e-01 3.04440826e-01
-8.96435678e-01 1.19399846e+00 4.80190277e+00 9.94405568e-01
-1.60349321e+00 -7.88468271e-02 6.09679282e-01 2.34696250e-02
-1.31948128e-01 -2.74253070e-01 -1.10850894e+00 4.65613246e-01
4.55294341e-01 5.96982539e-01 2.05897242e-01 1.21191525e+00
5.62416241e-02 3.35542470e-01 -1.13393223e+00 1.57199383e+00
1.34514943e-01 -1.23140931e+00 3.19146961e-01 1.59886464e-01
3.49431127e-01 -1.24165766e-01 5.58471829e-02 3.10201705e-01
-2.65427172e-01 -1.31174862e+00 7.41999567e-01 4.28061157e-01
9.05531824e-01 -8.81866753e-01 5.42169213e-01 4.69214708e-01
-1.01147544e+00 1.77849159e-01 -8.33215892e-01 7.61857778e-02
-1.85750946e-01 5.66754758e-01 -1.06285691e+00 5.74926734e-01
1.00049853e+00 5.32418072e-01 -6.58834875e-01 7.23381042e-01
4.13948074e-02 2.34866560e-01 -7.38585114e-01 -1.15775548e-01
1.05445914e-01 -2.46586800e-01 4.03050989e-01 7.68253624e-01
6.19648397e-01 6.16002195e-02 -3.74520272e-02 1.29946899e+00
-6.31947443e-02 8.37844517e-03 -9.41763341e-01 1.93251252e-01
4.87977594e-01 1.31915593e+00 -9.16006446e-01 1.38140887e-01
-2.56863683e-01 8.91132176e-01 2.52033234e-01 9.02376473e-02
-6.19818211e-01 -3.11158389e-01 7.71685004e-01 3.94383281e-01
4.31881547e-01 -4.63198096e-01 -2.02915534e-01 -9.35331523e-01
5.97895943e-02 -4.55699146e-01 -2.20100790e-01 -7.77194679e-01
-1.34514236e+00 8.25291991e-01 1.54397562e-01 -1.07848883e+00
-7.19068050e-02 -9.95542586e-01 -6.60214722e-01 1.00589931e+00
-1.54917049e+00 -1.38007796e+00 -4.22606081e-01 6.65699542e-01
5.82861662e-01 -3.25373471e-01 6.61603332e-01 2.77342528e-01
-4.69603270e-01 5.74198961e-01 -5.55076063e-01 3.51430506e-01
1.93889827e-01 -9.26566780e-01 6.89404964e-01 2.43669331e-01
6.89175650e-02 5.38185179e-01 1.11082740e-01 -5.92092991e-01
-1.68580198e+00 -1.19589794e+00 3.87821347e-01 -7.21137285e-01
-2.95228332e-01 -8.68821383e-01 -9.75297511e-01 4.01264936e-01
-4.89482671e-01 2.64627099e-01 3.17112446e-01 -2.03123882e-01
-2.97132343e-01 -4.99879032e-01 -1.80127299e+00 4.41659033e-01
1.39380252e+00 -2.94674695e-01 -4.80913371e-01 2.00952053e-01
5.55659056e-01 -4.98437583e-01 -4.32215035e-01 6.89956486e-01
3.47354054e-01 -1.18545508e+00 1.13598287e+00 -4.09352630e-01
1.93896219e-01 -3.66490692e-01 -2.77521223e-01 -1.14764047e+00
1.04755722e-01 -2.82568634e-01 1.55766323e-01 1.24801219e+00
9.60437581e-02 -6.83383405e-01 1.48063684e+00 4.27269310e-01
-2.15026051e-01 -7.93218255e-01 -1.34163594e+00 -7.83408821e-01
2.03457415e-01 -4.28715974e-01 1.35155177e+00 9.26225126e-01
-8.36794734e-01 6.88662156e-02 3.36299807e-01 3.66554111e-01
7.68208027e-01 6.98775575e-02 1.12803292e+00 -1.66758513e+00
4.77584183e-01 -3.60794067e-01 -9.50613320e-01 -1.13858378e+00
5.68398297e-01 -1.04109478e+00 -2.32812196e-01 -1.20426333e+00
-5.30170202e-01 -8.89860630e-01 4.71089393e-01 4.43613321e-01
3.45683485e-01 3.19100142e-01 1.26373589e-01 -3.59448418e-02
2.20262185e-02 7.88403869e-01 1.27174377e+00 -1.39074802e-01
-3.42108250e-01 1.52971506e-01 -3.51910710e-01 8.16700757e-01
7.68638909e-01 -3.27270269e-01 -4.67274487e-01 -7.02699304e-01
-2.57035524e-01 -1.62441999e-01 5.99347055e-01 -1.12380731e+00
-2.02220064e-02 -2.04593316e-01 9.81829286e-01 -9.53671634e-01
8.99491608e-01 -1.14459741e+00 1.61795527e-01 8.88094679e-02
3.77706230e-01 -2.66154259e-02 3.70409638e-01 1.59673452e-01
1.12723060e-01 -1.86803311e-01 8.35477889e-01 -4.12224054e-01
-4.85719562e-01 7.29031682e-01 4.10353363e-01 -4.71827269e-01
1.04406738e+00 -8.38294864e-01 2.17475682e-01 1.43768713e-01
-4.56002504e-01 2.13468890e-03 6.74811900e-01 5.62184095e-01
1.04567039e+00 -1.48460770e+00 -6.42505050e-01 9.68029380e-01
-6.98574111e-02 6.79183543e-01 2.09811375e-01 1.18777730e-01
-8.00520599e-01 3.57927471e-01 -2.51971006e-01 -1.17951691e+00
-1.06558335e+00 3.75072181e-01 6.63874745e-01 7.01501966e-01
-4.54745263e-01 1.06668484e+00 1.56121463e-01 -1.13036180e+00
1.40171438e-01 -4.54760611e-01 -6.37131333e-02 -1.06944077e-01
2.58226067e-01 1.33641556e-01 2.85525531e-01 -8.84362698e-01
-6.87759876e-01 1.12123144e+00 5.31174466e-02 2.32029557e-01
1.52941573e+00 2.92367578e-01 -2.11109519e-01 1.02083959e-01
1.46263111e+00 -2.30380446e-01 -1.28441536e+00 2.89009035e-01
-2.65884221e-01 -8.09778333e-01 -1.20795622e-01 -4.00700301e-01
-1.41763651e+00 1.13913238e+00 8.34263802e-01 3.22664045e-02
8.39415193e-01 1.35794461e-01 5.31314015e-01 3.56589556e-01
6.52987301e-01 -7.26803243e-01 -1.52936429e-01 4.47687864e-01
1.15068185e+00 -1.31570661e+00 -1.50178447e-01 -7.86710620e-01
-7.42081776e-02 1.31924546e+00 9.45036709e-01 -3.24303746e-01
8.16838145e-01 8.61487463e-02 4.64116372e-02 -4.83330369e-01
-2.41461068e-01 3.89844961e-02 1.72116429e-01 9.25496399e-01
-6.66668639e-02 -1.34650856e-01 3.46124411e-01 5.16660631e-01
-3.95332664e-01 -9.99089144e-03 -9.07501727e-02 8.69616747e-01
-3.57758164e-01 -1.02504098e+00 -7.53026843e-01 3.18947464e-01
-1.12050623e-01 4.56254900e-01 -4.67468381e-01 7.94349670e-01
5.48135638e-01 6.23861074e-01 4.30281192e-01 -4.16357726e-01
4.34565097e-01 1.92956045e-01 8.08590710e-01 -6.07578814e-01
-2.08030060e-01 -2.95040667e-01 -5.09636879e-01 -6.82806373e-01
-2.76216596e-01 -4.69317555e-01 -1.16887271e+00 -1.96689829e-01
-2.58193016e-01 9.20323506e-02 1.26915848e+00 7.00087368e-01
5.67105770e-01 -5.80303073e-02 9.07856107e-01 -1.52761686e+00
-4.52160835e-01 -7.77757883e-01 -5.18207908e-01 2.91435748e-01
2.77156234e-01 -1.00351405e+00 -4.65415806e-01 -2.85303801e-01] | [13.3261079788208, 0.20587414503097534] |
da92315a-da8d-4c5d-bb7b-5f3b41a74f0c | humaniflow-ancestor-conditioned-normalising | 2305.06968 | null | https://arxiv.org/abs/2305.06968v1 | https://arxiv.org/pdf/2305.06968v1.pdf | HuManiFlow: Ancestor-Conditioned Normalising Flows on SO(3) Manifolds for Human Pose and Shape Distribution Estimation | Monocular 3D human pose and shape estimation is an ill-posed problem since multiple 3D solutions can explain a 2D image of a subject. Recent approaches predict a probability distribution over plausible 3D pose and shape parameters conditioned on the image. We show that these approaches exhibit a trade-off between three key properties: (i) accuracy - the likelihood of the ground-truth 3D solution under the predicted distribution, (ii) sample-input consistency - the extent to which 3D samples from the predicted distribution match the visible 2D image evidence, and (iii) sample diversity - the range of plausible 3D solutions modelled by the predicted distribution. Our method, HuManiFlow, predicts simultaneously accurate, consistent and diverse distributions. We use the human kinematic tree to factorise full body pose into ancestor-conditioned per-body-part pose distributions in an autoregressive manner. Per-body-part distributions are implemented using normalising flows that respect the manifold structure of SO(3), the Lie group of per-body-part poses. We show that ill-posed, but ubiquitous, 3D point estimate losses reduce sample diversity, and employ only probabilistic training losses. Code is available at: https://github.com/akashsengupta1997/HuManiFlow. | ['Roberto Cipolla', 'Ignas Budvytis', 'Akash Sengupta'] | 2023-05-11 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sengupta_HuManiFlow_Ancestor-Conditioned_Normalising_Flows_on_SO3_Manifolds_for_Human_Pose_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sengupta_HuManiFlow_Ancestor-Conditioned_Normalising_Flows_on_SO3_Manifolds_for_Human_Pose_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-human-pose-and-shape-estimation', 'normalising-flows'] | ['computer-vision', 'methodology'] | [-7.16380775e-02 3.88633311e-01 -2.30193600e-01 -3.33646983e-01
-9.72956061e-01 -3.91013712e-01 5.20473123e-01 -5.54149747e-01
-1.96496770e-01 7.39359140e-01 4.15605426e-01 3.27844024e-01
-1.57702312e-01 -2.42661446e-01 -9.64468002e-01 -4.61700559e-01
-1.94472373e-01 1.00901461e+00 -2.05378118e-03 1.38317078e-01
-2.01136246e-01 6.44115448e-01 -1.58142173e+00 -9.99533460e-02
4.24064547e-01 8.28523457e-01 -1.39540508e-01 1.04444921e+00
4.39692199e-01 3.33160549e-01 -3.02992344e-01 -4.60101157e-01
5.63018680e-01 -5.80435634e-01 -5.01537144e-01 4.43682283e-01
8.10553253e-01 -2.23360971e-01 -3.83637100e-01 1.00020289e+00
5.76668799e-01 1.44617692e-01 1.10456061e+00 -1.42763400e+00
-1.83819786e-01 -2.58390039e-01 -6.35595679e-01 -1.68332949e-01
8.46314311e-01 3.40101600e-01 7.70533144e-01 -9.53325868e-01
1.06830716e+00 1.62139058e+00 1.01706028e+00 7.20014691e-01
-1.52992558e+00 -4.47939724e-01 -1.22022055e-01 -1.69909492e-01
-1.42777586e+00 -3.71155500e-01 6.93208873e-01 -7.40184903e-01
5.42840481e-01 3.17419827e-01 1.19526958e+00 1.10222781e+00
7.52353966e-01 7.43604481e-01 8.83268118e-01 -2.51001745e-01
8.67428780e-02 -1.59383714e-01 -1.73636153e-01 8.56768191e-01
2.29289085e-01 3.08131784e-01 -7.87183642e-01 -3.63686979e-01
1.16000211e+00 -1.32508039e-01 -3.07826698e-01 -1.05335987e+00
-1.20173216e+00 6.93445086e-01 2.78310269e-01 -3.76018316e-01
-4.97595608e-01 4.21592385e-01 3.21472660e-02 -4.81660701e-02
5.27182877e-01 1.16664112e-01 -3.44758511e-01 9.76569857e-03
-7.90974200e-01 8.99188995e-01 7.45006919e-01 1.14643896e+00
6.75310850e-01 -8.39936659e-02 1.30671889e-01 5.02408266e-01
5.63460946e-01 9.29135919e-01 1.54308960e-01 -1.61873496e+00
1.02691084e-01 2.44638249e-01 8.82074535e-02 -1.01375699e+00
-5.15400529e-01 -4.41328168e-01 -8.34541500e-01 5.20940959e-01
6.22221291e-01 -4.58327346e-02 -8.39336395e-01 2.01046371e+00
9.11146700e-01 -2.82828838e-01 -4.37686503e-01 1.23077512e+00
3.77964497e-01 4.30821538e-01 5.10140695e-02 1.70242582e-02
1.22316492e+00 -3.39276373e-01 -3.27404499e-01 -4.40149993e-01
1.27576590e-01 -7.20525146e-01 7.84103036e-01 3.50869566e-01
-1.41091907e+00 -5.70088923e-01 -8.06302428e-01 -5.78354448e-02
2.97368616e-01 -7.49006942e-02 1.95215821e-01 6.01124406e-01
-7.99890816e-01 6.23577118e-01 -9.91529167e-01 -2.80179411e-01
2.49503121e-01 2.44568214e-01 -7.89503098e-01 -9.64086279e-02
-8.01940620e-01 1.08818972e+00 8.99626017e-02 2.60471646e-02
-6.61040664e-01 -8.93383741e-01 -1.15251422e+00 -4.47719514e-01
1.99308917e-01 -1.44936991e+00 1.15464771e+00 -6.84076309e-01
-1.14131320e+00 1.27223897e+00 -2.53404498e-01 -3.44019055e-01
8.84076178e-01 -6.12957299e-01 1.59194350e-01 3.90901685e-01
9.75397751e-02 9.55001116e-01 1.04422140e+00 -1.32813978e+00
-2.56274164e-01 -7.08630800e-01 -4.62161720e-01 6.56676173e-01
6.91471756e-01 -4.57729965e-01 -3.23454827e-01 -6.25669718e-01
5.54242790e-01 -1.11301470e+00 -3.05942446e-01 7.57553875e-01
-6.30660117e-01 3.06607913e-02 1.66987777e-01 -1.05099547e+00
6.61733091e-01 -1.89416504e+00 6.36616647e-01 3.92984629e-01
3.26718152e-01 -3.18121880e-01 1.83619618e-01 1.14765227e-01
-2.61056051e-02 -6.74019828e-02 -3.74084830e-01 -4.95729834e-01
2.92897433e-01 1.20495059e-01 -1.17711648e-02 1.24801838e+00
9.22464058e-02 7.55902052e-01 -6.68377161e-01 -6.57378256e-01
6.33018315e-01 7.02815533e-01 -7.50750899e-01 2.63666153e-01
-2.36907423e-01 7.57543445e-01 -3.00483137e-01 4.75625813e-01
6.74511313e-01 -3.90298374e-04 -1.42040715e-01 -2.91159630e-01
3.22184086e-01 -5.22440746e-02 -1.47629535e+00 1.83968306e+00
1.27015561e-02 1.71236396e-01 2.19165087e-01 -5.43679535e-01
7.24006474e-01 2.24389449e-01 5.96098959e-01 -9.51257125e-02
3.50211859e-02 2.46123552e-01 -3.18001181e-01 -1.92250341e-01
2.74168730e-01 -5.59498966e-01 -2.20671803e-01 5.17780632e-02
3.25368375e-01 -5.56000710e-01 -3.78170997e-01 1.44378036e-01
6.78481221e-01 8.39359343e-01 6.19729638e-01 -2.43926033e-01
1.21402264e-01 -1.36164594e-02 5.33103645e-01 5.58907092e-01
-2.87578881e-01 1.11073267e+00 4.19558167e-01 -3.21064085e-01
-1.35393643e+00 -1.75963366e+00 -1.82630524e-01 3.19459081e-01
2.47582883e-01 -1.88328356e-01 -7.01455414e-01 -3.11910480e-01
4.01083469e-01 7.20254421e-01 -6.48605824e-01 -1.87957570e-01
-5.91725290e-01 -1.60029709e-01 2.10523874e-01 3.19902480e-01
9.60138962e-02 -7.13929236e-01 -1.00609660e+00 -9.83831212e-02
-3.39211106e-01 -8.18961322e-01 -7.68905163e-01 -6.79298714e-02
-9.07553792e-01 -1.22554266e+00 -1.10488987e+00 -4.01970416e-01
6.54953063e-01 -2.78331161e-01 1.25994718e+00 -2.13095233e-01
-5.93346596e-01 9.01632726e-01 8.12220201e-02 -1.52989894e-01
-4.27016109e-01 -5.61847329e-01 4.27661687e-01 -1.95822597e-01
5.34458756e-02 -5.29658675e-01 -6.28157556e-01 4.12242204e-01
-2.07420722e-01 9.28896442e-02 1.89064488e-01 8.36397886e-01
1.15397954e+00 -3.79235536e-01 -4.64958772e-02 -3.54589999e-01
-8.02709088e-02 -2.97261983e-01 -3.55731726e-01 3.03571168e-02
1.18939318e-01 6.57950267e-02 -7.82636751e-04 -4.94191498e-01
-9.15516198e-01 6.09389603e-01 -1.85546234e-01 -8.52402031e-01
-4.07281905e-01 3.91070433e-02 -2.77156681e-01 1.95000559e-01
9.22351539e-01 1.28409818e-01 4.28917259e-01 -4.89723146e-01
5.63733101e-01 8.70846510e-02 9.58122790e-01 -7.24467933e-01
8.51899922e-01 5.57764590e-01 5.03453255e-01 -1.06011832e+00
-7.47800648e-01 -5.55607080e-01 -9.83437002e-01 -5.66756070e-01
1.10619354e+00 -1.05458128e+00 -6.33319736e-01 3.33190024e-01
-1.12489808e+00 -2.43158415e-01 -5.90366244e-01 7.93100357e-01
-1.35015929e+00 4.82724398e-01 -3.22002500e-01 -1.20508981e+00
-1.76889688e-01 -8.69435847e-01 1.55252337e+00 -1.83985189e-01
-9.64098930e-01 -7.93699265e-01 9.97112170e-02 2.10572511e-01
-3.86956751e-01 8.15541685e-01 5.65670013e-01 -1.44085556e-01
-4.01000410e-01 -4.73971099e-01 3.53532135e-01 2.67898172e-01
-2.11000666e-01 -3.36051881e-01 -8.43809307e-01 -2.57464141e-01
1.68711022e-01 -3.86212319e-01 3.56559366e-01 1.07081318e+00
6.93371415e-01 -3.43805343e-01 -3.18042189e-01 7.25321531e-01
1.04355657e+00 -4.93626863e-01 4.40574527e-01 -3.94154973e-02
6.73713386e-01 1.03527188e+00 6.89089417e-01 4.79089528e-01
2.51571894e-01 7.88712680e-01 3.70239884e-01 2.45000452e-01
-3.98692340e-01 -7.11049497e-01 2.85153866e-01 4.01875675e-01
1.25896302e-03 5.26148863e-02 -7.16764152e-01 4.28804845e-01
-1.72170258e+00 -1.07007647e+00 -6.56039044e-02 2.41786313e+00
5.81630647e-01 2.42116362e-01 6.14499271e-01 -4.69103083e-02
7.36003995e-01 -6.42380044e-02 -8.49331260e-01 1.79218203e-01
-7.31857717e-02 4.65437770e-03 4.82261717e-01 8.27489078e-01
-9.16993082e-01 3.90357852e-01 6.22941923e+00 7.56184161e-01
-3.83960336e-01 6.37796940e-03 4.59834963e-01 -3.22565138e-01
-2.97664195e-01 -2.26303309e-01 -8.00778627e-01 3.04664552e-01
5.69825351e-01 -1.84051082e-01 7.83490315e-02 8.23067009e-01
-2.69035939e-02 -1.99804515e-01 -1.23471749e+00 1.22228837e+00
1.76426575e-01 -1.05759954e+00 -2.60147583e-02 3.84631336e-01
3.04996401e-01 -2.57104844e-01 1.42930420e-02 -2.25401789e-01
2.10676059e-01 -8.98342550e-01 1.52109134e+00 8.28044116e-01
1.03182435e+00 -7.19632745e-01 1.52746931e-01 7.48786449e-01
-1.08909047e+00 3.59567225e-01 -2.87803441e-01 1.91676140e-01
5.72510481e-01 5.76213121e-01 -6.51085496e-01 4.95447159e-01
7.97858775e-01 6.34375095e-01 -5.78501783e-02 1.09055507e+00
-2.78906792e-01 1.91650510e-01 -7.06162274e-01 2.17505425e-01
-3.30792993e-01 -1.22821718e-01 1.14931929e+00 9.11849499e-01
3.88859510e-01 1.12255618e-01 2.35012949e-01 1.00253272e+00
3.40257317e-01 -3.14748079e-01 -6.68184757e-01 4.65645969e-01
3.30521673e-01 7.30189502e-01 -4.33563828e-01 -6.12452067e-02
1.42749473e-01 1.07788002e+00 -5.39610162e-02 2.08314359e-01
-5.51530659e-01 6.11783639e-02 9.51596439e-01 4.42709833e-01
1.13542989e-01 -3.59852940e-01 -4.84964460e-01 -1.10632360e+00
2.60314494e-01 -7.00625420e-01 5.14265835e-01 -1.15940428e+00
-1.42343616e+00 3.01550269e-01 5.55729330e-01 -1.33834171e+00
-7.22790599e-01 -7.01618969e-01 -4.00413983e-02 1.11382079e+00
-5.27469277e-01 -1.08595824e+00 8.30887929e-02 4.39311028e-01
5.05389810e-01 2.19252288e-01 7.08761692e-01 -1.68037102e-01
1.16024300e-01 3.87575299e-01 -3.26214671e-01 -3.27871114e-01
5.38711429e-01 -1.38407469e+00 6.51892066e-01 5.70106924e-01
7.58045390e-02 4.60766733e-01 1.17911339e+00 -1.02845585e+00
-1.44810021e+00 -7.65602887e-01 8.04422021e-01 -1.08154774e+00
2.43154109e-01 -4.36374515e-01 -6.20408714e-01 9.17489946e-01
-4.58646029e-01 -5.85178751e-03 2.65353024e-01 -1.31010950e-01
-2.69750148e-01 3.04945409e-01 -1.45030963e+00 7.04246640e-01
1.33372772e+00 -3.12228113e-01 -7.49155164e-01 2.32543111e-01
4.51492667e-01 -8.60175431e-01 -8.80355060e-01 3.54842395e-01
9.89611626e-01 -9.77496386e-01 1.49483633e+00 -3.83090556e-01
3.38497996e-01 -1.75871328e-01 -4.53670293e-01 -1.21865392e+00
-2.45280698e-01 -7.41712511e-01 -2.86070138e-01 4.92048919e-01
3.15559767e-02 -3.34265798e-01 1.10764003e+00 7.58406103e-01
-1.10280529e-01 -8.27239215e-01 -1.39360261e+00 -8.57731164e-01
3.28411281e-01 -6.06041968e-01 2.05638304e-01 3.49389851e-01
-3.76965702e-01 -5.25187217e-02 -6.48509562e-01 1.09055854e-01
1.29624009e+00 -1.84574634e-01 1.08707356e+00 -1.14535105e+00
-4.90786672e-01 -1.41798154e-01 -8.02265286e-01 -1.33099425e+00
7.30176643e-02 -7.87912607e-01 2.55915523e-01 -1.32619047e+00
1.16135128e-01 -5.81951439e-02 5.86217999e-01 6.31502345e-02
3.97996455e-02 2.79923290e-01 3.19790959e-01 3.23210061e-01
-2.49307558e-01 5.88690758e-01 1.37114608e+00 3.24232876e-01
-4.03397046e-02 5.04541695e-02 -2.51654834e-01 1.06580901e+00
3.90039831e-01 -4.21356261e-01 -3.11834097e-01 -5.47442846e-02
2.78997775e-02 4.58620846e-01 1.00946379e+00 -1.10722470e+00
-6.58224523e-02 -6.95576668e-02 1.04240704e+00 -8.68060470e-01
9.46188688e-01 -7.23566830e-01 8.21340740e-01 6.22999012e-01
-3.94232661e-01 -5.96187562e-02 -9.80651379e-03 7.62892425e-01
3.70331436e-01 4.06283811e-02 9.56835568e-01 -4.16570008e-01
-4.55963403e-01 4.75543261e-01 -1.34815097e-01 3.52700412e-01
8.08296621e-01 -6.84894860e-01 2.55159944e-01 -6.55786037e-01
-1.07839990e+00 6.81932494e-02 8.42488229e-01 7.72721544e-02
7.05128014e-01 -1.59182048e+00 -9.06473935e-01 3.29992712e-01
-1.56080455e-03 1.87672451e-01 4.03417259e-01 6.71869040e-01
-5.70996404e-01 4.80430573e-03 -1.90415055e-01 -1.07838607e+00
-1.20829308e+00 2.09511787e-01 6.53841138e-01 1.10904083e-01
-1.07801306e+00 8.79981935e-01 4.47868705e-01 -8.41252565e-01
1.97633162e-01 1.12797640e-01 5.50149202e-01 -3.41090083e-01
2.55331188e-01 5.79050839e-01 -4.35515493e-01 -1.32401276e+00
-4.15049493e-01 1.05215216e+00 6.45980716e-01 -5.42037725e-01
1.04496479e+00 -3.30841243e-01 3.00244659e-01 5.02575576e-01
1.37588406e+00 -1.23414591e-01 -1.68637145e+00 1.40808225e-02
-4.51834053e-01 -6.61054134e-01 -3.85720044e-01 -7.14892089e-01
-5.19958496e-01 7.86243558e-01 5.48694849e-01 -3.35183382e-01
6.43583417e-01 4.76915509e-01 5.42873621e-01 -1.29667953e-01
5.23485363e-01 -6.70477509e-01 -1.18150480e-01 1.17405467e-01
1.42478466e+00 -9.10748839e-01 3.59194010e-01 -7.20678270e-01
-6.90920532e-01 8.60763133e-01 4.26930726e-01 -4.91537392e-01
8.24160993e-01 -1.71848476e-01 -4.77332063e-02 -2.80136526e-01
-3.85131657e-01 5.36558479e-02 8.44738126e-01 1.06001294e+00
2.40963101e-01 2.15477109e-01 1.08254664e-01 3.83021981e-01
-7.31772065e-01 -1.64885134e-01 1.59179866e-01 6.61040485e-01
-3.96788657e-01 -5.95909178e-01 -8.91599596e-01 3.51647198e-01
-1.73872724e-01 4.45056111e-01 -2.59664148e-01 9.54379559e-01
-1.51001634e-02 5.02853870e-01 8.92083272e-02 -2.45173082e-01
5.53732336e-01 9.61938724e-02 1.07085776e+00 -5.75471580e-01
1.15152098e-01 3.82320821e-01 1.50652915e-01 -8.73707712e-01
-3.37779582e-01 -1.06042373e+00 -1.05817723e+00 -3.28330547e-01
-6.62706867e-02 -1.15314767e-01 4.69772547e-01 6.32906914e-01
1.78318515e-01 -7.51267374e-02 1.03777058e-01 -1.56206107e+00
-9.42428470e-01 -7.08143234e-01 -8.11030805e-01 5.85258365e-01
5.57824135e-01 -9.48421299e-01 -6.67166412e-01 1.61283553e-01] | [6.992441654205322, -1.0190507173538208] |
658d84cf-8271-425f-ba97-e43447c65ab9 | small-object-detection-for-near-real-time | 2106.06403 | null | https://arxiv.org/abs/2106.06403v1 | https://arxiv.org/pdf/2106.06403v1.pdf | Small Object Detection for Near Real-Time Egocentric Perception in a Manual Assembly Scenario | Detecting small objects in video streams of head-worn augmented reality devices in near real-time is a huge challenge: training data is typically scarce, the input video stream can be of limited quality, and small objects are notoriously hard to detect. In industrial scenarios, however, it is often possible to leverage contextual knowledge for the detection of small objects. Furthermore, CAD data of objects are typically available and can be used to generate synthetic training data. We describe a near real-time small object detection pipeline for egocentric perception in a manual assembly scenario: We generate a training data set based on CAD data and realistic backgrounds in Unity. We then train a YOLOv4 model for a two-stage detection process: First, the context is recognized, then the small object of interest is detected. We evaluate our pipeline on the augmented reality device Microsoft Hololens 2. | ['Martin Ruskowski', 'Christiane Plociennik', 'Parsha Pahlevannejad', 'Snehal Walunj', 'Hooman Tavakoli'] | 2021-06-11 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [ 3.98306459e-01 -5.09712212e-02 3.65780145e-01 -1.09295659e-01
-7.53851593e-01 -5.03274441e-01 1.71026096e-01 -1.80867091e-02
-2.28068680e-01 3.78627121e-01 -2.11448148e-01 -3.04039299e-01
3.61530989e-01 -4.81625229e-01 -9.25511777e-01 -2.31098354e-01
6.71229586e-02 5.82642138e-01 6.84694469e-01 -1.84746668e-01
7.48917460e-02 5.87857842e-01 -1.95742846e+00 3.97257328e-01
4.11529183e-01 1.03255284e+00 7.51751721e-01 1.20593977e+00
5.04359901e-01 7.98587680e-01 -7.94309199e-01 9.53990892e-02
5.43927789e-01 -2.23587081e-01 -1.02887534e-01 5.82960308e-01
7.72277892e-01 -7.11258888e-01 -4.13353562e-01 9.89812136e-01
7.80883610e-01 1.74105331e-01 1.92556843e-01 -1.30059910e+00
7.22992867e-02 -7.84165785e-02 -5.93915284e-01 3.41655284e-01
1.06286848e+00 4.45979834e-01 5.25558233e-01 -1.12245417e+00
7.41497993e-01 1.15739644e+00 3.68910789e-01 5.03102899e-01
-1.04686153e+00 -5.77880621e-01 2.17745468e-01 -1.34169692e-02
-1.27919400e+00 -6.73717320e-01 5.07025719e-01 -5.31604826e-01
8.40984464e-01 1.95028707e-01 9.68060076e-01 1.28414583e+00
1.48102492e-01 9.52716887e-01 7.79600322e-01 -4.69823897e-01
5.23349166e-01 1.49463266e-01 -2.26909876e-01 6.09489381e-01
2.40139022e-01 2.57930011e-01 -5.04141510e-01 5.68917692e-02
1.21809459e+00 3.04971606e-01 -4.94380146e-01 -6.95235789e-01
-1.51433122e+00 2.48907998e-01 3.24206859e-01 -1.42128542e-01
-6.08884037e-01 -3.30963135e-02 2.10812837e-01 1.29090667e-01
2.33446017e-01 2.20452949e-01 -4.75529343e-01 -1.55380800e-01
-8.12956214e-01 4.24942821e-01 6.30856693e-01 1.51141250e+00
3.77927542e-01 5.82292639e-02 8.82781744e-02 4.76854950e-01
2.14595273e-01 4.29502726e-01 2.53576130e-01 -1.09428585e+00
5.53941011e-01 3.18655789e-01 5.31182349e-01 -6.59064531e-01
-2.92047024e-01 -2.32957229e-01 -1.15792237e-01 7.20147252e-01
6.46241426e-01 -3.02838594e-01 -9.83337224e-01 1.09242427e+00
6.17365241e-01 4.07981575e-01 -3.65726985e-02 1.36062455e+00
6.94071174e-01 3.47577423e-01 -1.34740368e-01 -9.95224565e-02
1.58281052e+00 -7.90782511e-01 -6.50482118e-01 -7.18479514e-01
5.69119930e-01 -8.92444491e-01 1.24146986e+00 6.84543133e-01
-1.13363004e+00 -6.50875926e-01 -1.15278637e+00 -2.07417575e-03
-1.27181023e-01 4.46344912e-01 7.32310593e-01 6.71147168e-01
-5.96321523e-01 3.28838229e-01 -9.41937268e-01 -5.77343464e-01
3.30042452e-01 2.92453974e-01 -3.85409653e-01 -5.39664209e-01
-4.00473893e-01 6.35334492e-01 4.91514057e-01 1.52044356e-01
-1.13410544e+00 -5.72343588e-01 -1.16430092e+00 2.76628137e-02
8.16933990e-01 -9.03478742e-01 1.47706425e+00 -8.89059126e-01
-1.45802641e+00 8.09473932e-01 1.39330909e-01 -1.68264270e-01
7.72786796e-01 -4.16044414e-01 -3.70945901e-01 3.15704793e-02
-1.13715967e-02 3.80589038e-01 8.55684876e-01 -1.25706315e+00
-9.18303907e-01 -6.02934241e-01 4.98158932e-01 5.09049773e-01
3.71419787e-01 1.46695957e-01 -4.59384382e-01 -4.19687659e-01
3.22434604e-01 -9.97373581e-01 -3.70787442e-01 1.34318069e-01
-2.28979602e-01 3.40147227e-01 8.64067137e-01 -5.79161823e-01
6.44062996e-01 -2.28989148e+00 -2.38695130e-01 -5.96987344e-02
8.55336525e-03 -2.78240945e-02 1.17974311e-01 -1.76611125e-01
-1.84985828e-02 -6.73519731e-01 1.96656078e-01 -3.05983216e-01
-2.88702548e-01 3.56594622e-02 -1.65365398e-01 2.99195141e-01
-2.62486767e-02 4.66481298e-01 -1.26098609e+00 -3.48194361e-01
6.37897193e-01 4.16154325e-01 -7.66928732e-01 4.56448317e-01
-2.93793112e-01 5.42355359e-01 -1.70228586e-01 1.09933078e+00
6.02161288e-01 -4.63551469e-02 7.49411583e-02 -1.41457006e-01
1.11073911e-01 1.15277611e-01 -1.66477108e+00 1.75949144e+00
-6.24324203e-01 7.35117912e-01 1.63508147e-01 -2.48570293e-01
5.14228225e-01 2.35024035e-01 2.48655841e-01 -3.22781324e-01
2.31997564e-01 1.04668438e-01 5.79826981e-02 -5.56690216e-01
9.52609837e-01 9.45531800e-02 1.00772232e-01 1.44928619e-01
-6.11556657e-02 -5.37149489e-01 6.39354512e-02 1.03013940e-01
1.46974444e+00 5.65928400e-01 4.32032436e-01 4.34212536e-01
-6.50301203e-02 1.82829514e-01 5.65193772e-01 6.58762872e-01
-9.76445600e-02 1.31112170e+00 -2.97682267e-02 -3.96365911e-01
-1.03176212e+00 -1.23148680e+00 2.41271093e-01 1.03721976e+00
3.16517681e-01 -3.57669562e-01 -6.42383099e-01 -6.30594194e-01
-3.65465403e-01 7.96554506e-01 -2.00031996e-01 7.02622980e-02
-3.34442168e-01 -2.94642270e-01 -2.04529107e-01 7.09162712e-01
-9.44166854e-02 -1.16258991e+00 -1.40524125e+00 1.80631697e-01
9.25012827e-02 -1.66344786e+00 -5.45456469e-01 1.69986427e-01
-8.96472037e-01 -1.24454618e+00 -6.45639002e-01 -6.49042606e-01
1.03316879e+00 7.75993884e-01 1.33720815e+00 -1.12141214e-01
-6.15926087e-01 4.38316524e-01 -1.41650751e-01 -7.60749519e-01
-1.60140902e-01 -5.06301641e-01 2.58841962e-01 -1.38951391e-01
3.33083600e-01 -1.97580040e-01 -8.85424793e-01 5.66218555e-01
-4.98583853e-01 5.30545935e-02 4.26108420e-01 6.04373574e-01
6.22232914e-01 -5.15872389e-02 6.15518764e-02 -6.91815913e-01
1.88752532e-01 -2.03926384e-01 -8.58389616e-01 -1.12212598e-01
1.20530576e-01 -6.69995129e-01 2.55562335e-01 -7.93351591e-01
-1.03902614e+00 7.43147850e-01 1.11125365e-01 -9.92195010e-01
-4.59015667e-01 -1.06045134e-01 -4.14315253e-01 4.25748825e-02
7.74878263e-01 -2.13289529e-01 -3.93222958e-01 -1.65981814e-01
2.30570391e-01 7.31896639e-01 6.34895980e-01 -3.14904779e-01
3.50403547e-01 4.77416396e-01 -4.03481334e-01 -9.05538619e-01
-4.89804447e-01 -5.67688346e-01 -6.51520789e-01 -3.45335782e-01
5.76543689e-01 -1.25986719e+00 -6.81516945e-01 1.30693883e-01
-1.17345881e+00 -5.22790492e-01 -6.37593925e-01 9.13111091e-01
-7.09559143e-01 1.36534974e-01 -2.59972364e-01 -9.63023663e-01
-3.69804585e-03 -1.23824275e+00 1.50896144e+00 1.96334019e-01
-1.67197064e-01 -3.06551218e-01 -5.62530458e-01 4.60455418e-01
-2.11929128e-01 2.03588039e-01 1.61661848e-01 -4.13460851e-01
-9.94810104e-01 -6.72041535e-01 -4.63958904e-02 -2.73405463e-02
8.05966333e-02 1.52428113e-02 -1.18313456e+00 -1.86084256e-01
6.41637444e-02 -1.78026080e-01 1.53492361e-01 4.76059914e-01
9.83222663e-01 1.31835252e-01 -4.46646273e-01 3.71967435e-01
9.27953660e-01 3.53388548e-01 5.73363006e-01 1.01855487e-01
8.02197278e-01 5.27139485e-01 1.33015859e+00 5.60122609e-01
3.53629291e-01 9.34847653e-01 4.20370966e-01 7.18215480e-02
-1.83577374e-01 -7.43408501e-02 3.80215615e-01 4.45344567e-01
-6.41267747e-02 -6.59468397e-02 -8.37772667e-01 4.44896340e-01
-1.73478711e+00 -8.64395916e-01 -1.66049078e-01 2.44858694e+00
1.91670999e-01 4.47048038e-01 1.88114733e-01 2.70684034e-01
6.75655186e-01 -3.97423267e-01 -5.61042726e-01 2.53320843e-01
1.38847336e-01 -1.61423266e-01 4.80888397e-01 1.98241502e-01
-9.34141874e-01 8.71059120e-01 5.86462212e+00 6.73605055e-02
-1.06817603e+00 2.07277268e-01 1.61213666e-01 -6.08653426e-01
3.16759229e-01 -1.09608836e-01 -7.19250023e-01 3.16006452e-01
6.06713951e-01 1.44398883e-01 7.92097300e-02 1.36681104e+00
2.67213672e-01 -5.30074894e-01 -1.51433587e+00 1.46238840e+00
1.17812701e-01 -8.43784392e-01 -5.47848344e-01 1.53564751e-01
4.81112123e-01 6.38577193e-02 -4.68208604e-02 3.36638600e-01
3.96521360e-01 -5.50793409e-01 8.50986660e-01 1.70455396e-01
6.32482946e-01 -5.97626448e-01 5.01569211e-01 6.19014919e-01
-9.88754869e-01 -3.28505039e-01 -6.16198003e-01 -1.41247258e-01
4.13024694e-01 4.13556725e-01 -1.40661383e+00 1.03165425e-01
7.55375862e-01 3.96815807e-01 -4.65130389e-01 1.50901937e+00
-1.82324678e-01 1.56771451e-01 -4.05683100e-01 2.34176502e-01
-3.15420479e-01 -4.11207676e-02 7.18413591e-01 7.88804829e-01
6.15060210e-01 4.36148435e-01 4.37320083e-01 4.39689159e-01
1.48114473e-01 -1.43370673e-01 -9.90209699e-01 4.13366497e-01
2.78781265e-01 1.30115676e+00 -9.92994428e-01 -3.96368116e-01
-4.56669748e-01 1.27936327e+00 -1.21094599e-01 3.50722134e-01
-6.56940758e-01 -2.09678397e-01 6.10476375e-01 4.70574856e-01
4.92928118e-01 -3.99959326e-01 3.20031345e-02 -1.42786288e+00
2.24567607e-01 -1.11819661e+00 3.02740276e-01 -1.48934853e+00
-5.45125902e-01 6.97846830e-01 -2.50521809e-01 -1.73549533e+00
-5.71763158e-01 -8.30513299e-01 -2.95648962e-01 4.39884067e-01
-9.05143678e-01 -1.13290095e+00 -8.72799456e-01 4.42556947e-01
1.05455637e+00 -7.46277869e-02 8.23243201e-01 4.16836441e-01
-3.35070550e-01 2.30644628e-01 -2.52441227e-01 -1.00690886e-01
5.56466401e-01 -1.12317133e+00 6.72015905e-01 8.79350245e-01
2.98291683e-01 3.82424921e-01 9.86409307e-01 -7.22881138e-01
-1.66748118e+00 -8.25808585e-01 2.59189531e-02 -1.05648172e+00
3.41648638e-01 -9.20809805e-01 -6.90131843e-01 8.74005139e-01
-3.49713176e-01 6.49450779e-01 3.55943203e-01 -1.66629076e-01
7.84464180e-02 6.25345558e-02 -1.16529489e+00 7.62963176e-01
1.31164587e+00 -4.23516929e-01 -5.21108329e-01 6.44438386e-01
6.04731500e-01 -1.15347481e+00 -5.81293285e-01 1.71086773e-01
8.18600476e-01 -9.68266726e-01 1.12916911e+00 -4.27799553e-01
-8.01923964e-03 -6.53070986e-01 -3.05387646e-01 -1.35406148e+00
1.84794128e-01 -7.09608972e-01 -4.11111861e-01 6.82781518e-01
1.12835839e-02 -6.26911670e-02 9.52252686e-01 7.40559578e-01
-5.03471121e-02 -2.18188941e-01 -8.01235378e-01 -6.56419694e-01
-1.14621854e+00 -8.28410804e-01 4.67369258e-01 5.78693569e-01
-6.94720224e-02 4.95214760e-01 -1.11301757e-01 5.18858671e-01
5.37688971e-01 -4.86305244e-02 1.36921799e+00 -1.22569883e+00
-4.46801573e-01 1.82393044e-01 -8.86254191e-01 -1.12797403e+00
-2.92718530e-01 -1.05834700e-01 3.58834177e-01 -1.35962427e+00
-2.31611565e-01 -5.00338256e-01 7.04064593e-02 -7.00198486e-02
-1.46100685e-01 2.87120163e-01 3.17757696e-01 -1.15026860e-02
-5.19072115e-01 1.75157726e-01 1.15203762e+00 6.91416785e-02
-4.93838131e-01 6.44705832e-01 -2.43362680e-01 9.96463060e-01
2.07980201e-01 -2.13048607e-01 -5.17947733e-01 -2.41471067e-01
6.73445612e-02 4.30293381e-01 5.21931767e-01 -1.37548566e+00
2.49882221e-01 3.66518274e-02 6.79403365e-01 -8.09148312e-01
7.97984302e-01 -1.25276875e+00 1.45655870e-01 3.14464599e-01
2.73408264e-01 3.10966037e-02 1.94595039e-01 7.00951219e-01
1.41236842e-01 -1.85443729e-01 4.45948273e-01 -3.58048856e-01
-8.40832829e-01 1.76846385e-01 -2.32649207e-01 -2.51154244e-01
1.24879754e+00 -6.28918469e-01 1.07547909e-01 -4.15221900e-01
-8.48060191e-01 -1.74170788e-02 6.45919144e-01 7.35173285e-01
1.00830686e+00 -1.19943941e+00 -4.33296174e-01 5.49767554e-01
4.00327593e-01 5.58328688e-01 1.66126594e-01 6.63740456e-01
-6.66468024e-01 -1.31022692e-01 -1.45072371e-01 -9.89279985e-01
-1.57884264e+00 9.18884456e-01 2.73319513e-01 4.93233919e-01
-8.07443202e-01 7.84602463e-01 6.69977367e-01 -2.47581616e-01
3.52338493e-01 -6.80734217e-01 5.88123426e-02 -1.28618494e-01
9.07301307e-01 3.89215380e-01 2.38338485e-01 -5.96799850e-01
-3.13787729e-01 3.35780531e-01 -6.82584792e-02 -1.84312075e-01
1.20711374e+00 -1.56394854e-01 6.06872559e-01 5.31971574e-01
6.01302147e-01 -1.00817410e-02 -1.70332325e+00 -1.55449852e-01
-3.39337766e-01 -1.01338303e+00 1.31473839e-02 -4.71792579e-01
-8.75874758e-01 8.27458978e-01 1.00178719e+00 -2.06788987e-01
9.88258660e-01 -5.89753613e-02 4.30434227e-01 4.82995600e-01
9.17501867e-01 -1.02374744e+00 3.62386793e-01 2.28003159e-01
8.88145208e-01 -1.39266920e+00 5.09108789e-02 -6.39658451e-01
-8.05549860e-01 9.76254761e-01 9.23108101e-01 2.91006546e-02
3.55689496e-01 7.20422029e-01 2.89558321e-01 -2.91382164e-01
-4.42214757e-01 -1.64218739e-01 5.07359169e-02 8.18426430e-01
1.17126949e-01 -6.75066337e-02 5.80779850e-01 3.72199714e-01
-3.73665065e-01 1.69686779e-01 8.34389985e-01 1.31233120e+00
-2.45147765e-01 -5.02153039e-01 -8.26586783e-01 2.52142817e-01
-4.25006747e-01 4.56062295e-02 7.78417278e-04 5.57878315e-01
1.16568811e-01 8.10478449e-01 2.02574477e-01 -2.68194854e-01
7.92464972e-01 -2.58410215e-01 7.96975195e-01 -1.10324490e+00
-3.07999521e-01 3.01332891e-01 1.88837312e-02 -7.40617394e-01
1.13273382e-01 -8.69527876e-01 -1.02518272e+00 3.30530763e-01
-6.34826541e-01 -4.59752649e-01 1.02300632e+00 7.62452781e-01
3.93849790e-01 8.28532338e-01 3.04297030e-01 -1.60515594e+00
-2.17101797e-01 -8.66043448e-01 -4.96376723e-01 1.12169705e-01
4.46620733e-01 -6.83881700e-01 -2.89802980e-02 2.55178303e-01] | [7.228487491607666, -2.1761057376861572] |
f836023e-f1ab-4416-bd14-a17effb72137 | a-probit-tensor-factorization-model-for | 2111.03943 | null | https://arxiv.org/abs/2111.03943v2 | https://arxiv.org/pdf/2111.03943v2.pdf | A Probit Tensor Factorization Model For Relational Learning | With the proliferation of knowledge graphs, modeling data with complex multirelational structure has gained increasing attention in the area of statistical relational learning. One of the most important goals of statistical relational learning is link prediction, i.e., predicting whether certain relations exist in the knowledge graph. A large number of models and algorithms have been proposed to perform link prediction, among which tensor factorization method has proven to achieve state-of-the-art performance in terms of computation efficiency and prediction accuracy. However, a common drawback of the existing tensor factorization models is that the missing relations and non-existing relations are treated in the same way, which results in a loss of information. To address this issue, we propose a binary tensor factorization model with probit link, which not only inherits the computation efficiency from the classic tensor factorization model but also accounts for the binary nature of relational data. Our proposed probit tensor factorization (PTF) model shows advantages in both the prediction accuracy and interpretability | ['Yanghua Xiao', 'Wenbin Lu', 'Rui Song', 'Ye Liu'] | 2021-11-06 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [-2.94820487e-01 2.66714618e-02 -5.00826776e-01 -2.51888484e-01
1.73789784e-01 -3.38021100e-01 4.86161351e-01 6.35979950e-01
9.50270966e-02 7.08912551e-01 6.19424954e-02 -4.30389553e-01
-8.38528693e-01 -1.11344826e+00 -3.17565411e-01 -4.41394687e-01
-2.63120860e-01 6.47966802e-01 3.49387437e-01 -2.83885539e-01
-9.07459855e-02 4.08188105e-01 -1.34862232e+00 2.87552446e-01
9.51349139e-01 1.03733051e+00 -2.43990466e-01 1.21847250e-01
-2.90805459e-01 1.22115076e+00 1.30426195e-02 -1.18837535e+00
-1.21521726e-01 -3.32008712e-02 -9.91147816e-01 -2.98961610e-01
4.32448694e-03 3.23458575e-02 -8.29015553e-01 9.12378192e-01
-1.12515345e-01 -8.84441473e-03 3.42811733e-01 -1.51131403e+00
-6.66022897e-01 8.81192386e-01 -6.29493654e-01 2.75211334e-01
2.19851092e-01 -6.63312435e-01 1.66272771e+00 -6.44730449e-01
3.50727826e-01 1.25407338e+00 5.38383782e-01 -1.50262266e-01
-1.44174671e+00 -5.47135890e-01 6.80552796e-02 6.55983329e-01
-1.51340425e+00 5.20654842e-02 6.61404371e-01 -5.76395988e-01
6.54055297e-01 4.07275647e-01 6.31305337e-01 3.70975584e-01
1.47675812e-01 7.57393837e-01 1.04896009e+00 -2.41090164e-01
-2.05457315e-01 -5.31439809e-03 3.91391635e-01 7.05606997e-01
4.89151657e-01 -6.84362054e-02 -6.68008804e-01 -2.10286051e-01
5.88730991e-01 2.50165433e-01 -8.55366960e-02 -5.25370955e-01
-1.16106808e+00 7.72048414e-01 6.87643826e-01 5.20553291e-01
-3.18075210e-01 7.13062361e-02 3.80498648e-01 3.03252935e-01
5.91034770e-01 2.08339766e-01 -4.31805938e-01 -9.63711217e-02
-4.55741972e-01 8.80522653e-02 6.44255579e-01 5.85671604e-01
8.50858986e-01 -3.04592997e-01 1.01138704e-01 8.26681435e-01
3.11161399e-01 -2.44819763e-04 1.61750942e-01 -4.13196981e-01
6.17980242e-01 1.30542266e+00 -2.15609848e-01 -1.57480347e+00
-3.53312135e-01 -8.07480931e-01 -1.08909440e+00 -3.67138326e-01
2.58088052e-01 3.40866178e-01 -3.63929629e-01 1.42555654e+00
4.23651189e-01 5.67103475e-02 5.75091802e-02 6.48164213e-01
8.16011965e-01 3.81508380e-01 1.00414716e-01 -2.55702674e-01
1.31755555e+00 -5.89108586e-01 -8.82434487e-01 1.73222944e-01
7.25542843e-01 -8.18221211e-01 4.67569381e-01 2.12056667e-01
-5.76949835e-01 -3.57146591e-01 -7.70396471e-01 -1.71125472e-01
-3.08468878e-01 1.14007778e-01 1.56547546e+00 4.33274269e-01
-5.70756614e-01 4.94805068e-01 -8.32045436e-01 -2.24213958e-01
2.91025788e-01 4.27272677e-01 -6.89892828e-01 -4.17847753e-01
-1.27663159e+00 7.33093023e-01 5.15214026e-01 2.23427892e-01
-5.32961600e-02 -6.22421622e-01 -5.17525434e-01 2.71811128e-01
8.06455135e-01 -5.81950963e-01 6.47286177e-01 -5.11166573e-01
-9.15580928e-01 2.21885756e-01 -1.94517732e-01 -1.70367166e-01
3.10268909e-01 -3.12348336e-01 -6.75185978e-01 -1.15011290e-01
1.60845071e-02 -1.38064831e-01 3.40587378e-01 -1.12486064e+00
-5.15366733e-01 -6.84178054e-01 4.51784164e-01 -1.27901696e-02
-5.71757555e-01 -2.95277178e-01 -5.22966683e-01 -6.26887321e-01
6.64191663e-01 -1.00559115e+00 2.02633850e-02 -2.69469380e-01
-6.26295507e-01 -5.48564434e-01 7.66305029e-01 -4.87311274e-01
1.64436042e+00 -2.00219083e+00 4.44700539e-01 5.13643444e-01
6.30000174e-01 1.77177444e-01 2.30806798e-01 7.63510883e-01
-2.51343042e-01 7.12839887e-02 2.15470523e-01 -4.85170595e-02
-3.69925052e-01 4.32797700e-01 -2.80244470e-01 -1.43269468e-02
1.55270159e-01 8.84474695e-01 -8.94466996e-01 -5.97975850e-01
-2.31898408e-02 4.18474913e-01 -4.62344617e-01 5.79305738e-02
-9.90091786e-02 2.01215222e-01 -5.79572797e-01 4.88037467e-01
5.08956671e-01 -5.21245897e-01 7.61216342e-01 -6.99627042e-01
5.82220964e-02 2.85968691e-01 -1.27707338e+00 1.13855457e+00
-3.12409885e-02 1.21103592e-01 -3.34350735e-01 -1.13943553e+00
8.58869314e-01 3.61090422e-01 9.60082769e-01 -4.21061039e-01
-9.95663553e-02 1.38324216e-01 4.27967906e-01 -4.30640876e-01
6.82000995e-01 -1.32942632e-01 3.21693569e-01 4.21388030e-01
-2.05495432e-02 4.13728952e-01 4.85097021e-01 5.70414484e-01
1.12271893e+00 1.82608457e-03 2.69615501e-01 -9.34021845e-02
6.95623815e-01 -1.36514947e-01 5.85285246e-01 2.17310581e-02
4.48714584e-01 -1.08538352e-01 9.87488151e-01 -5.68044901e-01
-6.10067070e-01 -8.17208171e-01 -6.80292249e-02 7.41248727e-01
6.64373711e-02 -9.51132238e-01 5.85203059e-02 -6.09625399e-01
4.56589431e-01 9.77774262e-02 -6.85065687e-01 -1.55106008e-01
-2.61347055e-01 -7.61512697e-01 2.32306644e-01 4.95406419e-01
2.71955132e-01 -3.75148922e-01 2.77352810e-01 1.10556185e-01
-4.39952761e-01 -1.31994152e+00 1.52928695e-01 9.42915231e-02
-1.18768883e+00 -1.38536739e+00 7.35020041e-02 -2.57635415e-01
6.54926181e-01 5.83575666e-01 9.54265118e-01 3.94072294e-01
2.65845284e-02 3.09460144e-02 -6.35137439e-01 1.06658861e-01
-1.24917984e-01 2.53205329e-01 1.23324007e-01 4.25961852e-01
2.52861142e-01 -5.74970305e-01 -2.85870135e-01 4.75180954e-01
-1.04479158e+00 1.48706630e-01 7.20147967e-01 7.83797264e-01
5.56702256e-01 7.23219872e-01 5.66386282e-01 -9.69681084e-01
5.87216556e-01 -5.34430146e-01 -4.05455798e-01 6.01058364e-01
-9.65446413e-01 3.22137177e-01 4.45152670e-01 -1.35572627e-01
-7.78144836e-01 -3.00752401e-01 2.78972536e-01 -2.53097445e-01
5.22552490e-01 1.31762362e+00 -1.78815231e-01 -1.28331944e-01
2.16784924e-01 -1.76221114e-02 6.33922815e-02 -5.72322726e-01
3.98660749e-01 2.58346647e-01 3.84836346e-02 -6.70991540e-01
9.43308592e-01 3.02873462e-01 7.31487572e-01 -5.61120450e-01
-8.00686300e-01 -5.01369178e-01 -8.81181479e-01 -1.33412689e-01
4.97389138e-01 -7.76934683e-01 -1.12790990e+00 1.06360890e-01
-9.55128074e-01 4.35602278e-01 1.53153554e-01 6.42354667e-01
3.67988274e-02 5.41970372e-01 -6.22229397e-01 -8.76854897e-01
3.82652879e-02 -9.20831680e-01 3.82453382e-01 -1.70788795e-01
-9.77972299e-02 -9.39252198e-01 -1.38726577e-01 6.87576592e-01
1.36339322e-01 8.83129388e-02 1.63178504e+00 -4.07343477e-01
-8.94453108e-01 -4.50937718e-01 -6.04983568e-01 1.73742384e-01
2.61527270e-01 -6.30008196e-03 -4.29592162e-01 2.05825120e-02
-6.09953463e-01 -1.94148600e-01 7.13119984e-01 -6.09125942e-02
7.76896775e-01 -1.64796278e-01 -5.01461744e-01 1.09369203e-01
1.32437193e+00 4.19760086e-02 5.22442043e-01 5.69877736e-02
1.13849473e+00 7.31989861e-01 6.52442694e-01 2.36195073e-01
7.42497385e-01 8.63353431e-01 4.84364092e-01 8.37409645e-02
1.62971497e-01 -4.73862469e-01 -1.94927603e-01 1.02203357e+00
-5.30669332e-01 9.21889488e-03 -9.29485738e-01 3.17843467e-01
-2.26514316e+00 -8.89065385e-01 -6.75150871e-01 2.21228790e+00
7.91488171e-01 1.33093908e-01 1.85414135e-01 6.23086095e-01
4.64023709e-01 -1.13598099e-02 -8.44032019e-02 -8.19122270e-02
-2.76565552e-02 -6.76916018e-02 1.62817478e-01 3.96044850e-01
-8.53621840e-01 7.80239940e-01 5.29318571e+00 6.20146453e-01
-9.16450262e-01 -6.73417002e-03 3.57691139e-01 3.55955273e-01
-5.18856108e-01 2.71671653e-01 -5.32776475e-01 1.46754175e-01
5.13047874e-01 -5.94439059e-02 4.82640326e-01 6.21946335e-01
-6.90812990e-02 -1.23827383e-01 -1.12168038e+00 9.10351694e-01
-2.38060310e-01 -1.15172017e+00 2.26570874e-01 4.08128589e-01
5.60844600e-01 -3.42800021e-01 -1.57679185e-01 1.90750167e-01
1.90422848e-01 -8.59595299e-01 3.91329229e-01 6.22177660e-01
4.42387432e-01 -8.78145635e-01 9.34103847e-01 1.56160966e-01
-1.40226603e+00 -9.87204313e-02 -1.84030443e-01 -3.07390362e-01
-4.63826954e-02 1.25777292e+00 -6.38571203e-01 1.21308756e+00
5.64633548e-01 8.55382740e-01 -6.00715280e-01 1.05872345e+00
-2.65675068e-01 4.30844933e-01 -2.09209964e-01 9.96574238e-02
-1.86194301e-01 -4.57821786e-01 1.57600775e-01 5.48089802e-01
1.31390750e-01 1.81390673e-01 3.04726124e-01 4.80791062e-01
-1.30588546e-01 3.53048027e-01 -3.61530930e-01 -4.50262874e-01
3.96161079e-01 1.20319068e+00 -6.71262622e-01 -1.99780725e-02
-6.38452947e-01 2.98845142e-01 5.85323036e-01 1.57721385e-01
-4.28939730e-01 5.91821857e-02 6.47871256e-01 2.78233677e-01
1.12450533e-01 -5.27740359e-01 -3.99619311e-01 -1.25621760e+00
2.12962866e-01 -4.98894006e-01 5.37582576e-01 -3.46405953e-01
-1.39128685e+00 5.34234464e-01 9.22200084e-02 -1.03227437e+00
-9.26191732e-03 -5.09214461e-01 5.72390035e-02 7.06169605e-01
-1.42636883e+00 -1.43908441e+00 -3.23911250e-01 5.79708517e-01
-2.42801949e-01 -1.04771485e-03 7.81988740e-01 6.58423364e-01
-7.00801969e-01 4.52637196e-01 -5.00262193e-02 1.71896040e-01
3.77223402e-01 -1.10210657e+00 -9.89513621e-02 6.00608885e-01
3.25496316e-01 8.94727588e-01 5.01577079e-01 -8.06944132e-01
-1.62922025e+00 -8.58953059e-01 1.28157139e+00 -3.41397434e-01
1.14408469e+00 -8.08900595e-02 -1.00088251e+00 6.60346925e-01
-5.29370964e-01 2.04694241e-01 9.99436498e-01 9.36318576e-01
-7.64123321e-01 -6.08646274e-01 -6.83742583e-01 4.76788133e-01
1.00385547e+00 -6.57181084e-01 -2.65151203e-01 1.56561479e-01
4.47959155e-01 4.21332270e-02 -1.39178979e+00 6.50679410e-01
7.82942235e-01 -8.80276978e-01 9.67982113e-01 -6.34349704e-01
6.62061691e-01 -5.79209745e-01 -2.04364702e-01 -9.84845459e-01
-6.11552060e-01 -2.01260269e-01 -6.06021941e-01 1.36941254e+00
4.21910226e-01 -6.10385716e-01 8.00427675e-01 6.35608852e-01
3.58906299e-01 -1.02003586e+00 -8.63022149e-01 -7.60052204e-01
-4.58074927e-01 -4.88968939e-01 5.65280974e-01 1.19339633e+00
3.00546080e-01 8.55841994e-01 -5.43518126e-01 1.09308921e-01
5.13899803e-01 5.03669679e-01 7.33038068e-01 -1.79452348e+00
-4.19874936e-01 -1.85843229e-01 -8.14354002e-01 -7.88647592e-01
-6.06518947e-02 -1.11246192e+00 -6.62093937e-01 -1.69395804e+00
4.16980237e-01 -9.65568066e-01 -4.08703893e-01 6.69672906e-01
-2.49271765e-01 1.91322595e-01 1.50995716e-01 6.45556092e-01
-4.84694898e-01 5.91413736e-01 1.23976672e+00 -8.32054839e-02
-9.07553360e-02 1.78085640e-01 -7.58870125e-01 2.61945933e-01
3.09803933e-01 -3.37792754e-01 -6.68651879e-01 -3.53964090e-01
8.93122673e-01 2.81413406e-01 2.43069395e-01 -6.47989452e-01
2.48923153e-01 -1.63132787e-01 5.82080334e-02 -5.57363629e-01
3.84006381e-01 -1.11531734e+00 5.51741838e-01 4.03378278e-01
-8.97103399e-02 2.59677976e-01 -1.76112399e-01 8.01840723e-01
-5.09269834e-01 1.60294741e-01 1.52338356e-01 3.05389225e-01
-5.25473952e-01 4.77525353e-01 2.16019943e-01 -5.14817953e-01
8.96768987e-01 1.14087604e-01 -5.32296598e-01 -1.43913195e-01
-7.13955104e-01 1.14778772e-01 -1.24692721e-02 6.51444077e-01
5.30380607e-01 -1.49242437e+00 -4.74898219e-01 -7.36957341e-02
3.57487142e-01 -1.44793272e-01 1.79717362e-01 1.15342295e+00
-1.54345229e-01 6.69530869e-01 1.30199566e-01 -3.71532857e-01
-1.25145233e+00 7.00097740e-01 -2.23086193e-01 -7.38392830e-01
-4.51978207e-01 5.08976340e-01 -7.36654177e-02 -1.65704086e-01
-7.23182857e-02 -1.53890297e-01 -3.63713771e-01 2.67695814e-01
1.94219336e-01 5.22637129e-01 2.18000725e-01 -7.19529510e-01
-4.83758032e-01 3.77755493e-01 -2.37334102e-01 4.15073037e-01
1.35935128e+00 1.15501545e-01 -8.70234191e-01 5.27238965e-01
8.26095223e-01 9.14659277e-02 -4.08554226e-01 -6.66860580e-01
3.09540033e-01 -6.19243026e-01 1.41072646e-01 -8.18253100e-01
-1.19982874e+00 6.46906674e-01 3.38986181e-02 5.89182556e-01
8.01684499e-01 1.74417436e-01 5.93841136e-01 3.63338292e-01
6.01400077e-01 -5.59644997e-01 1.15169086e-01 5.50726593e-01
6.59966528e-01 -1.19546330e+00 5.17994583e-01 -1.10818803e+00
-4.41036910e-01 1.10501444e+00 5.02234697e-01 2.29815558e-01
9.69669640e-01 -6.11729771e-02 -3.65165234e-01 -3.96939725e-01
-6.81925833e-01 -2.57623374e-01 5.91714978e-01 3.16789716e-01
7.75353968e-01 4.43811446e-01 -6.49309695e-01 5.28010190e-01
-2.08619818e-01 2.73496695e-02 4.84211892e-02 5.70164979e-01
1.16552236e-02 -1.52475405e+00 -2.05800593e-01 6.75810099e-01
-4.32369739e-01 -1.06620140e-01 -4.12956476e-01 6.10454857e-01
1.25856712e-01 1.10579145e+00 -3.16312939e-01 -8.70962262e-01
2.21583858e-01 -3.90651822e-02 3.59646201e-01 -5.27533889e-01
-3.13166887e-01 -1.55693471e-01 2.00190380e-01 -3.71190429e-01
-6.45867527e-01 -5.70402801e-01 -1.14613712e+00 -6.56489193e-01
-6.24607325e-01 5.25442839e-01 6.40270591e-01 1.11398911e+00
4.21263874e-01 7.02217937e-01 5.23545802e-01 -1.23252915e-02
-2.11574942e-01 -9.03035045e-01 -7.78926790e-01 3.60610217e-01
1.32241324e-02 -1.12307727e+00 4.33614440e-02 -2.80792028e-01] | [8.640690803527832, 7.7816481590271] |
8ebf78fb-ce40-41a2-86de-52d47b348283 | newsnet-a-novel-dataset-for-hierarchical | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_NewsNet_A_Novel_Dataset_for_Hierarchical_Temporal_Segmentation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_NewsNet_A_Novel_Dataset_for_Hierarchical_Temporal_Segmentation_CVPR_2023_paper.pdf | NewsNet: A Novel Dataset for Hierarchical Temporal Segmentation | Temporal video segmentation is the get-to-go automatic video analysis, which decomposes a long-form video into smaller components for the following-up understanding tasks. Recent works have studied several levels of granularity to segment a video, such as shot, event, and scene. Those segmentations can help compare the semantics in the corresponding scales, but lack a wider view of larger temporal spans, especially when the video is complex and structured. Therefore, we present two abstractive levels of temporal segmentations and study their hierarchy to the existing fine-grained levels. Accordingly, we collect NewsNet, the largest news video dataset consisting of 1,000 videos in over 900 hours, associated with several tasks for hierarchical temporal video segmentation. Each news video is a collection of stories on different topics, represented as aligned audio, visual, and textual data, along with extensive frame-wise annotations in four granularities. We assert that the study on NewsNet can advance the understanding of complex structured video and benefit more areas such as short-video creation, personalized advertisement, digital instruction, and education. Our dataset and code is publicly available at: https://github.com/NewsNet-Benchmark/NewsNet. | ['Bernard Ghanem', 'Chia-Wen Lin', 'Raghavendra Ramachandra', 'Wentian Zhang', 'Mengmeng Xu', 'Bo Ren', 'Liangsheng Xu', 'Bei Gan', 'Xiujun Shu', 'Ruizhi Qiao', 'Bing Li', 'Mingchen Zhuge', 'Haozhe Liu', 'Keyu Chen', 'Haoqian Wu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-semantic-segmentation'] | ['computer-vision'] | [ 1.35611400e-01 -2.36530617e-01 -7.41569340e-01 -3.70843679e-01
-5.55201709e-01 -5.85892320e-01 2.50721008e-01 3.11314672e-01
-2.25232840e-01 5.10363102e-01 6.72603846e-01 -1.49000049e-01
-2.25574709e-02 -5.13307333e-01 -8.48941624e-01 -4.83384281e-01
-1.90225661e-01 -5.17809093e-02 9.26199853e-01 -1.74399152e-01
3.39862257e-01 -1.11332603e-01 -2.01485920e+00 8.15180302e-01
5.98042667e-01 1.23897660e+00 3.28110993e-01 7.18444228e-01
-2.53421187e-01 1.28072941e+00 -6.36279285e-01 -2.71245658e-01
-6.31289780e-02 -6.01509809e-01 -1.01912093e+00 4.23245639e-01
6.82532072e-01 -4.47256058e-01 -5.42564571e-01 1.10037696e+00
1.41309977e-01 3.80821645e-01 5.95832653e-02 -1.48867774e+00
-1.53779298e-01 1.23467791e+00 -6.20865047e-01 7.51952171e-01
8.17868710e-01 4.83056381e-02 1.05252731e+00 -5.41560829e-01
1.02384949e+00 1.08403099e+00 5.10161459e-01 1.54825643e-01
-6.19229555e-01 -5.34251094e-01 5.30916035e-01 8.07756305e-01
-1.10077572e+00 -2.63215721e-01 6.10208690e-01 -8.89858723e-01
6.04155660e-01 2.28848919e-01 1.07201171e+00 1.27732182e+00
5.51333353e-02 1.25319719e+00 7.13860214e-01 -9.71057639e-03
-1.59102287e-02 -4.72472340e-01 5.71328402e-01 6.11262739e-01
-1.72697976e-01 -5.13515413e-01 -9.31430638e-01 3.88694853e-01
6.33551538e-01 1.68889686e-01 -4.89822388e-01 -1.50604650e-01
-1.48352945e+00 5.08079290e-01 -9.62030962e-02 4.68770623e-01
-2.17480183e-01 -4.78346236e-02 9.15573180e-01 3.81485611e-01
4.38886881e-01 1.84929013e-01 -4.42640871e-01 -8.38082433e-01
-1.22229624e+00 1.59689158e-01 7.60051012e-01 1.38654983e+00
6.43338382e-01 2.62886044e-02 -3.52693647e-01 5.52425206e-01
-1.35528818e-01 -2.02350304e-01 6.62644863e-01 -1.39470327e+00
5.16660273e-01 4.97620463e-01 -2.87587285e-01 -8.32256854e-01
-1.42362073e-01 -1.80610381e-02 -3.57118636e-01 -5.88018477e-01
4.34786022e-01 4.11743820e-02 -8.78376245e-01 1.36191928e+00
4.86940920e-01 8.24869692e-01 -3.29018384e-01 7.97518253e-01
1.43284845e+00 9.06133950e-01 9.41181369e-03 -4.75841343e-01
1.78424978e+00 -1.47502041e+00 -9.97810304e-01 5.69695458e-02
5.70412755e-01 -8.61684382e-01 1.10970473e+00 6.17109179e-01
-1.15782750e+00 -7.08789706e-01 -9.85959768e-01 -2.83682078e-01
-3.95245671e-01 -3.17759722e-01 3.20435047e-01 2.54642010e-01
-8.48560631e-01 6.34503186e-01 -7.86937475e-01 -5.03257275e-01
2.84746826e-01 -2.06402779e-01 -1.80095941e-01 -2.27369845e-01
-1.21549571e+00 8.46585333e-02 6.07449114e-01 -3.60405385e-01
-1.15338778e+00 -9.05821860e-01 -9.88864005e-01 -2.54292823e-02
1.01785946e+00 -1.76650643e-01 1.39482629e+00 -9.86513913e-01
-1.30347919e+00 7.62284875e-01 -5.48464358e-02 -4.47765380e-01
3.37387711e-01 -3.80580664e-01 -3.32626998e-01 5.75570345e-01
2.60080427e-01 6.74177170e-01 7.69866943e-01 -7.43846893e-01
-9.62128937e-01 -8.27081352e-02 3.90989780e-01 2.18102038e-01
-4.70726311e-01 3.42515975e-01 -1.19084179e+00 -9.74434435e-01
-1.26230299e-01 -7.23014474e-01 2.57283747e-02 -6.98442340e-01
-3.91528249e-01 -1.77535251e-01 9.37407553e-01 -7.00658798e-01
1.79821348e+00 -2.45488334e+00 4.07557368e-01 -4.00803477e-01
5.01885533e-01 -3.65524471e-01 5.41295763e-03 3.87257576e-01
-7.82015622e-02 2.72069693e-01 1.86118737e-01 -5.91388009e-02
-1.19473837e-01 -3.70814018e-02 -2.92524904e-01 1.62052199e-01
-4.93284315e-01 6.30946875e-01 -1.06599534e+00 -9.70288932e-01
9.86183807e-03 1.61392644e-01 -5.49798250e-01 2.73319304e-01
-3.23486447e-01 3.80967855e-01 -6.07768655e-01 9.00261521e-01
1.25241846e-01 -2.29301199e-01 -5.35975397e-02 -4.26277936e-01
-4.14253145e-01 2.83934236e-01 -9.63076115e-01 1.87141764e+00
1.48854718e-01 1.10041034e+00 3.03702690e-02 -1.00987041e+00
2.61110157e-01 5.01257122e-01 1.11999810e+00 -6.48541093e-01
2.16573536e-01 -1.79228336e-01 -3.04211289e-01 -8.72609556e-01
9.36837554e-01 4.24363762e-01 -4.59421098e-01 2.46764719e-01
3.20632875e-01 -1.80444438e-02 1.07527936e+00 4.71932918e-01
1.22582531e+00 3.07899952e-01 2.52177089e-01 -2.89433122e-01
1.63852334e-01 3.28897685e-01 8.05419683e-01 4.68332082e-01
-5.65820456e-01 5.75768769e-01 8.45772624e-01 -3.99707437e-01
-7.52754271e-01 -8.51218939e-01 -5.88232502e-02 1.70635891e+00
5.90968549e-01 -1.05010819e+00 -1.08797967e+00 -5.79563439e-01
-3.40157837e-01 2.64916331e-01 -4.21973914e-01 2.10708648e-01
-7.16453850e-01 -3.95991862e-01 3.62379640e-01 4.12617385e-01
4.31564242e-01 -1.07479942e+00 -5.27934194e-01 1.33355558e-01
-8.15341532e-01 -1.58625436e+00 -9.47004318e-01 -1.35518676e-02
-7.42599547e-01 -1.24913359e+00 -5.37736535e-01 -9.44655776e-01
2.63360709e-01 6.31492674e-01 1.38176882e+00 -6.10623276e-03
-1.84365660e-01 7.79021740e-01 -9.11440849e-01 -9.65114608e-02
1.56777464e-02 2.76718643e-02 -7.51859769e-02 -9.21883881e-02
3.44161838e-01 -4.87822086e-01 -5.99714160e-01 7.17445970e-01
-1.07701397e+00 3.70164037e-01 9.54188779e-02 2.50868350e-01
8.66957843e-01 2.48765185e-01 1.94976211e-01 -9.72499013e-01
3.45444530e-01 -8.42146933e-01 -2.86752939e-01 1.31322145e-01
4.68385033e-02 -5.50197184e-01 5.43837130e-01 -7.04877079e-01
-9.36590970e-01 -2.81997144e-01 1.26672238e-01 -5.60451925e-01
-3.78476024e-01 7.09165096e-01 -9.36373547e-02 2.36486480e-01
2.92899847e-01 1.71569958e-01 -6.02930784e-01 -3.59708071e-01
2.91712046e-01 3.15848768e-01 6.34396195e-01 -8.55113387e-01
3.27497840e-01 2.31559217e-01 -4.46429342e-01 -1.01908171e+00
-1.04538572e+00 -8.24453771e-01 -8.04167211e-01 -8.58284831e-01
1.28196454e+00 -1.15466917e+00 -4.98084009e-01 4.29387271e-01
-7.18817651e-01 -6.15928173e-01 -3.98877710e-01 4.51995611e-01
-7.41528690e-01 4.98878449e-01 -1.21174145e+00 9.00937691e-02
4.03029174e-01 -1.55523396e+00 9.20026243e-01 3.21496934e-01
-2.24170446e-01 -7.30189323e-01 -2.59757757e-01 8.76535177e-01
-1.09290108e-01 2.85560697e-01 5.48458755e-01 -6.05398238e-01
-7.87047744e-01 1.42704278e-01 1.13389276e-01 6.30948916e-02
-1.55012757e-01 5.20714045e-01 -6.80290103e-01 -2.06133872e-01
1.49599448e-01 -5.10097921e-01 9.28569198e-01 5.09101510e-01
1.65143764e+00 -3.64087194e-01 -2.08919019e-01 5.60295641e-01
1.06693172e+00 5.33378661e-01 5.13093174e-01 4.07568991e-01
1.08520448e+00 6.51647091e-01 1.02131081e+00 4.94941920e-01
5.57352781e-01 5.13456285e-01 3.07971627e-01 3.28017056e-01
-1.27490312e-01 -8.93313587e-02 6.73667192e-01 1.67963386e+00
-2.13950783e-01 -4.40527916e-01 -9.35260296e-01 6.70581043e-01
-1.93992352e+00 -1.28760839e+00 -4.68493365e-02 1.73043013e+00
8.92126739e-01 3.97425592e-01 4.29980218e-01 1.53558552e-02
9.14226294e-01 6.12283945e-01 -5.02987146e-01 9.01806429e-02
1.83211595e-01 -4.07228559e-01 2.20593870e-01 1.50492772e-01
-1.27510834e+00 9.70829606e-01 5.86837339e+00 1.32559586e+00
-9.64850485e-01 1.38944343e-01 7.86818862e-01 -2.95208842e-01
-1.79749846e-01 -8.26116949e-02 -8.60562325e-01 6.68992519e-01
1.07448411e+00 -3.76123309e-01 1.89324081e-01 7.90695012e-01
2.94079334e-01 -2.34207347e-01 -1.29249394e+00 8.21318805e-01
1.59693554e-01 -1.57169461e+00 3.29669900e-02 -2.47062683e-01
9.95046973e-01 -1.02871008e-01 -9.93832201e-02 4.96220559e-01
1.77164465e-01 -6.58569157e-01 1.12885821e+00 3.55352968e-01
5.83217204e-01 -4.75946009e-01 4.55111623e-01 1.56970054e-01
-1.92537022e+00 -8.26657563e-03 1.21197395e-01 5.30810617e-02
2.48380437e-01 4.06548321e-01 -7.41384774e-02 4.44784284e-01
1.12178981e+00 1.30250573e+00 -5.16140461e-01 1.11577857e+00
1.24634944e-01 8.04864883e-01 6.06831834e-02 2.88431972e-01
3.05392444e-01 -2.24444762e-01 5.41600347e-01 1.35567951e+00
3.74146938e-01 3.59803110e-01 7.88725913e-01 1.35423154e-01
-2.95440495e-01 1.25277519e-01 -2.79278845e-01 -2.60461837e-01
5.65725505e-01 1.10029447e+00 -1.43378222e+00 -7.45650291e-01
-7.17785120e-01 6.20091677e-01 -1.89804047e-01 4.67515290e-01
-1.30559254e+00 -3.10329288e-01 7.09802151e-01 4.32745427e-01
3.35809827e-01 -3.06550354e-01 1.81032792e-01 -1.34360600e+00
-4.92646880e-02 -1.04700625e+00 7.13473320e-01 -8.49920511e-01
-7.83280969e-01 5.63363016e-01 4.71258432e-01 -1.45139074e+00
-5.19936718e-03 -2.25391179e-01 -4.58656460e-01 -2.90499926e-01
-1.05211902e+00 -5.37931859e-01 -6.67283118e-01 6.82604969e-01
1.47975266e+00 1.21875808e-01 -6.81161359e-02 6.90132380e-01
-9.71596599e-01 2.00587511e-01 -1.74559057e-01 2.29912773e-01
8.55548561e-01 -1.09375751e+00 1.61614008e-02 9.78355050e-01
9.35415998e-02 2.32948422e-01 7.56656885e-01 -7.28110850e-01
-1.25692093e+00 -1.07659864e+00 4.23716396e-01 -3.21214229e-01
1.14093769e+00 -1.74003348e-01 -7.73136079e-01 8.45306516e-01
5.56558788e-01 -1.10288180e-01 8.05453062e-01 -1.14710413e-01
-8.53954256e-02 4.02137302e-02 -6.24605119e-01 6.35786176e-01
1.37808990e+00 -6.45380259e-01 -5.88817835e-01 3.37464064e-01
1.46645534e+00 -7.00149119e-01 -1.24311924e+00 3.71349841e-01
4.09535438e-01 -1.25611186e+00 9.41384256e-01 -4.57278877e-01
9.80446696e-01 -1.40009180e-01 -1.89372689e-01 -8.66830826e-01
-1.15600832e-01 -7.75775850e-01 -2.74192601e-01 1.27456474e+00
1.20604269e-01 1.03906065e-01 8.19434941e-01 1.36479124e-01
-6.88719928e-01 -8.56718719e-01 -6.21417999e-01 -6.48772299e-01
-6.06145620e-01 -8.21148217e-01 4.47732568e-01 1.19482696e+00
1.54460296e-01 1.34893179e-01 -3.54815871e-01 -1.54038355e-01
3.97043586e-01 2.96635002e-01 5.55772185e-01 -9.48696554e-01
3.91784497e-02 -6.82465792e-01 -4.58137929e-01 -1.45218444e+00
1.20611072e-01 -4.79758352e-01 1.20559514e-01 -1.38733530e+00
3.69078100e-01 2.76823789e-01 -2.91859716e-01 3.38768005e-01
-3.19920368e-02 2.77806491e-01 2.67060876e-01 3.41484487e-01
-1.49238002e+00 2.89498836e-01 1.55677164e+00 -2.24092200e-01
-8.18009600e-02 -2.62579769e-01 -2.76716858e-01 1.01991487e+00
5.34237623e-01 -3.12999249e-01 -6.19749784e-01 -2.87241668e-01
2.22062260e-01 4.39396650e-01 -1.97872162e-01 -1.22984684e+00
3.55204642e-01 -6.14431202e-01 -1.65871158e-01 -6.59635484e-01
1.26570731e-01 -6.84904277e-01 1.57498732e-01 1.43709987e-01
-4.86726969e-01 1.34726316e-01 1.38190940e-01 4.86770302e-01
-9.08255458e-01 -1.73492543e-02 4.60003465e-01 -2.49358818e-01
-1.51030374e+00 6.73779905e-01 -6.05908811e-01 4.85554516e-01
1.36547971e+00 -5.81350803e-01 -4.42255765e-01 -4.12097603e-01
-1.02738881e+00 4.58660513e-01 4.26955193e-01 6.32940412e-01
5.03626287e-01 -1.33872378e+00 -2.73418188e-01 -1.65329650e-01
1.72636777e-01 1.02798857e-01 6.67690992e-01 9.45763767e-01
-6.19572222e-01 1.92114517e-01 -3.52214336e-01 -8.81311178e-01
-1.54135966e+00 7.43873775e-01 -2.18825668e-01 -4.83891554e-02
-7.00189710e-01 8.31597388e-01 4.83762085e-01 3.56405199e-01
5.78465939e-01 -6.94160163e-01 -6.89567208e-01 8.17356288e-01
5.93829572e-01 4.90761220e-01 -3.53128850e-01 -7.76446104e-01
1.32029140e-02 5.46875477e-01 2.36668419e-02 3.88740063e-01
1.17861533e+00 -4.42130953e-01 -1.89223960e-01 9.45451558e-01
1.10888660e+00 4.37404178e-02 -1.48804462e+00 -2.58256644e-01
9.38870460e-02 -4.73038942e-01 -1.63636938e-01 5.33939339e-02
-1.23912489e+00 5.41691720e-01 6.96502700e-02 7.49619186e-01
1.28921175e+00 2.26926789e-01 1.09140599e+00 7.04403743e-02
2.87856996e-01 -1.25205040e+00 5.48785210e-01 7.20054865e-01
5.17793357e-01 -1.08714759e+00 1.29501194e-01 -7.19612062e-01
-6.42379344e-01 1.04633594e+00 9.49763060e-01 2.86008209e-01
6.95556819e-01 4.14173365e-01 -2.33441979e-01 -3.88705134e-01
-8.33192050e-01 -2.41994441e-01 4.44974512e-01 -4.71412018e-02
5.99622011e-01 1.05868407e-01 -1.12627894e-01 7.70807743e-01
-2.90476114e-01 -9.01883021e-02 8.79689634e-01 1.10263586e+00
-7.30723381e-01 -7.05338597e-01 -2.87818819e-01 4.43849415e-01
-7.35219359e-01 -4.19730283e-02 -1.31264612e-01 7.68362641e-01
2.43199825e-01 8.06579828e-01 3.98282886e-01 -5.10608971e-01
1.38121068e-01 -8.34078714e-02 2.73493201e-01 -6.90787256e-01
-3.91614288e-01 2.49746487e-01 1.93285525e-01 -1.04552603e+00
-7.83826768e-01 -6.99301481e-01 -1.14908791e+00 -4.59955841e-01
2.20668659e-01 2.69800186e-01 1.90400437e-01 8.45086396e-01
5.80435097e-02 1.03005052e+00 3.59779745e-01 -1.02020860e+00
4.31784362e-01 -6.92127526e-01 -3.22014332e-01 5.76152563e-01
2.68853247e-01 -5.59289455e-01 -2.25902885e-01 6.84858322e-01] | [9.341115951538086, 0.5558005571365356] |
570af657-8b27-4dec-ad67-a800ac590b4d | modeling-context-in-referring-expressions | 1608.00272 | null | http://arxiv.org/abs/1608.00272v3 | http://arxiv.org/pdf/1608.00272v3.pdf | Modeling Context in Referring Expressions | Humans refer to objects in their environments all the time, especially in
dialogue with other people. We explore generating and comprehending natural
language referring expressions for objects in images. In particular, we focus
on incorporating better measures of visual context into referring expression
models and find that visual comparison to other objects within an image helps
improve performance significantly. We also develop methods to tie the language
generation process together, so that we generate expressions for all objects of
a particular category jointly. Evaluation on three recent datasets - RefCOCO,
RefCOCO+, and RefCOCOg, shows the advantages of our methods for both referring
expression generation and comprehension. | ['Alexander C. Berg', 'Tamara L. Berg', 'Licheng Yu', 'Shan Yang', 'Patrick Poirson'] | 2016-07-31 | null | null | null | null | ['referring-expression-generation'] | ['computer-vision'] | [ 1.37669966e-01 2.70166636e-01 1.57253563e-01 -5.71178675e-01
-4.77393150e-01 -7.00618923e-01 9.29245651e-01 2.77306795e-01
-2.87378550e-01 6.48784995e-01 8.57664227e-01 -5.87661155e-02
3.37584257e-01 -5.70529044e-01 -4.86025691e-01 -1.63642555e-01
3.37286294e-01 4.56505477e-01 3.85467559e-02 -3.87877107e-01
3.48778427e-01 6.23677373e-01 -1.34841633e+00 5.37046373e-01
5.33426583e-01 5.48354387e-01 1.17437258e-01 6.33906782e-01
-3.22323889e-01 1.18780506e+00 -1.06095433e+00 -6.66230798e-01
-1.55946195e-01 -6.61913216e-01 -1.21618629e+00 5.11381209e-01
7.24258363e-01 -2.08183289e-01 4.13286798e-02 7.79795051e-01
5.31521499e-01 5.15447617e-01 8.54094982e-01 -1.28325856e+00
-1.31033397e+00 5.86150289e-01 -6.76235437e-01 2.58567274e-01
1.06859112e+00 3.19736540e-01 8.38258684e-01 -8.81334066e-01
1.05653775e+00 1.91009784e+00 2.08974451e-01 7.30568767e-01
-1.05156147e+00 -2.91195959e-01 4.68164861e-01 1.15672678e-01
-1.37955427e+00 -5.43687224e-01 5.27096033e-01 -4.13047940e-01
1.27341938e+00 3.00762892e-01 4.70944613e-01 1.08337998e+00
-5.38784973e-02 8.20047677e-01 1.04748487e+00 -6.06472015e-01
6.45106938e-03 -6.07116371e-02 3.69344950e-02 6.66838765e-01
-6.36700541e-02 -2.33833611e-01 -6.57293618e-01 -1.11406386e-01
8.95682454e-01 -3.10365260e-01 -3.49553883e-01 -9.05328169e-02
-1.55979609e+00 7.49262452e-01 7.26119816e-01 2.89472371e-01
-5.83719790e-01 4.76884246e-01 2.90178299e-01 -7.42341429e-02
5.39462090e-01 7.85708904e-01 2.11654618e-01 -2.62979686e-01
-2.83560842e-01 6.30569279e-01 5.63335419e-01 1.56684017e+00
5.19045413e-01 -2.98984438e-01 -7.67898083e-01 9.69739914e-01
1.69676900e-01 3.68616164e-01 -6.59276247e-02 -1.44796157e+00
6.17650628e-01 5.59417546e-01 4.38486457e-01 -1.22509229e+00
-3.23355854e-01 2.90777460e-02 -4.42339808e-01 -6.69948608e-02
3.04811835e-01 -2.09181577e-01 -6.64256692e-01 1.90047646e+00
2.56547600e-01 -4.43138868e-01 8.49285796e-02 1.09732342e+00
1.14707398e+00 5.51905036e-01 6.93387628e-01 1.31739140e-01
1.79646659e+00 -1.10344267e+00 -9.24962997e-01 -4.78572845e-01
7.91555405e-01 -7.12110400e-01 1.35390770e+00 2.77734157e-02
-1.36007369e+00 -6.01264775e-01 -6.67113423e-01 -6.42467141e-01
-4.42104757e-01 3.45111042e-02 7.21643329e-01 5.19261472e-02
-1.16684508e+00 2.47499626e-02 -3.63390654e-01 -6.47680402e-01
4.94813621e-01 -6.09585270e-02 -4.35773581e-01 -2.78978683e-02
-9.19953227e-01 1.26825666e+00 1.80695176e-01 2.87745567e-03
-4.34313178e-01 -3.99172783e-01 -1.05702746e+00 -2.46314242e-01
2.32514232e-01 -1.14359736e+00 1.58152962e+00 -9.93957639e-01
-9.38601315e-01 1.41285276e+00 -6.91476941e-01 -2.70402312e-01
4.47120965e-01 -5.12037396e-01 -1.16431579e-01 4.34885442e-01
3.46736103e-01 1.51386762e+00 4.30736721e-01 -1.66405618e+00
-3.27050954e-01 -2.22640336e-01 6.81391537e-01 6.02762461e-01
3.52591634e-01 5.08451819e-01 -6.44229293e-01 -7.96727419e-01
-6.62687793e-03 -7.09664702e-01 -1.09788731e-01 4.31462824e-01
-4.25585717e-01 -7.69920111e-01 7.10587561e-01 -4.26215649e-01
8.51449490e-01 -1.89409590e+00 3.30472261e-01 -1.71214864e-01
2.43598968e-01 -3.70219685e-02 -7.98720866e-02 4.06495988e-01
-1.81143761e-01 4.26690012e-01 2.68642716e-02 -4.79297400e-01
-4.56801727e-02 4.82221365e-01 -3.32609922e-01 5.74393906e-02
5.40184617e-01 1.39862955e+00 -1.11978221e+00 -9.78095293e-01
1.28202826e-01 3.79956454e-01 -4.25985515e-01 2.34324932e-01
-4.07043159e-01 4.63501871e-01 -3.92639548e-01 6.23405695e-01
4.15890753e-01 -5.34354568e-01 -2.54345000e-01 -2.30950207e-01
2.23794490e-01 1.57636598e-01 -6.45011306e-01 1.65797973e+00
-7.53295779e-01 7.28029907e-01 -2.65466273e-01 -2.78880447e-01
8.03100288e-01 1.72444493e-01 -2.29315504e-01 -7.22392619e-01
6.36610687e-02 -3.00427854e-01 -9.27132294e-02 -7.88719356e-01
8.18469822e-01 -2.64116585e-01 -2.13894382e-01 6.43788040e-01
-8.74852389e-02 -6.44888520e-01 4.15433049e-01 6.67568505e-01
4.67260182e-01 3.62674057e-01 6.96351767e-01 -1.45503685e-01
3.87986124e-01 1.39302045e-01 -1.96598440e-01 9.15782392e-01
-2.29763627e-01 7.50549018e-01 6.57212377e-01 -1.59461260e-01
-9.02487874e-01 -1.06334758e+00 5.98022640e-02 1.44167018e+00
3.22021395e-01 -3.90054524e-01 -7.34083295e-01 -5.27033865e-01
-2.03676701e-01 1.16547108e+00 -7.47083187e-01 9.44903195e-02
-6.78067327e-01 -2.23791435e-01 4.92695093e-01 1.13524520e+00
5.57473063e-01 -1.52922952e+00 -1.08591843e+00 -4.13279645e-02
-6.01315141e-01 -1.40862799e+00 -5.85004151e-01 -2.21836135e-01
-4.59037483e-01 -7.21264482e-01 -8.79669666e-01 -7.44129419e-01
9.06440020e-01 3.71103108e-01 1.72318745e+00 5.05939126e-01
-2.06455767e-01 8.43934119e-01 -6.74412310e-01 -6.14764392e-01
-3.51428181e-01 -3.71147454e-01 -3.19791019e-01 -5.01102924e-01
3.86390209e-01 6.35238178e-03 -2.56187290e-01 4.20651406e-01
-5.39572954e-01 3.94467533e-01 4.51131761e-02 5.48482358e-01
3.58865321e-01 -7.72496700e-01 2.82662809e-01 -7.03761101e-01
9.69864011e-01 -3.77933085e-01 -1.35891885e-01 4.41444516e-01
2.59229183e-01 -9.06350315e-02 9.46460944e-03 -3.48491311e-01
-1.31385362e+00 -1.97997063e-01 1.11742966e-01 -2.20042646e-01
-3.34452331e-01 1.87606633e-01 -1.06477745e-01 3.43127489e-01
8.63763213e-01 -2.39954397e-01 -7.61008486e-02 -2.71588899e-02
9.92709637e-01 3.14757645e-01 8.18298280e-01 -1.02140248e+00
4.55145568e-01 5.27106643e-01 -3.93142775e-02 -8.66478324e-01
-9.77236748e-01 -2.87548006e-01 -7.79692709e-01 -3.23005736e-01
1.18734384e+00 -1.00812972e+00 -7.93926358e-01 1.75027996e-01
-1.82401705e+00 -2.83037513e-01 -3.45106661e-01 7.96987861e-02
-6.76961184e-01 3.19699466e-01 -3.69300693e-01 -8.26163352e-01
-7.43901655e-02 -1.01228547e+00 1.51674700e+00 4.65965509e-01
-9.03514802e-01 -1.13496208e+00 -4.66924489e-01 2.48229682e-01
4.47800934e-01 3.45573068e-01 8.37577581e-01 -3.90798777e-01
-5.71835756e-01 2.59548724e-01 -8.70224655e-01 -1.43455744e-01
1.75667286e-01 -3.55923958e-02 -7.56447792e-01 2.16895133e-01
-2.65308917e-01 -6.68780446e-01 5.90462089e-01 6.79279566e-02
1.26346195e+00 -3.49025220e-01 -4.20988142e-01 2.69786835e-01
9.29773152e-01 2.94746310e-01 6.42374277e-01 2.50155348e-02
7.08485842e-01 1.14877689e+00 9.85781491e-01 2.75748551e-01
7.18524873e-01 9.73943532e-01 3.11539352e-01 -3.71894419e-01
-3.57916296e-01 -1.40688926e-01 -5.97511195e-02 1.43110931e-01
-1.98818237e-01 -7.81829834e-01 -1.07730806e+00 7.44743943e-01
-1.71015143e+00 -9.68220055e-01 -3.84854913e-01 1.44454277e+00
8.55722427e-01 -2.96003401e-01 -4.99217473e-02 -4.91849184e-01
5.52699327e-01 1.47184432e-01 -2.45727986e-01 -6.02929115e-01
-3.46681893e-01 2.00434566e-01 5.19000590e-02 5.82443118e-01
-9.15649951e-01 1.26995659e+00 7.50104761e+00 2.96316266e-01
-7.76159108e-01 -1.19092464e-01 8.09127212e-01 2.26408616e-02
-3.43092382e-01 -6.76856637e-02 -6.63618743e-01 -1.65921316e-01
3.62205207e-01 -1.95035398e-01 2.77886033e-01 6.87589407e-01
3.05745602e-01 -6.12962604e-01 -1.39491737e+00 1.13015413e+00
6.46689177e-01 -1.08565903e+00 5.03404260e-01 -4.23966914e-01
6.19735241e-01 -4.84862149e-01 -7.83786625e-02 3.83784361e-02
4.20157254e-01 -1.30581129e+00 9.14965510e-01 8.81773174e-01
5.21713316e-01 -5.40702343e-01 7.34192848e-01 -3.71316560e-02
-7.69566655e-01 3.92497212e-01 -1.04016282e-01 -1.93649963e-01
6.32228255e-01 -2.51083933e-02 -9.63778913e-01 4.15541470e-01
6.93495274e-01 6.07964516e-01 -1.12596548e+00 6.54120147e-01
-6.35074019e-01 6.64147586e-02 -8.57264400e-02 -2.83044577e-01
1.82762384e-01 5.99319302e-02 3.44152272e-01 1.38111186e+00
8.82212073e-02 4.72809911e-01 1.15116462e-01 1.43466973e+00
2.02213917e-02 4.13530990e-02 -7.97704577e-01 -7.85112679e-02
5.72539568e-01 1.07877779e+00 -6.36311650e-01 -6.99210823e-01
-2.94460058e-01 1.10463846e+00 3.23986411e-01 5.16823471e-01
-8.09886694e-01 -3.27352464e-01 4.81234491e-01 1.42312497e-01
-3.84762660e-02 -3.50176692e-01 -2.30838686e-01 -8.30898046e-01
-4.62269317e-03 -8.71132612e-01 2.72770464e-01 -1.88084936e+00
-1.37468266e+00 5.20261049e-01 6.98702812e-01 -9.86147225e-01
-6.37598157e-01 -7.42748260e-01 -5.17754912e-01 1.20979857e+00
-1.12787366e+00 -1.36983180e+00 -7.64554441e-01 5.64065754e-01
6.47120357e-01 3.18402588e-01 7.75676727e-01 -2.66150177e-01
-4.37654369e-02 3.06349874e-01 -1.10649323e+00 1.66769683e-01
7.07879186e-01 -1.24570453e+00 5.79406500e-01 6.19938374e-01
2.51000851e-01 1.17421889e+00 8.26972067e-01 -5.82057595e-01
-7.96569765e-01 -6.85889065e-01 1.02545559e+00 -8.05396318e-01
4.48399782e-01 -3.00105095e-01 -9.74371850e-01 9.79639351e-01
8.63479137e-01 -5.22076711e-02 4.37292248e-01 9.95076522e-02
-3.60157132e-01 6.63844943e-01 -1.05504858e+00 9.94885802e-01
1.51307654e+00 -7.25864112e-01 -1.05367589e+00 5.31578064e-01
8.61684561e-01 -8.40783834e-01 -5.93888998e-01 2.30257511e-02
2.04262838e-01 -9.14665759e-01 1.11663508e+00 -9.45278466e-01
7.42089450e-01 -3.95853639e-01 -1.57469839e-01 -1.32098222e+00
3.11761014e-02 -3.94913375e-01 2.63167381e-01 1.47109604e+00
2.08748937e-01 -2.99052775e-01 2.94287443e-01 7.70518184e-01
7.03056753e-02 -3.12273800e-01 -6.57825232e-01 -4.24323857e-01
2.73088124e-02 -2.50665486e-01 5.77648640e-01 8.16437900e-01
1.59727082e-01 5.61599016e-01 -1.79553643e-01 -1.91218406e-01
9.87020880e-02 -7.71893859e-02 1.07090342e+00 -7.64066935e-01
-1.63725495e-01 -3.83877814e-01 -3.66253942e-01 -1.34556413e+00
3.67235899e-01 -7.39017785e-01 6.45425767e-02 -1.85303593e+00
3.31600279e-01 -1.93053022e-01 4.30064291e-01 7.57113755e-01
-4.94489878e-01 1.79091915e-01 6.34790897e-01 3.26153263e-02
-7.90190399e-01 3.46855581e-01 1.71938002e+00 -1.00572459e-01
1.44367963e-02 -5.00932992e-01 -1.07900286e+00 8.85904968e-01
5.52845120e-01 -5.41833639e-02 -4.17687058e-01 -7.86425352e-01
-1.73758231e-02 1.01804920e-01 9.18779790e-01 -7.41215527e-01
2.17488594e-03 -2.43313447e-01 4.69048500e-01 -7.59174347e-01
5.29380202e-01 -4.80909705e-01 -1.46558285e-01 -2.58337408e-02
-7.25854516e-01 6.46818161e-01 2.77319282e-01 2.34017804e-01
-2.53805280e-01 -3.27378273e-01 6.50888443e-01 -4.14170712e-01
-8.40184808e-01 -4.45578039e-01 -3.61385703e-01 4.56005633e-01
9.37015593e-01 -2.32410312e-01 -7.61808217e-01 -1.00814378e+00
-8.99108112e-01 4.96399343e-01 5.81772625e-01 7.15781391e-01
7.64579833e-01 -1.46552896e+00 -6.44415736e-01 -3.67214322e-01
6.11821651e-01 1.42376930e-01 1.39930407e-02 6.70917273e-01
-5.86032152e-01 3.32203567e-01 -2.04668820e-01 -7.72821665e-01
-1.35182762e+00 6.17829204e-01 4.31483060e-01 1.81674317e-01
-5.04729867e-01 1.08258164e+00 4.82332766e-01 -2.26348668e-01
4.93848100e-02 -4.99137461e-01 -3.58535826e-01 1.10704713e-01
5.83179176e-01 3.66273597e-02 -4.22111392e-01 -1.19313598e+00
-3.18088800e-01 7.71548390e-01 -9.43675190e-02 -3.61971736e-01
7.78907180e-01 -3.84807169e-01 -3.56209308e-01 5.32865047e-01
9.66902852e-01 5.50046787e-02 -8.64792705e-01 -2.34456107e-01
3.83608565e-02 -5.69620013e-01 -4.60501075e-01 -1.02264345e+00
-6.40818238e-01 8.93901169e-01 1.68151706e-01 -1.40557349e-01
1.02061439e+00 6.52538538e-01 9.63925868e-02 5.40781677e-01
5.27648211e-01 -7.72266388e-01 5.60874879e-01 4.51101124e-01
1.67803049e+00 -1.06884551e+00 2.87516087e-01 -6.76510692e-01
-1.30856323e+00 9.98480976e-01 9.35847759e-01 -1.33698992e-03
4.97471020e-02 1.21118128e-03 4.21137750e-01 -4.89542514e-01
-8.93699110e-01 -3.33264232e-01 2.12548986e-01 8.63644123e-01
8.92097950e-01 -1.30218081e-02 -3.50071967e-01 1.52032003e-01
-3.41197371e-01 -1.19197026e-01 5.78179181e-01 9.98395801e-01
-1.45752490e-01 -9.65232730e-01 -5.48340082e-01 1.50962621e-01
-1.41247213e-01 -1.16311960e-01 -9.43183422e-01 1.14411426e+00
-2.16969624e-02 1.13291395e+00 3.42368722e-01 2.60060579e-01
6.47103667e-01 5.55351153e-02 7.17754126e-01 -7.62495577e-01
-7.91478992e-01 -1.06986359e-01 6.12882853e-01 -5.88236928e-01
-9.72308517e-01 -7.97387779e-01 -1.57889855e+00 4.34350371e-02
-1.42430186e-01 -9.51538384e-02 1.63814485e-01 7.63254166e-01
3.94944310e-01 5.60171962e-01 -4.52429354e-02 -9.32101011e-01
-1.84375137e-01 -1.03224826e+00 -6.94607347e-02 9.90656137e-01
1.01518124e-01 -8.45454931e-01 -1.39491916e-01 2.51232773e-01] | [10.611825942993164, 1.5999025106430054] |
95314cb6-fb7f-4f7a-ada2-e4f2eb76d033 | adaattn-revisit-attention-mechanism-in | 2108.03647 | null | https://arxiv.org/abs/2108.03647v2 | https://arxiv.org/pdf/2108.03647v2.pdf | AdaAttN: Revisit Attention Mechanism in Arbitrary Neural Style Transfer | Fast arbitrary neural style transfer has attracted widespread attention from academic, industrial and art communities due to its flexibility in enabling various applications. Existing solutions either attentively fuse deep style feature into deep content feature without considering feature distributions, or adaptively normalize deep content feature according to the style such that their global statistics are matched. Although effective, leaving shallow feature unexplored and without locally considering feature statistics, they are prone to unnatural output with unpleasing local distortions. To alleviate this problem, in this paper, we propose a novel attention and normalization module, named Adaptive Attention Normalization (AdaAttN), to adaptively perform attentive normalization on per-point basis. Specifically, spatial attention score is learnt from both shallow and deep features of content and style images. Then per-point weighted statistics are calculated by regarding a style feature point as a distribution of attention-weighted output of all style feature points. Finally, the content feature is normalized so that they demonstrate the same local feature statistics as the calculated per-point weighted style feature statistics. Besides, a novel local feature loss is derived based on AdaAttN to enhance local visual quality. We also extend AdaAttN to be ready for video style transfer with slight modifications. Experiments demonstrate that our method achieves state-of-the-art arbitrary image/video style transfer. Codes and models are available. | ['Errui Ding', 'Qian Li', 'Zhengxing Sun', 'Xin Li', 'Meiling Wang', 'Fu Li', 'Dongliang He', 'Tianwei Lin', 'Songhua Liu'] | 2021-08-08 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Liu_AdaAttN_Revisit_Attention_Mechanism_in_Arbitrary_Neural_Style_Transfer_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Liu_AdaAttN_Revisit_Attention_Mechanism_in_Arbitrary_Neural_Style_Transfer_ICCV_2021_paper.pdf | iccv-2021-1 | ['video-style-transfer'] | ['computer-vision'] | [ 3.06859344e-01 -5.64600050e-01 -6.73565492e-02 -5.30396760e-01
-6.06101811e-01 -5.36143959e-01 4.17706966e-01 -5.42203523e-02
-3.81072372e-01 6.40544236e-01 3.17413837e-01 2.98514038e-01
-1.27831459e-01 -9.58555639e-01 -6.51993692e-01 -8.82148564e-01
4.53175336e-01 4.18342475e-04 1.49751067e-01 -2.98891693e-01
4.20618534e-01 6.73243940e-01 -1.43425894e+00 2.82708704e-01
1.01378632e+00 1.23268855e+00 3.56697708e-01 4.54708278e-01
-3.43194753e-01 3.95761818e-01 -6.11086309e-01 -4.04669374e-01
3.53804857e-01 -7.69807637e-01 -5.50826490e-01 4.53368425e-02
8.26969624e-01 -5.00338435e-01 -4.05235052e-01 1.38267100e+00
6.25881076e-01 3.79603475e-01 7.40737021e-01 -1.20339966e+00
-1.25140190e+00 -2.22088136e-02 -8.02118361e-01 2.30042070e-01
2.35813066e-01 1.56670466e-01 1.01681185e+00 -9.42808032e-01
3.80292654e-01 1.26483786e+00 4.10054386e-01 4.82993245e-01
-1.21599913e+00 -6.62442863e-01 2.83548534e-01 3.66460055e-01
-1.40478921e+00 -2.15600193e-01 1.09359634e+00 -9.06706899e-02
7.23012149e-01 2.98882097e-01 7.33478546e-01 8.61248612e-01
4.04667169e-01 7.57384956e-01 8.20390284e-01 -2.37285867e-01
7.83527270e-02 1.02425903e-01 -3.31461728e-01 6.35766029e-01
-1.33327216e-01 -2.64714062e-01 -5.40692091e-01 1.56485334e-01
1.22047269e+00 3.77106875e-01 -3.30184489e-01 -3.93052936e-01
-1.06609654e+00 6.57315969e-01 7.92918742e-01 3.04209262e-01
-4.20363367e-01 6.69637620e-02 2.71800727e-01 3.71346712e-01
5.44210255e-01 1.37994528e-01 -4.47002470e-01 -1.50550127e-01
-6.87076688e-01 1.57879159e-01 1.60324574e-01 1.06749201e+00
1.11435151e+00 1.83463931e-01 -5.89434803e-01 1.13005602e+00
-7.37139629e-03 6.43595278e-01 6.47954464e-01 -1.09735692e+00
3.05917084e-01 6.42099798e-01 -1.70888662e-01 -1.30694032e+00
-2.85936723e-04 -4.77378517e-01 -1.03822875e+00 3.64503503e-01
2.61076719e-01 5.86613305e-02 -6.86755836e-01 1.83855867e+00
-3.43853584e-03 -3.28779556e-02 -3.06183249e-01 1.08800232e+00
6.60188735e-01 7.79276192e-01 9.05710608e-02 -3.60835381e-02
1.24848092e+00 -8.81028116e-01 -6.82045937e-01 -9.71975997e-02
1.63424104e-01 -9.90952313e-01 1.58560574e+00 1.80364475e-01
-1.21482766e+00 -8.72223437e-01 -8.29154551e-01 -2.95622468e-01
-2.64514923e-01 -5.77501357e-02 2.44590074e-01 3.79993856e-01
-1.06728721e+00 7.39536226e-01 -4.96976733e-01 -4.37045157e-01
5.33831537e-01 4.19745475e-01 -1.58527449e-01 2.99012158e-02
-8.67355347e-01 5.00780344e-01 4.44642119e-02 -2.22279161e-01
-5.07311702e-01 -7.35230803e-01 -8.60374749e-01 2.57111162e-01
5.19526303e-02 -8.31409931e-01 9.54771817e-01 -1.45024645e+00
-1.70200348e+00 6.35150433e-01 -2.31001511e-01 2.45274752e-01
2.91443139e-01 -3.09490591e-01 -3.74942511e-01 8.47527236e-02
1.93048999e-01 8.33525181e-01 9.76440966e-01 -1.07155323e+00
-6.50609076e-01 -2.86787868e-01 1.04865268e-01 4.90364015e-01
-9.44977283e-01 2.82994062e-02 -5.26035786e-01 -1.21266818e+00
4.11916114e-02 -5.01856029e-01 9.21030268e-02 5.07783115e-01
-1.39301434e-01 -1.82507724e-01 1.10881507e+00 -4.88670737e-01
1.06879604e+00 -2.29559588e+00 2.80893296e-01 -9.75421350e-03
2.45635360e-01 1.71074450e-01 -5.97443700e-01 -3.60304341e-02
7.46045038e-02 -6.74770176e-02 -2.95800120e-01 -7.33300671e-02
-1.64419413e-01 8.34174827e-02 -2.37844005e-01 3.06287795e-01
4.08940941e-01 8.80343497e-01 -1.03097236e+00 -3.91664416e-01
4.81607050e-01 5.68187892e-01 -1.04704642e+00 3.29104602e-01
1.89448614e-02 4.78780001e-01 -5.50418615e-01 6.16978228e-01
7.76249886e-01 8.52295533e-02 -3.50604683e-01 -5.74705422e-01
1.08377670e-03 -2.63955772e-01 -7.88732171e-01 1.77652287e+00
-6.83650851e-01 6.60284102e-01 -7.77833015e-02 -9.01190221e-01
1.05381131e+00 1.23463459e-02 4.66790974e-01 -7.82058835e-01
3.22347909e-01 5.52617759e-02 -6.45464808e-02 -3.02736849e-01
6.24440610e-01 -2.69607097e-01 -1.55612245e-01 1.70397803e-01
3.27816516e-01 -2.43684217e-01 -2.13806525e-01 2.72381473e-02
7.39983857e-01 2.67273873e-01 1.24158546e-01 -3.78430575e-01
8.38878512e-01 -5.32003164e-01 5.82827568e-01 5.24269938e-01
-3.30328315e-01 1.10472000e+00 3.28006148e-01 -3.62455040e-01
-1.23711610e+00 -1.36780751e+00 -1.49033898e-02 1.42377460e+00
4.60176319e-01 -2.67863005e-01 -8.54644120e-01 -8.24818194e-01
-1.28740445e-01 4.83750850e-01 -6.37547195e-01 -4.40113276e-01
-6.11592591e-01 -2.58861929e-01 1.66766718e-01 6.06528401e-01
8.66284490e-01 -1.42224646e+00 -1.61913589e-01 1.80893540e-01
-1.05322234e-01 -7.66643703e-01 -1.18129432e+00 -2.68945128e-01
-6.79239035e-01 -5.43786943e-01 -1.24207163e+00 -8.70050728e-01
7.49911964e-01 4.86053139e-01 8.64392221e-01 -7.26254657e-02
-3.01165044e-01 3.53513151e-01 -3.25926661e-01 -7.69497901e-02
-6.47777990e-02 -1.12279402e-02 1.19482800e-01 3.94876629e-01
2.63242424e-01 -7.89304018e-01 -1.01868880e+00 3.56541008e-01
-1.09906936e+00 -1.56873569e-01 6.25150979e-01 8.41820180e-01
5.85490942e-01 -2.52762109e-01 5.54094017e-01 -4.26014960e-01
6.16262376e-01 -2.82051414e-01 -2.40173906e-01 9.57284868e-02
-1.20802365e-01 3.36338067e-03 1.00329590e+00 -5.00333309e-01
-1.22009349e+00 -2.45445415e-01 -2.46220052e-01 -6.92087114e-01
-2.59817064e-01 -5.96714579e-02 -5.30598938e-01 -2.56140053e-01
4.36852515e-01 4.16019171e-01 -8.03429983e-04 -4.47358549e-01
4.41279709e-01 5.62225878e-01 5.58772445e-01 -6.28385305e-01
8.73294353e-01 4.47273254e-01 -1.82119727e-01 -6.80371642e-01
-8.74345124e-01 -1.82288811e-01 -5.93247950e-01 -3.26487392e-01
8.46224666e-01 -6.14012599e-01 -7.28672028e-01 6.94116414e-01
-1.04817843e+00 -1.35668203e-01 -4.68296856e-01 3.92035425e-01
-7.00327098e-01 5.23251176e-01 -5.53032994e-01 -3.90844166e-01
-3.48831713e-01 -1.14885890e+00 1.12004173e+00 5.69615245e-01
-2.59893566e-01 -9.90419090e-01 -2.17873603e-01 -5.50866053e-02
8.50907981e-01 -2.13572220e-03 8.82811069e-01 -7.47409090e-02
-5.17526031e-01 -8.86805058e-02 -6.86914325e-01 5.40992916e-01
6.89403176e-01 -1.80494174e-01 -8.62896264e-01 -4.27393943e-01
-2.37088129e-02 -1.86562464e-01 7.64824748e-01 4.44714963e-01
1.61693347e+00 -3.83873492e-01 1.96875110e-01 9.33313549e-01
1.47976077e+00 7.83432424e-02 5.68343997e-01 3.29121262e-01
1.13466489e+00 3.85390192e-01 3.52191955e-01 5.16695142e-01
-4.83764969e-02 8.31426919e-01 3.06957096e-01 -3.04914922e-01
-3.72251332e-01 -7.24144652e-02 3.49765450e-01 7.37128735e-01
-3.04602385e-01 -5.36181219e-02 -2.66896099e-01 4.74186748e-01
-1.59446025e+00 -9.69730198e-01 3.42673779e-01 2.09064293e+00
8.51369321e-01 8.76809135e-02 9.72918198e-02 -5.91831878e-02
8.10082018e-01 2.99381167e-01 -6.83203459e-01 -5.65298259e-01
-2.50037372e-01 2.36810297e-01 2.44145706e-01 3.29690278e-01
-8.84977758e-01 1.06774402e+00 5.13531446e+00 1.32202864e+00
-1.25248110e+00 9.71145332e-02 7.45975435e-01 -2.72566795e-01
-5.43944657e-01 -3.45853150e-01 -4.87753630e-01 6.04810297e-01
2.93311447e-01 -3.07217926e-01 7.03705728e-01 8.59170496e-01
1.71333879e-01 1.55765295e-01 -8.72786105e-01 1.12375200e+00
2.09676489e-01 -9.97952402e-01 6.93327248e-01 -1.99328680e-02
8.21265638e-01 -3.53718638e-01 4.82242048e-01 3.51522088e-01
-5.12260385e-02 -6.48508728e-01 8.88660014e-01 7.23148048e-01
1.06114268e+00 -9.89169002e-01 6.32206559e-01 -1.34579718e-01
-1.25929880e+00 -1.31278545e-01 -4.86499190e-01 -1.45253688e-01
1.08539328e-01 4.95950252e-01 2.68254783e-02 5.04773200e-01
7.75786340e-01 9.28981543e-01 -5.49925506e-01 1.01106155e+00
1.08412184e-01 2.61435628e-01 -7.78428186e-03 -4.27426286e-02
3.10502976e-01 -3.19342554e-01 6.68491423e-01 1.03646624e+00
6.00067496e-01 -3.67083810e-02 -2.69714072e-02 9.62137163e-01
-2.36523584e-01 3.25348318e-01 -5.35417795e-01 4.57205892e-01
2.85150439e-01 1.32413173e+00 -6.60136104e-01 -4.28226292e-01
-5.14454603e-01 1.50727296e+00 3.10975283e-01 5.13707578e-01
-7.53581166e-01 -7.19648957e-01 1.08412719e+00 -3.64308618e-02
3.03371131e-01 -4.40805331e-02 -4.71937776e-01 -1.23076630e+00
8.32673907e-02 -5.46413422e-01 7.81588778e-02 -9.29386020e-01
-1.58577478e+00 7.75889516e-01 -1.52492329e-01 -1.55253768e+00
1.58402085e-01 -5.94155967e-01 -8.87292385e-01 1.01607800e+00
-1.28054523e+00 -1.01293015e+00 -6.76626503e-01 7.72714317e-01
8.52585316e-01 -3.18425089e-01 6.10026300e-01 4.06696975e-01
-5.48964262e-01 1.00222611e+00 1.13846324e-01 8.34999979e-02
1.02507937e+00 -1.15916586e+00 1.91157684e-01 5.92470527e-01
-2.04052985e-01 5.69640458e-01 5.24369895e-01 -4.99492556e-01
-1.02143824e+00 -1.35770810e+00 5.48368752e-01 -2.24319473e-01
5.12015104e-01 -2.71452576e-01 -1.08637834e+00 3.26538920e-01
3.80135983e-01 2.09076747e-01 3.19700062e-01 -2.46571600e-01
-2.64384061e-01 -4.95038182e-01 -1.26630068e+00 9.70557988e-01
1.21962869e+00 -5.60880780e-01 -3.90036285e-01 5.58757782e-02
6.88308716e-01 -8.82318169e-02 -8.27965558e-01 5.30332446e-01
5.83021402e-01 -1.04470932e+00 9.36146438e-01 -1.98547676e-01
6.48554027e-01 -3.24348003e-01 -3.05938184e-01 -1.36558473e+00
-8.81045818e-01 -4.16286200e-01 1.72571987e-01 1.44133961e+00
3.00830696e-02 -5.54515600e-01 6.81817591e-01 3.66245002e-01
-3.22264880e-01 -7.32748210e-01 -7.73020267e-01 -7.70632267e-01
1.06162541e-01 -1.40635790e-02 6.56241179e-01 9.08495963e-01
-3.39196801e-01 2.54305542e-01 -3.84541422e-01 -1.56498656e-01
5.96792698e-01 1.07884340e-01 5.05406499e-01 -8.39682102e-01
-1.47836909e-01 -9.21834230e-01 -6.58845365e-01 -1.09075153e+00
4.75200377e-02 -8.17449868e-01 -4.15570587e-02 -1.25741851e+00
4.48314369e-01 -2.37421393e-01 -7.08365202e-01 2.95543194e-01
-3.57755542e-01 8.56037915e-01 2.87937760e-01 1.55471548e-01
-6.04406118e-01 9.80700374e-01 1.67711473e+00 -2.22826153e-01
-2.11200848e-01 -1.92938417e-01 -7.77270198e-01 7.50493169e-01
9.00441647e-01 -1.97269350e-01 -3.27559233e-01 -5.16532958e-01
-2.46451378e-01 -2.99033105e-01 3.44483882e-01 -1.20578003e+00
-3.29120457e-02 -2.42138028e-01 8.32786620e-01 -4.73889053e-01
3.56096894e-01 -7.96559870e-01 -3.16722542e-01 8.92276466e-02
-3.23095709e-01 1.39455020e-01 2.07034871e-01 4.61307138e-01
-2.89933503e-01 3.51871364e-02 9.89213943e-01 3.00652254e-02
-7.49049067e-01 5.86604297e-01 -2.62560457e-01 -7.74926543e-02
8.86842072e-01 -3.64379883e-01 -1.72633193e-02 -4.96423334e-01
-5.05155861e-01 -1.18062124e-01 6.78473532e-01 5.52782178e-01
6.99923396e-01 -1.71297407e+00 -6.68287754e-01 4.71478343e-01
1.16752625e-01 -1.32749572e-01 6.63998544e-01 4.90952432e-01
-3.79364282e-01 1.14999816e-01 -7.44804561e-01 -6.12940133e-01
-9.42335069e-01 6.38630986e-01 1.84902638e-01 2.10557804e-01
-5.46182752e-01 7.96272814e-01 8.80487621e-01 -3.40676755e-01
1.51976630e-01 -2.83548832e-01 -1.48577347e-01 -1.45648554e-01
6.27904594e-01 2.03624308e-01 -6.33811504e-02 -7.30832338e-01
-1.22720405e-01 1.18981099e+00 -1.17623955e-01 3.88776362e-02
1.27842605e+00 -3.53318572e-01 -1.70811281e-01 1.92127466e-01
1.60402441e+00 2.11727545e-01 -1.63780046e+00 -4.00818378e-01
-6.51979566e-01 -8.20699513e-01 6.74294904e-02 -5.41587651e-01
-1.41909063e+00 9.02793169e-01 7.91622221e-01 1.53687438e-02
1.60379636e+00 -8.00692737e-02 9.78995085e-01 2.54721250e-02
1.34692490e-01 -7.76763976e-01 4.10563678e-01 6.05037689e-01
1.20801902e+00 -1.05457282e+00 -2.21487239e-01 -2.37869129e-01
-5.14191031e-01 1.00995028e+00 9.47362721e-01 -4.68164861e-01
4.96824473e-01 -1.62526257e-02 7.89601505e-02 2.38943279e-01
-1.79490820e-01 -5.77667803e-02 4.87855256e-01 6.65515363e-01
4.69546884e-01 -1.97108448e-01 -4.50928397e-02 5.67051649e-01
-1.12636112e-01 -2.14451239e-01 9.86774564e-02 6.32397354e-01
-4.63632822e-01 -9.63710546e-01 -3.34313750e-01 5.54515600e-01
-3.78001481e-01 -2.77587712e-01 -2.91796587e-02 3.66144091e-01
2.04071447e-01 5.38072646e-01 4.34232473e-01 -4.67045248e-01
4.90138859e-01 -3.35920721e-01 5.66538930e-01 -3.67557973e-01
-4.52110559e-01 1.15694724e-01 -5.23250341e-01 -5.08468151e-01
-1.25167146e-01 -4.95499730e-01 -9.88898158e-01 -4.96307105e-01
-2.49581132e-02 -1.58142805e-01 3.09931159e-01 7.62226045e-01
3.20639193e-01 7.66606867e-01 9.11398530e-01 -1.41255355e+00
-1.05168037e-01 -9.29357052e-01 -6.52214110e-01 7.19142199e-01
2.88132966e-01 -6.99510396e-01 -2.01757252e-01 1.12189882e-01] | [11.525264739990234, -0.7820078730583191] |
28d1a0be-351d-4813-bef3-eedc90cd8c92 | source-free-domain-adaptation-via-avatar | 2106.15326 | null | https://arxiv.org/abs/2106.15326v1 | https://arxiv.org/pdf/2106.15326v1.pdf | Source-free Domain Adaptation via Avatar Prototype Generation and Adaptation | We study a practical domain adaptation task, called source-free unsupervised domain adaptation (UDA) problem, in which we cannot access source domain data due to data privacy issues but only a pre-trained source model and unlabeled target data are available. This task, however, is very difficult due to one key challenge: the lack of source data and target domain labels makes model adaptation very challenging. To address this, we propose to mine the hidden knowledge in the source model and exploit it to generate source avatar prototypes (i.e., representative features for each source class) as well as target pseudo labels for domain alignment. To this end, we propose a Contrastive Prototype Generation and Adaptation (CPGA) method. Specifically, CPGA consists of two stages: (1) prototype generation: by exploring the classification boundary information of the source model, we train a prototype generator to generate avatar prototypes via contrastive learning. (2) prototype adaptation: based on the generated source prototypes and target pseudo labels, we develop a new robust contrastive prototype adaptation strategy to align each pseudo-labeled target data to the corresponding source prototypes. Extensive experiments on three UDA benchmark datasets demonstrate the effectiveness and superiority of the proposed method. | ['Mingkui Tan', 'Qing Du', 'Yanxia Liu', 'Shuaicheng Niu', 'Hongbin Lin', 'Yifan Zhang', 'Zhen Qiu'] | 2021-06-18 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 5.39184213e-01 6.02402166e-02 -3.71974081e-01 -6.60405993e-01
-7.18774915e-01 -6.36161566e-01 6.84143245e-01 1.08059250e-01
-3.19327742e-01 7.64167845e-01 -4.50350121e-02 3.28270383e-02
1.56312287e-01 -6.59710884e-01 -6.25571787e-01 -7.32110441e-01
4.66845781e-01 7.46987343e-01 -9.54606663e-03 -4.12614532e-02
4.04601917e-02 9.32701826e-02 -1.26763630e+00 -2.30015395e-03
1.29822695e+00 1.08826458e+00 2.00039238e-01 1.12695687e-01
-2.74377793e-01 3.28345329e-01 -6.15115762e-01 -2.34838799e-01
3.44998807e-01 -8.63180518e-01 -8.33403170e-01 3.81387085e-01
1.13114566e-01 -1.97941318e-01 1.29453361e-01 1.10856640e+00
3.48663568e-01 3.09500039e-01 9.46228623e-01 -1.61020160e+00
-8.64169359e-01 5.34219742e-01 -6.13445520e-01 -1.89971834e-01
1.74417756e-02 1.33252740e-01 7.89200246e-01 -1.07497883e+00
7.85591543e-01 9.90182936e-01 3.58033836e-01 1.06858766e+00
-1.27657533e+00 -1.04903710e+00 2.18163997e-01 1.97930649e-01
-1.54736698e+00 -6.35783613e-01 1.19085681e+00 -5.32154202e-01
1.39891952e-01 -1.84615448e-01 3.90114009e-01 1.37144828e+00
-5.56594849e-01 6.87069118e-01 1.07697093e+00 -5.58467329e-01
6.90643728e-01 7.71367013e-01 1.86203644e-01 2.95284748e-01
1.06619395e-01 7.42874593e-02 -2.34720871e-01 -4.17748600e-01
5.55203795e-01 -1.92039162e-01 -1.78475916e-01 -1.02833092e+00
-1.22235978e+00 8.39913309e-01 3.33594948e-01 -8.82662013e-02
-4.91723329e-01 -7.46418893e-01 4.19871300e-01 3.78374785e-01
2.65285820e-01 3.38426679e-01 -4.23179150e-01 3.59847188e-01
-4.28357631e-01 2.37709597e-01 6.56640887e-01 1.25436795e+00
9.52608287e-01 -3.88992503e-02 1.23643931e-02 1.21436286e+00
4.14392501e-01 6.44275188e-01 7.92423010e-01 -5.92030764e-01
6.41690016e-01 7.64294982e-01 2.64949262e-01 -6.27209842e-01
-3.33760655e-03 -1.58238545e-01 -1.00542080e+00 1.32042482e-01
5.31715035e-01 -2.23040268e-01 -8.77594411e-01 1.99795699e+00
8.26756418e-01 1.79012552e-01 6.64240301e-01 8.49259675e-01
7.23291636e-01 6.93170428e-01 2.72190422e-01 -3.69996578e-01
1.00270963e+00 -1.02735043e+00 -3.93195659e-01 -4.27926868e-01
4.55132425e-01 -5.95230639e-01 1.10353601e+00 -3.74740735e-02
-5.78332961e-01 -6.96695864e-01 -1.26446784e+00 1.78660080e-01
-2.18804821e-01 1.67905554e-01 1.33437127e-01 4.74629939e-01
-3.09077561e-01 1.31250337e-01 -4.84600812e-01 -5.04032493e-01
7.57747412e-01 2.83957690e-01 -4.72828507e-01 -2.27550253e-01
-1.12525034e+00 4.33407843e-01 7.89576113e-01 -1.58047557e-01
-8.83618236e-01 -5.70906341e-01 -1.07623672e+00 -2.68869877e-01
4.27262664e-01 -6.32036030e-01 1.34792149e+00 -1.39241457e+00
-1.63319457e+00 8.85618567e-01 -1.46396220e-01 -2.60629803e-01
5.17153382e-01 1.88893512e-01 -6.42320335e-01 -1.97218642e-01
2.63299972e-01 6.57365501e-01 1.02122176e+00 -1.46199369e+00
-8.57591271e-01 -3.68297011e-01 -3.80278200e-01 3.27934772e-01
-5.59087813e-01 -3.07922453e-01 -1.57914072e-01 -4.28612858e-01
-1.25918046e-01 -8.23101342e-01 -2.29201108e-01 6.57766610e-02
-3.82028520e-01 -2.33311042e-01 8.42398763e-01 -5.23251355e-01
1.01654601e+00 -2.26257300e+00 5.31270169e-02 4.77307826e-01
5.32449409e-02 6.00976050e-01 -3.33537281e-01 9.35666170e-03
-5.96065670e-02 -2.21168488e-01 -6.27523005e-01 -7.74704367e-02
-9.24726501e-02 1.30919516e-01 -4.72140700e-01 1.18555114e-01
2.43002847e-01 5.35208106e-01 -1.04321432e+00 -6.79117799e-01
5.74146062e-02 2.48132527e-01 -3.49688202e-01 6.32924795e-01
-3.06323767e-01 7.26996183e-01 -7.94861853e-01 5.55702209e-01
8.12778890e-01 -2.74105489e-01 2.29385436e-01 4.60840017e-02
2.00338900e-01 3.57465893e-02 -1.26274145e+00 1.64142084e+00
-4.32824343e-01 1.96780920e-01 -1.05148286e-01 -1.20936358e+00
1.52264106e+00 2.54262030e-01 2.24867448e-01 -6.42349958e-01
1.42355233e-01 4.30772573e-01 -5.96714579e-02 -3.62107009e-01
1.97848246e-01 -1.40924156e-01 -3.10767502e-01 6.70759857e-01
2.11957410e-01 1.89581826e-01 -1.33213431e-01 4.35158499e-02
6.81183159e-01 2.68691152e-01 6.47617817e-01 4.21497673e-02
8.30069661e-01 4.16331261e-01 1.03200424e+00 4.62281734e-01
-3.79207343e-01 5.87559342e-01 1.57515705e-01 -4.08628523e-01
-1.35967326e+00 -1.09495604e+00 9.42513049e-02 8.92734289e-01
1.71946228e-01 -4.67375405e-02 -7.80455053e-01 -1.10951066e+00
-2.05589145e-01 7.20385909e-01 -6.87237144e-01 -5.82436085e-01
-5.49497128e-01 -4.37510788e-01 2.31996059e-01 5.25988936e-01
7.04412878e-01 -1.15368629e+00 -3.77715439e-01 2.35313699e-01
-2.50116706e-01 -9.45943534e-01 -7.37268329e-01 -1.01013362e-01
-7.26743519e-01 -9.97046888e-01 -8.32365930e-01 -1.21664727e+00
9.68130112e-01 2.68251121e-01 8.21716964e-01 -3.59651715e-01
2.03951672e-01 -5.30281849e-02 -4.51486111e-01 -4.14031535e-01
-8.27206850e-01 1.15052015e-01 1.58228815e-01 3.97347480e-01
6.22039914e-01 -5.97886741e-01 -4.18292910e-01 4.85816717e-01
-6.90038264e-01 1.72669202e-01 8.40233028e-01 1.10974813e+00
8.47827315e-01 -1.06549658e-01 1.05904639e+00 -1.35015690e+00
7.64529943e-01 -7.81240821e-01 -6.66676342e-01 4.49001461e-01
-8.35301876e-01 -7.57251382e-02 1.07235861e+00 -9.52390373e-01
-1.28215456e+00 5.31434298e-01 1.27696082e-01 -7.02555835e-01
-4.59109515e-01 3.89664143e-01 -8.13157737e-01 2.33571902e-01
1.02065074e+00 4.99970376e-01 2.24129051e-01 -4.68183160e-01
4.81032223e-01 1.20542574e+00 8.29707503e-01 -4.90275741e-01
1.08592355e+00 1.68118894e-01 -4.09664869e-01 -6.34268165e-01
-5.78798056e-01 -4.76412445e-01 -8.85710001e-01 1.86808035e-01
6.56516075e-01 -9.78468597e-01 -1.51709765e-01 5.31212211e-01
-9.02827680e-01 -3.08482975e-01 -5.26541471e-01 4.38057274e-01
-6.33315802e-01 2.30939791e-01 1.39778838e-01 -5.39926171e-01
-5.36242843e-01 -8.95484209e-01 5.20195186e-01 5.44848204e-01
-2.41310224e-01 -8.19483817e-01 2.13599667e-01 1.65627852e-01
1.75391927e-01 3.18657964e-01 9.31275308e-01 -1.29391134e+00
-1.20468296e-01 -1.74040914e-01 -1.74594089e-01 4.93235618e-01
4.54262584e-01 -3.27998787e-01 -7.88828850e-01 -2.93473780e-01
-1.12196304e-01 -4.91672218e-01 2.13488623e-01 -9.66788158e-02
8.45416069e-01 -6.19724035e-01 -4.70932126e-01 5.30161321e-01
1.02505648e+00 3.96584004e-01 1.09631263e-01 3.16107064e-01
8.22509766e-01 6.10820889e-01 1.09649217e+00 5.88323534e-01
4.17068779e-01 6.89150870e-01 -7.60666952e-02 1.03738286e-01
-7.90456012e-02 -7.42272973e-01 1.90707669e-01 7.76736736e-01
3.78011823e-01 1.11010559e-02 -9.68080938e-01 7.62506008e-01
-1.87095749e+00 -6.48054779e-01 2.60101944e-01 2.50057578e+00
1.03499961e+00 -4.71357033e-02 4.76324260e-01 -1.14565328e-01
1.08375335e+00 -2.49218076e-01 -1.13508725e+00 -1.86542943e-01
8.69405344e-02 -2.95967925e-02 7.91308656e-02 7.70517811e-02
-1.13600910e+00 8.62666309e-01 5.03239822e+00 7.40945339e-01
-1.09509838e+00 1.74172565e-01 4.42687750e-01 3.42566520e-01
-2.26411745e-01 3.86228524e-02 -5.97396731e-01 6.07009888e-01
7.15831161e-01 -5.96839190e-01 2.72803277e-01 1.44148147e+00
-1.92687958e-01 4.14733410e-01 -1.24695086e+00 9.70309913e-01
1.47838280e-01 -1.08264852e+00 1.98534340e-01 -4.15724777e-02
8.45836401e-01 -4.43218470e-01 -1.52302340e-01 5.88580966e-01
5.45239866e-01 -5.45126438e-01 5.51403880e-01 4.08158042e-02
9.64354217e-01 -9.45833683e-01 4.41785604e-01 5.16405702e-01
-1.13751590e+00 -1.47140875e-01 -6.40930295e-01 3.05931777e-01
-1.02253310e-01 3.14625576e-02 -1.12076628e+00 5.28391004e-01
4.21941578e-01 7.94030666e-01 -4.79371279e-01 1.14851201e+00
-3.19366783e-01 5.36168516e-01 -1.62469313e-01 1.02509275e-01
-3.92589010e-02 -2.29215488e-01 4.45359796e-01 8.30087304e-01
2.71799922e-01 1.24436453e-01 2.67752171e-01 9.70486522e-01
-2.67045051e-01 4.91780549e-01 -5.42031050e-01 -1.70248616e-02
1.10260034e+00 1.14004517e+00 -3.34603190e-01 -3.94126266e-01
-3.87198776e-01 1.08000839e+00 4.22589660e-01 4.32841927e-01
-5.57712018e-01 -5.54796517e-01 6.48353100e-01 -1.60987258e-01
2.16555953e-01 2.98898071e-01 -1.97354659e-01 -1.30224657e+00
1.44576967e-01 -1.02940142e+00 7.83404827e-01 -4.67676461e-01
-1.60975468e+00 7.38355160e-01 -2.34284513e-02 -1.73373055e+00
-5.39979637e-01 2.13662209e-03 -7.92950869e-01 1.02663052e+00
-1.37149680e+00 -1.42145705e+00 -5.10675490e-01 9.47955251e-01
5.38571000e-01 -4.64046061e-01 9.74707007e-01 1.03089452e-01
-7.14082658e-01 1.11567307e+00 3.99647593e-01 3.84783000e-01
9.45482910e-01 -9.28504765e-01 5.70788383e-01 8.24516118e-01
-9.80382711e-02 5.83345294e-01 4.59617674e-01 -5.96351743e-01
-9.84389901e-01 -1.44677317e+00 7.37270057e-01 -1.71642944e-01
3.63138527e-01 -4.79323268e-01 -1.30696738e+00 7.50987172e-01
-1.14627697e-01 1.42066851e-01 8.12237024e-01 -5.27194925e-02
-6.62550211e-01 -2.19317138e-01 -1.41028500e+00 4.39303815e-01
8.16588223e-01 -2.09262103e-01 -8.05008233e-01 2.26593763e-01
6.60631895e-01 -2.79840916e-01 -7.61237264e-01 3.47901762e-01
4.05814171e-01 -3.98199469e-01 9.00105417e-01 -6.36301816e-01
2.17475340e-01 -4.28195298e-01 1.24394752e-01 -1.53111303e+00
-1.69327363e-01 -4.30144310e-01 -1.58139374e-02 1.79776227e+00
4.10423279e-01 -7.67814577e-01 7.14028001e-01 7.68995523e-01
1.66205540e-02 -3.01135302e-01 -8.26204598e-01 -1.01714385e+00
1.55698672e-01 1.16439477e-01 1.00245678e+00 1.35938144e+00
9.26735345e-03 6.91345572e-01 -4.00499463e-01 1.46694928e-01
8.17286193e-01 5.44053733e-01 1.19294941e+00 -1.27418351e+00
-1.29166141e-01 -1.90846533e-01 -1.37195900e-01 -9.36740279e-01
2.72820622e-01 -8.57253850e-01 3.46413553e-01 -1.19957280e+00
2.28684962e-01 -1.05762601e+00 -3.33784342e-01 5.53966880e-01
-3.22578222e-01 -1.67307202e-02 -3.09024891e-03 6.53946519e-01
-5.03368199e-01 6.89513147e-01 1.00741208e+00 -1.19769335e-01
-5.87386847e-01 2.26600900e-01 -9.57553148e-01 4.75462079e-01
8.53935122e-01 -5.48146665e-01 -7.02054322e-01 -1.81768373e-01
-4.33881015e-01 2.78498717e-02 1.98705196e-01 -9.31895435e-01
1.00305155e-01 -5.08609593e-01 4.42666203e-01 -2.25042135e-01
1.31922988e-02 -8.59873474e-01 4.44215536e-02 2.71505266e-01
-4.77160782e-01 -5.02102315e-01 -9.44265202e-02 6.53675795e-01
-1.93906710e-01 -2.03717008e-01 1.01852560e+00 1.12206690e-01
-1.16041565e+00 4.50841337e-01 1.48027837e-01 2.09727377e-01
1.26592243e+00 -2.78797776e-01 -3.66378486e-01 -9.89568755e-02
-5.71908832e-01 4.36780185e-01 7.38672078e-01 5.91648579e-01
5.28919280e-01 -1.65891051e+00 -8.16012383e-01 4.87654775e-01
8.40403736e-01 4.27580863e-01 2.61165082e-01 3.07036191e-01
1.38174906e-01 -1.45972706e-02 -5.84721267e-01 -5.68499982e-01
-1.10862863e+00 1.12091660e+00 6.91410750e-02 1.27233058e-01
-5.93541384e-01 7.52359033e-01 5.05283058e-01 -8.02604377e-01
1.60194770e-01 2.46645421e-01 -4.46574062e-01 -1.01479784e-01
6.95069909e-01 1.43403104e-02 -2.73301899e-01 -9.34024870e-01
-3.69611919e-01 5.11551440e-01 -4.07887489e-01 4.03379165e-02
1.07815599e+00 -3.51080507e-01 3.74450713e-01 2.76074111e-01
1.20663869e+00 -2.58074969e-01 -1.41872954e+00 -8.18263113e-01
-7.83424675e-02 -5.12614250e-01 -6.31372452e-01 -8.55892777e-01
-8.80032539e-01 7.78125286e-01 7.15362310e-01 -3.07530969e-01
1.30607522e+00 -1.09718703e-02 8.76183867e-01 3.83913547e-01
1.82962954e-01 -1.17004466e+00 6.76705837e-02 1.26865283e-01
5.96230149e-01 -1.46495175e+00 -2.64236808e-01 -3.46270084e-01
-1.04861259e+00 8.56115699e-01 9.86306906e-01 7.22889602e-02
3.48793030e-01 -4.45237070e-01 3.90178263e-01 2.68503696e-01
-4.47950274e-01 -1.37880805e-03 1.77170962e-01 1.06523156e+00
-2.66809016e-01 6.21856749e-02 -1.25062436e-01 9.68974531e-01
-1.21661723e-01 1.23876520e-01 4.43859041e-01 9.91062224e-01
-2.73536354e-01 -1.49169970e+00 -3.07672948e-01 2.24634841e-01
1.09278269e-01 2.69095153e-01 -5.82416773e-01 6.91523969e-01
1.08903989e-01 9.88950253e-01 -1.62158564e-01 -6.51216507e-01
5.08511364e-01 2.55231351e-01 9.26076099e-02 -7.26882219e-01
-1.73780367e-01 -2.32931614e-01 -1.75128385e-01 -1.07339345e-01
-1.99493781e-01 -8.02263021e-01 -1.18606830e+00 1.47679076e-02
-2.50407517e-01 4.62849200e-01 6.13754213e-01 9.17080283e-01
5.02307951e-01 5.97555563e-02 9.72491562e-01 -3.40518892e-01
-7.82173634e-01 -8.86984348e-01 -3.33098382e-01 8.76580298e-01
1.86417952e-01 -7.31207430e-01 -6.13571964e-02 3.72350544e-01] | [10.38912296295166, 3.0901811122894287] |
3f9202f6-6a2b-4300-a52a-405802e963a9 | not-all-tokens-are-equal-human-centric-visual | 2204.08680 | null | https://arxiv.org/abs/2204.08680v3 | https://arxiv.org/pdf/2204.08680v3.pdf | Not All Tokens Are Equal: Human-centric Visual Analysis via Token Clustering Transformer | Vision transformers have achieved great successes in many computer vision tasks. Most methods generate vision tokens by splitting an image into a regular and fixed grid and treating each cell as a token. However, not all regions are equally important in human-centric vision tasks, e.g., the human body needs a fine representation with many tokens, while the image background can be modeled by a few tokens. To address this problem, we propose a novel Vision Transformer, called Token Clustering Transformer (TCFormer), which merges tokens by progressive clustering, where the tokens can be merged from different locations with flexible shapes and sizes. The tokens in TCFormer can not only focus on important areas but also adjust the token shapes to fit the semantic concept and adopt a fine resolution for regions containing critical details, which is beneficial to capturing detailed information. Extensive experiments show that TCFormer consistently outperforms its counterparts on different challenging human-centric tasks and datasets, including whole-body pose estimation on COCO-WholeBody and 3D human mesh reconstruction on 3DPW. Code is available at https://github.com/zengwang430521/TCFormer.git | ['Wanli Ouyang', 'Xiaogang Wang', 'Ping Luo', 'Chen Qian', 'Wentao Liu', 'Sheng Jin', 'Wang Zeng'] | 2022-04-19 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zeng_Not_All_Tokens_Are_Equal_Human-Centric_Visual_Analysis_via_Token_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zeng_Not_All_Tokens_Are_Equal_Human-Centric_Visual_Analysis_via_Token_CVPR_2022_paper.pdf | cvpr-2022-1 | ['2d-human-pose-estimation'] | ['computer-vision'] | [-2.57320613e-01 9.13291499e-02 -1.24269687e-01 -1.67281032e-01
-3.72809231e-01 -2.28659764e-01 3.97115767e-01 -5.66276396e-03
-3.22462738e-01 3.75291288e-01 2.78917462e-01 3.79382819e-01
2.56145507e-01 -6.57073975e-01 -6.52779639e-01 -8.34762335e-01
5.26533246e-01 8.67057502e-01 7.14362562e-01 -9.97776687e-02
-1.09980308e-01 4.76566218e-02 -1.42683291e+00 1.82995170e-01
9.32655871e-01 8.35964024e-01 5.93147457e-01 6.13013916e-02
-2.03299627e-01 4.70856756e-01 -5.77550232e-01 -3.95675242e-01
3.01727772e-01 -1.40715733e-01 -6.90928042e-01 3.83894652e-01
6.24117196e-01 6.97776675e-02 -1.64321661e-01 1.23318648e+00
5.49320161e-01 6.99518919e-02 6.57760143e-01 -1.14517510e+00
-4.95872736e-01 6.12353206e-01 -1.12960398e+00 -1.47057191e-01
-2.33171917e-02 2.70220608e-01 6.97195947e-01 -1.30434835e+00
6.18226230e-01 1.53638053e+00 6.00165486e-01 5.80062091e-01
-1.16899419e+00 -5.56968093e-01 4.53907669e-01 4.07970577e-01
-1.34143686e+00 -1.94247454e-01 9.39411640e-01 -5.00750721e-01
3.49949181e-01 2.15347543e-01 1.11726832e+00 7.87961960e-01
2.11839139e-01 1.03338611e+00 1.00510144e+00 -2.13459328e-01
2.98899002e-02 -2.34774247e-01 2.01244634e-02 8.47118199e-01
3.39807898e-01 -3.14921886e-01 -5.53055644e-01 4.45599901e-03
1.09241426e+00 2.74892390e-01 -4.37342644e-01 -7.01869309e-01
-1.73547077e+00 5.93732774e-01 6.36811376e-01 7.07563534e-02
-5.23438096e-01 2.45685145e-01 4.94735837e-01 -2.92511106e-01
4.01314050e-01 -1.66241437e-01 -2.75632113e-01 3.26865852e-01
-7.06550062e-01 3.00405115e-01 2.58481324e-01 1.00002980e+00
8.25092673e-01 4.96779978e-02 -4.77686524e-01 1.14365590e+00
5.03998697e-01 6.04258716e-01 5.11212349e-01 -1.00173354e+00
1.63538277e-01 5.44893920e-01 1.78347640e-02 -7.60550976e-01
-2.05008879e-01 -4.80366826e-01 -1.14238703e+00 2.06962928e-01
5.75264573e-01 3.39320898e-02 -1.41106069e+00 1.42782843e+00
1.00036037e+00 6.43598288e-02 -5.48765957e-01 1.28571224e+00
1.18633986e+00 5.28202593e-01 3.84997413e-03 4.73726094e-02
1.90790796e+00 -1.28956902e+00 -4.90800560e-01 -2.46413991e-01
-2.95225233e-02 -8.54338586e-01 1.04063118e+00 4.81751651e-01
-1.27713859e+00 -7.30863810e-01 -5.65289617e-01 -2.56298244e-01
6.37944043e-02 2.05352649e-01 3.67001086e-01 1.79909840e-01
-8.87556016e-01 -4.03482392e-02 -8.00071239e-01 -1.58191875e-01
3.52886021e-01 -8.97815526e-02 -2.88665444e-01 -3.69572729e-01
-9.70518589e-01 6.72381043e-01 2.64607310e-01 3.92797738e-01
-8.72459769e-01 -6.67739034e-01 -9.91316378e-01 -1.51913673e-01
4.98353779e-01 -1.05379570e+00 1.14468801e+00 -8.47940207e-01
-1.20812297e+00 1.07360768e+00 -3.72684926e-01 -1.00504845e-01
8.80583584e-01 -2.73379125e-02 3.01951051e-01 1.69743046e-01
3.26871455e-01 9.78845119e-01 1.13561261e+00 -1.51606584e+00
-6.72891378e-01 -5.38966775e-01 -1.59354717e-01 3.06864113e-01
-9.75097343e-03 -1.58000097e-01 -1.16875970e+00 -1.04130363e+00
4.03249919e-01 -9.11431491e-01 -3.88291508e-01 2.96788126e-01
-6.04570210e-01 -5.47205091e-01 6.11113608e-01 -4.61258084e-01
7.63965845e-01 -1.89195430e+00 3.76935750e-01 1.43840924e-01
6.49778962e-01 8.61227438e-02 1.31724387e-01 -5.48127014e-03
2.50484556e-01 -2.79197663e-01 -2.33455583e-01 -6.31374836e-01
-3.99029925e-02 3.98611903e-01 1.99228317e-01 4.96723741e-01
-2.25536749e-01 1.02443242e+00 -9.08961117e-01 -9.95257556e-01
5.20560265e-01 5.84357262e-01 -4.54236448e-01 -1.87655777e-01
-2.63920546e-01 5.44412851e-01 -5.49743772e-01 9.92786348e-01
8.52601886e-01 -2.47759312e-01 -5.06880879e-02 -4.70158547e-01
-7.10691437e-02 -2.36157089e-01 -1.13434923e+00 1.75554752e+00
-2.02254821e-02 3.02698463e-01 4.71594572e-01 -9.69582975e-01
6.60699487e-01 1.40567347e-01 7.07902730e-01 -5.35629213e-01
8.77858326e-02 -1.31422862e-01 -2.27277726e-01 -3.78200889e-01
3.59941930e-01 -1.36458814e-01 1.08332776e-01 -6.43882379e-02
-2.28549376e-01 -1.11246683e-01 5.17302714e-02 5.78636453e-02
5.15895009e-01 5.81887411e-03 6.05700053e-02 -3.79211277e-01
3.64860952e-01 2.43381456e-01 1.21505129e+00 5.53137660e-01
-3.72593552e-01 1.21728289e+00 4.40613627e-02 -4.45447922e-01
-7.21740842e-01 -1.35350657e+00 -1.33716643e-01 8.57729912e-01
6.78509533e-01 -4.09608543e-01 -8.80868137e-01 -5.62354088e-01
1.14242062e-01 1.02561064e-01 -7.17831135e-01 1.04956785e-02
-6.24069691e-01 -4.55847114e-01 1.65729895e-01 5.91544092e-01
6.59031808e-01 -1.28548539e+00 -7.13117898e-01 1.23051979e-01
-8.16610575e-01 -9.27049577e-01 -9.46683228e-01 -9.34675187e-02
-7.13505328e-01 -1.20694959e+00 -1.29319274e+00 -1.16151536e+00
1.02473986e+00 6.74924612e-01 1.25555730e+00 1.60432354e-01
-5.53321004e-01 4.36563998e-01 -4.68982100e-01 -5.77071488e-01
1.70073882e-01 -1.53098777e-01 7.20342323e-02 7.96633363e-02
8.20257142e-02 -2.50002593e-01 -1.03905451e+00 5.56717217e-01
-6.22141242e-01 3.44178140e-01 5.18539369e-01 9.90293145e-01
1.01782489e+00 -3.56832817e-02 6.80715218e-02 -6.29488945e-01
1.95181102e-01 -2.72958100e-01 -2.49019936e-01 1.84119165e-01
-1.51200399e-01 -4.14265037e-01 3.23469371e-01 -5.86069882e-01
-1.01147199e+00 1.88313156e-01 -6.17482699e-02 -8.75590026e-01
-1.86765403e-01 3.16584796e-01 -2.75561512e-01 6.12752214e-02
2.41491094e-01 4.04908329e-01 7.01277778e-02 -5.71252167e-01
3.43405098e-01 2.00855315e-01 7.25510895e-01 -9.32197213e-01
7.31919587e-01 6.52580857e-01 -2.29394972e-01 -8.88198555e-01
-5.93498647e-01 -5.80976188e-01 -6.84442639e-01 -5.40284514e-01
9.76430058e-01 -9.45763350e-01 -5.03113568e-01 9.02228117e-01
-1.10764611e+00 -4.07514542e-01 -4.86614436e-01 3.80909592e-01
-3.99821967e-01 5.12439132e-01 -7.02529192e-01 -3.93148571e-01
-4.90763098e-01 -1.18782818e+00 1.21599901e+00 4.26605999e-01
-1.13473304e-01 -7.99731851e-01 -1.40136138e-01 6.22092843e-01
-3.22844461e-02 2.77128279e-01 6.87744141e-01 1.86982617e-01
-4.86173570e-01 2.12268770e-01 -1.04484968e-01 2.21261680e-01
2.38306016e-01 3.55818793e-02 -6.77540481e-01 -3.00741345e-01
-5.29582910e-02 -2.74722993e-01 1.13776851e+00 9.59622383e-01
1.24468505e+00 -6.31502494e-02 -3.44996959e-01 6.04581058e-01
9.73648548e-01 -1.96976349e-01 6.06226087e-01 1.21533968e-01
1.14622009e+00 6.81670129e-01 9.50261533e-01 4.73955482e-01
9.04664099e-01 7.87347674e-01 5.60902894e-01 -5.29573441e-01
-4.02150780e-01 -1.66774064e-01 2.62520283e-01 6.87564790e-01
-3.85180771e-01 1.36303991e-01 -9.25630927e-01 7.42240012e-01
-2.04927683e+00 -7.87899733e-01 -3.20942611e-01 2.06874347e+00
9.60718453e-01 6.72501996e-02 2.95686245e-01 -1.07909933e-01
8.60947490e-01 4.29904908e-02 -6.69020236e-01 3.92101735e-01
1.16492562e-01 -6.38432056e-02 3.38129282e-01 2.75204271e-01
-9.71855581e-01 1.09358239e+00 5.17984295e+00 1.09347987e+00
-9.93163168e-01 3.09494197e-01 5.85107625e-01 -9.29109901e-02
-1.59502149e-01 -1.33055776e-01 -6.07552171e-01 7.43305445e-01
-3.27270508e-01 -1.53922206e-02 -1.95899591e-01 7.95456231e-01
4.39055413e-01 -3.03299457e-01 -7.12783456e-01 1.33037174e+00
-4.17759158e-02 -1.14186978e+00 3.40231866e-01 4.73786257e-02
5.93076885e-01 1.23349689e-02 -8.59880634e-03 -1.63656529e-02
2.90475309e-01 -8.95897448e-01 1.22295833e+00 5.08182049e-01
6.17632568e-01 -6.33853316e-01 4.50953037e-01 4.60463434e-01
-1.69887209e+00 2.69531995e-01 -6.85988545e-01 1.92391723e-01
2.35345494e-02 9.08773661e-01 -4.40373749e-01 5.61869204e-01
1.19529700e+00 8.65749300e-01 -4.06127691e-01 1.25539553e+00
-1.55806318e-01 3.04134578e-01 -2.16061875e-01 3.10242474e-01
4.45081703e-02 -3.55277151e-01 5.55427670e-01 1.04979038e+00
1.21162899e-01 1.70705676e-01 7.32983530e-01 7.73707926e-01
1.61468565e-01 2.30983850e-02 -1.35870367e-01 6.72513545e-01
5.66640675e-01 1.47596300e+00 -1.04758358e+00 -5.00145555e-01
-2.87939250e-01 9.30076122e-01 1.75916508e-01 4.22888100e-01
-8.62607241e-01 -9.78119075e-02 7.82767236e-01 4.95148152e-01
4.07194793e-01 -2.81509012e-01 -3.50580394e-01 -1.37126899e+00
2.54211962e-01 -7.63230324e-01 4.69081819e-01 -7.40152180e-01
-1.29776144e+00 2.40459636e-01 1.39871135e-01 -1.48977292e+00
2.88439184e-01 -3.16221625e-01 -6.80783689e-01 6.31314933e-01
-1.23543763e+00 -1.43160224e+00 -7.02339292e-01 9.26904023e-01
9.89272475e-01 3.33229929e-01 1.86569273e-01 2.49045029e-01
-5.65715909e-01 4.76702988e-01 -2.45395586e-01 2.81120390e-01
9.49294448e-01 -1.28668225e+00 2.91794777e-01 7.00158834e-01
4.20054495e-02 6.53584301e-01 6.03268802e-01 -8.47553730e-01
-1.30362535e+00 -1.15705526e+00 4.68602866e-01 -3.83207113e-01
1.20958596e-01 -2.78422952e-01 -8.73704851e-01 4.73518997e-01
1.78760603e-01 2.73476452e-01 2.69781742e-02 -1.03185035e-01
-1.38540223e-01 -1.45582780e-01 -1.14268434e+00 5.79467714e-01
1.18867409e+00 2.39504501e-02 -6.03987515e-01 2.93305606e-01
4.17942643e-01 -7.51344264e-01 -7.93158412e-01 4.71830517e-01
5.20050526e-01 -9.40416694e-01 1.26724470e+00 9.94144902e-02
3.69710743e-01 -7.42591321e-01 2.25636527e-01 -1.41440070e+00
-5.46067357e-01 -1.73238054e-01 -1.21398434e-01 1.21394646e+00
-2.08207533e-01 -6.03883386e-01 8.65747631e-01 3.30710828e-01
-2.95550823e-01 -8.95775855e-01 -9.91104603e-01 -5.53508282e-01
3.68746892e-02 -1.19850688e-01 3.29820633e-01 9.76186156e-01
-2.45740011e-01 4.88687642e-02 -4.05578494e-01 -5.95460050e-02
1.17554665e+00 2.11545244e-01 8.60138774e-01 -1.34908545e+00
-2.40758806e-02 -4.22143638e-01 -3.20473850e-01 -1.30603945e+00
-1.82940513e-01 -7.30143309e-01 3.05403352e-01 -1.87787092e+00
5.22762954e-01 -5.67176282e-01 -1.30551398e-01 7.72599936e-01
-3.74256313e-01 5.16354322e-01 2.74135262e-01 3.86937559e-01
-5.39175868e-01 8.42926264e-01 1.78313398e+00 -4.90308493e-01
1.64243087e-01 -6.17210381e-02 -4.95043248e-01 1.02821136e+00
5.83115458e-01 -2.28912488e-01 -4.36963379e-01 -5.73874772e-01
-3.38685364e-01 -1.65672168e-01 7.52732575e-01 -9.75046039e-01
3.63563895e-01 -3.35160404e-01 7.34784484e-01 -1.01535547e+00
4.00893778e-01 -7.13626027e-01 1.92755327e-01 6.44605339e-01
1.42780721e-01 1.01228161e-02 -1.04902647e-01 6.41892731e-01
-2.41344661e-01 1.76838130e-01 1.00876284e+00 -5.82074642e-01
-8.95545304e-01 6.35292292e-01 -2.68103361e-01 1.60895050e-01
1.05644381e+00 -5.65937996e-01 -4.25380021e-02 -7.19897524e-02
-8.62170577e-01 7.67452776e-01 7.93798923e-01 5.75710595e-01
1.06822693e+00 -1.40935993e+00 -8.35590065e-01 3.39892544e-02
1.35019228e-01 7.92281926e-01 5.37419140e-01 1.03377008e+00
-4.19369072e-01 3.65165360e-02 -2.79646039e-01 -1.14656627e+00
-1.60572541e+00 3.12063098e-01 4.64327604e-01 2.72387657e-02
-1.24416363e+00 8.55277121e-01 9.28382039e-01 -2.49729201e-01
3.57940108e-01 -5.45944512e-01 -1.67596698e-01 1.38644189e-01
3.39544147e-01 2.56538808e-01 -1.79995611e-01 -8.36803854e-01
-4.38698232e-01 1.06231833e+00 -9.90466848e-02 1.01399079e-01
9.75802898e-01 -2.66460598e-01 -1.04961917e-01 2.50088304e-01
5.63760877e-01 4.80806269e-02 -1.47671413e+00 -5.14496565e-01
-4.84557539e-01 -6.43448055e-01 -1.08172081e-01 -3.61457586e-01
-1.65628242e+00 6.94932461e-01 4.18351263e-01 -2.09416226e-01
1.01321054e+00 3.10011327e-01 9.90066767e-01 -1.14839420e-01
7.44363725e-01 -1.15512300e+00 3.75763834e-01 5.16843438e-01
9.92707849e-01 -1.20503199e+00 -6.07796535e-02 -6.66940272e-01
-8.16264808e-01 8.60684931e-01 9.65591729e-01 -1.07859328e-01
5.70693851e-01 1.14312813e-01 2.59340167e-01 -4.51746732e-01
-3.84813368e-01 -2.98693120e-01 3.36340606e-01 8.30332518e-01
1.58144712e-01 2.37330511e-01 -3.13428730e-01 3.87447655e-01
-4.40812074e-02 -2.75181025e-01 2.38366053e-01 7.65955031e-01
-5.27540267e-01 -9.40922558e-01 -8.86044502e-01 3.79993290e-01
-1.78781256e-01 5.93131445e-02 -3.53345752e-01 5.61538994e-01
5.60847819e-01 7.50968456e-01 -1.16610527e-02 -2.72061557e-01
2.45242104e-01 -3.03901911e-01 5.53980768e-01 -7.32759297e-01
-5.26606023e-01 5.17832816e-01 -2.50699461e-01 -4.92956311e-01
-4.83194411e-01 -8.43817532e-01 -1.57085419e+00 -2.60385096e-01
-1.30127877e-01 4.78823408e-02 1.62519187e-01 7.55586922e-01
4.26925309e-02 6.64318085e-01 3.19089472e-01 -1.27965200e+00
-1.19623058e-01 -8.46033633e-01 -6.33774757e-01 3.56835932e-01
1.73232034e-01 -1.10559022e+00 7.16876686e-02 3.59956115e-01] | [7.9678473472595215, -0.47895488142967224] |
d0103744-bb7a-4b88-9cf4-101b287e0eb4 | a-fixed-point-model-for-pancreas-segmentation | 1612.08230 | null | http://arxiv.org/abs/1612.08230v4 | http://arxiv.org/pdf/1612.08230v4.pdf | A Fixed-Point Model for Pancreas Segmentation in Abdominal CT Scans | Deep neural networks have been widely adopted for automatic organ
segmentation from abdominal CT scans. However, the segmentation accuracy of
some small organs (e.g., the pancreas) is sometimes below satisfaction,
arguably because deep networks are easily disrupted by the complex and variable
background regions which occupies a large fraction of the input volume. In this
paper, we formulate this problem into a fixed-point model which uses a
predicted segmentation mask to shrink the input region. This is motivated by
the fact that a smaller input region often leads to more accurate segmentation.
In the training process, we use the ground-truth annotation to generate
accurate input regions and optimize network weights. On the testing stage, we
fix the network parameters and update the segmentation results in an iterative
manner. We evaluate our approach on the NIH pancreas segmentation dataset, and
outperform the state-of-the-art by more than 4%, measured by the average
Dice-S{\o}rensen Coefficient (DSC). In addition, we report 62.43% DSC in the
worst case, which guarantees the reliability of our approach in clinical
applications. | ['Yuyin Zhou', 'Elliot K. Fishman', 'Wei Shen', 'Lingxi Xie', 'Alan L. Yuille', 'Yan Wang'] | 2016-12-25 | null | null | null | null | ['pancreas-segmentation'] | ['medical'] | [ 1.10559694e-01 3.65870923e-01 -7.41576180e-02 -3.18043888e-01
-4.22870815e-01 -6.02639019e-01 4.24415208e-02 4.04570460e-01
-6.89720094e-01 7.58701205e-01 -2.00921968e-01 -2.35048130e-01
1.03313498e-01 -6.16243362e-01 -5.93749940e-01 -8.94861042e-01
-9.28551778e-02 6.01763248e-01 3.73354316e-01 2.08475024e-01
-8.09624617e-04 4.84410971e-01 -5.60411274e-01 -1.70651942e-01
1.32391059e+00 1.09070003e+00 1.07253762e-02 4.37427402e-01
-2.09276512e-01 4.66643423e-01 -4.36503947e-01 -4.68664765e-01
4.63338435e-01 -7.58574545e-01 -6.91520095e-01 4.89483736e-02
1.58941463e-01 -3.62249404e-01 -1.27145082e-01 1.23546982e+00
3.47225338e-01 9.68195349e-02 5.34090936e-01 -7.32936859e-01
-1.20788351e-01 7.77323425e-01 -5.86905181e-01 1.04549445e-01
-2.31471419e-01 2.81453609e-01 5.40821254e-01 -3.82025003e-01
5.79978108e-01 6.00422859e-01 8.84626031e-01 3.53750378e-01
-1.28492963e+00 -5.19895434e-01 9.83855724e-02 -5.13509512e-01
-1.33624625e+00 -6.84881359e-02 5.53082883e-01 -4.87797230e-01
3.35142046e-01 2.46415824e-01 9.17908251e-01 4.43737566e-01
5.53228319e-01 5.48233807e-01 7.45088458e-01 -1.79270640e-01
2.67404646e-01 -8.24073106e-02 -1.12303486e-03 7.03385949e-01
5.66178977e-01 -1.88465476e-01 3.55133504e-01 -9.70985857e-04
1.12784457e+00 4.74529602e-02 -5.36966801e-01 -5.51167250e-01
-1.41065812e+00 8.24872911e-01 8.99593294e-01 4.18070763e-01
-6.58186674e-01 4.51657884e-02 4.36957598e-01 -1.33497268e-01
3.19493830e-01 5.57892621e-01 -3.22292477e-01 1.34806275e-01
-1.19153845e+00 1.31524317e-02 9.38329041e-01 5.40545106e-01
2.39896044e-01 -1.70004666e-01 -2.03661188e-01 7.47411847e-01
3.92823935e-01 6.01045713e-02 5.38373351e-01 -9.67638612e-01
2.91601837e-01 6.56434894e-01 -7.27788871e-03 -7.54881322e-01
-7.52988279e-01 -7.17239738e-01 -1.19961262e+00 9.77053568e-02
9.80772972e-01 -3.85481089e-01 -1.26084018e+00 1.48889458e+00
6.21927619e-01 6.19879737e-03 -9.28847939e-02 1.28550112e+00
7.80816615e-01 3.56433541e-01 2.27450788e-01 -2.99173445e-01
1.09059596e+00 -1.07698488e+00 -6.16295934e-01 -1.05527632e-01
6.96263313e-01 -4.89278644e-01 7.00833976e-01 3.50608528e-01
-1.11901295e+00 -1.77135348e-01 -8.88631523e-01 3.46596956e-01
1.49050906e-01 1.89838260e-02 4.77992803e-01 5.48068225e-01
-9.47410524e-01 8.93648982e-01 -1.28460360e+00 -1.82049125e-01
7.63457417e-01 5.71859598e-01 -1.95134684e-01 6.71654865e-02
-8.98272336e-01 7.59004474e-01 6.76535845e-01 2.89809734e-01
-4.19971049e-01 -8.46620977e-01 -7.50736773e-01 1.54743701e-01
4.18673456e-01 -7.03000546e-01 1.11476481e+00 -9.43098485e-01
-1.71440232e+00 7.60265291e-01 3.60068023e-01 -5.82859039e-01
1.10445273e+00 1.17449515e-01 1.76143453e-01 2.31727719e-01
-2.56707788e-01 7.81707823e-01 3.68148565e-01 -1.25451958e+00
-2.42342785e-01 -3.14910561e-01 1.34346271e-02 1.64540127e-01
8.89967009e-02 -2.15086177e-01 -6.23722076e-01 -6.67331696e-01
5.33043087e-01 -1.23210871e+00 -7.66306520e-01 3.76282364e-01
-5.70943475e-01 2.74085164e-01 2.25269899e-01 -8.25723827e-01
1.10725451e+00 -2.20479727e+00 -3.63097079e-02 5.10716915e-01
5.20802438e-01 2.06578702e-01 2.11988777e-01 -3.91013235e-01
2.65124184e-03 9.44949687e-02 -7.60130405e-01 -1.35398507e-01
-2.86545962e-01 3.18089306e-01 2.96361655e-01 7.63732255e-01
-1.60180300e-01 9.57771182e-01 -9.11346614e-01 -8.76276970e-01
2.77276844e-01 3.90974939e-01 -6.01195455e-01 8.54727551e-02
-9.20754671e-02 8.72821391e-01 -4.69255090e-01 5.12656212e-01
7.25527346e-01 -5.19753695e-01 3.73184651e-01 -1.35400936e-01
-7.86678046e-02 -1.46634370e-01 -1.05333710e+00 1.76513672e+00
-1.34141862e-01 3.79312277e-01 2.21620470e-01 -9.36296701e-01
7.74415016e-01 3.22353721e-01 9.47723508e-01 -4.41384226e-01
3.83269519e-01 4.05681521e-01 5.57247400e-01 -3.08309108e-01
1.45735800e-01 -3.47985506e-01 2.21374065e-01 3.13888848e-01
-1.51521251e-01 -2.42879033e-01 2.43201360e-01 -1.00534245e-01
7.93327510e-01 2.01642141e-02 4.28263187e-01 -4.81397897e-01
4.17413890e-01 1.39541551e-01 6.85051322e-01 5.51333487e-01
-5.55841923e-01 1.01958656e+00 7.82945871e-01 -6.08987987e-01
-9.76979196e-01 -8.18559051e-01 -4.07163292e-01 4.15265769e-01
4.46327299e-01 2.26497769e-01 -1.21603417e+00 -9.59122062e-01
-1.97534606e-01 5.20298243e-01 -5.51635802e-01 8.95349830e-02
-7.75499940e-01 -1.06473207e+00 5.15295088e-01 5.15832484e-01
3.73620927e-01 -1.16327977e+00 -1.00493908e+00 5.18647671e-01
-1.00221403e-01 -1.08438599e+00 -4.56219733e-01 2.28023559e-01
-1.32674611e+00 -1.04132605e+00 -1.26004791e+00 -6.28138661e-01
1.08674502e+00 -3.99104446e-01 1.15847027e+00 4.37354863e-01
-1.32192686e-01 -3.63589674e-01 -3.58451679e-02 -1.31153777e-01
-6.02450788e-01 2.25408614e-01 -3.52005184e-01 -3.80115598e-01
-2.47213259e-01 -2.15713367e-01 -9.50701118e-01 3.53846729e-01
-9.99845207e-01 2.05071107e-01 5.46534836e-01 9.08376992e-01
8.89449298e-01 -1.85318768e-01 2.39188686e-01 -9.04841125e-01
3.77067387e-01 -2.83167183e-01 -8.55248988e-01 2.15332851e-01
-5.15065968e-01 -8.02943856e-02 6.52361929e-01 -6.55245960e-01
-6.55147016e-01 3.44892353e-01 -1.93012834e-01 -3.54850948e-01
-1.72967821e-01 5.78264475e-01 3.72045845e-01 -3.95496458e-01
4.82208073e-01 -5.20354062e-02 1.98674053e-01 -1.35911375e-01
1.44580640e-02 9.84210595e-02 4.78372753e-01 -2.90748298e-01
3.53436649e-01 2.52257109e-01 1.69020101e-01 -3.48630905e-01
-6.19096220e-01 -9.68840569e-02 -8.49306107e-01 -2.78128147e-01
9.84273732e-01 -4.76758331e-01 -5.33294857e-01 3.10302824e-01
-1.02664804e+00 -7.43110001e-01 -4.26307648e-01 7.71259785e-01
-3.80052030e-01 4.05804664e-01 -8.40343297e-01 -3.87133807e-01
-6.08542800e-01 -1.65421200e+00 6.95598066e-01 4.46979940e-01
-2.45665133e-01 -1.11263227e+00 -1.37285385e-02 -2.96386797e-02
4.92506504e-01 8.06090236e-01 8.38802576e-01 -9.00441766e-01
-4.53955770e-01 -2.00714350e-01 -3.37582678e-01 2.77868003e-01
-5.37741855e-02 -5.67861646e-02 -4.55409378e-01 -1.51701748e-01
3.77449319e-02 1.02403939e-01 6.08713448e-01 9.66034234e-01
1.45277691e+00 -6.81082308e-02 -2.91319013e-01 8.44128549e-01
1.38801801e+00 3.31937373e-01 4.60466117e-01 1.67366117e-01
5.94657183e-01 3.62610042e-01 3.47944558e-01 2.96677172e-01
8.47343132e-02 3.57675403e-01 6.29193485e-01 -4.90978777e-01
-3.92538942e-02 2.04443887e-01 -3.69120359e-01 7.75607347e-01
-7.05945417e-02 -3.37111443e-01 -1.12149239e+00 4.40545499e-01
-1.58126163e+00 -2.62101352e-01 -2.27366224e-01 2.17597651e+00
8.69017005e-01 1.95763856e-01 2.31062621e-03 -2.37872690e-01
6.70206785e-01 -1.16282590e-01 -6.70792520e-01 -1.31392837e-01
3.20589691e-01 1.24164015e-01 7.05860496e-01 4.40741628e-01
-1.11401200e+00 5.89189589e-01 6.05938196e+00 3.97560209e-01
-1.43308103e+00 1.99948642e-02 1.12831497e+00 1.53016364e-02
5.15607819e-02 -2.84500062e-01 -1.78137735e-01 6.99000537e-01
5.73554218e-01 7.60513619e-02 3.33074331e-01 5.90253115e-01
1.00855850e-01 -2.75800496e-01 -9.67911124e-01 7.68370330e-01
-2.01144949e-01 -1.10674751e+00 -3.03173423e-01 1.95809603e-01
8.87593746e-01 2.48556569e-01 -3.01959664e-01 9.32593569e-02
-2.22196840e-02 -1.06044233e+00 5.00406265e-01 3.69443625e-01
8.30286264e-01 -5.58929145e-01 1.09390497e+00 3.38458687e-01
-8.13666284e-01 2.39646360e-01 -1.75719500e-01 5.67201495e-01
3.29663187e-01 8.20207238e-01 -8.45670402e-01 3.58651280e-01
5.38965762e-01 2.03856528e-01 -2.16681436e-01 1.65346706e+00
7.48769343e-02 4.63858724e-01 -7.01008201e-01 1.06771827e-01
4.88168031e-01 -6.84658051e-01 4.10853982e-01 1.14096808e+00
3.94479305e-01 3.07210326e-01 2.05140904e-01 1.09082973e+00
-3.50778252e-01 2.95594901e-01 -1.29809916e-01 2.14308530e-01
-4.26843157e-03 1.33341694e+00 -1.45417917e+00 -4.42270607e-01
-1.33521006e-01 8.90275478e-01 -5.70107847e-02 2.57947654e-01
-9.81144845e-01 -1.10862784e-01 1.66181430e-01 1.97337121e-02
1.94936499e-01 8.29127878e-02 -6.75260305e-01 -1.02703226e+00
-2.44023465e-02 -5.57598650e-01 3.25942457e-01 -2.19403610e-01
-9.59555924e-01 7.44368434e-01 -1.96963876e-01 -1.15445387e+00
-1.18213013e-01 -4.43631858e-01 -5.36315322e-01 7.58309126e-01
-1.32526433e+00 -5.40954173e-01 -6.24616683e-01 5.21886609e-02
3.69684458e-01 4.36844140e-01 5.87901831e-01 4.30858105e-01
-5.51704764e-01 5.53276122e-01 1.38250142e-01 4.13393378e-01
4.23381180e-01 -1.31195331e+00 2.01107040e-01 7.16788948e-01
-2.40081802e-01 4.56926465e-01 5.83413899e-01 -6.30997300e-01
-9.35025632e-01 -9.88892853e-01 4.90695566e-01 1.98867507e-02
5.03259182e-01 2.20683077e-03 -1.07578838e+00 5.93411684e-01
5.75457700e-02 6.35635734e-01 4.70145524e-01 -4.50128466e-01
4.92643148e-01 1.53986707e-01 -1.57022631e+00 6.06675208e-01
6.46429956e-01 4.05789912e-01 -2.62669176e-01 2.23492607e-01
5.72000623e-01 -1.15525472e+00 -1.26166916e+00 6.13587260e-01
6.62400842e-01 -8.92945290e-01 8.71365488e-01 -1.67323396e-01
5.13575196e-01 -2.61219084e-01 3.18426937e-01 -1.36996210e+00
-4.27058637e-02 -3.19520682e-01 -7.33558908e-02 6.57495797e-01
5.61517596e-01 -6.42497122e-01 9.57168400e-01 8.93513918e-01
-1.69861302e-01 -1.13761127e+00 -1.00128078e+00 -3.83356214e-01
2.75145739e-01 -5.65558895e-02 4.44215924e-01 1.12593448e+00
-2.90572852e-01 -2.98630863e-01 1.54004619e-01 8.52066278e-03
7.74066687e-01 1.22316755e-01 3.16025496e-01 -1.29671323e+00
9.63158999e-03 -7.21525252e-01 -1.99890286e-01 -9.73089397e-01
-1.31327823e-01 -8.17468643e-01 3.42272162e-01 -1.52414989e+00
2.43726343e-01 -7.34875858e-01 -4.25944835e-01 3.84500593e-01
-3.73121023e-01 4.14511859e-01 2.91046560e-01 1.37697786e-01
-4.91159827e-01 2.69179106e-01 1.76743543e+00 3.31348330e-02
-5.15937567e-01 6.57534227e-02 -4.38911468e-01 7.55826950e-01
9.46599305e-01 -5.31766176e-01 -9.11336672e-03 -3.27345222e-01
-1.70365408e-01 2.30784699e-01 2.02880934e-01 -9.37036753e-01
1.21885002e-01 -7.51937553e-02 7.01553226e-01 -4.14801776e-01
-1.59977049e-01 -1.19378853e+00 2.58375198e-01 9.47902262e-01
-3.14962208e-01 -8.42885897e-02 1.61544427e-01 1.04732625e-01
-1.08272433e-01 -4.17290837e-01 1.18382418e+00 -2.83142596e-01
-5.77884959e-03 4.93137896e-01 3.03661870e-03 6.57850504e-02
1.07582068e+00 -2.63469189e-01 8.73069316e-02 -1.16354257e-01
-7.93848991e-01 3.38721722e-01 4.93415803e-01 -1.52529851e-01
3.16348374e-01 -1.03676057e+00 -6.18396282e-01 2.43361518e-01
-3.17875803e-01 7.00797498e-01 1.93007231e-01 1.52001595e+00
-1.21199119e+00 1.64563566e-01 -2.54174904e-03 -9.60841477e-01
-9.01543498e-01 3.23589027e-01 8.18549871e-01 -5.03552794e-01
-1.06753099e+00 6.63246870e-01 2.98215657e-01 -3.62826973e-01
2.05835074e-01 -9.30391192e-01 -6.64813593e-02 -3.24345142e-01
1.74625650e-01 2.06724092e-01 2.20287545e-03 -4.78142798e-01
-3.72372538e-01 3.66796702e-01 5.47231846e-02 2.66816765e-01
1.33253801e+00 6.25730082e-02 -1.48883462e-01 7.96881784e-03
1.03111506e+00 -2.36588418e-01 -1.44929183e+00 -4.05166782e-02
-1.37868777e-01 -3.38490039e-01 2.20033869e-01 -9.62678373e-01
-1.57402980e+00 5.75332761e-01 8.09386611e-01 2.88631976e-01
9.63185906e-01 -1.26704216e-01 1.01659906e+00 -3.53490226e-02
3.58383469e-02 -8.08877170e-01 -4.57721174e-01 3.39975983e-01
6.04880691e-01 -1.29920053e+00 8.77523497e-02 -4.65948582e-01
-6.95852995e-01 1.19343126e+00 6.21957541e-01 -2.48128951e-01
4.18948472e-01 3.93226355e-01 3.94491345e-01 -1.74582861e-02
-6.07251488e-02 3.72169942e-01 2.75280774e-01 2.47226302e-02
5.64523876e-01 1.27711356e-01 -4.36051369e-01 3.77644837e-01
-6.74714521e-02 1.04755320e-01 2.61450171e-01 6.76516950e-01
-2.88299143e-01 -6.46441936e-01 -4.48915690e-01 6.29484534e-01
-8.57294738e-01 -4.28190688e-03 1.97530165e-02 9.85642135e-01
-1.83403455e-02 3.63463938e-01 7.74588510e-02 2.61764526e-01
1.86030880e-01 -1.63861603e-01 4.80354816e-01 -2.76888639e-01
-9.00024533e-01 3.67218763e-01 -4.03419316e-01 -5.65935612e-01
-3.23101819e-01 -5.75581789e-01 -1.73576891e+00 -7.53366277e-02
-3.29549819e-01 5.79207763e-02 7.08974659e-01 9.10990477e-01
-9.23066810e-02 8.39284599e-01 2.68267751e-01 -8.73682857e-01
-5.54517686e-01 -9.71621513e-01 -3.60378653e-01 3.96031767e-01
3.60468298e-01 -4.15792078e-01 -2.30626896e-01 -4.82266210e-02] | [14.520011901855469, -2.6318166255950928] |
2ddb0534-5647-42aa-a9fc-f715a835d820 | processing-long-legal-documents-with-pre | 2211.00974 | null | https://arxiv.org/abs/2211.00974v2 | https://arxiv.org/pdf/2211.00974v2.pdf | Processing Long Legal Documents with Pre-trained Transformers: Modding LegalBERT and Longformer | Pre-trained Transformers currently dominate most NLP tasks. They impose, however, limits on the maximum input length (512 sub-words in BERT), which are too restrictive in the legal domain. Even sparse-attention models, such as Longformer and BigBird, which increase the maximum input length to 4,096 sub-words, severely truncate texts in three of the six datasets of LexGLUE. Simpler linear classifiers with TF-IDF features can handle texts of any length, require far less resources to train and deploy, but are usually outperformed by pre-trained Transformers. We explore two directions to cope with long legal texts: (i) modifying a Longformer warm-started from LegalBERT to handle even longer texts (up to 8,192 sub-words), and (ii) modifying LegalBERT to use TF-IDF representations. The first approach is the best in terms of performance, surpassing a hierarchical version of LegalBERT, which was the previous state of the art in LexGLUE. The second approach leads to computationally more efficient models at the expense of lower performance, but the resulting models still outperform overall a linear SVM with TF-IDF features in long legal document classification. | ['Ilias Chalkidis', 'Ion Androutsopoulos', 'Petros Tsotsi', 'Dimitris Mamakas'] | 2022-11-02 | null | null | null | null | ['document-classification'] | ['natural-language-processing'] | [-2.40410015e-01 -1.23215988e-02 -2.61212260e-01 -1.25109598e-01
-1.00965834e+00 -9.41141367e-01 8.26726675e-01 2.46763110e-01
-6.84902906e-01 7.85040855e-01 4.95171279e-01 -9.15866971e-01
-4.05657530e-01 -9.45640028e-01 -6.95440531e-01 -5.88794172e-01
-9.13741440e-02 8.89415085e-01 2.25858897e-01 -4.46227789e-01
6.64346442e-02 5.46072960e-01 -1.73608899e+00 5.39875865e-01
9.31019068e-01 1.00840342e+00 1.15985304e-01 6.66175723e-01
-5.45724630e-01 7.57127345e-01 -7.76837945e-01 -6.74283803e-01
2.51699984e-01 1.36116847e-01 -1.04656887e+00 -3.00621003e-01
6.86863840e-01 -3.46585333e-01 -5.74692488e-01 4.45754915e-01
2.82614142e-01 2.29121432e-01 7.31671870e-01 -8.63876402e-01
-6.63318753e-01 9.91157591e-01 -2.33691186e-01 4.64290857e-01
3.72928441e-01 -2.29845688e-01 1.44544172e+00 -9.49804425e-01
3.96189898e-01 1.24211085e+00 8.83047462e-01 9.59631205e-02
-1.02067912e+00 -4.48312670e-01 2.72277921e-01 4.37644154e-01
-1.24427998e+00 -4.21551049e-01 8.70032012e-02 -2.73986846e-01
1.66876853e+00 5.29189050e-01 6.17263198e-01 1.03038096e+00
8.72319937e-02 7.83166289e-01 9.30321813e-01 -4.28901732e-01
-1.05748735e-01 -2.90480889e-02 2.62010604e-01 4.88447636e-01
2.32832402e-01 -8.37036371e-02 -3.01259041e-01 -3.97216856e-01
1.76652491e-01 -1.03641264e-01 -2.78554946e-01 1.37161747e-01
-1.10830045e+00 1.30123043e+00 2.96754271e-01 7.35789001e-01
-1.50733694e-01 3.47133689e-02 8.56741428e-01 6.18947744e-01
7.40300298e-01 6.49539530e-01 -8.91189039e-01 -3.03276181e-01
-1.11961555e+00 3.90234321e-01 1.05005336e+00 8.96817982e-01
6.21814072e-01 1.20318919e-01 -2.38165349e-01 1.05326724e+00
-1.78367347e-01 5.40435016e-01 5.21301985e-01 -4.87463087e-01
8.93170059e-01 3.47550720e-01 -1.56078339e-01 -5.50377786e-01
-4.94824558e-01 -6.08954787e-01 -6.18402362e-01 -2.79190063e-01
6.03488147e-01 -1.04875565e-01 -1.03223443e+00 1.04751420e+00
1.22984178e-01 -4.24836457e-01 -3.52956094e-02 6.44023120e-01
8.30280423e-01 1.12351656e+00 -1.42441884e-01 -5.85696213e-02
1.36980343e+00 -1.04253471e+00 -5.02486944e-01 -2.70046800e-01
1.01671529e+00 -8.64094853e-01 1.18197417e+00 5.61537564e-01
-8.48035336e-01 -3.71632785e-01 -7.76046515e-01 -4.89186466e-01
-9.09928501e-01 2.03359649e-01 9.56459880e-01 9.37614858e-01
-9.66744840e-01 6.53846204e-01 -4.22367603e-01 -4.00190264e-01
2.00103015e-01 3.39923650e-01 -3.73859257e-01 -1.47741824e-01
-1.54405296e+00 1.22971356e+00 4.70790535e-01 -2.42751068e-03
-6.51054144e-01 -8.76366258e-01 -9.98233080e-01 4.22365963e-01
4.58738208e-01 -1.25028640e-01 1.16505969e+00 -4.93474990e-01
-1.34454906e+00 5.83607078e-01 4.08493401e-03 -5.97447038e-01
4.98111129e-01 -2.18902886e-01 -1.10450290e-01 -1.63385898e-01
3.89716104e-02 5.26929379e-01 5.48376381e-01 -7.29486763e-01
-7.73209989e-01 -1.08719021e-01 3.41369987e-01 2.42860317e-01
-7.14428306e-01 6.98707178e-02 -2.82990217e-01 -5.85370123e-01
-3.89691472e-01 -8.07637095e-01 -1.14314973e-01 -5.35216391e-01
-1.36451319e-01 -6.57969236e-01 6.72641277e-01 -7.37617850e-01
1.39242649e+00 -1.79527593e+00 8.37673768e-02 1.99760310e-02
-6.98740557e-02 5.16528666e-01 -2.61595935e-01 9.29098248e-01
4.58768159e-02 5.57030924e-02 1.48704290e-01 -3.21949810e-01
1.81482956e-01 7.14470208e-01 -5.37274301e-01 5.74399769e-01
-1.84636325e-01 8.10627937e-01 -6.22313380e-01 -4.06943977e-01
2.88420588e-01 2.77010918e-01 -5.11665881e-01 -1.95675269e-01
-3.72422040e-01 -2.90047616e-01 -2.90074110e-01 4.40350771e-01
4.06381309e-01 7.36780390e-02 -8.38771388e-02 -1.12409494e-03
-4.23610300e-01 9.95006442e-01 -7.30834723e-01 1.51923418e+00
-9.57773149e-01 8.07124555e-01 3.96121331e-02 -1.30507290e+00
8.62009585e-01 3.75602901e-01 5.03884375e-01 -4.82330501e-01
2.13546515e-01 2.57312477e-01 1.62845608e-02 -5.18824518e-01
7.34641254e-01 -3.17422599e-02 -3.28264505e-01 3.29373240e-01
2.27670565e-01 -1.96844086e-01 7.29748785e-01 1.14109665e-01
1.09151828e+00 -1.64668281e-02 1.23492800e-01 -4.69958544e-01
4.17174190e-01 6.05893508e-03 2.70710438e-01 9.36433256e-01
2.60238588e-01 3.64669323e-01 6.77573085e-01 -7.69330144e-01
-1.18936646e+00 -6.25352681e-01 -4.30796921e-01 1.67195165e+00
-5.60242534e-01 -7.03212142e-01 -4.60927725e-01 -1.09966683e+00
3.12484145e-01 8.82909656e-01 -5.02432644e-01 2.47332603e-01
-6.28024518e-01 -8.24793816e-01 8.76042068e-01 5.20859778e-01
1.81423977e-01 -8.91462803e-01 -3.54099572e-01 3.12372804e-01
-3.25776786e-01 -1.00087523e+00 -2.76602536e-01 8.64659488e-01
-5.74229598e-01 -6.85409129e-01 -9.07038569e-01 -5.27439594e-01
1.50651768e-01 -1.88558400e-01 1.15647447e+00 1.01416476e-01
-2.01792851e-01 -3.40502858e-02 -7.63932467e-01 -3.87366265e-01
-2.29968771e-01 8.15268338e-01 -1.86003238e-01 -3.34953368e-01
3.68055731e-01 -3.21890980e-01 -6.73835725e-02 1.57436445e-01
-7.22880483e-01 -3.45619142e-01 7.88895607e-01 1.08943701e+00
2.90762895e-04 4.06777672e-02 4.82367009e-01 -9.98272955e-01
4.53868717e-01 -4.78351623e-01 -4.11149114e-01 2.46840388e-01
-5.32144427e-01 1.64128289e-01 1.20902908e+00 -5.21846950e-01
-7.65030146e-01 -2.21226469e-01 -7.70194650e-01 -3.06501061e-01
-1.07779779e-01 8.50046396e-01 1.25654027e-01 3.44889909e-02
8.31327736e-01 2.68049508e-01 -3.16232443e-01 -7.43939757e-01
3.62986177e-01 1.05096662e+00 1.62140504e-01 -5.92176259e-01
6.94640458e-01 4.57302555e-02 -2.58479923e-01 -1.08476162e+00
-1.00850558e+00 -5.83313942e-01 -5.70217431e-01 2.41884395e-01
4.20645684e-01 -7.54843891e-01 -5.80825925e-01 1.23340443e-01
-9.55139637e-01 -3.63984764e-01 -3.01433593e-01 4.13590610e-01
-1.97101325e-01 4.00508106e-01 -8.97962749e-01 -6.69407964e-01
-4.12178099e-01 -7.99703598e-01 1.12961209e+00 -3.97470534e-01
-1.48259610e-01 -8.83646727e-01 -2.11682394e-01 4.29851443e-01
6.44217134e-01 -2.66534001e-01 1.28315210e+00 -9.50571597e-01
2.27974579e-02 -3.63289982e-01 -2.19633430e-01 2.96891421e-01
-1.20468445e-01 -1.27898157e-01 -7.60598958e-01 -4.19346839e-01
-2.42444128e-01 -4.98503268e-01 1.12594795e+00 1.73260301e-01
1.08233106e+00 -4.75510240e-01 -3.21432918e-01 5.21448612e-01
1.14367986e+00 1.73107818e-01 6.59216940e-01 4.86506045e-01
5.42560816e-01 6.64574087e-01 5.45404375e-01 3.58528405e-01
3.97245228e-01 7.05588520e-01 3.22273821e-01 -1.61390021e-01
7.55983347e-04 -1.68263271e-01 4.65647936e-01 8.42296064e-01
-1.39028281e-01 -5.68052232e-01 -1.11050498e+00 7.24326134e-01
-1.55454028e+00 -8.83183360e-01 -4.05185044e-01 2.04813957e+00
7.19992459e-01 2.88329959e-01 2.00925231e-01 5.67645490e-01
1.48030147e-01 3.51985782e-01 -1.39647365e-01 -8.80820632e-01
-2.62731519e-02 4.10963148e-01 8.43170106e-01 4.82661694e-01
-1.16511333e+00 9.95932639e-01 6.36707640e+00 1.40629447e+00
-1.14756095e+00 2.58207291e-01 4.66076851e-01 -2.94432312e-01
-2.40957633e-01 -2.93862987e-02 -1.18558860e+00 5.32922804e-01
1.27839601e+00 -9.20948908e-02 4.79066551e-01 9.77869153e-01
-2.35728636e-01 -1.00914873e-01 -9.90997255e-01 6.13614976e-01
2.61912763e-01 -1.09378386e+00 2.68222652e-02 3.08753550e-01
3.48950744e-01 3.42908740e-01 -1.64398640e-01 9.54810917e-01
3.15620720e-01 -1.12254667e+00 7.76938260e-01 -1.12527266e-01
1.02333140e+00 -8.97166908e-01 9.93771970e-01 7.89818823e-01
-1.12819242e+00 -3.15487474e-01 -8.53827655e-01 -2.44575456e-01
-5.75119108e-02 7.64814734e-01 -6.22913837e-01 7.76204944e-01
8.85811150e-01 5.28884530e-01 -4.35347438e-01 7.28897214e-01
1.20137157e-02 8.15538466e-01 -7.86630869e-01 -5.59614003e-01
7.85399675e-01 8.12214687e-02 1.64376810e-01 1.41430426e+00
5.01360834e-01 -1.13884047e-01 1.99082300e-01 2.07886264e-01
8.58229473e-02 3.72409046e-01 -6.31471097e-01 6.10300973e-02
3.17243993e-01 1.05414367e+00 -5.15230775e-01 -4.84858721e-01
-6.85360372e-01 6.57621264e-01 6.75578475e-01 2.14958265e-01
-9.24698114e-01 -6.51865125e-01 3.35137874e-01 4.36021864e-01
5.73577344e-01 -1.57965034e-01 8.64857957e-02 -1.14807105e+00
-9.38116759e-02 -1.01082659e+00 6.09252512e-01 -3.57241601e-01
-1.26868343e+00 8.50642979e-01 2.37049982e-01 -9.10924435e-01
-2.40751609e-01 -8.86350989e-01 -2.09015623e-01 7.73463249e-01
-1.38496709e+00 -1.25377953e+00 8.54363069e-02 5.55820942e-01
8.33267510e-01 -3.98563929e-02 1.03909767e+00 7.04338968e-01
-2.73069412e-01 6.70552015e-01 5.21456063e-01 1.37538360e-02
5.19566000e-01 -1.37426150e+00 4.29512680e-01 3.04290026e-01
4.77597535e-01 4.94972378e-01 5.44759274e-01 -4.35960352e-01
-1.43204081e+00 -8.68367255e-01 1.54879570e+00 -3.22495192e-01
8.22145224e-01 -8.32181692e-01 -8.47505748e-01 9.27602172e-01
2.26064563e-01 -1.16509870e-01 5.78512669e-01 7.55738854e-01
-4.45920795e-01 -1.63102716e-01 -7.58661151e-01 3.62187475e-01
8.89907062e-01 -4.02703345e-01 -8.56124282e-01 6.59401417e-01
4.31589842e-01 -5.65270782e-01 -9.83464718e-01 1.93633363e-01
5.05532086e-01 -9.13413584e-01 9.18381929e-01 -4.41760570e-01
2.42483214e-01 2.33652532e-01 4.71279360e-02 -1.45159006e+00
-5.18025756e-01 -4.35361773e-01 2.11711273e-01 1.14489794e+00
6.87969863e-01 -8.81175280e-01 8.42867136e-01 1.75443083e-01
-3.11358154e-01 -1.07149613e+00 -1.29675734e+00 -1.09812236e+00
6.24978542e-01 -5.06194830e-01 6.82333231e-01 8.16546321e-01
2.54493564e-01 3.75134289e-01 -5.49992263e-01 -3.48507077e-01
2.63756374e-03 1.32026836e-01 4.73537087e-01 -1.18980026e+00
-2.41867885e-01 -4.43017811e-01 -1.87658444e-01 -1.40528417e+00
3.40852052e-01 -1.15487432e+00 -1.38724465e-02 -1.61463189e+00
-2.17616960e-01 -7.76372790e-01 1.38730258e-01 7.91885912e-01
5.33294529e-02 4.30693477e-02 4.79590334e-02 5.60910515e-02
-1.99380249e-01 5.28091550e-01 8.38475645e-01 -2.37133428e-01
1.05792508e-01 -6.25430271e-02 -4.35481995e-01 5.16502261e-01
5.81041753e-01 -5.55750012e-01 -2.84481883e-01 -6.17232323e-01
2.49354094e-01 -3.71267414e-03 -1.76791534e-01 -7.82268226e-01
1.47285581e-01 5.95614361e-03 4.58644450e-01 -7.23636150e-01
3.84843439e-01 -8.49482179e-01 -1.86909765e-01 4.47155595e-01
-2.98327088e-01 4.49186526e-02 2.32386962e-01 1.09190553e-01
-3.03972214e-01 -6.27681315e-01 4.20702219e-01 -2.21389621e-01
-1.28819004e-01 1.79023385e-01 -1.00671411e+00 -3.97722162e-02
8.57703447e-01 -1.47526890e-01 -5.79973280e-01 -3.48623663e-01
-4.57146138e-01 2.44828150e-01 -3.85759696e-02 3.44031125e-01
1.00278541e-01 -9.07654464e-01 -9.27836835e-01 2.33358026e-01
-5.24927899e-02 -9.52181220e-02 8.19967240e-02 8.49403024e-01
-6.64896905e-01 9.57029045e-01 2.69493639e-01 -2.80306756e-01
-1.25530791e+00 5.75071871e-01 1.08400099e-01 -8.81475806e-01
-7.21565127e-01 9.20374632e-01 -1.05185777e-01 -5.42824686e-01
3.86898935e-01 -5.67076802e-01 -3.41920227e-01 5.09467125e-01
2.53978431e-01 3.58657509e-01 6.56944454e-01 -6.23350441e-01
-2.84617335e-01 5.44183910e-01 -3.63364667e-01 9.91232023e-02
1.41541600e+00 2.13459074e-01 -2.93205529e-02 1.46554574e-01
1.33231544e+00 2.08756536e-01 -6.09149814e-01 -1.31765142e-01
9.26944613e-02 -4.45156604e-01 1.88475654e-01 -7.76991010e-01
-8.17977548e-01 9.13791537e-01 -9.04778466e-02 5.19340217e-01
8.81368577e-01 -2.61650775e-02 1.00741947e+00 7.35660791e-01
4.85257864e-01 -9.23663080e-01 -3.87451589e-01 9.98156667e-01
9.28333879e-01 -8.60604763e-01 1.91924259e-01 -2.96772361e-01
-4.88241225e-01 1.22698390e+00 2.88565546e-01 -5.52918762e-02
4.44760144e-01 6.68922842e-01 1.71927661e-02 -1.09199226e-01
-1.12012398e+00 -4.39725161e-01 3.02388698e-01 3.63196999e-01
5.59370518e-01 -2.57657707e-01 -5.56864738e-01 3.76870245e-01
-6.61812425e-01 -2.61517435e-01 2.83763170e-01 7.96730936e-01
-5.60332775e-01 -1.07404518e+00 -4.93539214e-01 1.05560994e+00
-8.63504052e-01 -3.84440184e-01 -2.40841180e-01 1.05534303e+00
2.11219892e-01 1.09167552e+00 1.01757281e-01 -7.88700134e-02
4.06973213e-01 2.37557247e-01 3.71167421e-01 -7.05381095e-01
-1.09650147e+00 9.29186717e-02 6.01420641e-01 -2.45264217e-01
1.83588132e-01 -6.98620141e-01 -7.62648165e-01 -6.35872304e-01
-6.16329610e-01 5.47963023e-01 5.49856007e-01 9.45245266e-01
-4.24932167e-02 3.20468009e-01 3.11100781e-01 -9.55494761e-01
-8.96829963e-01 -1.40995610e+00 -6.43958151e-01 1.51273727e-01
2.44055972e-01 -6.69151127e-01 -7.10817635e-01 -2.61405736e-01] | [10.56991195678711, 8.921576499938965] |
649ecc5d-1640-4afc-bb8e-738dcf3479a9 | application-of-symmetric-uncertainty-and | 1211.0613 | null | https://arxiv.org/abs/1211.0613v2 | https://arxiv.org/pdf/1211.0613v2.pdf | Application of Symmetric Uncertainty and Mutual Information to Dimensionality Reduction and Classification of Hyperspectral Images | Remote sensing is a technology to acquire data for disatant substances, necessary to construct a model knowledge for applications as classification. Recently Hyperspectral Images (HSI) becomes a high technical tool that the main goal is to classify the point of a region. The HIS is more than a hundred bidirectional measures, called bands (or simply images), of the same region called Ground Truth Map (GT). But some bands are not relevant because they are affected by different atmospheric effects; others contain redundant information; and high dimensionality of HSI features make the accuracy of classification lower. All these bands can be important for some applications; but for the classification a small subset of these is relevant. The problematic related to HSI is the dimensionality reduction. Many studies use mutual information (MI) to select the relevant bands. Others studies use the MI normalized forms, like Symmetric Uncertainty, in medical imagery applications. In this paper we introduce an algorithm based also on MI to select relevant bands and it apply the Symmetric Uncertainty coefficient to control redundancy and increase the accuracy of classification. This algorithm is feature selection tool and a Filter strategy. We establish this study on HSI AVIRIS 92AV3C. This is an effectiveness, and fast scheme to control redundancy. | ['Driss Aboutajdine', 'Ahmed Hammouch', 'Elkebir Sarhrouni'] | 2012-11-03 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 5.28454423e-01 -3.85424882e-01 4.12674667e-03 -4.20708388e-01
-3.99113357e-01 -3.85883152e-01 3.89782101e-01 6.37514815e-02
-2.81180233e-01 9.74991202e-01 1.52411625e-01 8.84289294e-02
-8.88972878e-01 -1.19961452e+00 -1.31397069e-01 -9.44603264e-01
-2.12760061e-01 3.33055556e-01 2.19061032e-01 -3.97174925e-01
4.39494550e-01 7.27917671e-01 -1.99948895e+00 4.38538492e-01
1.11843026e+00 1.02258933e+00 4.43794429e-01 3.06243896e-01
-1.54514730e-01 3.87639970e-01 -3.37369531e-01 6.00160539e-01
4.75169420e-01 -5.47644436e-01 -8.09867203e-01 4.59400900e-02
-2.49795303e-01 8.02458152e-02 3.60907823e-01 1.43492031e+00
4.13576156e-01 2.11348414e-01 1.23543096e+00 -1.03497815e+00
-3.01713228e-01 6.65125966e-01 -8.02489102e-01 3.53572667e-01
-6.10535331e-02 -2.63732255e-01 5.96577764e-01 -5.17854631e-01
3.37532312e-01 9.12876964e-01 4.78918284e-01 -9.17244330e-02
-9.02273476e-01 -5.49283803e-01 -3.48987728e-01 4.29670513e-01
-1.58224964e+00 -5.42371497e-02 6.01563871e-01 -6.95815682e-01
3.78219783e-01 9.00731385e-01 7.90818632e-01 7.38854036e-02
4.70842928e-01 2.19609197e-02 1.50757170e+00 -5.30295491e-01
1.62872061e-01 4.11460370e-01 6.43589139e-01 2.97406226e-01
5.81175447e-01 1.15535140e-01 -7.44481906e-02 -1.15620002e-01
2.21475586e-01 2.65260011e-01 -7.01168358e-01 1.53584808e-01
-9.14174378e-01 9.91204143e-01 5.70704103e-01 8.86275589e-01
-5.47514558e-01 -5.10280013e-01 8.29548314e-02 3.48325431e-01
3.19095910e-01 5.24167061e-01 -2.93136358e-01 3.81047308e-01
-8.28754246e-01 1.80838704e-02 5.53713799e-01 4.04786855e-01
1.09347355e+00 -1.80702463e-01 -1.68465003e-01 6.27572060e-01
2.94335514e-01 7.03163981e-01 6.68821871e-01 -3.28396469e-01
9.19276029e-02 9.73116577e-01 -2.60683388e-01 -1.32359612e+00
-6.65417552e-01 -6.10021353e-01 -9.84753609e-01 4.58322972e-01
-2.58630425e-01 -1.02227978e-01 -1.11899245e+00 1.10981679e+00
1.78709120e-01 -2.83918798e-01 5.52830882e-02 1.03195846e+00
9.15711582e-01 1.18990695e+00 -1.63032681e-01 -5.02800882e-01
1.32869816e+00 -2.15923429e-01 -8.07243049e-01 3.80971953e-02
2.38266483e-01 -1.00430727e+00 6.73603714e-01 6.76569402e-01
-3.60770762e-01 -3.02180350e-01 -1.14342046e+00 4.89736974e-01
-8.34455490e-01 3.00929248e-01 5.93414485e-01 7.49859631e-01
-6.47887170e-01 7.71519184e-01 -3.43291432e-01 -4.73150551e-01
3.17663066e-02 3.16966325e-01 -4.70463485e-01 1.41591467e-02
-1.24565065e+00 1.11140501e+00 8.63257706e-01 1.70054018e-01
-2.06285194e-01 -4.03234899e-01 -5.76012492e-01 -1.10697257e-03
1.90778837e-01 -1.20334193e-01 3.15107375e-01 -9.96958077e-01
-1.01889372e+00 6.05737329e-01 2.27194563e-01 -1.44084066e-01
2.18804106e-01 1.86929196e-01 -5.80171049e-01 2.06900731e-01
1.06127016e-01 -1.05393017e-02 5.45800090e-01 -1.09386563e+00
-7.86970794e-01 -7.20161378e-01 -4.43275988e-01 3.06047082e-01
-1.79165944e-01 -2.36744322e-02 2.11625919e-01 -3.66131991e-01
9.48701978e-01 -8.98295939e-01 -2.23491237e-01 -4.34554338e-01
-4.39408541e-01 2.21001193e-01 9.66745853e-01 -8.36027145e-01
1.04137444e+00 -2.04768610e+00 -2.25689895e-02 9.90863264e-01
-3.43522280e-01 1.76692903e-01 1.93126857e-01 3.26182306e-01
-2.79618859e-01 4.78563339e-01 -7.63067901e-01 6.84899688e-01
-5.89986920e-01 -3.17328870e-02 -3.79446079e-04 6.83409035e-01
2.89329467e-03 -5.94504587e-02 -3.41597289e-01 -5.98170340e-01
3.94914001e-01 3.83545667e-01 -9.38738436e-02 -1.06757700e-01
1.12274319e-01 3.61123681e-01 -6.79634809e-01 7.02763498e-01
1.11124003e+00 1.65717795e-01 -2.08446622e-01 -8.14378798e-01
-5.05841374e-01 -4.27124083e-01 -1.40919590e+00 1.08486867e+00
-2.28520040e-03 4.53763217e-01 -1.92945063e-01 -1.30463564e+00
1.14328325e+00 2.37880468e-01 7.80003846e-01 -1.99026853e-01
2.99382001e-01 2.26670429e-01 1.48117617e-01 -7.98276067e-01
2.46289685e-01 -1.93830520e-01 5.28012037e-01 7.12706894e-02
-2.93737829e-01 -6.04689300e-01 1.31455407e-01 -2.18130440e-01
4.25934434e-01 -2.12336019e-01 6.66134477e-01 -7.17094898e-01
7.25663424e-01 3.36594462e-01 5.92742205e-01 3.68135273e-01
6.82937056e-02 5.97086847e-01 2.04138473e-01 -2.37781674e-01
-8.50365996e-01 -5.56845963e-01 -8.81924033e-01 2.40290955e-01
1.15165673e-01 1.42555460e-01 -3.22938263e-01 -1.92331195e-01
2.21410375e-02 5.94776928e-01 -4.40678895e-01 -1.78638354e-01
1.03926897e-01 -1.48199141e+00 5.89195415e-02 -3.44718695e-01
1.06355619e+00 -6.99715257e-01 -5.06303370e-01 1.59174725e-01
-2.23732919e-01 -3.46594572e-01 3.08714300e-01 1.98052332e-01
-1.01357257e+00 -1.15011156e+00 -3.66976231e-01 -2.10093752e-01
4.47818846e-01 4.16803956e-01 7.44899511e-01 -3.61924153e-03
-6.78044736e-01 -1.30173415e-01 -7.90963292e-01 -7.23732173e-01
-3.02641988e-01 -1.75998032e-01 -2.35595480e-01 1.77723542e-01
2.56975532e-01 -5.63536346e-01 -3.60033870e-01 2.27665007e-01
-1.12831771e+00 -1.50557682e-01 7.25084484e-01 6.51094437e-01
8.01705241e-01 8.15076113e-01 2.88607359e-01 -9.22203600e-01
3.43353868e-01 -6.06923938e-01 -7.28066683e-01 3.62326384e-01
-6.84376895e-01 2.98114792e-02 1.90865621e-01 2.68668443e-01
-1.07811189e+00 9.11862329e-02 2.02293620e-01 1.16718262e-01
-9.51279774e-02 1.06378794e+00 -1.44509852e-01 -3.79847229e-01
1.12439489e+00 1.58259958e-01 1.28426746e-01 -4.89997178e-01
-2.21495718e-01 1.05981159e+00 -5.95063120e-02 -1.97759673e-01
7.44426072e-01 4.34702218e-01 6.22255027e-01 -1.28956091e+00
-5.70098519e-01 -6.98054135e-01 -5.09145856e-01 -2.31914937e-01
1.01852143e+00 -5.98301470e-01 -3.25061768e-01 -8.39659292e-03
-8.19458604e-01 4.55699444e-01 5.58748655e-02 9.48210239e-01
-2.49288782e-01 3.15181196e-01 -1.58085242e-01 -1.12026477e+00
-4.76396203e-01 -1.08660090e+00 5.99230349e-01 4.58144337e-01
4.03222859e-01 -4.26078111e-01 -1.43905148e-01 1.69460237e-01
3.72551471e-01 5.97619593e-01 8.40011060e-01 -4.63639349e-01
-4.46126997e-01 -1.27548710e-01 -3.43520373e-01 5.99246383e-01
5.70889652e-01 3.42935830e-01 -8.52026403e-01 -1.86227243e-02
3.78666461e-01 2.45365098e-01 1.10621607e+00 5.72051585e-01
1.22962439e+00 -9.95322466e-02 -4.68254715e-01 7.23963082e-01
1.96244812e+00 7.21037865e-01 9.32956159e-01 3.80089670e-01
3.57239753e-01 8.17403138e-01 1.19018781e+00 4.20851052e-01
-5.25274038e-01 2.99397767e-01 4.88096058e-01 -1.48567915e-01
2.27925137e-01 5.41806996e-01 -1.44981012e-01 5.62894106e-01
-6.47258341e-01 -4.08616886e-02 -1.02335823e+00 1.43330619e-01
-1.75159502e+00 -1.29025829e+00 -6.08483672e-01 2.37253189e+00
5.74457109e-01 -2.87059188e-01 -2.64841467e-01 6.94718122e-01
1.00732934e+00 -8.38157162e-02 -1.15705401e-01 5.78366257e-02
-3.74187201e-01 1.82283267e-01 8.20048869e-01 5.46145082e-01
-1.30285573e+00 3.76696467e-01 5.82024384e+00 7.85633802e-01
-1.29281437e+00 -1.23191342e-01 7.01871574e-01 3.57496142e-01
-3.01879376e-01 6.22279663e-03 -5.10628581e-01 3.40139091e-01
5.25277317e-01 -1.82847366e-01 1.83017209e-01 6.36995494e-01
4.40610796e-01 -6.50717378e-01 -4.00326788e-01 1.10845280e+00
1.75257802e-01 -9.60210145e-01 2.92651713e-01 1.96787730e-01
8.21123183e-01 -7.17959106e-02 -1.16542630e-01 -3.32483530e-01
-2.18944281e-01 -9.80920553e-01 2.14812487e-01 9.80794489e-01
5.12122154e-01 -1.00435603e+00 1.18558478e+00 2.87330836e-01
-1.07872474e+00 -1.49403706e-01 -9.60723639e-01 2.46144459e-02
-3.05584252e-01 1.16686666e+00 -7.36228347e-01 1.07193172e+00
7.19631553e-01 6.36780441e-01 -6.95677817e-01 1.43870032e+00
1.73594639e-01 2.95195341e-01 -4.25406039e-01 -1.87648550e-01
1.95597127e-01 -8.73091936e-01 7.32032597e-01 8.36472452e-01
8.96112978e-01 5.87326646e-01 1.43564373e-01 8.80392671e-01
7.36893415e-01 7.08203137e-01 -9.65473473e-01 -8.09153095e-02
4.49517012e-01 1.15316629e+00 -7.80663729e-01 -2.34983340e-01
6.88256770e-02 6.42588496e-01 -5.89662015e-01 1.80706233e-01
-6.08976364e-01 -6.47096217e-01 2.75967658e-01 -5.00164330e-02
-3.22835416e-01 5.98249584e-02 -2.29832113e-01 -9.07010317e-01
-1.71082318e-01 -7.67227530e-01 6.65882707e-01 -8.30468118e-01
-1.19909596e+00 6.71750546e-01 2.78242439e-01 -1.69801581e+00
2.57953703e-02 -7.05772519e-01 -2.57522643e-01 1.30008805e+00
-1.47514904e+00 -9.21136022e-01 -6.55540347e-01 7.04032958e-01
1.69014812e-01 -3.86259317e-01 8.24289441e-01 1.85598791e-01
-3.64092588e-01 -3.87732744e-01 5.65291524e-01 -2.84949452e-01
5.34853041e-01 -1.08717823e+00 -1.02930212e+00 9.04377401e-01
-3.22035551e-01 2.80513674e-01 8.94774139e-01 -8.89858723e-01
-1.09609270e+00 -8.62830460e-01 4.63561356e-01 2.74213850e-01
-6.21460006e-03 5.21529198e-01 -6.97358429e-01 3.40062827e-01
-1.22321770e-01 -2.68607557e-01 8.22778940e-01 -1.15361819e-02
2.54267097e-01 -6.00668669e-01 -1.55065989e+00 2.72237062e-01
3.74539882e-01 -8.18363875e-02 -6.51434779e-01 7.02303231e-01
3.82379949e-01 1.18895575e-01 -8.42382669e-01 5.98941863e-01
3.46865088e-01 -1.38679993e+00 7.87867427e-01 -1.34282917e-01
1.16479114e-01 -7.73307502e-01 -3.36996198e-01 -1.35427213e+00
-5.52613497e-01 1.61391690e-01 1.07903767e+00 9.85175014e-01
4.40160215e-01 -7.58821964e-01 5.38751364e-01 3.04860115e-01
-9.30571780e-02 -9.92514193e-02 -7.16999114e-01 -7.43554235e-01
-3.47139239e-01 6.44120947e-02 8.06305051e-01 1.36505973e+00
4.77182586e-03 3.92625555e-02 -1.58892155e-01 4.29538399e-01
8.10090601e-01 2.25141257e-01 1.87822029e-01 -1.96873093e+00
-6.03698976e-02 -3.88221383e-01 -7.98838377e-01 2.52472907e-01
-3.27137023e-01 -7.20201731e-01 -7.23919123e-02 -1.72777152e+00
1.92588717e-01 -5.96847296e-01 -3.12858462e-01 3.98977578e-01
4.78248745e-02 2.09210724e-01 4.69480157e-02 4.51964825e-01
5.80662012e-01 1.99410662e-01 1.07661188e+00 -3.79255801e-01
-4.56784695e-01 2.09346876e-01 -3.54308635e-01 6.30208135e-01
9.99071419e-01 -3.61953557e-01 -5.76384902e-01 1.28709063e-01
2.11113453e-01 3.97223793e-02 -6.73198700e-02 -1.48291719e+00
1.67493131e-02 -5.45216680e-01 5.36474645e-01 -8.02394807e-01
1.94068685e-01 -1.41680253e+00 8.73187065e-01 8.06810975e-01
9.94323567e-02 -4.10442084e-01 4.97632893e-03 1.30067021e-01
-4.70043838e-01 -8.30078065e-01 1.08455157e+00 -3.81816328e-01
-8.27993095e-01 3.70949298e-01 -2.11829334e-01 -8.07550967e-01
1.15804565e+00 -2.87290424e-01 -2.31267393e-01 -3.70061696e-01
-7.16554165e-01 -1.94865584e-01 1.23018257e-01 -1.91027328e-01
4.85590070e-01 -1.15118778e+00 -8.53046954e-01 6.90438226e-02
1.96281940e-01 -2.03090355e-01 3.54647934e-01 6.99029744e-01
-1.06617224e+00 2.80038506e-01 -7.23986983e-01 -5.68210423e-01
-1.46624315e+00 2.79186368e-01 5.99054456e-01 1.46039695e-01
-2.97053784e-01 4.90340680e-01 -3.36468741e-02 -1.19484767e-01
-4.19987321e-01 -3.14302981e-01 -1.10240960e+00 4.01738435e-01
6.90009594e-01 5.14391065e-01 2.38422111e-01 -7.05052197e-01
-3.02182883e-01 9.42389488e-01 5.64806581e-01 -2.92721957e-01
1.35298097e+00 -2.13653464e-02 -8.14083397e-01 5.92471123e-01
1.01255226e+00 5.00516333e-02 -3.12456757e-01 5.18152229e-02
-1.62643433e-01 -7.36638367e-01 4.71842855e-01 -9.07352746e-01
-1.03433657e+00 6.71151876e-01 1.15024686e+00 5.65290213e-01
1.50085247e+00 -4.88903522e-01 -1.81114137e-01 6.13060713e-01
4.52959508e-01 -1.10991299e+00 -7.08986938e-01 3.03848535e-01
1.17730165e+00 -1.35620809e+00 3.68150800e-01 -9.86223102e-01
-5.07249057e-01 1.46062267e+00 1.22609489e-01 1.19508445e-01
1.15474081e+00 -1.13260344e-01 -1.88724786e-01 -3.88227880e-01
9.14727673e-02 -5.89557052e-01 3.95684272e-01 5.60052097e-01
6.45569086e-01 3.23017269e-01 -1.06602263e+00 3.54685515e-01
-2.82975912e-01 4.74168248e-02 5.04420459e-01 8.80095005e-01
-1.16103876e+00 -7.93715417e-01 -1.06477439e+00 9.70348179e-01
-3.22639912e-01 -9.77576822e-02 -3.38201225e-02 8.15450430e-01
5.06356180e-01 1.17460895e+00 -1.03288420e-01 -4.62250859e-01
1.98841646e-01 1.05080921e-02 1.74120352e-01 -5.17683804e-01
-2.35887170e-01 -4.39878479e-02 4.68423665e-02 -2.45459348e-01
-7.30532467e-01 -5.33310771e-01 -1.12607253e+00 -2.48454154e-01
-6.68517709e-01 6.12645805e-01 1.06822014e+00 8.73779535e-01
-9.79798511e-02 3.78742248e-01 9.38175023e-01 -5.40708840e-01
-3.86877924e-01 -1.19761276e+00 -1.48444748e+00 3.27815235e-01
1.25710472e-01 -6.68977261e-01 -6.29288554e-01 -4.18619663e-02] | [9.725162506103516, -1.8372302055358887] |
23d072e2-f5af-4579-8842-d81e291e4809 | 190600424 | 1906.00424 | null | https://arxiv.org/abs/1906.00424v1 | https://arxiv.org/pdf/1906.00424v1.pdf | Plain English Summarization of Contracts | Unilateral contracts, such as terms of service, play a substantial role in modern digital life. However, few users read these documents before accepting the terms within, as they are too long and the language too complicated. We propose the task of summarizing such legal documents in plain English, which would enable users to have a better understanding of the terms they are accepting. We propose an initial dataset of legal text snippets paired with summaries written in plain English. We verify the quality of these summaries manually and show that they involve heavy abstraction, compression, and simplification. Initial experiments show that unsupervised extractive summarization methods do not perform well on this task due to the level of abstraction and style differences. We conclude with a call for resource and technique development for simplification and style transfer for legal language. | ['Laura Manor', 'Junyi Jessy Li'] | 2019-06-02 | plain-english-summarization-of-contracts | https://aclanthology.org/W19-2201 | https://aclanthology.org/W19-2201.pdf | ws-2019-6 | ['unsupervised-extractive-summarization'] | ['natural-language-processing'] | [ 3.60039115e-01 4.20092225e-01 -4.60099697e-01 -5.57981491e-01
-1.31442583e+00 -9.49502170e-01 6.92544639e-01 3.57488602e-01
-2.20643476e-01 8.03531110e-01 9.33706820e-01 -6.66275561e-01
1.92813153e-04 -4.02152836e-01 -5.19909143e-01 1.24982305e-01
3.54556441e-01 5.65501511e-01 -4.31346707e-02 -5.09239614e-01
5.56353569e-01 2.00241178e-01 -1.10087562e+00 5.46717584e-01
1.63883424e+00 3.13757747e-01 -1.37122124e-01 5.41076243e-01
-5.63929737e-01 6.93415463e-01 -1.12031603e+00 -1.21835160e+00
2.76341051e-01 -3.88869941e-01 -9.34717417e-01 -8.08054581e-02
6.85358047e-01 -7.66950011e-01 -3.58008325e-01 9.95598197e-01
1.30213991e-01 -2.76398867e-01 9.03125525e-01 -8.59877110e-01
-6.10013664e-01 1.19993913e+00 -3.81131470e-01 -1.13272950e-01
6.22160554e-01 -7.36155659e-02 1.37744105e+00 -2.94845521e-01
1.15146363e+00 9.94909108e-01 5.80666900e-01 5.73437214e-01
-1.24464738e+00 -3.90943885e-01 1.29212543e-01 -1.75279692e-01
-9.52814043e-01 -7.72188485e-01 6.59530997e-01 -4.08038527e-01
1.15314507e+00 7.34247923e-01 4.89431649e-01 7.75529206e-01
3.53840888e-01 9.00037467e-01 5.38981557e-01 -4.33651716e-01
1.73718706e-01 1.81743219e-01 3.08963388e-01 2.82696694e-01
7.38517165e-01 -5.88553369e-01 -3.43580842e-01 -4.07652467e-01
2.60533635e-02 -2.20109612e-01 -1.43215150e-01 9.68595371e-02
-5.93952596e-01 8.22620690e-01 -2.75500804e-01 3.84021074e-01
-2.00105339e-01 1.56867281e-01 7.90842533e-01 4.06034410e-01
3.14584553e-01 5.73069870e-01 -3.47456217e-01 -5.51328540e-01
-1.43015492e+00 8.09645236e-01 1.22749424e+00 1.53498948e+00
2.78013706e-01 -3.61709297e-01 -1.93715438e-01 8.31480682e-01
-1.41249835e-01 5.14545739e-01 7.88338408e-02 -1.46409714e+00
7.42930293e-01 5.73239803e-01 1.80447146e-01 -6.60460234e-01
1.61777213e-01 -1.63572714e-01 -4.91723031e-01 1.14874378e-01
2.22610906e-01 -5.65168038e-02 -7.79190660e-01 1.19810975e+00
-4.30644780e-01 -9.86073613e-01 5.88685237e-02 1.78009987e-01
6.21762037e-01 5.92245758e-01 -1.41410321e-01 -4.40746576e-01
1.32244480e+00 -6.80988789e-01 -1.03746688e+00 -1.83819056e-01
9.56211567e-01 -1.00768137e+00 1.12751794e+00 5.49397588e-01
-1.48973036e+00 1.22593949e-02 -1.00292408e+00 -6.81824565e-01
-2.45098814e-01 6.61017820e-02 7.16609597e-01 8.83876204e-01
-6.15301847e-01 8.26050103e-01 -4.93540019e-01 -2.17179373e-01
6.18318081e-01 -1.54266478e-02 -1.58346027e-01 -9.25949365e-02
-8.96367311e-01 7.82445312e-01 2.63900962e-02 -2.45558918e-01
1.20289885e-02 -6.86126471e-01 -9.51548040e-01 3.96928281e-01
5.70796967e-01 -5.58412373e-01 1.85471308e+00 -5.72708488e-01
-1.12161887e+00 7.00304329e-01 -4.80896473e-01 -5.15061677e-01
8.40580225e-01 -4.92617637e-01 -2.33327851e-01 1.10322386e-01
4.92188841e-01 2.49076843e-01 5.16109586e-01 -1.19912446e+00
-8.73153031e-01 -1.58778161e-01 3.64422947e-01 3.58537398e-02
-2.55057871e-01 4.68814105e-01 -6.10533834e-01 -9.53322530e-01
-9.38052908e-02 -7.61908174e-01 -1.95617035e-01 -2.77015805e-01
-4.80079949e-01 -1.37078956e-01 4.96119648e-01 -1.42722785e+00
1.83982897e+00 -1.95714140e+00 -1.56166060e-02 4.18629616e-01
1.93019345e-01 1.52740210e-01 1.14530839e-01 9.61893260e-01
4.40497875e-01 9.67996061e-01 -4.83412832e-01 -4.35722411e-01
4.03583258e-01 3.72510850e-01 -7.54040003e-01 1.52392959e-05
-2.22679183e-01 8.12727809e-01 -6.77939355e-01 -7.28428185e-01
-5.18556237e-01 -1.94738563e-02 -9.78425205e-01 -2.17093319e-01
-5.05531073e-01 6.61453754e-02 -3.40108573e-01 7.46295750e-01
4.93610829e-01 2.60103852e-01 3.23511362e-01 5.45224249e-02
-1.64884776e-01 7.85017610e-01 -8.91118944e-01 1.77018321e+00
-4.91619408e-01 7.44957566e-01 1.15976922e-01 -3.34566563e-01
3.76737684e-01 2.22174227e-01 2.58596420e-01 -6.68135941e-01
-8.95737559e-02 5.12454033e-01 5.71716670e-03 -5.82405925e-01
1.16281092e+00 -1.08622506e-01 -5.07208467e-01 7.18158662e-01
-5.36998749e-01 -7.59357274e-01 1.01707816e+00 9.54532146e-01
1.03352654e+00 -1.34901196e-01 1.50085852e-01 -5.51148923e-03
1.13731347e-01 8.81742090e-02 8.09085310e-01 7.28211045e-01
2.58980095e-01 5.63486099e-01 1.04925811e+00 1.44565897e-02
-1.37018645e+00 -7.63027966e-01 -5.73029444e-02 6.21649265e-01
-2.50735164e-01 -1.28283584e+00 -9.00314450e-01 -8.10710788e-01
1.72385737e-01 1.31649661e+00 -9.88467485e-02 2.69757491e-02
-7.53347635e-01 5.85991852e-02 7.39538014e-01 4.51843351e-01
1.90618351e-01 -6.53311789e-01 -4.89367396e-01 2.00326830e-01
-4.16251749e-01 -1.14677989e+00 -9.59828496e-01 -3.35347623e-01
-9.13227856e-01 -6.63002372e-01 -6.08971775e-01 -4.07391191e-01
6.93812072e-01 -1.37948200e-01 9.08429205e-01 3.73563290e-01
2.37585022e-03 1.75147370e-01 -3.95288765e-01 -6.37681127e-01
-8.82320940e-01 2.79619217e-01 -2.84586787e-01 -6.68502688e-01
1.67444155e-01 -5.32254755e-01 -1.61762387e-01 -2.81102061e-01
-1.28896749e+00 -3.79655659e-02 6.72337592e-01 5.00792325e-01
1.02116931e-02 9.48974714e-02 2.04379171e-01 -1.59689200e+00
1.38498437e+00 6.65547177e-02 -4.20830190e-01 4.32905555e-01
-6.96377993e-01 2.57271379e-01 6.11449778e-01 7.91956857e-02
-1.28220177e+00 -4.73940521e-01 -2.96503574e-01 4.87798065e-01
1.37206838e-01 9.15536344e-01 -1.02173403e-01 4.26018178e-01
6.46234274e-01 -1.04110613e-01 -6.43668175e-02 -5.98836660e-01
6.34772837e-01 1.08468652e+00 5.95877767e-01 -7.65164316e-01
7.34480739e-01 5.48951149e-01 -5.41088343e-01 -8.96720052e-01
-4.46769863e-01 -3.91168833e-01 -4.64081973e-01 1.98537365e-01
4.77350444e-01 -5.73009253e-01 -1.95577323e-01 8.86730403e-02
-1.47185183e+00 -1.17567435e-01 -5.67934871e-01 1.58689439e-01
-5.08250594e-01 1.04020727e+00 -9.86367166e-01 -6.69322252e-01
-5.11706114e-01 -9.72256064e-01 1.01513851e+00 1.21055149e-01
-1.03209066e+00 -5.71771145e-01 -1.48613483e-01 8.18529606e-01
3.41358602e-01 -3.96917872e-02 1.49859393e+00 -5.25014997e-01
-4.16642159e-01 -5.05832136e-01 -8.38336200e-02 5.00117242e-01
5.53318821e-02 5.86375713e-01 -4.07826155e-01 -1.29394695e-01
-1.83096975e-01 1.26798721e-02 7.94419348e-01 3.34742695e-01
8.57405841e-01 -7.63141751e-01 -2.80003726e-01 3.96618694e-01
1.02327859e+00 4.08913940e-01 9.21630502e-01 2.13128030e-01
4.43578839e-01 7.45376110e-01 4.24970031e-01 3.85650814e-01
3.48168552e-01 5.61971605e-01 -3.70129108e-01 1.44053146e-01
5.29620089e-02 -3.71336967e-01 3.94690067e-01 1.07799625e+00
-1.49859011e-01 -9.20880511e-02 -7.55088508e-01 5.75215936e-01
-1.77153563e+00 -1.18736494e+00 -2.70489808e-02 1.97932708e+00
1.00267303e+00 7.31033027e-01 1.42719179e-01 1.29133254e-01
2.60294467e-01 -3.20876986e-02 -4.18557882e-01 -1.14618373e+00
-2.06133962e-01 7.57465214e-02 7.74930298e-01 6.52873516e-01
-5.51332116e-01 8.06231797e-01 6.66887474e+00 7.90134847e-01
-7.63398170e-01 -2.71065444e-01 5.68539381e-01 4.83044889e-03
-1.12271690e+00 5.23619771e-01 -8.22893918e-01 5.54758012e-01
8.81912708e-01 -8.03533137e-01 3.36439610e-01 8.17934215e-01
4.66185361e-01 -2.71753609e-01 -1.19750285e+00 7.82178700e-01
8.75543579e-02 -1.60336149e+00 4.33345288e-01 3.22394073e-01
8.32226038e-01 -4.12841499e-01 -2.63353378e-01 3.31649631e-01
3.43578756e-01 -1.01158333e+00 1.14848697e+00 4.58824217e-01
7.84227014e-01 -6.03677690e-01 6.34702265e-01 3.77352327e-01
-5.94096541e-01 -1.80404425e-01 -2.65015841e-01 -7.71569461e-02
6.18340075e-01 5.48452199e-01 -3.74068499e-01 7.36581922e-01
3.14442277e-01 6.05293155e-01 -4.73573804e-01 9.99447048e-01
-4.11039978e-01 4.68820810e-01 -2.40882397e-01 -1.22737788e-01
2.59249974e-02 -4.76446271e-01 6.56254709e-01 1.34519374e+00
3.96997303e-01 2.26394311e-01 -7.48493373e-02 6.77715600e-01
-3.82354677e-01 2.41024479e-01 -6.39262319e-01 -4.46881562e-01
4.87008065e-01 6.97999775e-01 -5.88947058e-01 -8.10169697e-01
-6.01717234e-01 9.66423392e-01 -1.42037347e-01 5.21017909e-01
-5.81873357e-01 -8.56083333e-01 5.97242534e-01 3.39342564e-01
3.15788984e-01 -3.26056153e-01 -7.03109145e-01 -1.45310712e+00
7.02730477e-01 -1.42990613e+00 3.59859705e-01 -5.79628766e-01
-8.72005820e-01 3.05214465e-01 2.16312021e-01 -1.02389884e+00
-4.47888464e-01 -1.35053724e-01 -6.70304596e-01 5.94819009e-01
-1.18167233e+00 -9.35191333e-01 3.39970291e-01 -1.75288439e-01
7.93836236e-01 2.97181644e-02 4.44149077e-01 4.42381233e-01
-1.57697976e-01 6.89358234e-01 3.66368711e-01 7.48598799e-02
7.34151423e-01 -1.17238319e+00 6.90839827e-01 9.34642196e-01
9.36333388e-02 1.34905839e+00 1.09445453e+00 -1.07946420e+00
-1.18134141e+00 -5.12078822e-01 1.56620061e+00 -3.90133798e-01
5.34902394e-01 -4.67006862e-01 -7.62197793e-01 7.62817025e-01
5.64448476e-01 -1.19392240e+00 6.75279081e-01 2.07289696e-01
-3.40840220e-01 -1.85105994e-01 -1.04768717e+00 9.93392825e-01
1.12391353e+00 -8.38075280e-01 -1.06856394e+00 4.61238593e-01
7.02698052e-01 -2.39642933e-01 -7.70405591e-01 3.88583504e-02
6.98959470e-01 -7.98670709e-01 3.84901702e-01 -5.75196922e-01
7.30989158e-01 -5.23259342e-02 1.92075580e-01 -1.09011579e+00
1.42068893e-01 -1.30964291e+00 2.13925347e-01 1.58623981e+00
8.86669874e-01 -4.73715723e-01 7.10260272e-01 1.44168890e+00
-3.65539640e-01 -4.88978535e-01 -7.29621172e-01 -9.02580023e-01
3.72389913e-01 -5.12311757e-01 6.19884014e-01 6.34699106e-01
7.25869834e-01 3.88121396e-01 -1.29612327e-01 -4.76202130e-01
3.37813139e-01 2.72982866e-01 9.45870817e-01 -1.10315955e+00
-1.64165005e-01 -6.89011276e-01 -3.96573283e-02 -1.26915145e+00
-4.11621819e-04 -7.25333929e-01 -3.35193485e-01 -2.06207013e+00
2.78477311e-01 -1.93548635e-01 6.40180051e-01 2.24596560e-01
1.15707651e-01 -3.96436960e-01 2.43415594e-01 3.93029392e-01
-4.50343043e-01 2.85212159e-01 8.42829406e-01 -4.28362489e-01
-3.39664757e-01 1.16029441e-01 -1.25726759e+00 7.40940332e-01
6.04318321e-01 -3.69401425e-01 -4.36469972e-01 -5.88945210e-01
5.65430820e-01 1.14059664e-01 -3.43163848e-01 -6.04479432e-01
3.20136487e-01 -2.12194815e-01 -3.15381855e-01 -7.07542300e-01
1.02807693e-02 -6.43014133e-01 4.36418578e-02 3.99951339e-01
-6.00154221e-01 1.01148978e-01 4.27659191e-02 1.79701626e-01
-2.89977342e-01 -6.94039285e-01 1.96696967e-01 -2.72407860e-01
-1.51799276e-01 -9.37421396e-02 -5.97716093e-01 3.07825595e-01
6.30980492e-01 -4.77167338e-01 -3.96170706e-01 -8.62859786e-01
-4.83134419e-01 3.51721555e-01 9.27477539e-01 2.11471751e-01
4.16098773e-01 -8.40886354e-01 -8.77294302e-01 1.17355958e-02
-3.98162939e-02 -1.12058505e-01 3.74918766e-02 5.01296103e-01
-1.01345301e+00 6.04553640e-01 4.47239960e-03 7.56087005e-02
-1.56499434e+00 1.11936629e-01 -2.51845747e-01 -4.63208288e-01
-8.21493387e-01 2.20678940e-01 -2.45769188e-01 -9.98523459e-02
3.99921983e-01 -8.90791595e-01 -1.19082637e-01 3.21946919e-01
5.40094554e-01 2.92212009e-01 1.41279191e-01 -4.95299488e-01
4.23780195e-02 4.79361475e-01 -5.92453539e-01 -5.85202515e-01
1.49598420e+00 -1.92579180e-01 -4.22470272e-01 6.44304976e-02
1.08828568e+00 8.67914557e-01 -6.49614811e-01 8.57396945e-02
4.50765461e-01 -5.95344126e-01 -4.38105166e-01 -5.29896677e-01
-5.52589655e-01 5.52261949e-01 -3.88106287e-01 4.29278195e-01
8.58463407e-01 -9.61522907e-02 1.35310292e+00 6.78647518e-01
1.79301038e-01 -1.74773097e+00 -6.55064821e-01 4.64507014e-01
9.57404077e-01 -7.06827819e-01 5.89610338e-01 -6.52567744e-01
-7.09506989e-01 1.10756671e+00 -4.40831929e-02 2.04074666e-01
2.51388937e-01 3.26540798e-01 1.53012663e-01 -2.02928841e-01
-6.91247761e-01 8.72038528e-02 7.82506168e-02 3.58178198e-01
5.08686662e-01 -5.69639318e-02 -1.16138017e+00 5.58035135e-01
-5.83536327e-01 1.02507815e-01 1.14395869e+00 1.25032508e+00
-1.68237165e-01 -1.48703015e+00 -7.74825439e-02 7.95335948e-01
-9.09950137e-01 -3.03387731e-01 -7.46468961e-01 5.59354305e-01
-3.49253684e-01 1.04465687e+00 -2.41734445e-01 -6.36656880e-02
6.34173036e-01 1.02798104e-01 2.66676158e-01 -8.04217935e-01
-6.62132084e-01 2.05400616e-01 8.94941628e-01 -2.65322089e-01
3.46956030e-02 -1.03322303e+00 -1.26506019e+00 -6.08987749e-01
-8.61797780e-02 5.44795334e-01 6.33459687e-01 9.32652771e-01
2.22798929e-01 2.16087386e-01 1.59720346e-01 -2.17727140e-01
-8.00132632e-01 -7.73473918e-01 -5.09560823e-01 7.59312689e-01
1.16895504e-01 1.41437665e-01 -1.83838591e-01 5.85336745e-01] | [12.287991523742676, 9.539031982421875] |
ccd218d1-8e62-4ed9-8c3e-82b57c6288ee | multistage-stochastic-optimization-via | 2303.06515 | null | https://arxiv.org/abs/2303.06515v1 | https://arxiv.org/pdf/2303.06515v1.pdf | Multistage Stochastic Optimization via Kernels | We develop a non-parametric, data-driven, tractable approach for solving multistage stochastic optimization problems in which decisions do not affect the uncertainty. The proposed framework represents the decision variables as elements of a reproducing kernel Hilbert space and performs functional stochastic gradient descent to minimize the empirical regularized loss. By incorporating sparsification techniques based on function subspace projections we are able to overcome the computational complexity that standard kernel methods introduce as the data size increases. We prove that the proposed approach is asymptotically optimal for multistage stochastic optimization with side information. Across various computational experiments on stochastic inventory management problems, {our method performs well in multidimensional settings} and remains tractable when the data size is large. Lastly, by computing lower bounds for the optimal loss of the inventory control problem, we show that the proposed method produces decision rules with near-optimal average performance. | ['Kimberly Villalobos Carballo', 'Dimitris Bertsimas'] | 2023-03-11 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-2.35025659e-01 -3.45061980e-02 -3.56820852e-01 -2.92926848e-01
-9.95878518e-01 -6.07693255e-01 6.49897903e-02 2.58022118e-02
-4.37491864e-01 9.96419132e-01 -7.71564096e-02 -3.66868794e-01
-4.67514634e-01 -5.03382266e-01 -8.44631493e-01 -8.64759743e-01
-2.00102523e-01 6.64765596e-01 -2.29001239e-01 7.51187280e-02
4.23164189e-01 5.63846231e-01 -1.18404603e+00 -3.97337139e-01
1.03279793e+00 1.60681403e+00 4.46997620e-02 4.74456161e-01
1.60629690e-01 5.47626317e-01 -3.67645979e-01 -4.17594463e-01
8.48284960e-01 1.15772180e-01 -3.59519273e-01 5.88396907e-01
-1.01942189e-01 -3.37466955e-01 -1.30516976e-01 1.08329904e+00
3.74840647e-01 4.59173262e-01 9.73365903e-01 -1.13994706e+00
-6.34111106e-01 6.76054716e-01 -7.63151109e-01 5.41024357e-02
-8.44466165e-02 3.75139676e-02 9.89179909e-01 -1.21402097e+00
1.51748866e-01 1.22421587e+00 6.95672452e-01 -1.99167095e-02
-1.60663188e+00 -2.02822179e-01 2.60291010e-01 -3.66210133e-01
-1.30524147e+00 -3.14664274e-01 4.78645831e-01 -5.34327507e-01
6.22909844e-01 1.09867714e-01 2.43478358e-01 5.16568899e-01
2.85849631e-01 8.30226004e-01 1.20812261e+00 -2.86412209e-01
6.36984587e-01 7.40826011e-01 2.89029062e-01 8.31943154e-01
6.74904346e-01 1.92401275e-01 -3.21345299e-01 -5.12196362e-01
7.69093454e-01 8.89653787e-02 -1.41380474e-01 -8.46347272e-01
-7.58711040e-01 1.26550555e+00 -3.83207500e-01 -1.88533500e-01
-6.67012751e-01 1.74744487e-01 2.03080192e-01 6.01889968e-01
7.28724539e-01 2.38913253e-01 -4.97849196e-01 -2.94817630e-02
-1.03186190e+00 3.31696510e-01 1.35114086e+00 1.27569699e+00
3.68025661e-01 3.23119402e-01 -4.67364788e-01 5.37245512e-01
2.84084141e-01 8.35475028e-01 -1.28553761e-02 -1.14681351e+00
7.49702811e-01 -1.15133248e-01 1.00150859e+00 -6.24222279e-01
-2.89388031e-01 -6.23046815e-01 -3.70869637e-01 2.03005001e-01
6.52329028e-01 -6.32720470e-01 -4.65531468e-01 1.60049295e+00
2.36022279e-01 -3.11881542e-01 1.23616055e-01 6.60185397e-01
-5.64935923e-01 6.17244542e-01 -5.08250475e-01 -7.82573104e-01
8.74053836e-01 -7.60004282e-01 -9.18053269e-01 2.86333919e-01
4.34906691e-01 -4.92570400e-01 1.04382491e+00 6.35955095e-01
-1.44446111e+00 -8.97975080e-03 -9.71695125e-01 5.24700165e-01
-7.34089613e-02 2.74645090e-01 6.20475471e-01 7.95251429e-01
-9.20685530e-01 8.86346936e-01 -6.83312356e-01 1.14556536e-01
2.79386580e-01 5.13935328e-01 2.55228221e-01 1.16370864e-01
-9.05980647e-01 8.84710193e-01 2.33416542e-01 3.19108039e-01
-1.03179371e+00 -9.07970011e-01 -7.14768827e-01 1.11844212e-01
7.76148677e-01 -5.85639417e-01 1.26244628e+00 -3.83627295e-01
-1.79750264e+00 1.71281472e-01 1.21623047e-01 -6.74336374e-01
7.79701531e-01 -3.39463532e-01 -3.17422152e-02 1.02622740e-01
7.02004209e-02 -2.95743078e-01 1.16777170e+00 -1.10969591e+00
-6.69876575e-01 -5.00824809e-01 -1.06148034e-01 1.84855178e-01
-4.82533753e-01 -4.85964827e-02 -4.56095748e-02 -5.22836447e-01
-1.88908279e-01 -1.06228328e+00 -6.11328781e-01 -2.25640044e-01
-5.52185833e-01 1.16368696e-01 3.41945589e-01 -5.15847445e-01
1.26487887e+00 -1.90429580e+00 2.87248343e-01 4.54017788e-01
-2.32891992e-01 -4.35142457e-01 4.67279643e-01 4.90552008e-01
2.82456368e-01 2.91256933e-03 -4.84484375e-01 -6.39073312e-01
6.25465095e-01 -9.26344655e-03 -4.83746231e-01 8.83320808e-01
-1.60672382e-01 5.46204925e-01 -4.81020957e-01 -4.51137796e-02
1.49881408e-01 -1.28473133e-01 -4.22740251e-01 2.57277861e-02
-1.31801307e-01 -9.26072150e-02 -6.02883875e-01 6.62020624e-01
6.17179215e-01 -2.13191018e-01 -1.36412963e-01 1.17553025e-01
-1.58128530e-01 -4.57364291e-01 -1.57094669e+00 1.40460742e+00
-8.23161125e-01 1.83183849e-01 5.49342394e-01 -1.18114161e+00
5.86961806e-01 7.51494244e-02 5.89632094e-01 1.86768919e-01
7.79041648e-02 2.76779979e-01 -6.94473565e-01 -1.10661678e-01
3.95307422e-01 -6.39108717e-01 -5.13695538e-01 5.51063120e-01
-4.30805199e-02 -1.56034902e-01 5.51146902e-02 -1.32227093e-01
7.29376614e-01 -8.18705708e-02 2.33179763e-01 -8.48183155e-01
3.84246647e-01 2.33560745e-02 4.16278601e-01 9.52936471e-01
-4.28026132e-02 1.10499389e-01 7.84269810e-01 1.77078694e-03
-9.11202133e-01 -1.30736291e+00 -3.54058266e-01 9.82057989e-01
-1.13916971e-01 4.65125978e-01 -5.78433335e-01 -5.32241642e-01
9.95869219e-01 1.12530160e+00 -7.95897245e-01 -4.87872511e-02
6.27949387e-02 -9.39712048e-01 3.10744513e-02 3.89063627e-01
1.41000554e-01 -1.98536888e-01 -9.70164835e-02 4.01403338e-01
5.78577101e-01 -9.17130172e-01 -9.62500215e-01 4.29905146e-01
-1.00214577e+00 -7.50073135e-01 -1.03886509e+00 -2.92017072e-01
6.88982368e-01 1.11829318e-01 6.99890673e-01 -1.19898343e+00
-1.21387348e-01 8.19308519e-01 3.85444015e-01 -4.91555393e-01
2.21272297e-02 -1.19962476e-01 5.28102338e-01 3.23658645e-01
1.72135770e-01 -3.11361998e-01 -4.08361942e-01 3.17310363e-01
-6.97537839e-01 -6.20311379e-01 3.62832606e-01 8.56773317e-01
8.63405287e-01 4.02696788e-01 6.87215865e-01 -9.00604248e-01
1.07693756e+00 -6.87593937e-01 -1.37127745e+00 3.33349556e-01
-1.12919807e+00 5.59131801e-01 8.00498366e-01 -3.59813958e-01
-1.28629625e+00 -6.55555585e-03 7.69598305e-01 -7.51453519e-01
3.46084177e-01 5.52554905e-01 9.78310034e-02 -1.64129555e-01
1.48009166e-01 3.41093421e-01 2.86298215e-01 -6.19068325e-01
3.60702097e-01 4.30211574e-01 1.61149129e-01 -9.14405048e-01
8.39284301e-01 5.56771994e-01 2.81792313e-01 -6.61466777e-01
-1.03986168e+00 -2.97356516e-01 -3.28093171e-01 3.34136263e-02
5.42560399e-01 -8.19795668e-01 -1.09447086e+00 1.65773094e-01
-4.80390102e-01 -1.08228922e-01 -8.84915531e-01 7.56135464e-01
-1.28388381e+00 1.66956872e-01 -6.41251981e-01 -1.52500558e+00
-1.37117326e-01 -9.08738375e-01 7.96483159e-01 1.92353223e-02
3.72997373e-01 -1.29775512e+00 1.21374026e-01 2.52690375e-01
4.42823142e-01 2.49119461e-01 8.51210892e-01 -7.96510220e-01
-3.75342727e-01 -3.26166898e-01 1.53396204e-01 6.05005145e-01
1.36183649e-01 -3.38027239e-01 -4.19731826e-01 -5.17915606e-01
5.48992991e-01 -3.03916693e-01 7.06266165e-01 7.10322320e-01
1.14942312e+00 -5.77579260e-01 6.12677410e-02 5.26283741e-01
1.63953900e+00 3.02606840e-02 -1.38941109e-01 1.79713041e-01
2.68587917e-01 7.89355755e-01 1.02235019e+00 1.21016324e+00
3.89814019e-01 3.69113654e-01 6.67369664e-02 3.16510648e-01
9.47670102e-01 -1.45693615e-01 6.80169642e-01 5.62054694e-01
2.17651278e-01 -5.76866865e-02 -5.16259611e-01 4.59203422e-01
-2.12980247e+00 -8.09189141e-01 2.87650526e-01 2.66584182e+00
7.65225947e-01 1.10538572e-01 4.53539193e-01 -3.19332898e-01
7.12281287e-01 -1.73235267e-01 -1.12623632e+00 -7.23584771e-01
-2.37232164e-01 8.18337128e-02 1.39140630e+00 6.24568522e-01
-1.02084553e+00 5.58699310e-01 7.22587824e+00 7.32742190e-01
-3.81862581e-01 7.64852464e-02 7.00529397e-01 -5.82286775e-01
-3.82721454e-01 -2.08138257e-01 -9.49502230e-01 5.53718984e-01
1.14553511e+00 -6.60694718e-01 6.52516127e-01 1.06490791e+00
7.00295210e-01 -2.70999581e-01 -1.16364014e+00 7.70213127e-01
-3.36124837e-01 -1.17039919e+00 -4.06627327e-01 5.23213744e-01
9.73369777e-01 -2.97860295e-01 6.02252126e-01 3.57472897e-01
7.68786013e-01 -8.83569419e-01 6.46097362e-01 8.19119453e-01
5.17370939e-01 -1.35022116e+00 6.40270412e-01 2.82876104e-01
-8.86367142e-01 -5.66792846e-01 -4.87620056e-01 1.96763799e-01
3.14369231e-01 8.01173747e-01 -4.37275171e-01 3.72560501e-01
1.16989091e-01 4.18793857e-02 2.27087364e-01 8.34201515e-01
1.94217801e-01 4.11073178e-01 -6.73195601e-01 -1.91558495e-01
4.87744659e-01 -8.00186515e-01 5.51263034e-01 9.97648358e-01
4.00244623e-01 -1.40366644e-01 4.69098508e-01 8.81836832e-01
9.19659287e-02 2.81889200e-01 -7.42731273e-01 -2.91251183e-01
2.46900573e-01 7.48354137e-01 -4.01512474e-01 9.69807093e-04
-5.99965751e-01 1.10688043e+00 1.53426528e-01 7.58717895e-01
-6.74968719e-01 -5.16366005e-01 8.22314203e-01 -2.31319085e-01
6.87465429e-01 -4.74073499e-01 -3.33015025e-01 -1.11988556e+00
3.38855267e-01 -4.68720376e-01 4.89365876e-01 4.92292829e-02
-1.40305614e+00 -2.06725031e-01 7.79356956e-02 -7.58880317e-01
-5.50115347e-01 -6.41371071e-01 -2.48127118e-01 8.53711843e-01
-1.20190048e+00 -4.15701151e-01 7.68149912e-01 5.89915335e-01
4.28822458e-01 -2.83670872e-01 4.34699148e-01 -1.17076457e-01
-7.67998338e-01 5.55283785e-01 1.19019842e+00 -5.95942497e-01
2.42241487e-01 -1.69861054e+00 -3.21859390e-01 6.43937171e-01
-4.62126046e-01 5.52553296e-01 1.02497876e+00 -5.50793529e-01
-1.89348793e+00 -8.76765251e-01 2.22723156e-01 -8.87086615e-02
1.25160432e+00 -4.44343835e-01 -4.45432454e-01 5.29104531e-01
-2.26532012e-01 -2.92287581e-02 9.61571574e-01 8.18333700e-02
2.02792779e-01 -3.38204592e-01 -1.61089003e+00 2.97966570e-01
5.80702662e-01 -5.61174929e-01 -3.03022623e-01 6.99671745e-01
7.66374111e-01 9.37725138e-03 -1.47812772e+00 -7.28666931e-02
4.31306690e-01 -4.46785331e-01 7.02461123e-01 -1.03294122e+00
-3.15858983e-02 3.36539671e-02 -5.04055321e-01 -1.37730539e+00
-2.05027655e-01 -1.28389633e+00 -7.01825738e-01 1.00814056e+00
6.85813606e-01 -9.13904250e-01 8.04235339e-01 1.10725451e+00
1.83492944e-01 -1.08786488e+00 -9.90836978e-01 -1.25232840e+00
2.77649701e-01 -2.13104784e-01 2.42771029e-01 3.37393612e-01
4.53907102e-02 -1.69262692e-01 -6.11168146e-01 2.99710125e-01
1.33794260e+00 5.10295570e-01 3.26106936e-01 -8.79995286e-01
-6.91977859e-01 -3.10419083e-01 3.74522954e-02 -9.86886263e-01
3.86601329e-01 -5.71276546e-01 -6.84150308e-02 -1.07592046e+00
1.84839383e-01 -3.31121445e-01 -7.13006318e-01 -2.04397455e-01
1.56000197e-01 -5.73956728e-01 1.09009124e-01 -2.06410214e-02
-5.49301147e-01 9.33511436e-01 1.01990151e+00 9.48060304e-02
-3.50299418e-01 6.42092705e-01 -8.71686518e-01 6.25159502e-01
6.84184670e-01 -2.43168086e-01 -6.35364532e-01 7.48607218e-02
-3.80897596e-02 6.22997761e-01 -2.88411498e-01 -4.05103415e-01
4.49690111e-02 -4.82389927e-01 2.95457602e-01 -7.30866551e-01
2.98218220e-01 -8.89597178e-01 2.79275738e-02 3.81330848e-01
-5.14444292e-01 8.15823823e-02 -2.07023006e-02 1.04007649e+00
1.21715918e-01 -4.00889546e-01 6.75513566e-01 1.07233837e-01
-7.16964751e-02 6.32848442e-01 -5.08057892e-01 1.65646225e-01
1.22738492e+00 7.13452771e-02 2.82855093e-01 -6.14025533e-01
-1.12857342e+00 6.72556162e-01 4.99937326e-01 -1.43559456e-01
4.48632091e-01 -1.33793187e+00 -5.71962178e-01 -3.39187950e-01
-2.87000299e-01 -3.62932265e-01 4.09692014e-03 9.61153269e-01
-6.84839413e-02 8.42045784e-01 3.57335836e-01 -2.32202739e-01
-5.25263071e-01 7.93301940e-01 1.95655957e-01 -5.73761880e-01
-1.34907782e-01 8.28055501e-01 1.41856432e-01 -1.84418723e-01
3.86788994e-01 -3.60379815e-01 3.17642629e-01 3.55810583e-01
2.56116867e-01 7.88760602e-01 -1.23336636e-01 -2.89973374e-02
-3.36969569e-02 8.74695778e-02 -1.59479663e-01 -6.39910519e-01
1.49058938e+00 -6.08311892e-01 1.80542856e-01 9.59627390e-01
1.32292938e+00 -4.96743014e-03 -1.71108449e+00 -5.00101209e-01
3.49972159e-01 -4.26368207e-01 2.00954705e-01 -4.51188684e-01
-1.01703477e+00 4.12167251e-01 5.32327294e-01 2.68161535e-01
9.24141407e-01 -3.13647717e-01 4.96762216e-01 7.82741189e-01
6.93564236e-01 -1.69214392e+00 -2.59080976e-01 3.72424185e-01
8.07202220e-01 -1.13285422e+00 8.03816989e-02 -2.03454956e-01
-7.71912277e-01 1.05366361e+00 2.43823882e-02 -3.21682632e-01
9.58959103e-01 3.91728282e-01 -5.94182670e-01 2.14033220e-02
-7.68874943e-01 -1.58139348e-01 -1.04060897e-03 2.63200670e-01
-1.87145948e-01 3.76848876e-01 -5.07354856e-01 8.72763813e-01
-1.48790120e-03 -2.76856333e-01 5.72766602e-01 9.71802473e-01
-5.63606143e-01 -7.52654016e-01 -4.69521612e-01 7.96087146e-01
-7.28739560e-01 7.28778541e-02 3.46991457e-02 7.67919958e-01
-8.23081017e-01 7.68705666e-01 -1.42069265e-01 8.67765397e-02
3.69116426e-01 1.07163265e-01 5.73706210e-01 -4.71825063e-01
-1.25953853e-01 9.78535339e-02 1.66317686e-01 -6.79027319e-01
2.99692452e-01 -1.15455353e+00 -9.21551704e-01 -2.21996486e-01
-4.56547379e-01 3.40455592e-01 9.41449940e-01 7.83672810e-01
2.71960527e-01 1.17232032e-01 1.44553590e+00 -4.46256459e-01
-2.02149081e+00 -5.28451622e-01 -1.28964293e+00 -2.51710892e-01
2.99503267e-01 -8.92963111e-01 -7.06328750e-01 -5.20525336e-01] | [5.471462249755859, 3.728276252746582] |
cf48f5f1-d4b5-4988-aae4-bf5aa6778853 | black-lscdiscovery-shared-task-glossreader-at | null | null | https://aclanthology.org/2022.lchange-1.22 | https://aclanthology.org/2022.lchange-1.22.pdf | black[LSCDiscovery shared task] GlossReader at LSCDiscovery: Train to Select a Proper Gloss in English – Discover Lexical Semantic Change in Spanish | The contextualized embeddings obtained from neural networks pre-trained as Language Models (LM) or Masked Language Models (MLM) are not well suitable for solving the Lexical Semantic Change Detection (LSCD) task because they are more sensitive to changes in word forms rather than word meaning, a property previously known as the word form bias or orthographic bias. Unlike many other NLP tasks, it is also not obvious how to fine-tune such models for LSCD. In order to conclude if there are any differences between senses of a particular word in two corpora, a human annotator or a system shall analyze many examples containing this word from both corpora. This makes annotation of LSCD datasets very labour-consuming. The existing LSCD datasets contain up to 100 words that are labeled according to their semantic change, which is hardly enough for fine-tuning. To solve these problems we fine-tune the XLM-R MLM as part of a gloss-based WSD system on a large WSD dataset in English. Then we employ zero-shot cross-lingual transferability of XLM-R to build the contextualized embeddings for examples in Spanish. In order to obtain the graded change score for each word, we calculate the average distance between our improved contextualized embeddings of its old and new occurrences. For the binary change detection subtask, we apply thresholding to the same scores. Our solution has shown the best results among all other participants in all subtasks except for the optional sense gain detection subtask. | ['Nikolay Arefyev', 'Maxim Rachinskiy'] | null | null | null | null | lchange-acl-2022-5 | ['xlm-r'] | ['natural-language-processing'] | [ 2.31241807e-01 -7.89006576e-02 2.48658238e-03 -4.03353274e-01
-5.84292173e-01 -6.88837409e-01 5.85796595e-01 6.11043990e-01
-1.07873273e+00 5.81381321e-01 3.53991538e-01 -3.74868393e-01
1.82288930e-01 -8.11361551e-01 -5.45570076e-01 -5.26101410e-01
3.70912135e-01 3.73650432e-01 4.84720975e-01 -6.03999853e-01
2.86997080e-01 2.27141291e-01 -1.58669889e+00 4.08961356e-01
1.07618773e+00 6.32043898e-01 6.14703536e-01 1.44413710e-01
-6.07799828e-01 -1.82762489e-01 -5.47236204e-01 -3.43022913e-01
1.44477502e-01 -4.26999152e-01 -8.41770768e-01 -4.11464572e-01
6.10629559e-01 2.44555131e-01 2.41671294e-01 1.39167750e+00
6.35369718e-01 1.87208325e-01 5.55796266e-01 -6.77927136e-01
-1.20755136e+00 8.02999556e-01 -1.56509206e-01 6.10273778e-01
4.71487761e-01 4.38355692e-02 1.15729821e+00 -1.26398158e+00
8.66197586e-01 1.44412506e+00 5.87053657e-01 5.31849444e-01
-1.33739161e+00 -5.08489192e-01 4.24999148e-01 4.59409833e-01
-1.16018224e+00 -2.66051829e-01 5.54448843e-01 -4.16356444e-01
1.42563903e+00 9.66089517e-02 6.15787208e-01 1.20797586e+00
1.20669074e-01 3.52692187e-01 1.25036240e+00 -9.40129995e-01
2.69419014e-01 3.87457311e-01 2.02155203e-01 3.47303540e-01
4.62828606e-01 -1.05821170e-01 -5.88698328e-01 2.22524405e-01
2.28594065e-01 -1.39893129e-01 -2.61274248e-01 -2.20611677e-01
-1.33990872e+00 9.27567840e-01 5.09294808e-01 9.44454014e-01
-3.17935020e-01 -2.03460589e-01 5.07045984e-01 7.26278186e-01
5.23835301e-01 9.08396006e-01 -9.48150218e-01 -2.81327348e-02
-5.85578978e-01 5.91187999e-02 3.77423018e-01 5.30123532e-01
9.12702203e-01 -1.39198005e-01 -1.61557809e-01 1.21586978e+00
9.21264067e-02 3.76737595e-01 1.24010956e+00 -3.82911503e-01
3.48801613e-01 8.74786198e-01 5.43341786e-02 -9.11007941e-01
-3.96562368e-01 -7.37039000e-02 -3.46938461e-01 1.10458888e-01
3.27345759e-01 1.09467603e-01 -9.66799557e-01 2.16775656e+00
3.57195169e-01 -2.47387007e-01 6.70155063e-02 6.87365294e-01
7.01436460e-01 5.56707859e-01 3.46258879e-01 -2.01198936e-01
1.54419768e+00 -6.82740748e-01 -8.67114007e-01 -5.98510802e-01
9.57302034e-01 -8.39426696e-01 2.00531006e+00 9.08772871e-02
-7.64215589e-01 -7.43421614e-01 -1.17272055e+00 -2.24959180e-01
-1.17791772e+00 -4.03950550e-02 2.44019821e-01 5.19828022e-01
-9.55561042e-01 5.80470681e-01 -6.84012413e-01 -9.89251435e-01
-1.36939110e-02 1.94431823e-02 -4.32472169e-01 -8.56218562e-02
-1.73640227e+00 1.50935149e+00 6.76652968e-01 -8.99423808e-02
-4.14084822e-01 -6.79325104e-01 -9.97186005e-01 -9.52743366e-02
3.35831642e-01 -3.58869582e-01 1.05242765e+00 -9.63147759e-01
-1.20380032e+00 1.49328148e+00 -1.84941530e-01 -1.98089331e-01
1.33105852e-02 -9.61715505e-02 -8.99163067e-01 -4.44258660e-01
3.79403204e-01 5.64335585e-01 8.19641352e-01 -9.56243396e-01
-7.49432564e-01 -5.04074574e-01 -1.70907546e-02 2.72772819e-01
-5.23475826e-01 -4.76397052e-02 -9.70915928e-02 -9.07300353e-01
7.31849149e-02 -8.33201885e-01 -7.92519972e-02 -2.66663134e-01
5.04455250e-03 -7.00361609e-01 5.07035017e-01 -6.31012738e-01
1.51567411e+00 -2.26060963e+00 1.27543248e-02 -1.42776966e-01
-3.86808634e-01 6.42001450e-01 -4.32049632e-01 3.30278635e-01
-2.28247598e-01 3.35775375e-01 -3.13103825e-01 -9.94025618e-02
2.40775570e-01 3.82465601e-01 -2.03465164e-01 -1.74707435e-02
2.98916548e-01 7.66030908e-01 -1.04745245e+00 -1.25615329e-01
1.06196061e-01 9.06989276e-02 -5.70123851e-01 -1.23097591e-01
-2.92596698e-01 2.85059363e-02 1.64454967e-01 1.98605239e-01
5.51715851e-01 2.24546000e-01 2.43064880e-01 -2.58238286e-01
-1.33396342e-01 7.65223503e-01 -1.26525569e+00 1.96394503e+00
-8.29789340e-01 4.42560285e-01 -5.38517475e-01 -8.58567357e-01
8.88729692e-01 1.93426132e-01 -1.80544615e-01 -9.60657954e-01
-1.82705522e-01 5.06998897e-01 1.43315032e-01 -6.51622534e-01
6.99392796e-01 -5.80669641e-01 -4.20248151e-01 3.38097304e-01
1.79980606e-01 -1.84181169e-01 4.69618559e-01 -1.52280271e-01
1.16553128e+00 -6.82156757e-02 6.18726730e-01 -4.11294013e-01
5.44505060e-01 1.18168727e-01 6.78025961e-01 4.28738803e-01
-1.48870334e-01 3.50383371e-01 1.96149349e-01 -3.71029407e-01
-7.82975793e-01 -1.22628129e+00 -2.32012302e-01 1.41252327e+00
2.05176529e-02 -4.18008983e-01 -6.11459672e-01 -7.20101357e-01
-1.94958895e-02 1.28048778e+00 -6.78673923e-01 -4.50192630e-01
-4.71711010e-01 -7.41763592e-01 2.42636889e-01 3.74983251e-01
1.75679132e-01 -1.46319091e+00 -4.47571456e-01 4.52551752e-01
-1.39992326e-01 -9.59357440e-01 -5.26191413e-01 4.55703676e-01
-5.50659895e-01 -8.71348739e-01 -1.96901590e-01 -1.21503019e+00
4.59966213e-01 6.23555295e-02 1.07355571e+00 -1.95838302e-01
-2.50553161e-01 9.85984206e-02 -5.33156455e-01 -4.69744027e-01
-6.82539821e-01 1.53604701e-01 2.84195006e-01 -3.38618696e-01
1.03273451e+00 -3.45230997e-01 -2.71710157e-01 1.23819143e-01
-1.18042290e+00 -5.16521752e-01 3.66461277e-01 7.05017209e-01
6.46773219e-01 1.12011403e-01 6.71185553e-01 -8.50161910e-01
1.02124059e+00 -2.54874945e-01 -3.41540396e-01 3.96612912e-01
-9.04246807e-01 2.29690745e-01 4.53100860e-01 -6.09666705e-01
-1.05971849e+00 -2.39308029e-01 -2.90081263e-01 2.53502131e-01
-3.31346810e-01 6.00754499e-01 -2.93054432e-01 5.39414048e-01
9.85677481e-01 -7.05365092e-02 -3.51789832e-01 -6.30074978e-01
6.65593743e-01 6.53847694e-01 3.64664823e-01 -3.95582020e-01
7.28350282e-01 9.93689522e-02 -8.17810595e-01 -7.29310691e-01
-9.93780375e-01 -3.40979874e-01 -8.29067349e-01 3.59331727e-01
1.17487574e+00 -7.44445741e-01 6.41011372e-02 2.99625576e-01
-1.29436755e+00 -3.12346697e-01 -5.94495535e-01 4.66115505e-01
-8.24823827e-02 1.82071939e-01 -2.91887283e-01 -5.03631495e-02
-1.32467926e-01 -9.86025333e-01 8.05714309e-01 1.63165443e-02
-7.16870070e-01 -1.18475318e+00 4.69480515e-01 -9.13889036e-02
6.03868663e-01 1.18692769e-02 1.63185096e+00 -8.92124355e-01
2.62654066e-01 -8.71828794e-02 1.53662637e-01 6.14778399e-01
5.73607385e-01 -2.33187810e-01 -8.65782320e-01 -3.12628299e-01
-3.60157304e-02 -2.55679429e-01 9.30644333e-01 1.34314746e-01
6.77073956e-01 -1.38407990e-01 -2.12999925e-01 1.13389671e-01
1.47780979e+00 2.75416285e-01 4.75309044e-01 6.05796933e-01
4.47237909e-01 4.41017985e-01 5.94478607e-01 5.75616881e-02
3.13755006e-01 7.65725613e-01 6.09400570e-02 1.37092292e-01
-3.81265849e-01 -3.71928245e-01 7.89717257e-01 1.20121443e+00
3.86737853e-01 -1.94971547e-01 -7.39680946e-01 9.24622655e-01
-1.38754690e+00 -7.34084964e-01 -1.87166944e-01 2.30523968e+00
1.25580919e+00 2.60461688e-01 -3.58229607e-01 2.11234555e-01
8.16144764e-01 2.88819164e-01 -4.18575615e-01 -9.72082376e-01
-2.40000576e-01 6.08375072e-01 1.65264644e-02 5.83638847e-01
-7.31976151e-01 1.51525939e+00 5.44953775e+00 8.81965160e-01
-1.00071013e+00 5.01986265e-01 -1.29591241e-01 -6.75350353e-02
-6.13630533e-01 2.96676736e-02 -9.46066380e-01 5.24871647e-01
9.12467718e-01 -5.42929284e-02 4.88215566e-01 6.82314932e-01
2.29806110e-01 -2.65853316e-01 -1.26741660e+00 9.14750814e-01
1.85148731e-01 -9.30172563e-01 3.01294327e-01 -4.51686382e-01
7.50268102e-01 1.41768858e-01 -9.12048146e-02 6.65755153e-01
4.10915405e-01 -7.79668689e-01 6.06082618e-01 1.57586053e-01
8.41346383e-01 -4.11163628e-01 7.14471459e-01 4.45706099e-02
-9.25825953e-01 1.86559439e-01 -6.94720089e-01 -8.21336359e-02
1.36139572e-01 7.17251360e-01 -8.07467878e-01 1.00136645e-01
8.55343580e-01 4.42160517e-01 -9.02514100e-01 5.74071586e-01
-7.68473804e-01 5.02018034e-01 -1.44627035e-01 -1.03032172e-01
1.00461289e-01 3.27779017e-02 5.89727461e-01 1.28278887e+00
5.10678232e-01 -4.35098976e-01 -2.07065895e-01 7.37450361e-01
-1.60909995e-01 3.67292523e-01 -4.93037105e-01 -1.77270755e-01
6.66352034e-01 1.19632638e+00 -6.32192612e-01 -2.75914073e-01
-5.25835514e-01 1.35606480e+00 5.86447060e-01 2.54988968e-01
-4.58988100e-01 -5.16753554e-01 9.82418478e-01 8.34807903e-02
4.99547720e-01 -1.36262581e-01 -1.65957674e-01 -1.12839389e+00
1.25275716e-01 -7.98875213e-01 5.83724439e-01 -6.96912169e-01
-1.61598241e+00 6.42582417e-01 -2.64086574e-01 -9.66049910e-01
-1.79056078e-01 -8.63374472e-01 -4.44595546e-01 8.64071727e-01
-1.59716022e+00 -7.40203857e-01 -1.30922824e-01 5.38029611e-01
6.54436946e-01 -6.70172721e-02 1.17327821e+00 3.18926632e-01
-4.18612152e-01 5.38281024e-01 -1.43107936e-01 -3.70140672e-02
1.27956212e+00 -1.44227135e+00 5.48828423e-01 9.89963591e-01
3.23614120e-01 8.08869183e-01 6.81976795e-01 -6.16061091e-01
-6.43213034e-01 -1.08284748e+00 1.48390126e+00 -5.08452296e-01
8.74347746e-01 -3.24627310e-01 -1.40057850e+00 6.42337620e-01
3.41558456e-01 -9.18901265e-02 7.32261121e-01 3.22955817e-01
-4.75204557e-01 -1.86344907e-01 -9.80982065e-01 7.03123093e-01
1.25198460e+00 -6.46070182e-01 -1.47121489e+00 1.66269556e-01
1.08534849e+00 -4.94812243e-02 -7.21502841e-01 1.52783483e-01
-3.33581716e-02 -7.52446175e-01 5.03628969e-01 -8.77185881e-01
1.64621770e-01 -5.21288991e-01 -3.82591188e-01 -2.13829923e+00
-4.55600828e-01 -2.06030607e-01 6.02454066e-01 1.49470770e+00
6.66009843e-01 -7.07294881e-01 6.93049328e-03 2.60839909e-01
-2.26523459e-01 -1.74136564e-01 -1.05039036e+00 -9.92668211e-01
3.37489545e-01 -7.08811760e-01 8.03199291e-01 1.32990932e+00
2.01450940e-02 5.43094099e-01 3.17986429e-01 9.71573070e-02
-1.21159047e-01 -6.60318881e-02 3.42300385e-01 -1.12113357e+00
6.42114431e-02 -5.87709725e-01 -3.19341660e-01 -4.56369251e-01
3.67328763e-01 -1.41066337e+00 9.58395600e-02 -1.67094910e+00
-2.63600498e-02 -3.40832800e-01 -6.90306544e-01 7.01062083e-01
-4.86485809e-01 1.39778242e-01 2.31413096e-02 -1.80818304e-01
-3.63761365e-01 4.22137886e-01 9.10911798e-01 -4.64814082e-02
-4.35061574e-01 -4.61493760e-01 -7.39544630e-01 7.93269038e-01
8.16017687e-01 -7.36070335e-01 -2.20719829e-01 -6.38093233e-01
6.39856040e-01 -7.77100027e-01 2.05941051e-01 -7.11227000e-01
-1.60466447e-01 -1.35089993e-01 1.19377151e-01 -2.48075426e-01
-2.29727793e-02 -7.60981202e-01 -1.59824073e-01 5.49284816e-01
-4.21035230e-01 3.28222781e-01 3.69207531e-01 2.28009492e-01
-2.85520047e-01 -5.01535773e-01 9.07090902e-01 -2.74428576e-01
-1.26989627e+00 -1.28061801e-01 -5.93048692e-01 4.97188240e-01
8.26591015e-01 -5.72146513e-02 -3.15624952e-01 1.23204760e-01
-7.76028216e-01 -8.78852010e-02 5.55437684e-01 1.00651968e+00
3.03907245e-01 -1.69569230e+00 -4.95663702e-01 2.69973397e-01
6.82095885e-01 -2.51464635e-01 6.96203560e-02 5.58689654e-01
-1.12970836e-01 1.84702113e-01 -2.14616463e-01 -1.65868565e-01
-1.03213322e+00 7.97092855e-01 2.24712193e-01 -2.48685881e-01
-1.81424215e-01 8.09669673e-01 -6.38341159e-02 -9.55761254e-01
-1.42933264e-01 -7.94696510e-01 -1.96218193e-01 5.82268536e-01
4.91699964e-01 2.04116002e-01 5.26952982e-01 -4.34486747e-01
-5.88468015e-01 6.66668475e-01 -7.77444541e-02 -5.40242903e-02
1.44269502e+00 -3.28148305e-01 -2.59118885e-01 8.01220655e-01
1.29336154e+00 1.31531656e-01 -6.91310525e-01 -6.36310279e-01
4.91580427e-01 -2.16014981e-01 -6.58704191e-02 -9.73603666e-01
-5.89166284e-01 8.68166268e-01 8.65211964e-01 1.75056830e-01
8.74173760e-01 1.12608127e-01 8.60412419e-01 4.01141733e-01
2.23230317e-01 -1.65673780e+00 5.78997061e-02 7.60713935e-01
1.04477298e+00 -1.31802201e+00 -3.42444718e-01 2.10489538e-02
-6.13709748e-01 9.73684072e-01 5.69206119e-01 2.48548165e-02
4.52868700e-01 -9.52397138e-02 3.16696405e-01 -1.17407896e-01
-6.30357146e-01 -5.01032650e-01 3.04714113e-01 6.88605607e-01
4.50009823e-01 2.08195463e-01 -8.75186920e-01 5.54606438e-01
-4.71603841e-01 -3.83757144e-01 3.57151449e-01 6.26819551e-01
-7.12749004e-01 -1.35034454e+00 -2.24196494e-01 2.46099383e-01
-1.21366553e-01 -3.43130708e-01 -4.27368820e-01 8.27081501e-01
5.09608269e-01 8.68331015e-01 2.18079343e-01 -2.84315407e-01
7.14484930e-01 6.34345293e-01 3.69943649e-01 -1.06972373e+00
-4.62301612e-01 -3.80112559e-01 1.03445880e-01 -5.12540042e-01
-3.52853775e-01 -7.50192702e-01 -1.39913630e+00 6.38347641e-02
-1.08680025e-01 -1.23786405e-01 6.41362727e-01 1.09613919e+00
1.32362187e-01 4.32472527e-01 2.83126384e-01 -3.99995506e-01
-5.06645679e-01 -1.29188132e+00 -5.02743483e-01 9.47541654e-01
-2.45304629e-02 -6.64057970e-01 -4.62674797e-01 -2.29385234e-02] | [10.478604316711426, 9.15447998046875] |
d9939993-32a6-45be-b888-cd1d832e8346 | locking-and-quacking-stacking-bayesian-model | 2305.07334 | null | https://arxiv.org/abs/2305.07334v1 | https://arxiv.org/pdf/2305.07334v1.pdf | Locking and Quacking: Stacking Bayesian model predictions by log-pooling and superposition | Combining predictions from different models is a central problem in Bayesian inference and machine learning more broadly. Currently, these predictive distributions are almost exclusively combined using linear mixtures such as Bayesian model averaging, Bayesian stacking, and mixture of experts. Such linear mixtures impose idiosyncrasies that might be undesirable for some applications, such as multi-modality. While there exist alternative strategies (e.g. geometric bridge or superposition), optimising their parameters usually involves computing an intractable normalising constant repeatedly. We present two novel Bayesian model combination tools. These are generalisations of model stacking, but combine posterior densities by log-linear pooling (locking) and quantum superposition (quacking). To optimise model weights while avoiding the burden of normalising constants, we investigate the Hyvarinen score of the combined posterior predictions. We demonstrate locking with an illustrative example and discuss its practical application with importance sampling. | ['Yann McLatchie', 'Diego Mesquita', 'Luiz Max Carvalho', 'Yuling Yao'] | 2023-05-12 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 4.87796128e-01 1.15157031e-01 2.03050360e-01 -4.68204528e-01
-1.08744633e+00 -6.03587151e-01 1.04225767e+00 -2.23558426e-01
-3.63107324e-01 9.52531815e-01 1.84793264e-01 -4.29873466e-01
-6.15962267e-01 -5.62466979e-01 -5.15765786e-01 -1.20591450e+00
2.74051160e-01 8.46308351e-01 3.06866050e-01 3.09371471e-01
6.08823419e-01 5.09730399e-01 -1.44375992e+00 1.98974654e-01
9.84050989e-01 5.55254817e-01 1.35632724e-01 7.29504704e-01
3.56319733e-02 3.62888932e-01 -4.74749804e-01 -7.29493916e-01
-3.89321372e-02 -3.49367887e-01 -4.19085056e-01 -2.03341857e-01
3.30931425e-01 -1.61735445e-01 1.38860658e-01 1.06611192e+00
5.41308105e-01 3.40843201e-01 1.31394398e+00 -9.83700275e-01
-2.18204036e-02 4.71535832e-01 -4.86674190e-01 5.38053166e-04
2.02358067e-01 1.68840453e-01 8.83976877e-01 -6.04183435e-01
4.10591513e-02 1.53444743e+00 8.45985174e-01 1.82248935e-01
-1.83976460e+00 -6.14314914e-01 -1.68027937e-01 1.81150332e-01
-1.69890893e+00 -7.33506262e-01 4.10905480e-01 -4.80229706e-01
7.84843028e-01 5.52053809e-01 2.34420925e-01 1.03731799e+00
5.35835028e-01 6.50077701e-01 1.21010506e+00 -5.74975967e-01
5.03216982e-01 1.76390439e-01 6.66117966e-02 2.75470704e-01
1.68625325e-01 1.47427887e-01 -6.52900398e-01 -8.31714988e-01
5.66320419e-01 -1.02632500e-01 -1.10742033e-01 -3.17164421e-01
-1.04928994e+00 8.32800210e-01 -2.16290206e-01 -1.26911595e-01
-3.83719116e-01 2.32465848e-01 -1.89778611e-01 -2.20754430e-01
5.08254945e-01 3.65382373e-01 -3.56101781e-01 1.03647470e-01
-1.38042307e+00 7.58955002e-01 1.02098095e+00 7.43053675e-01
6.44466579e-01 -3.96204621e-01 -1.14693739e-01 9.08234358e-01
1.02270436e+00 7.77702093e-01 -4.11655270e-02 -1.21493185e+00
4.15804349e-02 -3.74431729e-01 3.74843627e-01 -3.99063915e-01
-3.69045466e-01 -1.54539034e-01 -8.41260910e-01 2.21586913e-01
7.06479669e-01 -1.29255965e-01 -9.68656063e-01 1.57131243e+00
2.15269238e-01 3.48260432e-01 -8.03701878e-02 5.48954964e-01
3.20210993e-01 9.25435960e-01 3.56325537e-01 -5.31772673e-01
1.23369455e+00 -4.05161649e-01 -7.02823281e-01 -2.36171246e-01
2.79299378e-01 -1.11720932e+00 1.78753495e-01 8.17438006e-01
-1.19133139e+00 -2.24443868e-01 -7.68965900e-01 2.04457015e-01
-1.94277111e-02 -2.26927102e-01 6.15618944e-01 9.99564588e-01
-7.51518846e-01 8.89122427e-01 -1.05784452e+00 -1.88322127e-01
1.45126790e-01 4.59187269e-01 -5.95835783e-02 -1.72650516e-01
-1.01166642e+00 1.38778973e+00 6.09766543e-01 9.52345803e-02
-4.77347553e-01 -7.22892165e-01 -4.57967490e-01 1.32258713e-01
3.35178703e-01 -8.38768780e-01 1.51792514e+00 -1.81239009e-01
-1.96055365e+00 2.89569765e-01 -2.93963552e-01 -2.34688193e-01
3.01560700e-01 -3.65763217e-01 -1.85328305e-01 -3.23858112e-01
-4.00323004e-01 5.16024768e-01 9.98206735e-01 -1.14426088e+00
-3.05631369e-01 -1.80078417e-01 -3.42367291e-01 1.81782454e-01
5.12684941e-01 2.57851154e-01 -8.64595473e-02 -1.79880723e-01
5.73071182e-01 -7.96837032e-01 -5.21602750e-01 -3.37451547e-01
-5.60427547e-01 -1.24556549e-01 1.99294835e-01 -7.65088201e-01
1.20759940e+00 -1.90644515e+00 3.43910009e-01 6.05265081e-01
-1.20779455e-01 -1.36769727e-01 2.17915416e-01 5.81554532e-01
3.54634412e-02 6.71912869e-03 -6.41466320e-01 -3.69421124e-01
2.76555896e-01 2.38250583e-01 -3.52913767e-01 5.68820655e-01
1.19250730e-01 5.25126696e-01 -6.30972683e-01 -5.25035381e-01
3.39314908e-01 5.69576621e-01 -7.40999222e-01 7.13839307e-02
-3.22696298e-01 4.37164724e-01 -1.77065983e-01 3.40136498e-01
9.12442863e-01 -1.26541689e-01 4.51779872e-01 -3.13202560e-01
-1.42612696e-01 4.84064341e-01 -1.54865217e+00 1.19209635e+00
-8.91887248e-02 1.16373725e-01 -1.47127211e-01 -9.21011567e-01
5.85107982e-01 3.83754939e-01 3.86591479e-02 9.64015648e-02
-2.73134917e-01 3.62919658e-01 2.35429823e-01 -2.19148874e-01
4.14632648e-01 -8.45257819e-01 -1.36309564e-01 4.09198254e-01
3.50952476e-01 -7.45232821e-01 -5.26070744e-02 7.78354108e-02
7.17134655e-01 6.11710906e-01 6.13516450e-01 -3.17954987e-01
5.13566583e-02 -3.09928864e-01 2.86163211e-01 1.16993630e+00
1.90322042e-01 7.60994017e-01 5.81601501e-01 1.44439444e-01
-1.15548813e+00 -1.45662117e+00 -5.40769994e-01 7.81152785e-01
-2.66487628e-01 -4.44934636e-01 -5.31149268e-01 -1.04661725e-01
-1.30067304e-01 1.23188388e+00 -1.22358620e-01 -1.28119513e-01
-1.83074668e-01 -1.52457798e+00 2.67111778e-01 7.77389929e-02
-5.93171827e-02 -5.49417675e-01 -4.81818557e-01 4.06885654e-01
-5.42187467e-02 -5.74852169e-01 3.36893164e-02 4.51286107e-01
-9.06536996e-01 -4.44464356e-01 -7.67483413e-01 3.57537419e-01
1.38140559e-01 -1.37677863e-01 9.95750368e-01 -5.87976158e-01
7.44183660e-02 2.86933392e-01 2.50684489e-02 -4.86894459e-01
-5.80136418e-01 -5.56684025e-02 3.00780147e-01 -1.65884390e-01
3.18768591e-01 -8.66188288e-01 -2.25994110e-01 1.61840409e-01
-7.67944098e-01 2.14946836e-01 8.86837542e-01 7.67270744e-01
3.30840141e-01 -9.06546935e-02 1.41005039e-01 -8.33504677e-01
5.72791338e-01 -4.98643249e-01 -7.14942694e-01 4.46410060e-01
-1.27167493e-01 2.49449641e-01 -4.24463823e-02 -4.27311331e-01
-1.51849997e+00 -6.73533231e-02 -1.23192206e-01 -2.05063552e-01
-3.35915953e-01 8.23897958e-01 -2.00652108e-01 2.13621631e-01
6.67601287e-01 -9.49152652e-03 5.11202551e-02 -6.24336660e-01
3.47959369e-01 7.29626715e-01 3.51549029e-01 -8.06812763e-01
4.73316371e-01 1.17435157e-01 2.95608014e-01 -8.71275306e-01
-9.84900296e-01 -2.93680429e-01 -5.42437375e-01 -2.28205085e-01
8.49205256e-01 -5.83367050e-01 -6.53088093e-01 5.05426586e-01
-1.37978816e+00 -6.31087944e-02 9.42384973e-02 9.02241409e-01
-7.03606367e-01 4.57071424e-01 -2.13788405e-01 -1.44996214e+00
1.87226027e-01 -1.11570263e+00 1.03482771e+00 3.82216990e-01
-5.86714506e-01 -8.82587731e-01 3.55870545e-01 3.05780470e-01
5.13065457e-01 -1.50851309e-01 7.90502310e-01 -8.43653142e-01
-5.39317429e-01 -3.01607519e-01 2.64941100e-02 4.26475555e-02
-2.85483927e-01 4.24494803e-01 -1.31889033e+00 1.12987505e-02
-2.38454901e-02 4.10185754e-02 9.65846181e-01 7.35371470e-01
9.25203502e-01 -1.49279237e-01 -6.19888306e-01 6.64744824e-02
1.12255251e+00 7.45621547e-02 7.33365417e-01 -3.24206054e-01
3.25940818e-01 8.08104157e-01 7.57059604e-02 5.12637556e-01
-6.45052269e-02 8.87631178e-01 -2.18529757e-02 6.37330234e-01
4.01398897e-01 1.10366888e-01 3.69266123e-01 8.79585862e-01
-4.75315750e-01 -1.33523807e-01 -1.01093006e+00 -2.56937575e-02
-1.91519630e+00 -1.31716466e+00 -2.25677878e-01 2.49408650e+00
9.00469422e-01 1.85601309e-01 -1.98753141e-02 -1.12378635e-01
6.95970595e-01 -1.13564022e-01 -3.19749087e-01 -2.05117345e-01
9.13341641e-02 4.91562754e-01 5.53590477e-01 8.51148903e-01
-1.03368521e+00 5.27337551e-01 7.16362762e+00 1.27254343e+00
-3.72497976e-01 3.06404769e-01 3.70505512e-01 -1.97892904e-01
-3.05739611e-01 4.54602271e-01 -1.00467074e+00 6.10366464e-01
1.37638390e+00 2.62077034e-01 3.50580066e-01 3.30354095e-01
1.87034588e-02 -8.43171120e-01 -9.98002648e-01 6.53483033e-01
-2.21169502e-01 -1.09374309e+00 -2.69190788e-01 1.96294203e-01
6.59608841e-01 -8.59770626e-02 -2.20723435e-01 4.37617779e-01
5.96527517e-01 -1.13189530e+00 7.03454792e-01 1.41637599e+00
2.90554672e-01 -6.31648660e-01 7.86699891e-01 5.33950388e-01
-4.21585053e-01 1.49059385e-01 -2.85009384e-01 -9.22277644e-02
7.00931489e-01 1.02041197e+00 -4.23984617e-01 6.06503010e-01
4.46768671e-01 1.32466078e-01 -1.72574297e-01 1.46331549e+00
1.92359909e-02 7.69705534e-01 -1.15424025e+00 -2.96057425e-02
-6.26943856e-02 -6.78013921e-01 6.21652663e-01 1.09986091e+00
5.93411982e-01 1.55395374e-01 -1.89922974e-01 1.08595705e+00
7.06802428e-01 -3.54336172e-01 -2.22260043e-01 -9.29066092e-02
6.19750202e-01 9.56282020e-01 -6.60708070e-01 -4.76395071e-01
-3.34723324e-01 3.82298589e-01 -9.07629877e-02 3.89599800e-01
-6.84584737e-01 8.64994654e-04 3.61124545e-01 -1.94897383e-01
3.12120020e-01 -1.84157580e-01 -4.02888447e-01 -9.87671912e-01
-4.58163410e-01 -6.40547872e-01 1.13945454e-01 -9.41241145e-01
-1.50971949e+00 -2.45796651e-01 9.93390977e-01 -9.13460732e-01
-4.91757780e-01 -7.21233070e-01 -6.78649724e-01 1.32722509e+00
-7.58995414e-01 -9.47519183e-01 4.59694952e-01 -4.82554287e-02
9.09026489e-02 8.03282559e-02 8.66048396e-01 -3.50995511e-02
-5.00006139e-01 5.11150099e-02 5.09215176e-01 -8.48901033e-01
7.80599773e-01 -1.17446017e+00 1.74439885e-02 4.89632130e-01
-4.85080443e-02 7.95144916e-01 1.40266621e+00 -6.05186999e-01
-9.48173285e-01 -2.59954631e-01 8.73700202e-01 -5.21029413e-01
7.67768800e-01 -1.72762692e-01 -1.00894129e+00 5.68733573e-01
1.28593102e-01 -5.67524731e-01 8.94650400e-01 4.11647677e-01
-1.61261991e-01 3.16388965e-01 -9.86796856e-01 6.71214283e-01
5.16249120e-01 -2.66378582e-01 -6.38791680e-01 4.97094870e-01
1.76045727e-02 -1.06951520e-01 -1.04837668e+00 5.85824072e-01
9.65252280e-01 -1.26588678e+00 1.06352675e+00 -5.23033977e-01
1.34884387e-01 -5.09539425e-01 -4.66923118e-01 -1.18535352e+00
-3.21067929e-01 -7.94110417e-01 -2.75677264e-01 1.15146267e+00
4.24734741e-01 -7.33952999e-01 4.12407607e-01 1.01577318e+00
-9.25686657e-02 -5.34173369e-01 -1.18296158e+00 -5.47391832e-01
3.04555655e-01 -7.38848209e-01 4.26741958e-01 6.25102222e-01
-4.13617603e-02 3.55639696e-01 -5.24397612e-01 3.19787145e-01
9.95363057e-01 -2.46613860e-01 6.94771051e-01 -1.59405243e+00
-7.11990356e-01 -6.77841187e-01 -2.54770041e-01 -1.06841433e+00
2.63661472e-03 -6.13451183e-01 3.12463641e-01 -1.26968479e+00
6.71368420e-01 -2.65090048e-01 -8.38298816e-03 1.12785079e-01
-2.57422715e-01 -5.85975051e-02 5.69123775e-02 2.00261831e-01
-3.69983882e-01 4.15604711e-01 8.54950190e-01 1.47356033e-01
6.61171079e-02 2.48937026e-01 -4.29102123e-01 8.89923453e-01
8.72109294e-01 -6.65917456e-01 -1.56907048e-02 2.16306299e-01
4.23961818e-01 3.01972240e-01 5.68080783e-01 -9.76587355e-01
3.55092376e-01 -3.46481353e-01 3.73234600e-01 -7.59730816e-01
7.65228331e-01 -3.68214458e-01 9.52977836e-01 1.02705091e-01
-1.22380212e-01 -5.60410559e-01 6.46052361e-02 6.06011152e-01
1.03593476e-01 -8.79874587e-01 8.15855622e-01 -1.34702697e-01
-1.15894057e-01 -3.21392387e-01 -8.17887366e-01 -5.84179819e-01
4.68789876e-01 -1.22266740e-01 -1.39871120e-01 -4.96402144e-01
-9.12471414e-01 -9.21075717e-02 2.24357471e-01 -3.03240716e-01
2.58879542e-01 -1.03214586e+00 -7.12063491e-01 -6.96498975e-02
-2.62060404e-01 5.37124500e-02 5.02092361e-01 1.37782431e+00
-2.36782551e-01 3.93271625e-01 2.25529641e-01 -6.52987719e-01
-9.09758508e-01 1.43629596e-01 3.97153348e-01 -4.29004878e-01
-9.57492366e-02 9.29119945e-01 1.70027062e-01 -5.82872152e-01
-1.14823073e-01 7.82001317e-02 6.47358410e-03 8.46183021e-03
4.42539632e-01 4.96926457e-01 -1.71616748e-01 -3.60692143e-01
-2.84960508e-01 4.65623289e-01 -6.45499304e-02 -6.79017305e-01
1.10732162e+00 2.10178178e-03 -4.31943923e-01 7.66752839e-01
4.54470009e-01 -1.30091205e-01 -1.32604003e+00 -1.27890453e-01
2.75812209e-01 -2.07048014e-01 1.16428308e-01 -7.15378821e-01
-3.24942395e-02 9.52842951e-01 4.21597183e-01 3.85730624e-01
5.70306301e-01 1.93167314e-01 3.08518782e-02 6.20956421e-01
3.22450995e-01 -8.57710958e-01 -5.52412987e-01 5.40089548e-01
6.38850927e-01 -9.81095910e-01 5.63467205e-01 -2.99393088e-01
-3.97001207e-01 1.03371775e+00 3.73318195e-02 8.45012814e-02
9.12214458e-01 2.48024851e-01 -4.47791278e-01 -2.40900200e-02
-7.15783477e-01 -1.11178547e-01 6.00071967e-01 3.85150015e-01
5.71319759e-01 2.25798786e-01 -2.27419823e-01 6.02451205e-01
-1.78692967e-01 -1.74244687e-01 3.51099283e-01 8.50878239e-01
-6.31116271e-01 -1.09471273e+00 -9.43425655e-01 8.22591186e-01
-3.68297070e-01 -3.05317014e-01 2.36506574e-02 4.52167988e-01
-1.33040383e-01 8.97496939e-01 1.65367424e-01 -2.65385062e-02
-5.24485298e-02 7.34999597e-01 1.01906693e+00 -5.45802236e-01
-7.72179365e-02 5.47528565e-01 2.06370518e-01 -2.58435328e-02
-8.30398381e-01 -1.12239003e+00 -4.12944525e-01 -4.64657038e-01
-7.68531740e-01 5.32889999e-02 7.09375441e-01 1.36191034e+00
4.05100919e-02 3.19601238e-01 -2.16102168e-01 -1.56183016e+00
-1.00967360e+00 -1.52246952e+00 -7.64407575e-01 -2.84916013e-01
-5.86857274e-02 -9.44303513e-01 -5.84250808e-01 2.75194887e-02] | [6.777148723602295, 3.9094014167785645] |
e94d5d7d-f666-4422-aeb1-0918bda8bf2a | deep-emotion-facial-expression-recognition | 1902.01019 | null | http://arxiv.org/abs/1902.01019v1 | http://arxiv.org/pdf/1902.01019v1.pdf | Deep-Emotion: Facial Expression Recognition Using Attentional Convolutional Network | Facial expression recognition has been an active research area over the past
few decades, and it is still challenging due to the high intra-class variation.
Traditional approaches for this problem rely on hand-crafted features such as
SIFT, HOG and LBP, followed by a classifier trained on a database of images or
videos.
Most of these works perform reasonably well on datasets of images captured in
a controlled condition, but fail to perform as good on more challenging
datasets with more image variation and partial faces.
In recent years, several works proposed an end-to-end framework for facial
expression recognition, using deep learning models.
Despite the better performance of these works, there still seems to be a
great room for improvement.
In this work, we propose a deep learning approach based on attentional
convolutional network, which is able to focus on important parts of the face,
and achieves significant improvement over previous models on multiple datasets,
including FER-2013, CK+, FERG, and JAFFE.
We also use a visualization technique which is able to find important face
regions for detecting different emotions, based on the classifier's output.
Through experimental results, we show that different emotions seems to be
sensitive to different parts of the face. | ['Amirali Abdolrashidi', 'Shervin Minaee'] | 2019-02-04 | null | null | null | null | ['image-variation'] | ['computer-vision'] | [-2.75279135e-01 -3.99298161e-01 -8.26903805e-02 -7.04489708e-01
-7.35564977e-02 -4.59505841e-02 5.80023468e-01 -3.27735394e-01
-4.00014609e-01 3.43589008e-01 5.12410738e-02 3.53287935e-01
1.26022607e-01 -5.92551053e-01 -3.93286645e-01 -6.68523192e-01
-2.70128638e-01 -2.37663537e-02 6.32829145e-02 -5.28519511e-01
2.90514588e-01 9.27704215e-01 -1.92645621e+00 6.52365029e-01
1.87423959e-01 1.46884573e+00 -2.83143938e-01 1.45749822e-01
-1.73041046e-01 9.91084874e-01 -8.24119329e-01 -5.36636829e-01
2.53595263e-02 -4.66475010e-01 -6.45596981e-01 1.71048999e-01
7.00009584e-01 -1.82325020e-01 -2.90485118e-02 1.10857642e+00
6.96693838e-01 -1.43784329e-01 5.36463141e-01 -1.58045661e+00
-6.03121102e-01 -1.21459171e-01 -7.45350003e-01 7.19844177e-02
4.24829274e-01 1.94095317e-02 5.27898073e-01 -1.28632557e+00
6.41769767e-01 1.59089565e+00 6.98720098e-01 6.11701488e-01
-9.79471505e-01 -1.09701419e+00 -1.45147949e-01 6.50729835e-01
-1.39077413e+00 -6.04122937e-01 1.04291987e+00 -4.26242143e-01
8.63021970e-01 2.88060997e-02 7.14470267e-01 1.22123694e+00
1.32489309e-01 7.16559827e-01 1.45930529e+00 -5.80440164e-01
1.82714030e-01 2.03239590e-01 -1.53379500e-01 7.59170771e-01
-4.37662184e-01 -4.57187649e-03 -6.57168627e-01 -1.29450187e-01
5.11473000e-01 2.07892492e-01 -3.29426825e-01 -4.40750808e-01
-5.91669559e-01 9.43184853e-01 5.09653986e-01 6.08727574e-01
-5.18469036e-01 -4.70575355e-02 7.20821559e-01 4.67792064e-01
7.17315674e-01 7.59408623e-02 -3.49531204e-01 -1.15898974e-01
-1.19381523e+00 2.50947326e-01 6.54359639e-01 4.47402239e-01
7.95493186e-01 -5.23497500e-02 -3.31946254e-01 9.99826908e-01
3.57754976e-01 1.11713998e-01 4.30733651e-01 -7.63165355e-01
-5.93751185e-02 6.91113472e-01 -1.46318465e-01 -1.47242415e+00
-4.66811270e-01 6.57731220e-02 -8.83652329e-01 8.44253063e-01
4.19991374e-01 -2.40049697e-02 -9.35827434e-01 1.81388247e+00
2.06997126e-01 9.29075778e-02 -1.49968535e-01 1.03568065e+00
9.18018639e-01 5.40415585e-01 1.82644531e-01 -1.49361148e-01
1.36341548e+00 -1.07278597e+00 -9.66909587e-01 3.37982401e-02
3.93979847e-01 -6.56570077e-01 1.00239754e+00 6.51535153e-01
-8.80806863e-01 -7.19435215e-01 -7.98316717e-01 1.59725189e-01
-7.11416483e-01 4.27739978e-01 4.47159111e-01 8.01865458e-01
-1.30336452e+00 4.38897878e-01 -4.24988300e-01 -6.47036076e-01
8.03468168e-01 4.36739206e-01 -8.26761603e-01 -1.75435513e-01
-1.02907455e+00 9.89097595e-01 9.29121673e-02 3.99418086e-01
-8.22320521e-01 -3.08560342e-01 -7.73164451e-01 1.23253606e-01
9.44491550e-02 -9.95347723e-02 9.31890011e-01 -1.63253534e+00
-1.73967922e+00 1.03374398e+00 -1.13948666e-01 -1.60871938e-01
4.64471608e-01 -2.02506617e-01 -5.49166262e-01 2.87885934e-01
-2.80949801e-01 7.83214390e-01 9.93036509e-01 -1.04288185e+00
-2.60306895e-01 -8.19098115e-01 -4.86913510e-02 -2.97400296e-01
-6.94980800e-01 7.87364244e-01 -5.00271320e-01 -5.07775426e-01
-3.60868096e-01 -9.01961565e-01 4.66003120e-02 3.56634527e-01
1.07855931e-01 -5.68004310e-01 1.36003911e+00 -6.22153759e-01
1.05814278e+00 -2.31292415e+00 1.17384940e-02 1.33689061e-01
-9.30863544e-02 6.61032736e-01 -1.52734101e-01 2.50919133e-01
-2.87119687e-01 7.23735169e-02 4.89710718e-02 -3.51487905e-01
-7.20664486e-02 1.22980662e-01 1.95468385e-02 5.93629241e-01
5.06339788e-01 6.64129674e-01 -5.13784230e-01 -4.17290390e-01
2.26513401e-01 8.05538177e-01 -3.55690241e-01 1.86496958e-01
3.51141728e-02 3.59561116e-01 -1.32271558e-01 9.28140998e-01
1.00362587e+00 1.37305692e-01 -6.20936323e-03 -1.67519644e-01
-6.99094555e-04 -5.10399878e-01 -1.02266121e+00 1.54233253e+00
-4.00706798e-01 9.23372686e-01 4.08014148e-01 -1.22793412e+00
1.17829609e+00 4.00311410e-01 3.03385466e-01 -9.82205927e-01
3.49353015e-01 2.03706697e-01 -7.25159422e-02 -8.43495429e-01
8.96657333e-02 -1.96831495e-01 2.79021591e-01 4.69018929e-02
2.96881199e-01 2.63263553e-01 1.54737353e-01 -1.51399389e-01
8.40138614e-01 7.37372190e-02 6.84686899e-02 -1.17213421e-01
8.36418688e-01 -4.30200487e-01 3.78905326e-01 1.48592874e-01
-5.54481626e-01 5.13204336e-01 6.97285652e-01 -7.61129320e-01
-6.02145195e-01 -3.15033227e-01 -2.46331498e-01 1.25874650e+00
-1.12237163e-01 -4.31372583e-01 -7.89456010e-01 -7.98557103e-01
5.18973693e-02 2.03182563e-01 -1.13962722e+00 -1.09389447e-01
-2.07934424e-01 -6.34688377e-01 5.59396327e-01 5.37641287e-01
8.33328426e-01 -1.55257452e+00 -7.65136600e-01 -4.58803885e-02
1.98711008e-01 -9.90779936e-01 1.95773110e-01 8.26491881e-03
-4.89138931e-01 -8.68468225e-01 -9.92128074e-01 -6.90519214e-01
5.05613625e-01 3.60406190e-02 1.03364050e+00 1.63169339e-01
-5.60308337e-01 2.04932287e-01 -4.96004134e-01 -6.21959269e-01
-6.94922656e-02 -4.06217426e-01 -6.97733164e-02 6.14046216e-01
7.27553904e-01 -2.57298559e-01 -4.63074535e-01 5.64640880e-01
-7.97644675e-01 -5.32291114e-01 5.47562659e-01 9.88114893e-01
2.27077886e-01 -1.36346310e-01 4.10587400e-01 -5.17583728e-01
5.62685251e-01 -3.77887338e-01 -2.90617198e-01 2.80104995e-01
-2.12132841e-01 -2.54315466e-01 5.47883570e-01 -3.10111791e-01
-9.52714860e-01 7.86845461e-02 -2.70927161e-01 -8.33982110e-01
-4.37559336e-01 2.61862487e-01 -3.20558608e-01 -5.35767496e-01
5.93333125e-01 -1.28954664e-01 4.32511687e-01 -3.85615796e-01
-9.74310189e-03 8.23453188e-01 8.09888393e-02 -2.27099344e-01
2.10678384e-01 6.00585938e-01 2.84465551e-02 -9.57515776e-01
-5.96231699e-01 -3.85207653e-01 -6.48580313e-01 -5.74471712e-01
7.76486158e-01 -8.26918185e-01 -7.37344086e-01 7.87495196e-01
-1.06645846e+00 -2.03978464e-01 2.16293618e-01 2.78420746e-01
-3.11258554e-01 2.71734506e-01 -4.56938982e-01 -9.43966866e-01
-2.08922282e-01 -1.22889352e+00 1.26225901e+00 3.72681141e-01
-1.15979038e-01 -8.34747434e-01 -6.15369305e-02 1.10908210e-01
7.42515326e-01 5.01860201e-01 5.26615024e-01 -3.46676379e-01
4.11623344e-02 -3.42251122e-01 -3.17880988e-01 5.94381392e-01
8.52256343e-02 4.30240750e-01 -1.29691505e+00 -2.71959871e-01
-1.04295008e-01 -9.27395344e-01 1.04389501e+00 1.84791118e-01
1.60202122e+00 -9.90672037e-02 -2.89742529e-01 5.65254807e-01
1.17946994e+00 1.94299325e-01 1.08263910e+00 3.24872762e-01
2.61979878e-01 9.15434837e-01 8.40682387e-01 4.55913037e-01
8.86166096e-03 1.03457594e+00 7.07064033e-01 -4.23641831e-01
-1.16541451e-02 2.74621785e-01 5.15432239e-01 2.51712054e-02
-4.03667718e-01 7.26487190e-02 -7.34939098e-01 3.38677198e-01
-1.72624493e+00 -1.28865075e+00 1.15903847e-01 1.86859167e+00
5.22568047e-01 -2.68975854e-01 3.66084307e-01 1.85362205e-01
4.71639663e-01 1.51131734e-01 -1.59516126e-01 -8.91992033e-01
-2.82589465e-01 5.41269898e-01 -1.15933105e-01 -5.78162335e-02
-1.12559223e+00 9.64998484e-01 6.16670084e+00 7.96270788e-01
-1.82309067e+00 3.72722605e-03 9.38576996e-01 -6.74441457e-02
4.90875125e-01 -4.83993709e-01 -5.65395296e-01 6.17486179e-01
8.35874856e-01 3.22856665e-01 3.83423343e-02 1.18427742e+00
7.59478584e-02 -1.62420630e-01 -7.09699035e-01 1.40841877e+00
5.24263501e-01 -9.59250927e-01 -2.11537242e-01 -3.61166783e-02
4.63486433e-01 -3.47439051e-01 2.44783178e-01 5.43194652e-01
-3.08362126e-01 -1.42929983e+00 4.98670489e-01 5.22161722e-01
6.41492724e-01 -9.01177943e-01 9.58506584e-01 -1.09149585e-03
-7.70765603e-01 -2.01849386e-01 -5.24698496e-01 -2.01912835e-01
-4.00285929e-01 2.66390830e-01 -6.52897656e-01 2.13543639e-01
1.31110823e+00 7.27668166e-01 -7.72842169e-01 8.77061367e-01
-1.96558937e-01 5.25576591e-01 -1.40412942e-01 -2.42958993e-01
1.51338920e-01 -5.50556295e-02 -8.95867422e-02 1.55295312e+00
3.77286047e-01 -1.36857584e-01 -3.00391465e-02 6.76949501e-01
-3.53307910e-02 4.47375745e-01 -7.86795437e-01 4.54307385e-02
-1.16406783e-01 1.71957946e+00 -6.02005959e-01 -3.37619931e-01
-6.10362887e-01 1.16397905e+00 4.22357917e-01 2.58051723e-01
-8.90659750e-01 -4.90262151e-01 6.88401163e-01 1.94428600e-02
4.48738068e-01 -7.52362460e-02 2.54030675e-01 -9.03900504e-01
9.76028591e-02 -1.03211486e+00 3.19109678e-01 -8.12641978e-01
-1.03524220e+00 7.66239464e-01 -2.79755741e-01 -9.92618263e-01
-2.27193281e-01 -9.65139687e-01 -5.63755572e-01 7.44106352e-01
-1.59017062e+00 -1.10143173e+00 -5.81072211e-01 9.19305503e-01
4.23264623e-01 -4.30756241e-01 9.94499207e-01 4.83882844e-01
-5.47491372e-01 7.16346264e-01 -1.57902926e-01 2.91793197e-01
1.08803594e+00 -8.83423269e-01 -3.22941750e-01 3.65929633e-01
2.77620167e-01 4.25287992e-01 5.27482688e-01 4.03901599e-02
-1.31367147e+00 -8.46886575e-01 7.41223216e-01 -2.73969918e-01
2.80166447e-01 -4.22093421e-01 -9.81905222e-01 3.32639933e-01
5.58287859e-01 5.42948127e-01 7.70713985e-01 1.91497371e-01
-4.69155848e-01 -3.70190710e-01 -1.28713870e+00 2.17828080e-01
8.73352110e-01 -3.54217768e-01 -1.94935068e-01 1.52940959e-01
-2.53199697e-01 -1.15237668e-01 -7.20902562e-01 6.60959899e-01
7.48686731e-01 -1.52702153e+00 6.48736835e-01 -6.82865322e-01
5.76568127e-01 -2.27641940e-01 -1.62096340e-02 -1.35039413e+00
-2.72637606e-01 -2.18498766e-01 1.08234175e-01 1.33461070e+00
1.91992186e-02 -3.06357026e-01 6.60412192e-01 4.23984438e-01
2.45320648e-01 -8.59884679e-01 -9.11306322e-01 -6.88680112e-01
-1.76365614e-01 -1.58372954e-01 4.47657168e-01 1.04328358e+00
-8.14438984e-02 1.98075190e-01 -5.67248523e-01 -2.15487063e-01
1.01453148e-01 2.28559688e-01 1.10671604e+00 -1.16267824e+00
2.05667555e-01 -6.79820836e-01 -1.10632825e+00 -3.33985895e-01
4.90710974e-01 -5.04548073e-01 -2.19307885e-01 -1.18706095e+00
1.39145300e-01 -6.36015832e-02 -4.63607728e-01 9.12861466e-01
3.05251516e-02 7.85656691e-01 2.25942552e-01 -1.66003495e-01
-6.26175702e-01 6.92885101e-01 1.09273696e+00 -2.48718932e-01
4.04554680e-02 -2.47908384e-01 -4.57941473e-01 7.83578694e-01
6.74598753e-01 -1.20018199e-01 4.47040983e-03 -1.84453383e-01
-1.93438858e-01 -3.05746049e-01 5.63920081e-01 -1.11615837e+00
5.60913570e-02 6.43323287e-02 1.01213384e+00 -2.38570735e-01
5.94542980e-01 -9.19633567e-01 -9.19446647e-02 3.11691076e-01
-4.38614003e-02 3.73349898e-02 5.04896402e-01 1.97136506e-01
-7.69993186e-01 1.16236813e-01 1.08439136e+00 -2.22271644e-02
-1.19023180e+00 2.23990992e-01 -1.40240625e-01 -4.53473002e-01
1.29396486e+00 -2.97058374e-01 1.17557876e-01 -4.23847377e-01
-6.76067770e-01 -8.54658037e-02 4.16923314e-01 7.10647941e-01
6.30334616e-01 -1.39721406e+00 -6.48317516e-01 4.16546613e-01
3.49513650e-01 -6.36085272e-01 3.07229817e-01 8.40511441e-01
-3.75562489e-01 1.04542941e-01 -9.56023693e-01 -8.17249835e-01
-1.74463570e+00 6.35371625e-01 4.72361237e-01 8.02901909e-02
-1.56024218e-01 9.62435961e-01 -4.32247110e-02 -1.69415995e-01
3.51332426e-01 -7.52622262e-02 -5.70298493e-01 4.77441758e-01
8.88676405e-01 2.94522848e-02 2.37426698e-01 -1.07498896e+00
-5.94121337e-01 8.72394681e-01 1.28041236e-02 1.97992608e-01
1.38622653e+00 3.25110137e-01 -2.66108871e-01 1.65651739e-01
1.47693741e+00 -4.64399531e-02 -1.14156282e+00 -3.28911692e-02
-1.17381141e-02 -8.83194864e-01 -1.68714058e-02 -8.41629803e-01
-1.63007343e+00 1.29565668e+00 1.08247447e+00 1.36868089e-01
1.60558426e+00 -1.23593017e-01 2.37803474e-01 1.10638507e-01
4.53854680e-01 -1.13405395e+00 4.29840326e-01 4.41506416e-01
1.20151556e+00 -1.54564154e+00 -2.81298578e-01 -1.06523484e-01
-6.64454579e-01 1.57232618e+00 7.95804679e-01 -5.88610098e-02
7.60587454e-01 1.84947520e-01 2.10534468e-01 -4.01013047e-01
-6.16855681e-01 -1.98087886e-01 1.51975095e-01 4.52093482e-01
6.73087597e-01 -2.14179903e-01 -2.14612141e-01 3.24596584e-01
9.16768238e-02 4.18331027e-01 -7.06073567e-02 8.18523526e-01
-3.86299253e-01 -1.05439043e+00 -4.72653687e-01 1.54455945e-01
-7.84042239e-01 4.12696898e-01 -7.53562689e-01 1.03631735e+00
3.65041196e-01 8.01585972e-01 7.50431716e-02 -3.44855994e-01
4.85774368e-01 2.75673300e-01 6.01215303e-01 -2.77743846e-01
-6.26361132e-01 -4.46340106e-02 -1.35023266e-01 -8.92091215e-01
-7.68921375e-01 -5.54038286e-01 -8.87348354e-01 -1.55942917e-01
-1.76145062e-02 -3.82570550e-03 7.18553424e-01 5.79392374e-01
5.41506529e-01 1.33177847e-01 7.45288253e-01 -1.03392220e+00
-2.27694556e-01 -1.20084214e+00 -7.38317192e-01 8.62004638e-01
1.98472083e-01 -1.06444323e+00 -2.38494679e-01 -1.03764877e-01] | [13.563329696655273, 1.8116379976272583] |
f765a076-119c-43ed-aad9-b2a6941e2c2d | geometry-based-adaptive-symbolic | 1403.0820 | null | http://arxiv.org/abs/1403.0820v2 | http://arxiv.org/pdf/1403.0820v2.pdf | Geometry-based Adaptive Symbolic Approximation for Fast Sequence Matching on Manifolds | In this paper, we consider the problem of fast and efficient indexing
techniques for sequences evolving in non-Euclidean spaces. This problem has
several applications in the areas of human activity analysis, where there is a
need to perform fast search, and recognition in very high dimensional spaces.
The problem is made more challenging when representations such as landmarks,
contours, and human skeletons etc. are naturally studied in a non-Euclidean
setting where even simple operations are much more computationally intensive
than their Euclidean counterparts. We propose a geometry and data adaptive
symbolic framework that is shown to enable the deployment of fast and accurate
algorithms for activity recognition, dynamic texture recognition, motif
discovery. Toward this end, we present generalizations of key concepts of
piece-wise aggregation and symbolic approximation for the case of non-Euclidean
manifolds. We show that one can replace expensive geodesic computations with
much faster symbolic computations with little loss of accuracy in activity
recognition and discovery applications. The framework is general enough to work
across both Euclidean and non-Euclidean spaces, depending on appropriate
feature representations without compromising on the ultra-low bandwidth, high
speed and high accuracy. The proposed methods are ideally suited for real-time
systems and low complexity scenarios. | ['Pavan Turaga', 'Rushil Anirudh'] | 2014-03-04 | null | null | null | null | ['dynamic-texture-recognition'] | ['computer-vision'] | [ 4.41721678e-01 -3.97193670e-01 9.41478088e-02 4.35845740e-02
-3.57499659e-01 -6.26590908e-01 5.92767835e-01 4.18407500e-01
-5.43843985e-01 6.86164081e-01 -9.46716145e-02 -2.62550831e-01
-7.43615746e-01 -8.28553140e-01 -3.24631602e-01 -8.14165056e-01
-5.10278940e-01 5.79426050e-01 4.85715508e-01 -2.08391249e-01
4.98903483e-01 1.07873058e+00 -1.81057227e+00 -2.23332748e-01
4.95088965e-01 1.01261914e+00 -3.55356872e-01 1.01428103e+00
7.66442344e-02 2.24125415e-01 -4.19291675e-01 -9.34638306e-02
2.95925438e-01 -2.68633604e-01 -8.41872931e-01 1.45919934e-01
2.31920972e-01 1.98397398e-01 -1.04417607e-01 6.69643223e-01
3.91728282e-01 4.87280369e-01 7.97923267e-01 -1.30742478e+00
-3.21820043e-02 -8.11870992e-02 -5.39328694e-01 3.03766221e-01
5.83203256e-01 -3.70930374e-01 6.50911510e-01 -8.64247322e-01
6.74840987e-01 8.21363032e-01 7.99907088e-01 6.35717716e-03
-1.17689204e+00 -5.69377467e-02 -1.36329696e-01 3.86809528e-01
-1.70911109e+00 -2.42124006e-01 5.20762503e-01 -3.61606151e-01
6.84134364e-01 5.96095979e-01 7.55274355e-01 8.13587248e-01
-1.23611622e-01 4.11151826e-01 6.60123944e-01 -5.29371560e-01
3.90985966e-01 -3.78125131e-01 2.63060629e-01 7.44584441e-01
3.37064773e-01 -3.53683978e-01 -4.54008371e-01 -4.37994748e-01
9.51180100e-01 3.05696189e-01 -2.80933768e-01 -6.92319036e-01
-1.70431280e+00 6.05148792e-01 -2.05224916e-01 7.10984826e-01
-3.49416673e-01 1.27815783e-01 4.48471278e-01 2.91657954e-01
2.69289672e-01 4.16568100e-01 -1.50312841e-01 -7.86003828e-01
-9.51961577e-01 3.36895615e-01 9.64204848e-01 8.90141368e-01
4.19424087e-01 -1.57423258e-01 2.27367759e-01 6.14174664e-01
-1.40940398e-01 1.53810456e-01 3.40716034e-01 -8.67244959e-01
2.16396973e-01 5.03318131e-01 1.56490520e-01 -1.34289646e+00
-6.87120020e-01 -2.45071426e-01 -9.13564324e-01 8.81822258e-02
7.53139436e-01 3.97816479e-01 -1.88481346e-01 1.48759294e+00
6.93741739e-01 5.35298824e-01 -1.92433998e-01 6.10861182e-01
-3.53884280e-01 4.58570033e-01 -3.18439037e-01 -4.42596525e-01
1.38554740e+00 -5.91555655e-01 -3.71047527e-01 7.14380145e-01
8.11964154e-01 -6.46301508e-01 9.67876136e-01 5.98793209e-01
-1.02080297e+00 -2.54435301e-01 -1.23633945e+00 -8.06231573e-02
-6.10105455e-01 -2.32885759e-02 5.82275212e-01 7.46968687e-01
-8.58381867e-01 8.34578931e-01 -1.03185356e+00 -6.88228548e-01
8.64243284e-02 5.05784512e-01 -2.57440597e-01 2.98789263e-01
-7.60728717e-01 6.67124987e-01 2.62593120e-01 1.36497945e-01
-2.18236893e-01 -4.24387038e-01 -6.61668062e-01 -1.03021428e-01
2.96485811e-01 -3.02146405e-01 8.46056700e-01 -7.30393350e-01
-1.29755962e+00 7.13865221e-01 4.20622192e-02 -4.99719918e-01
6.56113386e-01 -7.93928206e-02 -4.46668088e-01 2.91697681e-01
-2.39418060e-01 -3.29462796e-01 8.05314481e-01 -2.46851459e-01
-3.67720664e-01 -6.72807753e-01 2.38103140e-02 1.53150663e-01
-4.42078710e-01 -3.99051867e-02 -8.76971036e-02 -8.74565363e-01
2.19083175e-01 -1.02663469e+00 -9.86269563e-02 3.94824535e-01
4.60029803e-02 -2.47709483e-01 1.07826817e+00 -4.24993157e-01
1.23780620e+00 -2.03200698e+00 3.77226949e-01 6.67388797e-01
8.79766122e-02 1.97551563e-01 3.39831591e-01 7.17151523e-01
2.51952503e-02 -2.45581836e-01 -4.10119087e-01 1.65266752e-01
-7.09135607e-02 3.91047239e-01 -2.20163375e-01 1.02010918e+00
9.20140371e-02 4.39541370e-01 -8.77520025e-01 -6.10282421e-01
2.15731040e-01 4.97203678e-01 -3.71420354e-01 -2.61264324e-01
3.52683067e-02 3.65749091e-01 -6.19866729e-01 6.09892011e-01
2.47161061e-01 -2.12096199e-01 1.51757464e-01 1.33127630e-01
-4.52497214e-01 -9.20070335e-02 -1.82717812e+00 1.81789672e+00
-4.61476266e-01 6.23028457e-01 6.39701933e-02 -1.40937185e+00
1.01727581e+00 1.89136505e-01 8.49151731e-01 -4.35863018e-01
-1.17416412e-01 4.12361860e-01 -2.95927465e-01 -2.87233055e-01
6.94718063e-01 2.89387792e-01 -7.44471177e-02 7.59674966e-01
-3.79265994e-01 3.75207573e-01 4.32390243e-01 -4.69757095e-02
1.27986848e+00 2.03952566e-01 5.06370783e-01 -5.73904872e-01
9.22694325e-01 -1.11777140e-02 2.92286605e-01 2.91337222e-01
1.44282520e-01 3.70561510e-01 4.18883383e-01 -4.40249473e-01
-1.09030926e+00 -1.06285107e+00 -3.25531065e-01 1.10516179e+00
2.95185149e-01 -5.77699244e-01 -7.89463222e-01 -1.60543308e-01
-1.29570112e-01 9.16569754e-02 -4.45554346e-01 -4.67559248e-02
-9.47804689e-01 -6.69205368e-01 7.26942956e-01 3.88705313e-01
3.83193374e-01 -5.04762590e-01 -1.02803123e+00 4.84039783e-01
2.14442581e-01 -8.42675924e-01 -5.65937638e-01 -1.33324355e-01
-1.21469581e+00 -1.21549416e+00 -8.65528286e-01 -6.69575036e-01
6.24553978e-01 3.16513956e-01 6.61556482e-01 2.75690198e-01
-7.50414073e-01 8.11157644e-01 -3.11156750e-01 8.59715864e-02
1.48844589e-02 8.55036750e-02 3.96225870e-01 5.42057872e-01
2.42767632e-01 -8.97440910e-01 -7.18347967e-01 8.00757349e-01
-1.10462141e+00 -3.28902096e-01 1.49781346e-01 6.04511917e-01
7.26319730e-01 1.15726791e-01 2.72998542e-01 -4.54074442e-01
6.35436594e-01 -4.67303157e-01 -7.24626303e-01 3.23881358e-01
-5.93115687e-01 3.35454971e-01 6.84668839e-01 -5.74952483e-01
-5.35891175e-01 1.51220664e-01 2.04855680e-01 -2.02664256e-01
9.47722644e-02 2.22130567e-01 6.08678386e-02 -4.50085670e-01
6.54880941e-01 3.91417086e-01 -5.91505319e-02 -7.03242064e-01
2.75305390e-01 6.21851146e-01 6.09239995e-01 -7.99577534e-01
6.03121459e-01 7.04867542e-01 6.81177199e-01 -1.38004816e+00
2.18095109e-02 -8.00139964e-01 -1.08847547e+00 -9.99872908e-02
6.99896276e-01 -2.49440074e-01 -8.52351367e-01 3.71178478e-01
-9.13805425e-01 2.28978172e-02 -4.52660114e-01 4.57581758e-01
-9.59678233e-01 8.52787912e-01 -2.95279354e-01 -9.30513024e-01
-1.52207002e-01 -7.95910835e-01 1.02420294e+00 -1.52463898e-01
-4.51419502e-01 -1.10961998e+00 4.01957482e-01 3.58602218e-02
2.09833339e-01 6.59283936e-01 8.51153493e-01 -7.18283832e-01
-5.83976626e-01 -3.26732844e-01 4.20203209e-02 -2.37638116e-01
3.53906006e-02 1.46161661e-01 -4.49291348e-01 -1.02243416e-01
2.36643050e-02 1.51874602e-01 2.34969124e-01 -1.17123023e-01
1.20407104e+00 -2.81323612e-01 -3.38274330e-01 5.28848946e-01
1.20214927e+00 1.30147487e-01 4.86835420e-01 2.95490950e-01
3.95692676e-01 6.80540264e-01 8.13624740e-01 6.85207188e-01
2.94066519e-02 1.18963718e+00 1.52932173e-02 2.27012664e-01
2.68346339e-01 1.37275204e-01 1.86717287e-01 1.03437614e+00
-4.97799128e-01 1.32132217e-01 -1.01801908e+00 6.04646266e-01
-2.16746497e+00 -1.24547696e+00 -4.01904404e-01 2.59586620e+00
5.40319324e-01 -1.24351792e-01 6.69238985e-01 8.04638386e-01
7.64599383e-01 -4.74791266e-02 -2.71911263e-01 -5.53377688e-01
-3.38684693e-02 4.39838678e-01 4.45375293e-01 3.32476169e-01
-9.56618726e-01 1.47214726e-01 6.01481295e+00 7.96800852e-01
-8.01757157e-01 2.17830881e-01 8.67707208e-02 -9.98446718e-02
2.15677649e-01 -1.33676544e-01 -5.11877418e-01 4.19005692e-01
1.11731851e+00 -1.35207951e-01 3.83426696e-01 8.44953895e-01
1.61979318e-01 -2.41221055e-01 -1.11487079e+00 1.36367214e+00
-3.19913849e-02 -1.35884523e+00 -1.15061931e-01 4.57189798e-01
4.10458982e-01 -5.27995169e-01 -1.92530140e-01 -3.03222209e-01
-4.15870577e-01 -8.58781934e-01 4.73165363e-01 6.22860253e-01
6.03531778e-01 -8.49771917e-01 1.55846521e-01 3.20882708e-01
-1.49949443e+00 7.98345506e-02 -1.33731216e-01 -1.59527272e-01
1.58811972e-01 3.39085519e-01 -5.52585423e-01 5.25586009e-01
2.55787045e-01 6.12215638e-01 -3.49210709e-01 1.18290114e+00
5.59536278e-01 7.18377680e-02 -7.62100458e-01 -2.70095885e-01
3.03569913e-01 -6.34899378e-01 4.93839055e-01 1.35901523e+00
7.04723060e-01 1.94196403e-01 -4.73656273e-03 3.60440910e-01
3.84064913e-01 4.78294641e-01 -7.64377892e-01 -6.56955540e-02
1.82071194e-01 9.37217712e-01 -1.23208797e+00 -2.36442715e-01
-3.98290247e-01 1.09719813e+00 -2.53056344e-02 5.57449833e-02
-8.51893425e-01 -7.76920259e-01 7.45972753e-01 4.42137837e-01
2.34676316e-01 -1.03579903e+00 -9.99115035e-02 -1.15570331e+00
3.19590986e-01 -6.39474750e-01 6.23008311e-01 -2.52864987e-01
-7.09899247e-01 3.11680555e-01 2.22114533e-01 -1.42713165e+00
-5.24989486e-01 -7.90948272e-01 -4.66511607e-01 4.08563644e-01
-8.90127063e-01 -7.36658514e-01 -1.65469930e-01 1.12838888e+00
3.92777145e-01 7.91792497e-02 9.53590393e-01 4.51868355e-01
-3.52812290e-01 3.28197122e-01 4.03052896e-01 1.07021108e-02
1.21962324e-01 -1.12195790e+00 2.83275485e-01 8.76150608e-01
6.54440403e-01 5.91776192e-01 5.84409654e-01 -4.28377122e-01
-1.77614582e+00 -7.69298851e-01 7.64742911e-01 -3.74329895e-01
7.74653435e-01 -5.34258425e-01 -9.12026763e-01 2.66447067e-01
-6.17895663e-01 1.28836364e-01 7.19932437e-01 -9.12300050e-02
-1.41310036e-01 -3.79996561e-02 -1.10216165e+00 5.89726985e-01
1.49173284e+00 -6.05102599e-01 -3.82273197e-01 5.55417955e-01
1.26367882e-01 -5.88910207e-02 -1.20109069e+00 -3.19278985e-02
6.98817492e-01 -7.75049627e-01 1.18406785e+00 -6.01566911e-01
-5.47313869e-01 -6.18033886e-01 -3.31517547e-01 -4.46660668e-01
-7.45800883e-02 -1.19984496e+00 -4.27936465e-01 6.76154494e-01
5.37331961e-02 -5.87939382e-01 9.46845591e-01 3.89306873e-01
1.99253261e-01 -6.94512129e-01 -1.41001785e+00 -1.11482990e+00
-5.20844042e-01 -5.29015660e-01 5.51272631e-01 8.81916702e-01
3.82622212e-01 -5.49926125e-02 -3.20729554e-01 -6.44568279e-02
7.09265292e-01 2.86334008e-01 7.82259285e-01 -1.48521566e+00
-3.89441073e-01 -3.81713986e-01 -1.16251063e+00 -1.02794719e+00
-2.16488570e-01 -7.14009583e-01 -4.40488130e-01 -8.59022856e-01
-3.61475468e-01 -6.19615436e-01 -8.14926550e-02 -5.61005063e-03
6.60367429e-01 3.76276314e-01 -2.53814787e-01 3.79754871e-01
-7.06419408e-01 4.12745386e-01 7.70241439e-01 2.67979950e-01
-2.96472013e-01 2.55962968e-01 6.84012473e-02 6.52495503e-01
6.70591593e-01 -3.04728389e-01 -4.75534022e-01 -1.09502770e-01
2.74818420e-01 1.87457612e-04 4.39876467e-01 -1.23756158e+00
4.76548523e-01 -6.20128103e-02 4.28547300e-02 -2.91737884e-01
6.17412746e-01 -9.43569064e-01 4.97195184e-01 4.60686028e-01
-9.10149217e-02 5.00405967e-01 -2.77676791e-01 8.90105486e-01
-4.63031493e-02 -3.56438994e-01 6.69938445e-01 7.53354877e-02
-6.20275557e-01 3.42701405e-01 -4.54583615e-01 -5.04266098e-02
1.42189276e+00 -9.26983058e-01 1.85399368e-01 -1.99151695e-01
-1.02461791e+00 -2.57056594e-01 6.57758534e-01 2.51393974e-01
5.01055419e-01 -1.49703515e+00 -2.16786310e-01 2.18643814e-01
1.61306545e-01 -3.64091128e-01 -2.44901888e-03 1.27786493e+00
-9.14569378e-01 4.76760477e-01 -3.41328979e-01 -7.13464022e-01
-1.37873363e+00 5.64795792e-01 1.86836660e-01 -2.01970786e-01
-6.80487156e-01 1.64953053e-01 -3.16380590e-01 1.17721930e-01
3.56471926e-01 -4.95664597e-01 2.03933448e-01 1.90541476e-01
6.67716742e-01 1.09572875e+00 2.71471500e-01 -7.76626468e-01
-3.66971374e-01 1.07168639e+00 3.27796280e-01 -2.60082543e-01
1.19530213e+00 -1.05631068e-01 -5.28334789e-02 5.32282770e-01
1.16363442e+00 -3.18577327e-02 -8.86194468e-01 -1.97859749e-01
5.97212017e-01 -4.27589267e-01 -2.87923008e-01 4.97847907e-02
-5.38941860e-01 9.28316951e-01 5.00314653e-01 4.32416052e-01
1.11359227e+00 -2.46949077e-01 7.03228474e-01 5.86565137e-01
7.81648219e-01 -1.26098728e+00 -1.13151921e-02 3.10920209e-01
7.52983332e-01 -4.07657176e-01 1.69130657e-02 -3.20133656e-01
-4.13057655e-02 1.48593640e+00 4.15246747e-02 -2.67123759e-01
7.61167347e-01 5.46850748e-02 -6.48467124e-01 -2.41822973e-01
-3.64606321e-01 -1.41424105e-01 5.65522574e-02 6.43216074e-01
2.59021282e-01 -1.54980287e-01 -4.67546761e-01 7.80521613e-03
2.22157370e-02 -1.72632098e-01 4.99907225e-01 1.27158558e+00
-4.07325327e-01 -1.24304843e+00 -5.74954271e-01 2.19364524e-01
-5.08526027e-01 4.24654514e-01 -1.78687409e-01 9.36637282e-01
-2.36851703e-02 5.15192091e-01 1.76761881e-01 9.55883507e-03
3.54448736e-01 2.41271332e-01 7.19153047e-01 -3.18912357e-01
-2.83688366e-01 -1.42322794e-01 1.28579140e-01 -8.14706981e-01
-6.06159508e-01 -1.05385482e+00 -1.22040498e+00 -3.97759706e-01
-6.24330677e-02 3.82689118e-01 8.60100329e-01 6.15148246e-01
5.29829085e-01 -1.09147899e-01 4.28655356e-01 -6.39549673e-01
-5.72778046e-01 -4.90205497e-01 -9.30196524e-01 3.27281058e-01
1.54520199e-01 -8.98996115e-01 1.38082709e-02 3.38992029e-02] | [7.357458114624023, 3.56378173828125] |
ab7fd16e-033c-4d27-a951-856fda69c543 | gog-relation-aware-graph-over-graph-network | 2109.08475 | null | https://arxiv.org/abs/2109.08475v3 | https://arxiv.org/pdf/2109.08475v3.pdf | GoG: Relation-aware Graph-over-Graph Network for Visual Dialog | Visual dialog, which aims to hold a meaningful conversation with humans about a given image, is a challenging task that requires models to reason the complex dependencies among visual content, dialog history, and current questions. Graph neural networks are recently applied to model the implicit relations between objects in an image or dialog. However, they neglect the importance of 1) coreference relations among dialog history and dependency relations between words for the question representation; and 2) the representation of the image based on the fully represented question. Therefore, we propose a novel relation-aware graph-over-graph network (GoG) for visual dialog. Specifically, GoG consists of three sequential graphs: 1) H-Graph, which aims to capture coreference relations among dialog history; 2) History-aware Q-Graph, which aims to fully understand the question through capturing dependency relations between words based on coreference resolution on the dialog history; and 3) Question-aware I-Graph, which aims to capture the relations between objects in an image based on fully question representation. As an additional feature representation module, we add GoG to the existing visual dialogue model. Experimental results show that our model outperforms the strong baseline in both generative and discriminative settings by a significant margin. | ['Jie zhou', 'Peng Li', 'Fandong Meng', 'Xiuyi Chen', 'Feilong Chen'] | 2021-09-17 | null | https://aclanthology.org/2021.findings-acl.20 | https://aclanthology.org/2021.findings-acl.20.pdf | findings-acl-2021-8 | ['implicit-relations'] | ['natural-language-processing'] | [-8.33639503e-02 4.20103878e-01 -6.10632487e-02 -5.55500567e-01
-2.94717133e-01 -5.00710309e-01 8.85474384e-01 1.49864525e-01
-1.20707102e-01 2.78790146e-01 8.13388824e-01 -2.00133294e-01
2.08209634e-01 -6.14379585e-01 -5.50020218e-01 -3.54251415e-01
3.13081175e-01 7.10250497e-01 5.19078374e-01 -5.64298153e-01
-5.37473559e-02 2.48042122e-01 -1.06157839e+00 4.88984883e-01
5.14589787e-01 9.63533700e-01 4.87751275e-01 5.28358161e-01
-4.59156007e-01 1.53031516e+00 -5.58108807e-01 -4.89283383e-01
-3.31924379e-01 -9.77130234e-01 -1.13822579e+00 5.93525708e-01
2.73338735e-01 -6.34003520e-01 -8.18072855e-01 1.09284699e+00
1.01158246e-01 4.29653138e-01 5.37283301e-01 -1.44736135e+00
-8.91659379e-01 5.92873573e-01 -3.17176789e-01 1.49316818e-01
5.68482757e-01 4.52827185e-01 1.28902900e+00 -7.18787909e-01
8.62644970e-01 1.92885423e+00 2.58918735e-04 5.02525866e-01
-9.31697607e-01 -3.48551422e-01 6.46289229e-01 7.52187967e-01
-1.10776722e+00 -2.48641461e-01 1.07103276e+00 -3.15061271e-01
1.00831783e+00 2.77375311e-01 8.48217905e-01 1.11008704e+00
1.37366012e-01 8.95891428e-01 8.17411602e-01 -2.18873680e-01
-1.14337891e-01 1.01809196e-01 4.69788641e-01 9.77630377e-01
-4.49453562e-01 -1.06209569e-01 -4.59104747e-01 2.48722062e-02
8.97748947e-01 -5.11940978e-02 -4.33903784e-01 -4.76241261e-01
-8.34119201e-01 1.20774961e+00 1.00471520e+00 2.92218119e-01
-1.80346504e-01 -9.45785432e-04 1.39687076e-01 7.76659921e-02
-4.81651537e-02 1.75634369e-01 9.61189866e-02 3.92533064e-01
-2.16546759e-01 7.07317218e-02 9.53485727e-01 1.18325651e+00
1.03489196e+00 -1.40882462e-01 -7.38591254e-01 6.45526171e-01
5.78356445e-01 3.74541372e-01 2.25242764e-01 -9.92025971e-01
3.59126091e-01 1.28591740e+00 -1.47023186e-01 -1.10266674e+00
-4.66197252e-01 1.40862510e-01 -7.69138932e-01 -2.65563369e-01
3.17465752e-01 1.32678851e-01 -1.05356431e+00 1.87713802e+00
5.21140695e-01 -2.08922789e-01 2.86217313e-02 1.08938491e+00
1.77274621e+00 8.47668290e-01 1.71261400e-01 -3.06873053e-01
1.84515333e+00 -1.26551354e+00 -1.13214231e+00 -7.02632010e-01
1.43136024e-01 -5.77685595e-01 1.14055872e+00 -3.28556240e-01
-8.90684783e-01 -7.37362623e-01 -6.93099618e-01 -6.93457425e-01
-1.91366777e-01 -1.25303298e-01 4.72149044e-01 -4.91651893e-02
-1.03033376e+00 -2.83491015e-01 -2.44005948e-01 -6.82850897e-01
3.51358727e-02 1.36996984e-01 -3.19122374e-01 -2.62323469e-01
-1.34528327e+00 9.63580847e-01 5.23252070e-01 2.05206200e-01
-9.98671114e-01 -1.94326621e-02 -1.31843555e+00 3.68031144e-01
9.86561000e-01 -7.92153239e-01 1.20625937e+00 -9.01265681e-01
-1.42756522e+00 8.53084445e-01 -3.37781757e-01 -2.04550758e-01
1.54751226e-01 -3.18439491e-02 -1.43683955e-01 6.01811171e-01
-1.93443432e-01 8.26842308e-01 7.01907337e-01 -1.57412839e+00
-3.89341295e-01 -5.64450920e-01 6.45751774e-01 6.14381194e-01
1.71644613e-01 -2.03048393e-01 -1.18516803e+00 -2.47404441e-01
-3.34045850e-02 -1.08116579e+00 -3.64247151e-02 -1.12056844e-01
-5.95325172e-01 -6.00265682e-01 1.06967640e+00 -8.96938562e-01
9.41818535e-01 -1.99686611e+00 3.76871616e-01 -1.66840121e-01
3.79051536e-01 1.10879585e-01 -3.39868724e-01 5.91038764e-01
2.52403647e-01 -3.53424549e-01 1.08713232e-01 -1.34460762e-01
-4.80782561e-04 6.34119630e-01 -4.60279852e-01 2.00441837e-01
1.83689982e-01 1.45399547e+00 -7.87232399e-01 -8.59058082e-01
3.60737741e-01 3.23117137e-01 -2.47924373e-01 9.39273357e-01
-6.59365475e-01 5.36145806e-01 -5.41043341e-01 2.35736892e-01
4.60759372e-01 -6.98083699e-01 6.44443989e-01 -7.42701054e-01
2.65309870e-01 2.51046836e-01 -6.56029284e-01 1.84364676e+00
-2.53791064e-01 4.60672170e-01 1.90578148e-01 -7.63137639e-01
1.05795968e+00 2.32282028e-01 -5.13248704e-03 -9.18327034e-01
3.86713415e-01 -6.28255725e-01 7.18758330e-02 -7.48358369e-01
4.03393805e-01 -4.94839400e-02 -2.28069171e-01 3.02978426e-01
4.39700931e-01 -2.49116749e-01 2.09040150e-01 8.90092969e-01
6.06275797e-01 1.98815158e-03 4.22789007e-01 2.28369664e-02
8.20296764e-01 -1.66972786e-01 4.84219462e-01 6.95444763e-01
-3.60748410e-01 4.54225034e-01 8.25545013e-01 -2.75504112e-01
-4.68929917e-01 -1.00533855e+00 6.31038070e-01 1.13893712e+00
8.88098478e-01 -4.44819063e-01 -3.50040764e-01 -8.83292615e-01
-1.94092944e-01 8.64607632e-01 -8.48744094e-01 -2.34282225e-01
-6.76466763e-01 -8.28597397e-02 -6.82030469e-02 5.67123711e-01
7.92433321e-01 -1.47695374e+00 -3.96190017e-01 -1.73419476e-01
-7.17047393e-01 -1.56067002e+00 -9.24208701e-01 -1.89984336e-01
-4.24212664e-01 -1.37948084e+00 -3.47477049e-01 -9.75772500e-01
4.57866848e-01 4.75080699e-01 1.26333809e+00 3.85458946e-01
-9.57390815e-02 9.22704220e-01 -3.33836049e-01 -1.01334095e-01
-4.10911351e-01 -3.03162038e-01 -7.15769231e-01 1.83521658e-01
3.67444098e-01 -1.78686857e-01 -6.85414255e-01 4.57548916e-01
-8.35374951e-01 4.07352269e-01 5.87773621e-01 8.00100863e-01
4.20370013e-01 -4.88216966e-01 2.66796201e-01 -9.25197423e-01
6.08146727e-01 -3.28736931e-01 -4.12750214e-01 7.83927262e-01
-1.84440374e-01 1.78133950e-01 2.28401527e-01 -4.26398098e-01
-1.54327798e+00 5.61417043e-02 7.96372816e-03 -5.24208307e-01
6.45090500e-03 2.99028695e-01 -6.01890326e-01 2.48052359e-01
2.11813435e-01 3.53545666e-01 1.03086479e-01 -1.30765676e-01
9.22992229e-01 2.17805058e-01 8.23912740e-01 -3.87325853e-01
4.66469467e-01 5.27249873e-01 -7.34597817e-02 -8.70910406e-01
-1.18969572e+00 -7.47728229e-01 -5.80612242e-01 -4.71103579e-01
1.45350480e+00 -8.02929938e-01 -1.10570145e+00 8.54080692e-02
-1.41523004e+00 -3.79421622e-01 -2.25945413e-01 2.99348354e-01
-5.59218526e-01 6.46507323e-01 -6.57413661e-01 -7.77518988e-01
-2.92399079e-01 -1.19260681e+00 9.92787421e-01 5.47928870e-01
-4.45136093e-02 -1.13899660e+00 -7.06434622e-02 8.25030684e-01
-1.99401602e-02 2.48168752e-01 1.15289700e+00 -6.50380611e-01
-7.72891521e-01 3.16747487e-01 -7.32880116e-01 9.14447531e-02
2.14319810e-01 -4.92212981e-01 -7.63308227e-01 -2.10431322e-01
2.63938308e-01 -6.71076059e-01 7.82310367e-01 1.57237619e-01
6.48413479e-01 -5.10218501e-01 -3.54979068e-01 2.81463414e-01
9.98551428e-01 2.91767299e-01 5.40749609e-01 -2.15334907e-01
9.74743307e-01 1.00677013e+00 4.82035816e-01 9.97927636e-02
8.90184224e-01 7.30628490e-01 7.28857338e-01 -2.41598457e-01
-6.24311805e-01 -5.78483701e-01 2.14904308e-01 8.72943819e-01
1.77636504e-01 -4.97203916e-01 -7.63681531e-01 4.77879465e-01
-2.25721169e+00 -7.89332747e-01 -3.63662660e-01 1.60742617e+00
5.78578591e-01 -1.27438769e-01 4.43307422e-02 -6.96153402e-01
8.81627440e-01 5.16432822e-01 -7.28364408e-01 -7.12872818e-02
-1.99439391e-01 -3.48447591e-01 -2.95871109e-01 7.46164441e-01
-9.15841281e-01 1.20736337e+00 5.07377577e+00 3.80291194e-01
-7.21256137e-01 6.44640848e-02 3.67928207e-01 3.68782371e-01
-3.40651006e-01 3.44744503e-01 -6.31709337e-01 -9.17853788e-02
2.48637095e-01 -1.74712613e-01 2.26671666e-01 5.61730087e-01
-1.59087434e-01 -2.50151277e-01 -1.05716538e+00 1.01615143e+00
6.03313744e-01 -1.21078384e+00 5.57629943e-01 2.49778363e-03
3.62842530e-01 -4.67459619e-01 -2.29186490e-01 5.79516411e-01
6.67284787e-01 -1.03953433e+00 5.81508756e-01 5.32153606e-01
4.87390310e-01 -5.87562025e-01 6.50791645e-01 4.01515603e-01
-1.56927931e+00 -5.53400367e-02 -3.17794353e-01 2.38263682e-01
4.63582397e-01 -1.40741646e-01 -9.78149652e-01 8.31484377e-01
5.47184825e-01 5.59780717e-01 -6.79470122e-01 3.90626431e-01
-7.77655303e-01 2.27700666e-01 1.85789242e-01 -3.62642333e-02
4.97292578e-01 -2.71641612e-01 5.27873337e-01 9.99081671e-01
-1.48740977e-01 6.41786575e-01 5.39790511e-01 1.03923440e+00
-1.07255533e-01 -5.34602590e-02 -4.52585578e-01 -1.46958623e-02
1.37241319e-01 1.38590539e+00 -5.65530598e-01 -3.52773458e-01
-6.14576399e-01 1.00672996e+00 5.61048806e-01 6.46391690e-01
-6.25470698e-01 -3.70965488e-02 3.41232359e-01 -3.36453505e-02
5.03881395e-01 -2.79913127e-01 6.10078037e-01 -1.03689444e+00
-3.59845966e-01 -8.75565350e-01 1.01637197e+00 -1.31017017e+00
-1.22364557e+00 9.17223752e-01 1.52872249e-01 -6.70079589e-01
-4.72106785e-01 -3.46377343e-01 -5.86478829e-01 8.64448249e-01
-1.46594107e+00 -1.71581399e+00 -8.01572740e-01 1.17103851e+00
7.90027022e-01 1.44138902e-01 5.60840547e-01 -4.27591443e-01
-3.51027340e-01 1.08869679e-01 -7.79142916e-01 4.56537932e-01
6.61416531e-01 -1.09760797e+00 1.35348812e-01 6.69093966e-01
4.62999254e-01 4.42789704e-01 5.03875315e-01 -8.29304993e-01
-1.44003558e+00 -1.02013600e+00 8.38554680e-01 -3.89738023e-01
5.61679363e-01 -5.78575194e-01 -1.30233467e+00 8.78388464e-01
8.06766391e-01 -2.01761469e-01 4.99419183e-01 7.37537220e-02
-6.15502954e-01 1.10598721e-01 -6.90880418e-01 5.32178521e-01
9.83201921e-01 -8.16982031e-01 -1.16939569e+00 3.50385398e-01
9.35803890e-01 -5.37972569e-01 -4.64654148e-01 2.73047894e-01
2.18475148e-01 -1.02628553e+00 9.63985920e-01 -5.93547165e-01
3.56512189e-01 -9.31731164e-02 -9.52084735e-02 -9.36520875e-01
-5.48074663e-01 -5.06645441e-01 -1.91849276e-01 1.37332022e+00
-3.91592681e-02 -1.19698904e-01 4.08901960e-01 4.67936009e-01
8.19732174e-02 -3.58827710e-01 -7.29883254e-01 -3.12299609e-01
-3.29109401e-01 9.28909555e-02 4.49455351e-01 9.23656821e-01
-2.30577774e-03 1.33011150e+00 -6.31574154e-01 2.14990973e-01
3.08750868e-01 6.92374468e-01 9.91516709e-01 -1.11949670e+00
-3.49515051e-01 -2.02621862e-01 -1.31696522e-01 -1.55984485e+00
4.48987484e-01 -7.22921610e-01 1.41974971e-01 -2.19271970e+00
5.69226861e-01 2.80934930e-01 2.12073147e-01 3.41263860e-01
-5.40316820e-01 -1.84614122e-01 5.83731115e-01 3.15944999e-01
-9.78358030e-01 9.74501908e-01 1.79237211e+00 -3.95327896e-01
-1.52645648e-01 -4.15989548e-01 -6.37667358e-01 7.70834446e-01
2.80446112e-01 -5.06690182e-02 -8.40431333e-01 -2.61573046e-01
-9.30264816e-02 7.62705088e-01 6.07921183e-01 -4.18968767e-01
5.13015807e-01 -3.02730143e-01 2.08383515e-01 -7.25965977e-01
7.76733577e-01 -7.34063268e-01 -8.68205577e-02 3.44497055e-01
-3.96763206e-01 4.42622937e-02 6.94464594e-02 9.61053908e-01
-4.23682988e-01 2.00236905e-02 5.54955781e-01 -4.29618329e-01
-1.16734111e+00 4.18170482e-01 -1.77168339e-01 3.77573580e-01
8.43940079e-01 1.52014509e-01 -7.17733264e-01 -1.12001896e+00
-9.58380282e-01 7.29791701e-01 2.81693161e-01 7.18518615e-01
7.85197794e-01 -1.28178072e+00 -5.29970586e-01 -1.79285452e-01
3.86993349e-01 -1.67601220e-02 6.98176026e-01 5.41156650e-01
-1.14036247e-01 5.13361454e-01 -1.29584163e-01 -7.12109566e-01
-1.59472513e+00 8.77870977e-01 2.52567351e-01 -4.70753133e-01
-6.70858443e-01 8.71528029e-01 1.31220639e+00 -2.15378582e-01
3.18794072e-01 -2.28495628e-01 -7.83575058e-01 1.05522528e-01
4.02325928e-01 -2.83090502e-01 -4.86746699e-01 -1.19817626e+00
-3.70430470e-01 4.85992163e-01 -1.42042324e-01 -1.82340816e-01
7.90982902e-01 -4.96276289e-01 -1.54288188e-01 6.20708942e-01
9.22730207e-01 -4.10974294e-01 -1.46454668e+00 -6.63858831e-01
-4.06742424e-01 -1.90918803e-01 -1.67339996e-01 -7.86556602e-01
-1.04710126e+00 1.03688622e+00 2.46075556e-01 2.66470671e-01
1.14220035e+00 6.34785056e-01 6.00172520e-01 2.58992583e-01
7.01534888e-03 -6.10730410e-01 8.55465889e-01 6.36146247e-01
1.49355960e+00 -1.26076913e+00 9.01709870e-02 -6.70832992e-01
-1.17046154e+00 1.09339881e+00 1.03585112e+00 2.32364044e-01
3.64031225e-01 -5.11332393e-01 2.92517811e-01 -6.35751605e-01
-8.80292356e-01 -6.56633735e-01 7.50914335e-01 5.71365416e-01
6.83105811e-02 -2.01183364e-01 -3.87067255e-03 5.10883629e-01
4.61478494e-02 -4.42976117e-01 2.18514606e-01 5.75948298e-01
-3.92300874e-01 -7.56331801e-01 -1.51160821e-01 -3.21713015e-02
2.37451509e-01 -5.32220192e-02 -1.07760191e+00 1.07970273e+00
-5.74665926e-02 1.34274185e+00 2.26373896e-01 -2.58179605e-01
3.96031916e-01 6.88869655e-02 5.16071856e-01 -6.37293696e-01
-6.74340546e-01 1.55148819e-01 1.79514349e-01 -7.08174765e-01
-5.33408403e-01 -1.66824818e-01 -1.31615663e+00 8.62221196e-02
-3.21640044e-01 3.10269386e-01 2.23872602e-01 1.10928333e+00
2.87541956e-01 5.78195095e-01 3.11389774e-01 -3.53351891e-01
-1.60188645e-01 -1.06191289e+00 -4.90418196e-01 8.39447021e-01
1.85002565e-01 -6.76332593e-01 -1.66435793e-01 9.87122953e-02] | [10.80474853515625, 1.6134154796600342] |
02e18e13-f90d-4abe-9a5d-8441d633d65f | segment-based-fusion-of-multi-sensor-multi | 2211.15938 | null | https://arxiv.org/abs/2211.15938v1 | https://arxiv.org/pdf/2211.15938v1.pdf | Segment-based fusion of multi-sensor multi-scale satellite soil moisture retrievals | Synergetic use of sensors for soil moisture retrieval is attracting considerable interest due to the different advantages of different sensors. Active, passive, and optic data integration could be a comprehensive solution for exploiting the advantages of different sensors aimed at preparing soil moisture maps. Typically, pixel-based methods are used for multi-sensor fusion. Since, different applications need different scales of soil moisture maps, pixel-based approaches are limited for this purpose. Object-based image analysis employing an image object instead of a pixel could help us to meet this need. This paper proposes a segment-based image fusion framework to evaluate the possibility of preparing a multi-scale soil moisture map through integrated Sentinel-1, Sentinel-2, and Soil Moisture Active Passive (SMAP) data. The results confirmed that the proposed methodology was able to improve soil moisture estimation in different scales up to 20% better compared to pixel-based fusion approach. | ['Davood Akbarid', 'Saeid Niazmardi', 'Iman Khosravi', 'Hossein Bagheri', 'Reza Attarzadeh'] | 2022-11-29 | null | null | null | null | ['soil-moisture-estimation', 'data-integration'] | ['computer-vision', 'knowledge-base'] | [ 4.66917306e-01 -3.19378197e-01 -1.24808429e-02 -3.24605674e-01
-7.74281502e-01 -4.17838961e-01 6.60449982e-01 9.06355619e-01
-4.76510435e-01 9.55855668e-01 -2.07778722e-01 -4.47597295e-01
-4.62965995e-01 -1.50089657e+00 -2.76121199e-01 -1.11486828e+00
9.10737291e-02 1.32158035e-02 4.27807748e-01 -6.36336446e-01
1.76687866e-01 7.33070135e-01 -1.88571835e+00 1.88561548e-02
1.13359880e+00 8.24111879e-01 1.03796661e+00 6.15720630e-01
-2.63679415e-01 -2.76309729e-01 -7.29573891e-02 3.86691600e-01
3.38837802e-01 -1.41627982e-01 -2.23518729e-01 5.94831444e-02
2.00452998e-01 -5.29550910e-01 7.37826824e-01 1.11546683e+00
4.66211915e-01 -2.71451771e-01 5.32500446e-01 -7.08019793e-01
3.32389295e-01 3.44169378e-01 -9.13476288e-01 -1.75619826e-01
3.38393658e-01 -2.66281217e-02 5.69189370e-01 -5.98038256e-01
2.05888480e-01 1.02690125e+00 4.74095345e-01 -2.99271524e-01
-1.31841826e+00 -2.68173814e-01 -4.48539406e-02 1.59499973e-01
-1.12907946e+00 -3.15035850e-01 6.39333963e-01 -4.00873989e-01
6.34035230e-01 5.10082901e-01 1.07832015e+00 3.35589588e-01
2.98568100e-01 4.27999467e-01 1.56874835e+00 -5.47569215e-01
2.56738991e-01 2.23743781e-01 3.07199538e-01 1.01480357e-01
1.07851565e+00 -1.10172443e-01 -2.64900118e-01 -1.45614818e-01
2.76174992e-01 1.20719835e-01 -1.60881042e-01 -4.83147562e-01
-9.18436468e-01 6.66422427e-01 6.02532625e-01 6.26314461e-01
-9.53726828e-01 -3.63899142e-01 9.75278318e-02 5.50301000e-02
6.07203126e-01 2.64728218e-01 -2.67316848e-01 4.43039358e-01
-1.32467639e+00 4.01728690e-01 2.32628658e-01 2.38428727e-01
1.09326720e+00 4.50826883e-02 4.57285374e-01 4.47572470e-01
9.13462460e-01 1.45150220e+00 1.28594697e-01 -6.68448508e-01
2.16980845e-01 7.78378904e-01 4.10695702e-01 -1.18439806e+00
-5.25881350e-01 2.98963562e-02 -7.42318094e-01 6.87533438e-01
2.26221070e-01 -9.90653336e-02 -7.38567293e-01 1.05227256e+00
4.78186458e-01 -5.49444318e-01 3.90791684e-01 7.15293109e-01
6.34663761e-01 8.95876527e-01 4.86418664e-01 -2.81901300e-01
1.47163069e+00 -5.47077730e-02 -7.25483000e-01 -3.17234874e-01
5.00904024e-01 -9.09437835e-01 1.83008641e-01 6.58529475e-02
-7.88188219e-01 -4.98785168e-01 -1.36820233e+00 8.91597390e-01
-9.70361054e-01 8.20981041e-02 4.73662227e-01 7.82505751e-01
-9.76097107e-01 2.44525567e-01 -9.39032137e-01 -8.87569070e-01
-3.98544967e-02 5.94752505e-02 -5.48107207e-01 1.20238528e-01
-1.08448637e+00 1.06495321e+00 8.45138371e-01 4.74551350e-01
-8.09288621e-02 -4.03928220e-01 -7.59623945e-01 -1.96620729e-02
-4.22898643e-02 -2.24413484e-01 4.14363772e-01 -7.77198553e-01
-1.08490884e+00 8.26238334e-01 -2.48293906e-01 -5.48451543e-01
3.30385327e-01 -1.48887739e-01 -3.94353300e-01 4.26252693e-01
3.83590192e-01 4.17465001e-01 4.64608163e-01 -1.42424428e+00
-9.09319222e-01 -7.10442662e-01 -1.51168838e-01 3.37465227e-01
-1.94195375e-01 -1.36597782e-01 5.37109554e-01 4.45612147e-02
8.69008601e-01 -5.24778366e-01 -2.84090638e-01 -1.52405396e-01
1.44683897e-01 5.29232383e-01 1.01240921e+00 -6.96895838e-01
7.29122043e-01 -1.86986935e+00 -3.93612653e-01 3.95752490e-01
-1.70163840e-01 7.57102787e-01 -4.95002605e-02 5.66532254e-01
5.98256625e-02 8.60680565e-02 -4.14412826e-01 7.29516968e-02
-3.97363991e-01 1.80638447e-01 6.28839135e-02 3.90386432e-01
1.75447360e-01 3.94236296e-01 -7.31401801e-01 -5.84546626e-01
8.84095073e-01 4.44768310e-01 2.53872156e-01 7.93273095e-03
-8.76421258e-02 2.89523482e-01 -6.29603863e-01 8.21087122e-01
1.68866992e+00 4.85309303e-01 5.34313262e-01 -4.94092762e-01
-7.90957272e-01 -3.72876674e-01 -1.44281209e+00 1.13466561e+00
-3.68557721e-01 3.64066362e-01 5.74550688e-01 -1.28365052e+00
1.17734087e+00 5.73286295e-01 5.64219892e-01 -9.29338396e-01
-2.12450475e-01 7.07273304e-01 -2.51190722e-01 -5.97495794e-01
9.66818690e-01 -1.33086607e-01 4.63272721e-01 -8.32026675e-02
-2.51924038e-01 -3.01024854e-01 2.65322506e-01 -2.20459044e-01
4.06426162e-01 5.92413127e-01 4.04736817e-01 -5.38770497e-01
6.75502539e-01 1.07765764e-01 2.49150500e-01 6.57125652e-01
-2.91236371e-01 2.99074292e-01 -6.45248219e-02 -7.36041814e-02
-9.47860777e-01 -9.35314953e-01 -5.81571877e-01 5.40652037e-01
2.71220654e-01 2.27551490e-01 -3.47224712e-01 7.34213963e-02
2.66266644e-01 3.31737727e-01 -3.00749570e-01 5.61653435e-01
-9.44733471e-02 -1.36596239e+00 1.03945293e-01 3.68807502e-02
1.15709150e+00 -6.59661055e-01 -9.83209610e-01 5.41750789e-01
-1.84871867e-01 -7.20162213e-01 1.15839934e+00 2.48619169e-01
-1.41095436e+00 -9.35428321e-01 -9.52907324e-01 -7.07287863e-02
5.03465533e-01 9.72750008e-01 6.63375497e-01 -1.93807334e-01
-1.54613191e-02 2.82090485e-01 -6.40498817e-01 -5.28062344e-01
-3.76737028e-01 3.44169736e-02 -5.50676286e-01 1.68952160e-02
3.93991530e-01 -3.89761627e-01 -6.59371376e-01 1.61652476e-01
-1.16394782e+00 1.84869915e-01 5.67628860e-01 1.83623686e-01
4.67253119e-01 -6.34202883e-02 5.25874496e-01 -5.53873003e-01
2.86684215e-01 -7.56148875e-01 -9.47444201e-01 5.15742540e-01
-5.58666945e-01 -2.07012266e-01 -1.45795658e-01 3.98589045e-01
-1.31922984e+00 1.83091179e-01 1.80229254e-03 8.83644342e-01
-5.86935043e-01 1.05306661e+00 -2.99194038e-01 -2.44055152e-01
6.53381705e-01 7.33350217e-02 2.19736353e-01 -5.07374585e-01
-1.14649847e-01 8.17732513e-01 2.88921773e-01 -2.93992579e-01
6.10446572e-01 5.49715400e-01 4.03954297e-01 -1.17018533e+00
-2.48388976e-01 -8.78343940e-01 -7.14383781e-01 -5.79474211e-01
8.90369475e-01 -1.03224778e+00 -3.56783777e-01 9.03696418e-01
-9.82790172e-01 2.42795169e-01 2.80215919e-01 8.00326943e-01
-1.27793521e-01 4.53398019e-01 3.23338896e-01 -1.30288506e+00
-4.68696415e-01 -1.06390512e+00 8.78432810e-01 5.61102211e-01
2.45920658e-01 -1.02812469e+00 2.53186584e-01 2.22957149e-01
9.55451071e-01 6.76773667e-01 2.61939406e-01 -6.00637589e-03
-4.39504951e-01 -5.50782382e-01 -6.22446835e-01 1.79683685e-01
6.02389574e-01 3.46192986e-01 -1.02208769e+00 -2.49858439e-01
-7.62904435e-02 9.19522941e-02 1.02769995e+00 7.01466262e-01
1.24143653e-01 3.75919729e-01 -3.51778477e-01 1.44723192e-01
2.14255023e+00 1.89466387e-01 7.21472859e-01 7.78860211e-01
1.04350604e-01 6.66384995e-01 1.12175286e+00 6.11964464e-01
1.63683206e-01 4.21349883e-01 8.39454710e-01 -4.63642567e-01
9.82601121e-02 3.98823321e-01 1.66182399e-01 2.95595974e-01
-4.24439609e-01 -2.00388014e-01 -1.00811946e+00 6.32663369e-01
-2.01234794e+00 -1.28993547e+00 -6.71481311e-01 2.38283038e+00
2.77351499e-01 -6.81032687e-02 -1.56297565e-01 5.81037819e-01
8.44274342e-01 4.98704970e-01 -7.18109980e-02 -1.97182924e-01
-6.93342090e-01 3.66568506e-01 1.10107744e+00 5.22779465e-01
-1.25057435e+00 4.49195534e-01 5.43522596e+00 3.28025639e-01
-1.20451403e+00 -1.02946334e-01 -9.74320918e-02 5.60427904e-01
-5.40648937e-01 3.85341614e-01 -8.56882870e-01 2.67462730e-01
9.52514291e-01 7.75311934e-03 -3.16650301e-01 3.08471829e-01
7.40222454e-01 -1.41211605e+00 -1.19621260e-02 3.02303910e-01
-4.03748095e-01 -1.02642548e+00 -1.37567118e-01 1.53728381e-01
7.50817418e-01 4.36428301e-02 -3.50144982e-01 -5.43359220e-01
-1.38415378e-02 -2.01536402e-01 3.25449646e-01 9.83154655e-01
8.32512230e-02 -3.40696096e-01 1.14550340e+00 8.80488828e-02
-1.52188861e+00 3.11785847e-01 -3.83404613e-01 -1.53985098e-01
2.45076373e-01 8.44296634e-01 -4.98335987e-01 1.22279060e+00
6.67940676e-01 6.96331978e-01 -7.31650829e-01 1.38885331e+00
6.61828220e-02 4.17047948e-01 -8.79660070e-01 1.32673398e-01
3.02844286e-01 -4.60035264e-01 4.43579316e-01 9.61735904e-01
8.15964937e-01 -1.79226577e-01 -1.65370703e-01 5.22875011e-01
9.18087482e-01 4.09372479e-01 -7.42716134e-01 1.71209037e-01
5.25843799e-01 1.40079987e+00 -7.59737968e-01 -3.92162442e-01
-3.30074847e-01 4.11202818e-01 -6.29160225e-01 2.86150247e-01
-3.19225341e-01 -4.12211835e-01 2.77109087e-01 3.51306498e-01
2.96583056e-01 -5.28212845e-01 -3.58054608e-01 -6.47978485e-01
-6.71906546e-02 -4.06419516e-01 -2.72540096e-02 -9.70053613e-01
-4.29277748e-01 -1.36335984e-01 4.31867659e-01 -1.22843301e+00
1.73437968e-02 -5.78765213e-01 -5.71937740e-01 1.39775085e+00
-1.74656057e+00 -1.29624867e+00 -9.21995819e-01 5.99659756e-02
-8.17616954e-02 -1.27433941e-01 1.15332103e+00 3.96806523e-02
-2.33578399e-01 -4.36699420e-01 7.31655002e-01 -6.50798142e-01
5.79506934e-01 -1.11702991e+00 -1.76356927e-01 1.09754932e+00
-6.69700861e-01 8.57182220e-02 8.79012048e-01 -9.41408873e-01
-1.16430509e+00 -6.26181901e-01 8.95210207e-01 5.79496086e-01
5.02289474e-01 5.40405214e-01 -8.27755749e-01 9.58787054e-02
3.20279092e-01 -2.28869259e-01 4.49035883e-01 -7.76363909e-02
1.50711000e-01 -6.83914244e-01 -1.49986684e+00 2.47321889e-01
-1.60458479e-02 -2.73421943e-01 -2.09577262e-01 -5.79511784e-02
-1.09436378e-01 9.74975005e-02 -1.07345903e+00 8.88637662e-01
8.79943490e-01 -1.42769635e+00 9.62659717e-01 3.42069447e-01
1.76457286e-01 -5.31421244e-01 -5.74406803e-01 -1.12696338e+00
-1.54940173e-01 2.82551140e-01 6.74949825e-01 1.24859655e+00
3.13061565e-01 -9.38698649e-01 5.19749284e-01 3.70473891e-01
2.12729231e-01 2.15240270e-01 -7.12592483e-01 -3.34949315e-01
-1.94192693e-01 -8.37499946e-02 5.63379407e-01 5.27814984e-01
-3.22911322e-01 -4.50813711e-01 5.36518823e-03 5.76258659e-01
8.90329361e-01 6.04593456e-02 6.84829473e-01 -1.67277038e+00
3.18415642e-01 -4.41192240e-01 -6.62381291e-01 -1.54656544e-01
-3.78027737e-01 -2.45975956e-01 -8.37766975e-02 -1.89054871e+00
1.22242130e-01 -4.86794770e-01 -3.15106153e-01 4.74369228e-01
-4.18610662e-01 2.67728359e-01 2.34068796e-01 2.23924756e-01
1.00585580e-01 1.68840364e-01 8.19344759e-01 -3.06002617e-01
-2.04849824e-01 9.47716683e-02 -1.17815845e-01 3.82724047e-01
1.08823156e+00 -1.99367896e-01 -5.95209822e-02 -3.37822467e-01
4.39067870e-01 1.18779361e-01 5.33949971e-01 -1.33345199e+00
-1.57887235e-01 -3.95704001e-01 2.61091471e-01 -1.06430995e+00
2.24034324e-01 -1.05225492e+00 6.76527083e-01 6.45822287e-01
2.26660714e-01 -3.87384772e-01 4.71809477e-01 8.94888118e-02
-4.16815162e-01 -6.34798884e-01 8.31641376e-01 -2.17376918e-01
-1.10017037e+00 -2.92220592e-01 -7.11363554e-01 -1.04572237e+00
9.78529692e-01 -5.89268923e-01 -6.38907135e-01 1.84846055e-02
-7.20401108e-01 1.31533816e-01 4.51121837e-01 -1.75678334e-03
3.35476607e-01 -1.11607897e+00 -6.84375882e-01 1.00272588e-01
3.79245639e-01 -2.68213123e-01 4.26938027e-01 8.44227552e-01
-1.14868140e+00 4.52958494e-01 -9.59474027e-01 -8.93561125e-01
-1.48281169e+00 -8.21254998e-02 3.63007724e-01 -1.47617429e-01
8.15336555e-02 2.09855095e-01 -3.98432732e-01 -3.52433681e-01
-3.69409621e-01 -1.60432965e-01 -5.33422947e-01 6.73387706e-01
2.24677593e-01 4.08471912e-01 3.05645943e-01 -8.46967697e-01
-4.59684938e-01 8.83014381e-01 4.00900632e-01 -2.83235878e-01
1.34989989e+00 -5.32175064e-01 -1.56113997e-01 6.05491400e-01
7.23081052e-01 -5.25022559e-02 -8.38831961e-01 -2.73807913e-01
9.91656557e-02 -4.12600726e-01 5.68744063e-01 -6.49320424e-01
-8.08196187e-01 7.49998391e-01 1.16868854e+00 7.39485025e-01
1.29355764e+00 -6.45287812e-01 1.51192397e-01 4.61840481e-01
3.88558537e-01 -1.29221189e+00 -8.68317485e-01 -1.13104302e-02
6.88564599e-01 -1.77741110e+00 6.62206113e-01 -3.19781810e-01
3.09290644e-02 1.45512879e+00 8.94573554e-02 2.63150007e-01
8.60555291e-01 5.95021918e-02 1.63107857e-01 8.47183913e-03
9.95028466e-02 -9.52146947e-01 -7.65907988e-02 6.61820889e-01
7.05633223e-01 5.72558045e-01 -8.22635710e-01 -4.14790213e-01
4.72912282e-01 8.83661434e-02 2.78438896e-01 1.48522139e+00
-1.20378685e+00 -1.27477121e+00 -1.09384513e+00 3.91381234e-01
1.33075729e-01 2.67279536e-01 1.55453295e-01 7.65416920e-01
1.19559832e-01 1.16704285e+00 -1.05872877e-01 -6.90245628e-02
1.13684431e-01 -1.05463993e-02 3.84650946e-01 -8.48411024e-02
-2.41191387e-01 2.30905175e-01 -3.16512026e-02 -2.81770706e-01
-1.50409961e+00 -9.26945448e-01 -7.60012269e-01 -3.96637648e-01
-4.11021739e-01 8.60281289e-02 1.58187962e+00 9.05834317e-01
4.20692675e-02 1.00144945e-01 8.08969676e-01 -1.12469923e+00
-2.16649044e-02 -1.03435731e+00 -9.91335928e-01 -1.02791011e-01
3.53218764e-01 -7.69711673e-01 -2.27567062e-01 -2.17345014e-01] | [9.529847145080566, -1.6628894805908203] |
86e7120b-a5af-4a53-9751-a23aa69c671e | cross-view-hierarchy-network-for-stereo-image | 2304.06236 | null | https://arxiv.org/abs/2304.06236v1 | https://arxiv.org/pdf/2304.06236v1.pdf | Cross-View Hierarchy Network for Stereo Image Super-Resolution | Stereo image super-resolution aims to improve the quality of high-resolution stereo image pairs by exploiting complementary information across views. To attain superior performance, many methods have prioritized designing complex modules to fuse similar information across views, yet overlooking the importance of intra-view information for high-resolution reconstruction. It also leads to problems of wrong texture in recovered images. To address this issue, we explore the interdependencies between various hierarchies from intra-view and propose a novel method, named Cross-View-Hierarchy Network for Stereo Image Super-Resolution (CVHSSR). Specifically, we design a cross-hierarchy information mining block (CHIMB) that leverages channel attention and large kernel convolution attention to extract both global and local features from the intra-view, enabling the efficient restoration of accurate texture details. Additionally, a cross-view interaction module (CVIM) is proposed to fuse similar features from different views by utilizing cross-view attention mechanisms, effectively adapting to the binocular scene. Extensive experiments demonstrate the effectiveness of our method. CVHSSR achieves the best stereo image super-resolution performance than other state-of-the-art methods while using fewer parameters. The source code and pre-trained models are available at https://github.com/AlexZou14/CVHSSR. | ['Ming Tan', 'Zhongxin Yu', 'Mingchao Jiang', 'Yunchen Zhang', 'Liang Chen', 'Hongxia Gao', 'Wenbin Zou'] | 2023-04-13 | null | null | null | null | ['image-super-resolution', 'stereo-image-super-resolution'] | ['computer-vision', 'computer-vision'] | [ 1.82860404e-01 -1.60483703e-01 3.15004140e-02 -3.83892149e-01
-1.00515854e+00 -9.71698463e-02 2.38243759e-01 -4.28371787e-01
1.12446561e-01 6.13555193e-01 6.81326926e-01 1.47165194e-01
-1.58909902e-01 -8.09687912e-01 -7.12031424e-01 -7.11030304e-01
4.18505043e-01 -2.57138312e-01 2.94832587e-01 -3.54894698e-01
4.16552633e-01 4.32019234e-01 -1.66548514e+00 7.38580942e-01
9.46146071e-01 8.38147104e-01 7.39021599e-01 3.62879306e-01
9.70067084e-02 6.77567661e-01 -1.56161049e-02 -9.55464542e-02
2.97124535e-01 -2.89642960e-01 -7.55187154e-01 1.80028066e-01
6.33981228e-01 -8.24787617e-01 -4.11678135e-01 1.12495577e+00
5.64358473e-01 2.93691480e-03 1.46973342e-01 -3.62995684e-01
-7.73958802e-01 1.60273552e-01 -9.86832678e-01 5.15919328e-01
4.63688016e-01 1.49508594e-02 7.49001026e-01 -1.11866963e+00
5.28016806e-01 1.24350011e+00 2.55451649e-01 4.23076719e-01
-1.48375154e+00 -8.38588238e-01 1.09821677e-01 3.08951795e-01
-1.27553153e+00 -7.21128762e-01 7.49916911e-01 -2.76986480e-01
7.80167639e-01 1.96779236e-01 4.11233664e-01 1.03839481e+00
1.99594393e-01 5.24352431e-01 1.37184358e+00 -1.16149448e-01
-1.56227767e-01 1.38542149e-02 -1.59362763e-01 5.17644346e-01
9.65654477e-02 3.67118567e-01 -7.80601203e-01 1.18515559e-01
1.48618853e+00 1.56398997e-01 -7.42854536e-01 -2.94985145e-01
-1.21878278e+00 6.07715964e-01 6.55190289e-01 3.42719972e-01
-4.39711362e-01 -2.40490735e-01 1.82100296e-01 2.32542157e-02
5.62345624e-01 2.70393997e-01 -3.85606974e-01 3.93990129e-01
-5.74479282e-01 -1.67580675e-02 -5.97380921e-02 7.67766595e-01
7.56569386e-01 -3.24345566e-02 -8.47681090e-02 1.03435981e+00
7.71734342e-02 1.31067783e-01 3.51506829e-01 -1.31739175e+00
5.75134158e-01 5.45744359e-01 1.75857425e-01 -8.66390169e-01
-1.66664273e-01 -5.32477617e-01 -1.05914557e+00 2.29023859e-01
-1.03878319e-01 3.21525514e-01 -7.14307070e-01 1.52144396e+00
4.75449562e-01 2.59697773e-02 -1.07920825e-01 1.32748663e+00
1.01259255e+00 5.08046091e-01 -2.92712390e-01 -2.97924072e-01
1.25333178e+00 -8.83323669e-01 -5.90785921e-01 -1.31474137e-01
-5.82508892e-02 -8.74750495e-01 9.92024839e-01 3.96730065e-01
-1.28731644e+00 -8.64600599e-01 -1.01001751e+00 -4.72211123e-01
1.11113921e-01 8.59570689e-04 5.58164060e-01 -2.29900936e-03
-1.06914437e+00 4.91370976e-01 -6.60174012e-01 -7.97122940e-02
4.66793150e-01 3.07892263e-01 -5.19555688e-01 -4.44463581e-01
-9.64984059e-01 5.73146641e-01 1.69540197e-01 3.28796990e-02
-6.71165466e-01 -7.64564216e-01 -9.19913292e-01 2.67054774e-02
3.44937891e-01 -9.66067433e-01 9.85190809e-01 -9.07165527e-01
-1.40833282e+00 7.16700375e-01 -4.26266402e-01 8.12003091e-02
1.53270006e-01 -2.86912650e-01 -3.66287231e-01 3.52876812e-01
3.74315709e-01 4.76228356e-01 9.22335267e-01 -1.57309091e+00
-7.71911144e-01 -5.58120728e-01 2.63730101e-02 5.20161986e-01
-5.25958873e-02 1.35266250e-02 -6.80830479e-01 -6.29036844e-01
5.04249632e-01 -2.30876744e-01 -1.59622818e-01 -3.00821871e-01
-2.67230213e-01 2.08511174e-01 5.86811602e-01 -8.96643281e-01
1.11907411e+00 -2.10496163e+00 3.72061551e-01 -3.11686486e-01
5.18178999e-01 1.30537912e-01 -2.39792336e-02 8.96767750e-02
-2.27244452e-01 -1.43012747e-01 -3.00427917e-02 -1.42041489e-01
-6.86857522e-01 -2.81667151e-02 -3.19626838e-01 3.76671284e-01
2.33246654e-01 7.52565563e-01 -7.28021383e-01 -3.40788364e-01
5.12517571e-01 9.77286220e-01 -6.11490369e-01 3.68238270e-01
2.40946606e-01 1.16406727e+00 -4.94397730e-01 7.05897987e-01
9.74291325e-01 -6.53885901e-01 1.17251821e-01 -7.64653742e-01
-2.94477791e-01 2.36252278e-01 -1.04587066e+00 1.92784679e+00
-6.57929122e-01 4.06605184e-01 1.73870668e-01 -7.03906953e-01
7.58720279e-01 2.31362015e-01 4.40563768e-01 -1.22222471e+00
1.71795283e-02 1.70447141e-01 -3.79437715e-01 -5.44401348e-01
4.46613938e-01 -2.30099216e-01 2.89204776e-01 6.46469742e-02
-4.24049720e-02 3.09939533e-01 -1.56871602e-01 8.42420533e-02
7.05425262e-01 2.94541657e-01 4.58157301e-01 -2.13143532e-03
7.35038459e-01 -5.22118449e-01 6.55183971e-01 4.60358083e-01
5.60784414e-02 1.10678005e+00 -2.80623361e-02 -5.89024961e-01
-1.16414392e+00 -1.15807855e+00 -2.66627878e-01 7.59358227e-01
5.41335046e-01 -2.81641781e-01 -5.21059692e-01 -1.97467580e-01
-2.16186181e-01 2.56539226e-01 -6.22885942e-01 1.92286104e-01
-6.12274468e-01 -5.37395239e-01 -2.68342644e-01 5.46328366e-01
8.68553936e-01 -8.29871356e-01 -4.84843820e-01 7.15392157e-02
-7.86440611e-01 -1.30128360e+00 -5.31123757e-01 -2.75324613e-01
-1.00173807e+00 -1.00424230e+00 -7.46891439e-01 -4.25998539e-01
5.89470625e-01 1.03086352e+00 1.19041872e+00 1.27067417e-03
-3.25132847e-01 9.48604196e-02 -2.68044263e-01 4.39004540e-01
6.60206005e-02 -1.25070706e-01 -1.21223375e-01 2.40109280e-01
2.58958824e-02 -8.87362719e-01 -9.92240369e-01 4.21331853e-01
-7.84377038e-01 7.07758963e-01 7.56130755e-01 1.04391623e+00
8.99926901e-01 5.68717085e-02 2.57949471e-01 -7.01847970e-01
1.71522707e-01 -4.49997276e-01 -6.23256683e-01 3.08950208e-02
-3.49658936e-01 -3.84606831e-02 5.72267473e-01 -3.34828869e-02
-1.47572231e+00 -1.61245450e-01 -1.39814988e-02 -6.12605274e-01
-1.98646620e-01 1.24085046e-01 -4.87227261e-01 -1.19863488e-01
2.83828735e-01 4.95496929e-01 -2.28746697e-01 -7.33665228e-01
1.19209096e-01 5.56969583e-01 5.30010998e-01 -3.67170930e-01
5.05315423e-01 9.01723683e-01 -3.08328152e-01 -7.08258390e-01
-1.29392242e+00 -5.87522328e-01 -5.65888524e-01 -1.59217358e-01
9.25918698e-01 -1.46288562e+00 -6.66968763e-01 2.80185699e-01
-9.58222032e-01 -1.40716806e-01 1.57762930e-01 4.40241903e-01
-6.05212629e-01 4.10709858e-01 -8.28374863e-01 -5.74011564e-01
-4.10501927e-01 -1.23098302e+00 1.35027778e+00 4.68867809e-01
1.64709702e-01 -5.66716552e-01 -2.80670226e-01 1.03520966e+00
4.84156340e-01 1.89121943e-02 6.45494819e-01 2.44749859e-01
-1.14434171e+00 4.22886878e-01 -8.73524129e-01 2.81246513e-01
2.61294961e-01 -5.36273360e-01 -1.02213311e+00 -4.68042046e-01
2.16563001e-01 -1.64736241e-01 9.51110661e-01 6.14760220e-01
1.39268148e+00 -1.16776749e-01 -1.37660563e-01 1.16415787e+00
1.76574683e+00 -5.68521619e-02 9.02521849e-01 5.40397346e-01
9.85532403e-01 5.75831711e-01 6.53677642e-01 3.58115375e-01
6.45518839e-01 9.76992011e-01 3.54378164e-01 -3.67980272e-01
-2.71536767e-01 -2.30364174e-01 2.03528807e-01 7.75230587e-01
-3.12326461e-01 2.46461302e-01 -4.12411809e-01 4.09913242e-01
-1.69313502e+00 -1.00575829e+00 -5.04126064e-02 2.32916212e+00
6.58559144e-01 -8.59142393e-02 -1.89880773e-01 -1.93542778e-01
7.25633681e-01 2.33301580e-01 -6.75772250e-01 8.35970864e-02
-2.38835469e-01 8.78026262e-02 3.66881818e-01 6.34370327e-01
-9.40062225e-01 8.38977575e-01 5.20195293e+00 8.46098959e-01
-1.12877798e+00 2.35411301e-01 8.49945307e-01 -2.11415559e-01
-3.79407734e-01 -4.77074347e-02 -7.66708732e-01 4.81593072e-01
3.86462152e-01 2.69848078e-01 7.46539593e-01 3.45492184e-01
3.76723379e-01 -2.60167360e-01 -6.42943084e-01 1.17828798e+00
1.76338077e-01 -1.34726107e+00 1.15615502e-01 1.46886617e-01
8.55356336e-01 1.23190597e-01 5.51389977e-02 -1.16226256e-01
1.05666764e-01 -9.77241695e-01 3.45724791e-01 6.45875335e-01
1.06202102e+00 -8.34842443e-01 5.85295022e-01 3.33674178e-02
-1.39431143e+00 -2.35157326e-01 -3.63304257e-01 1.71374410e-01
1.83443621e-01 6.10161602e-01 -8.60397965e-02 9.68497157e-01
1.17218041e+00 8.55199814e-01 -4.27251995e-01 6.46606028e-01
-9.77225900e-02 -8.86328425e-03 1.71860412e-01 9.81140733e-01
-1.97510138e-01 -2.16278672e-01 6.03250861e-01 7.99150944e-01
2.99217284e-01 3.88173550e-01 -1.42605573e-01 1.02444136e+00
2.03981787e-01 -1.58528954e-01 -5.05410075e-01 4.06983227e-01
2.86399037e-01 1.38246834e+00 -3.59065533e-01 -3.34271312e-01
-6.97633028e-01 1.24206340e+00 4.90234375e-01 6.13683581e-01
-6.41815424e-01 2.00867858e-02 8.09342861e-01 4.69751388e-01
5.56868136e-01 -7.03821108e-02 -4.91748720e-01 -1.58608472e+00
2.09451944e-01 -9.76006746e-01 3.61284792e-01 -1.12997186e+00
-1.25583887e+00 7.07389057e-01 -2.65094668e-01 -1.43080747e+00
1.94978803e-01 -2.36447334e-01 -2.17812926e-01 1.12109363e+00
-1.80065346e+00 -1.28807724e+00 -6.25801146e-01 7.88951457e-01
8.10243845e-01 -6.22946536e-04 4.71110016e-01 4.49408919e-01
-5.34978986e-01 3.07988614e-01 4.91131051e-03 -1.35993704e-01
8.40571702e-01 -8.64188850e-01 2.24694297e-01 8.29892755e-01
-3.46787363e-01 6.41475379e-01 4.38635558e-01 -5.86212575e-01
-1.32055962e+00 -9.99294758e-01 5.62157333e-01 -3.45495611e-01
2.21901774e-01 -7.42364675e-02 -1.20676696e+00 4.11466271e-01
8.33197385e-02 1.11964084e-01 4.61321115e-01 4.04665805e-02
-5.83546758e-01 -2.84940630e-01 -9.56781387e-01 5.89025617e-01
1.14388096e+00 -6.98829651e-01 -2.32578292e-01 -2.11556986e-01
7.90099621e-01 -5.33100903e-01 -1.03484571e+00 6.84221625e-01
6.25647843e-01 -1.72011888e+00 1.27538490e+00 -1.30747229e-01
9.47397649e-01 -5.64221621e-01 -4.58820313e-01 -9.78586853e-01
-6.64319694e-01 -2.87475765e-01 -1.73929319e-01 1.00874436e+00
-9.00554135e-02 -7.08574653e-01 3.78368437e-01 3.36437970e-01
-1.00442611e-01 -9.12312746e-01 -8.02486539e-01 -3.74043733e-01
-3.11090499e-01 1.23140507e-03 5.32603025e-01 1.00803256e+00
-2.22489417e-01 4.69690502e-01 -5.82985461e-01 5.72813451e-01
9.85234976e-01 6.41527891e-01 6.57038331e-01 -8.61675382e-01
-5.20890892e-01 -3.06105196e-01 -1.05000928e-01 -1.08600307e+00
-1.66804031e-01 -6.07350409e-01 -1.81742877e-01 -1.59673619e+00
7.49443710e-01 -1.03809327e-01 -3.72920007e-01 1.06407695e-01
-4.15242195e-01 5.73452890e-01 1.50574166e-02 4.34640110e-01
-6.65951848e-01 5.99016368e-01 1.75183105e+00 3.43427509e-01
-1.87448457e-01 -3.40204775e-01 -1.17031562e+00 5.66086173e-01
7.32955158e-01 -2.29833126e-02 -3.04785609e-01 -7.08794415e-01
6.28769696e-02 5.65789104e-01 6.14157498e-01 -7.99387872e-01
6.56516626e-02 -1.68694243e-01 7.62784839e-01 -6.30330324e-01
4.57902551e-01 -5.85123420e-01 2.36998305e-01 -1.49659347e-02
-1.02723025e-01 -1.38783559e-01 6.41558915e-02 5.87423623e-01
-4.08723980e-01 4.23269868e-01 1.08892179e+00 -3.77103597e-01
-6.70492828e-01 4.31680143e-01 9.95036308e-03 -2.68597543e-01
5.97419262e-01 -3.24365497e-01 -3.40844750e-01 -2.71512717e-01
-6.00682616e-01 1.41356915e-01 8.85356367e-01 4.82152164e-01
9.31950748e-01 -1.27808213e+00 -7.45103896e-01 5.44204712e-01
1.59996033e-01 2.29933754e-01 1.01174486e+00 1.03538942e+00
-2.03703403e-01 3.68054658e-01 -4.40731615e-01 -6.85559571e-01
-1.28498006e+00 5.05546212e-01 4.11585748e-01 -1.98591143e-01
-9.77183640e-01 6.68345392e-01 1.00812435e+00 -2.24983364e-01
-1.58744752e-01 1.01191439e-02 -3.20054948e-01 -2.13893339e-01
8.65996659e-01 2.94699490e-01 -3.49537991e-02 -6.78472042e-01
2.70461142e-02 9.05456364e-01 -4.46769983e-01 6.72000945e-02
1.45078588e+00 -8.18652034e-01 -1.79321453e-01 1.97274491e-01
1.08307505e+00 -3.87713499e-02 -1.57428634e+00 -4.77867126e-01
-5.13687611e-01 -1.07214725e+00 3.49670529e-01 -6.74905360e-01
-1.31711853e+00 8.20717633e-01 7.68107235e-01 -3.32508057e-01
1.62738240e+00 1.33038824e-02 8.83726537e-01 -3.13524783e-01
5.89834332e-01 -8.49483013e-01 1.75165907e-01 3.06033432e-01
1.09571540e+00 -1.58906150e+00 1.54675081e-01 -5.82938254e-01
-6.44181848e-01 9.62646842e-01 8.44082713e-01 1.62419621e-02
3.83169591e-01 1.27094075e-01 -2.95164138e-02 -3.19043487e-01
-7.54773855e-01 -2.55601108e-01 2.98480481e-01 4.15362388e-01
5.59352279e-01 -2.51343269e-02 -7.56460130e-02 4.95859712e-01
1.76865220e-01 6.49165958e-02 3.45378906e-01 4.92403597e-01
-3.38922143e-01 -9.21512246e-01 -3.86011332e-01 3.52393448e-01
-5.26674569e-01 -3.04896772e-01 1.96136963e-02 4.45415854e-01
1.39673099e-01 8.78081083e-01 -2.36465279e-02 -4.73534614e-01
1.82241231e-01 -5.64609587e-01 5.85326016e-01 -4.70279038e-01
-2.79870421e-01 5.53381085e-01 -1.10635795e-01 -1.03245485e+00
-5.66718280e-01 -4.30503041e-01 -8.53224695e-01 -2.69880652e-01
-1.48097083e-01 -2.74031788e-01 2.46829733e-01 6.77904725e-01
6.92272305e-01 6.47210002e-01 8.84040415e-01 -1.19950306e+00
1.25054345e-02 -7.58258462e-01 -5.53170860e-01 3.07278842e-01
5.68330646e-01 -6.89965844e-01 -3.08179080e-01 -5.84150814e-02] | [10.726139068603516, -2.184730052947998] |
c05ad90c-d90d-46d9-83c6-2a9860eb0aa1 | sliding-line-point-regression-for-shape | 1801.09969 | null | http://arxiv.org/abs/1801.09969v1 | http://arxiv.org/pdf/1801.09969v1.pdf | Sliding Line Point Regression for Shape Robust Scene Text Detection | Traditional text detection methods mostly focus on quadrangle text. In this
study we propose a novel method named sliding line point regression (SLPR) in
order to detect arbitrary-shape text in natural scene. SLPR regresses multiple
points on the edge of text line and then utilizes these points to sketch the
outlines of the text. The proposed SLPR can be adapted to many object detection
architectures such as Faster R-CNN and R-FCN. Specifically, we first generate
the smallest rectangular box including the text with region proposal network
(RPN), then isometrically regress the points on the edge of text by using the
vertically and horizontally sliding lines. To make full use of information and
reduce redundancy, we calculate x-coordinate or y-coordinate of target point by
the rectangular box position, and just regress the remaining y-coordinate or
x-coordinate. Accordingly we can not only reduce the parameters of system, but
also restrain the points which will generate more regular polygon. Our approach
achieved competitive results on traditional ICDAR2015 Incidental Scene Text
benchmark and curve text detection dataset CTW1500. | ['Yixing Zhu', 'Jun Du'] | 2018-01-30 | null | null | null | null | ['curved-text-detection'] | ['computer-vision'] | [ 2.47479547e-02 -1.49152771e-01 1.43941313e-01 1.77762471e-02
-4.82650846e-01 -4.36772585e-01 4.38473374e-01 7.16596022e-02
-4.97245044e-01 1.33757755e-01 5.30677401e-02 -1.72360092e-01
3.42140228e-01 -8.11609626e-01 -6.53099120e-01 -5.21872282e-01
5.97854197e-01 4.96691287e-01 7.68017054e-01 -2.27233514e-01
6.07762098e-01 7.85477936e-01 -9.33642030e-01 1.22362971e-01
6.43364847e-01 6.91514015e-01 1.96236476e-01 5.27714491e-01
-3.51202965e-01 2.03747422e-01 -6.20401323e-01 -2.57897973e-01
3.96191835e-01 5.50863985e-03 -2.71668851e-01 3.20205420e-01
2.03309447e-01 -5.85832417e-01 -6.65575862e-01 7.96645045e-01
5.07801354e-01 -1.05172396e-01 8.58750880e-01 -9.35174644e-01
-6.48518324e-01 5.17149925e-01 -1.52824342e+00 1.30105149e-02
1.45088628e-01 -5.69649860e-02 8.08193564e-01 -1.34319961e+00
5.41686833e-01 1.35510683e+00 8.50496888e-01 4.27846342e-01
-9.03222561e-01 -6.58206820e-01 2.74520755e-01 -1.55389190e-01
-1.84099925e+00 -5.45798577e-02 8.56594026e-01 -2.54508138e-01
7.12902486e-01 2.88070470e-01 4.57094908e-01 5.32955050e-01
2.66733021e-01 1.10298645e+00 3.64931405e-01 -6.13099217e-01
-3.16556126e-01 -1.50402831e-02 7.30267540e-02 8.55722010e-01
2.89185673e-01 -4.33206588e-01 -1.58891141e-01 -4.41305339e-03
1.11994374e+00 3.63174438e-01 -2.54257500e-01 -4.19843465e-01
-1.45633018e+00 8.02758873e-01 4.05590206e-01 1.69278219e-01
-9.56397206e-02 -1.21690080e-01 4.34393287e-01 -2.85855055e-01
2.45269656e-01 1.43559473e-02 -2.39702553e-01 4.67029363e-01
-1.07059622e+00 1.88937277e-01 3.91202122e-01 1.24605858e+00
4.94373918e-01 -6.33523474e-03 -3.12732190e-01 9.06468868e-01
2.74528086e-01 6.12423539e-01 4.71067727e-01 -1.90194905e-01
1.06520426e+00 1.02708411e+00 -6.21967763e-02 -1.01171672e+00
-6.28580034e-01 -6.99248388e-02 -9.31699157e-01 4.71476652e-02
4.01346356e-01 -2.74534702e-01 -1.09996712e+00 7.77866244e-01
3.68141025e-01 -2.17551872e-01 -1.51624411e-01 9.39198554e-01
8.96598458e-01 8.49786043e-01 -3.79950583e-01 3.73971090e-02
1.41342866e+00 -9.88206208e-01 -3.82203758e-01 3.62878740e-02
8.20314109e-01 -9.13178861e-01 1.07809734e+00 3.42394173e-01
-1.01738596e+00 -4.47731227e-01 -1.06821203e+00 -5.38407922e-01
-2.90501565e-01 1.01310134e+00 -5.19347750e-02 1.50255784e-01
-6.39633298e-01 2.78313398e-01 -8.44452977e-01 -1.69263646e-01
4.62102473e-01 4.04183507e-01 -1.52880356e-01 2.82945424e-01
-6.01497352e-01 5.61477721e-01 3.05695623e-01 4.45737422e-01
-2.20592082e-01 -3.72088969e-01 -7.37117589e-01 1.63198724e-01
5.39341986e-01 -3.62037659e-01 8.42762113e-01 -5.64534724e-01
-1.33299184e+00 7.43815780e-01 -4.37920801e-02 -4.04385149e-01
9.67349946e-01 -2.87351102e-01 -2.15190202e-01 -8.69433582e-02
6.61064982e-02 8.41498554e-01 1.05109847e+00 -9.22164083e-01
-7.50237286e-01 -6.56967759e-01 -5.85705936e-01 4.98920649e-01
-2.22804815e-01 -4.05809050e-03 -9.07745719e-01 -8.76900733e-01
6.09390795e-01 -7.59494126e-01 -5.80937229e-02 4.05308872e-01
-9.19692457e-01 -6.55794978e-01 1.51002848e+00 -7.82072186e-01
1.20712912e+00 -2.17211866e+00 -7.11877272e-02 1.97867543e-01
3.51715326e-01 2.69904912e-01 1.33969635e-01 2.76844651e-01
1.51320070e-01 6.71215057e-02 -1.15756780e-01 -1.86185628e-01
-3.22455496e-01 -2.28093684e-01 -4.29061979e-01 7.43021369e-01
3.35091710e-01 8.86494398e-01 -2.58958787e-01 -7.77370512e-01
3.84968847e-01 4.08392578e-01 -3.69207978e-01 -2.03158185e-01
-2.93733895e-01 -4.91567999e-02 -7.77223110e-01 4.66090858e-01
9.75953400e-01 -3.38769592e-02 -3.15064907e-01 -2.86322802e-01
-3.82875592e-01 -1.58371896e-01 -1.36586750e+00 1.32498360e+00
-6.48346618e-02 9.98879492e-01 -2.64975846e-01 -6.69001341e-01
1.50047386e+00 1.59362853e-02 3.91144276e-01 -4.93067145e-01
1.93066731e-01 -1.18379764e-01 -1.77025497e-01 -4.81568068e-01
5.92075706e-01 3.60357612e-01 1.19946726e-01 2.19730109e-01
-6.90195978e-01 -2.62192726e-01 -1.44270301e-01 1.51295736e-01
5.59003353e-01 2.97431886e-01 3.39429379e-01 -5.08698821e-02
8.77251446e-01 6.77185431e-02 3.31178755e-01 5.12200892e-01
1.11021213e-01 8.86053145e-01 7.46093452e-01 -7.35397518e-01
-1.39860296e+00 -7.35063076e-01 -3.05712312e-01 9.83039558e-01
3.46416712e-01 -3.41361493e-01 -6.96415782e-01 -6.84624672e-01
7.31715262e-02 5.77596307e-01 -5.67922652e-01 3.67515147e-01
-1.13346517e+00 -4.58140343e-01 6.48752034e-01 6.66475117e-01
7.40503252e-01 -1.25352323e+00 -5.66389143e-01 1.93424448e-01
-2.06801351e-02 -9.77253377e-01 -9.57782924e-01 1.40121588e-02
-8.77656102e-01 -9.16030586e-01 -9.73829269e-01 -9.59385097e-01
8.76838386e-01 5.55583537e-01 3.95218492e-01 1.34150609e-01
-4.84912932e-01 -9.18058529e-02 -1.55943170e-01 -4.61494446e-01
-2.78687067e-02 2.94022597e-02 -3.93939108e-01 -8.03227872e-02
4.02885497e-01 2.55517751e-01 -7.22542703e-01 6.67109907e-01
-5.76209784e-01 3.30000341e-01 4.45822388e-01 6.01613522e-01
7.66521394e-01 4.70245555e-02 1.62086517e-01 -6.24910116e-01
5.00557065e-01 1.19443024e-02 -8.31465602e-01 2.56335944e-01
-2.66728163e-01 1.56534053e-02 7.11536229e-01 -4.95059580e-01
-7.03458071e-01 4.26399589e-01 1.60245802e-02 -5.22942781e-01
1.96346138e-02 -9.76622477e-02 2.46722978e-02 1.69983119e-01
6.11544967e-01 5.88160753e-01 -3.58585268e-01 -2.65648395e-01
4.32258189e-01 8.95358324e-01 4.97524083e-01 -2.27411121e-01
8.23113918e-01 7.81308234e-01 2.23080814e-01 -1.07992744e+00
-4.42285687e-01 -6.42996907e-01 -1.08021522e+00 5.07393330e-02
8.88525009e-01 -6.76644742e-01 -7.78881192e-01 2.95116127e-01
-1.46628439e+00 6.65949583e-02 3.33935246e-02 1.51372015e-01
-1.62706241e-01 7.29228914e-01 -4.95802402e-01 -7.77906239e-01
-7.85558999e-01 -9.11994994e-01 1.58674026e+00 2.57844537e-01
1.90532923e-01 -4.48026061e-01 -1.96064129e-01 -1.65802203e-02
-2.40275443e-01 1.20333828e-01 7.60137498e-01 -6.88395202e-01
-4.33516264e-01 -6.04223311e-01 -6.66634381e-01 -6.35401011e-02
-2.03380212e-01 3.90607417e-01 -5.44178724e-01 -2.70471066e-01
-2.50003815e-01 1.36583015e-01 9.18452501e-01 5.09008229e-01
1.41554403e+00 -1.83897376e-01 -7.28814363e-01 6.31996751e-01
1.27152336e+00 7.81227946e-02 6.10971212e-01 3.80940109e-01
1.09693646e+00 4.84726787e-01 5.76239645e-01 4.25158203e-01
4.07754868e-01 6.56237602e-01 7.24299178e-02 -3.57209355e-01
-1.18095927e-01 -3.15840691e-01 5.99508956e-02 5.67296267e-01
-1.08008072e-01 -3.65730703e-01 -9.32417452e-01 1.65073693e-01
-1.69260275e+00 -5.26203334e-01 -6.98308468e-01 2.08225560e+00
3.19335639e-01 4.43840772e-01 2.77678639e-01 1.36744574e-01
1.20387602e+00 -1.42104447e-01 -6.93835497e-01 -1.22711867e-01
-1.00575432e-01 -2.60299861e-01 7.28197157e-01 1.21733114e-01
-9.89074886e-01 1.15943730e+00 5.06713820e+00 9.70301986e-01
-1.22625709e+00 -3.28708380e-01 4.78630483e-01 9.87452269e-02
2.25592330e-01 -1.42727524e-01 -1.22950959e+00 2.42284551e-01
2.95471046e-02 3.78564969e-02 3.06082517e-01 9.29773569e-01
4.35053021e-01 8.64880607e-02 -1.02458906e+00 1.17270517e+00
1.93506718e-01 -1.17277491e+00 2.70387053e-01 -5.35155535e-02
5.25513291e-01 4.45754863e-02 -6.92308545e-02 1.86450854e-01
-1.57994777e-01 -8.17461014e-01 7.14922786e-01 4.10476416e-01
9.44604158e-01 -6.05624795e-01 4.54855770e-01 6.83519244e-01
-1.43938911e+00 -2.69877445e-02 -8.70870829e-01 5.13325930e-01
-2.38735318e-01 1.63914021e-02 -1.15990388e+00 3.21961761e-01
5.77998102e-01 5.98867357e-01 -6.38411045e-01 1.01681769e+00
-1.49677485e-01 2.41070494e-01 -6.18546426e-01 -4.07210320e-01
1.87578112e-01 -3.79809797e-01 4.74292845e-01 1.47845542e+00
3.31826508e-01 8.34065527e-02 7.58604780e-02 7.45625556e-01
-2.27097794e-01 6.36402547e-01 -5.44091940e-01 5.04628778e-01
4.45461124e-01 1.37317622e+00 -1.17521751e+00 -4.49363858e-01
-5.29761255e-01 8.72464418e-01 2.30285719e-01 1.56833410e-01
-8.55019569e-01 -7.55310237e-01 -5.20099580e-01 4.25770968e-01
6.25061452e-01 -2.69152552e-01 -6.15585744e-01 -9.52318907e-01
2.93986768e-01 -5.29658079e-01 3.91658068e-01 -1.15377665e+00
-6.45648837e-01 4.97929573e-01 -4.36989814e-01 -1.40165114e+00
3.82656544e-01 -5.67740142e-01 -8.75553310e-01 7.32305050e-01
-9.84387636e-01 -1.43471611e+00 -4.49910492e-01 7.13576198e-01
1.07651448e+00 -9.22001526e-03 2.80380249e-01 -2.54115731e-01
-8.02514136e-01 7.00370610e-01 2.96061277e-01 6.23708487e-01
7.63346910e-01 -1.06579673e+00 7.02055275e-01 5.30049860e-01
2.27669656e-01 4.73515362e-01 3.12596023e-01 -9.07239139e-01
-1.03093672e+00 -1.16282392e+00 4.57575351e-01 -3.91924649e-01
5.19179106e-01 -6.20366991e-01 -1.16582334e+00 7.40656257e-01
-1.49524763e-01 3.92694660e-02 -2.14540720e-01 -2.84581095e-01
-4.15704936e-01 1.18693206e-02 -1.03446150e+00 1.04652154e+00
5.95945776e-01 -4.24757339e-02 -5.35312891e-01 4.71601218e-01
6.32362902e-01 -6.42200828e-01 -4.28873420e-01 3.49182487e-01
6.59056127e-01 -6.98011935e-01 8.88056576e-01 -2.01474339e-01
4.11417633e-01 -5.36582589e-01 3.53152514e-01 -7.39018083e-01
-2.32870698e-01 -4.87430960e-01 2.69751251e-01 9.19744790e-01
4.19315636e-01 -3.97034287e-01 1.10751736e+00 8.00981447e-02
-2.90900081e-01 -9.37073588e-01 -8.82678151e-01 -1.82413980e-01
3.16746622e-01 -1.74838006e-01 4.76633579e-01 6.65086210e-01
3.72112878e-02 3.55982363e-01 -1.76070988e-01 2.27931455e-01
3.23191762e-01 -1.53425653e-02 1.00751889e+00 -1.18678534e+00
8.66292119e-02 -6.51983082e-01 -3.07084918e-01 -1.54689062e+00
-3.01149577e-01 -6.93222046e-01 8.29035137e-03 -1.44394028e+00
1.63648158e-01 -3.13126087e-01 1.25532642e-01 3.40620220e-01
-1.54181615e-01 7.64542967e-02 1.46719292e-01 4.78829056e-01
-3.90878350e-01 4.22273576e-01 1.61709571e+00 -6.93557411e-02
-5.91406047e-01 2.01000676e-01 -2.18177348e-01 1.04974008e+00
7.12339640e-01 -3.64407301e-01 4.60488955e-03 -3.50853235e-01
3.79314385e-02 2.97348708e-01 3.29488039e-01 -6.83126450e-01
6.27109826e-01 2.41810903e-02 1.09370840e+00 -1.67000711e+00
1.46647587e-01 -8.30987275e-01 -6.61529660e-01 4.20043051e-01
-2.31395766e-01 8.63844901e-02 2.84746319e-01 5.23997903e-01
4.52820778e-01 -4.10086066e-01 8.26027095e-01 1.16244443e-01
-9.14534256e-02 1.86106429e-01 -1.44811168e-01 -3.24110419e-01
1.21937680e+00 -5.98409295e-01 -3.94172251e-01 2.60317437e-02
-4.62798238e-01 4.21623588e-01 2.65018374e-01 4.35540199e-01
8.53986740e-01 -1.00875974e+00 -8.46316934e-01 4.13507372e-01
9.39738825e-02 5.37022293e-01 5.11026159e-02 6.82070076e-01
-9.45819914e-01 7.13051558e-01 1.49731264e-01 -8.71860325e-01
-1.40490150e+00 8.05984080e-01 3.93866390e-01 5.70359873e-03
-1.43667102e+00 5.22502780e-01 4.73170370e-01 -1.86853841e-01
2.83323586e-01 -5.16134262e-01 -3.73860121e-01 -3.39587301e-01
2.49627471e-01 4.49970067e-01 -1.03183538e-01 -5.75264454e-01
-2.02892616e-01 1.20056033e+00 -5.10451734e-01 -7.31637562e-03
9.84043002e-01 -6.43582568e-02 1.07910059e-01 8.27270523e-02
8.24796617e-01 3.51319492e-01 -1.37957227e+00 -3.29939902e-01
-6.01727776e-02 -3.02447408e-01 -1.73976198e-02 -2.98850387e-01
-8.84737849e-01 8.35726976e-01 4.45593864e-01 2.29716778e-01
7.79430747e-01 -2.28425879e-02 6.42108262e-01 5.59814870e-01
-2.09951475e-01 -1.18496919e+00 1.98852584e-01 3.32014889e-01
1.07670748e+00 -8.80812526e-01 4.08914238e-01 -6.50514066e-01
-7.29115784e-01 1.80608284e+00 8.80570829e-01 -5.28478563e-01
3.24766755e-01 1.91919103e-01 -1.97440833e-01 -2.43282184e-01
-2.20182300e-01 1.83915541e-01 4.26291168e-01 2.86233038e-01
4.83021915e-01 -8.77964944e-02 -3.68012786e-01 2.39739433e-01
-2.10114434e-01 -4.03038502e-01 6.61427259e-01 6.01108670e-01
-7.02712595e-01 -4.76365656e-01 -8.46742690e-01 6.55627966e-01
-3.21532816e-01 -3.49508077e-02 -3.97529215e-01 1.08570421e+00
5.21855243e-02 4.25047100e-01 5.21825194e-01 -2.86238436e-02
5.53726137e-01 -2.51387775e-01 2.65486360e-01 -5.46371043e-01
-3.34093720e-01 7.44154990e-01 -4.61845249e-01 6.55107573e-02
2.73821235e-01 -7.02464044e-01 -1.75965953e+00 -2.45618120e-01
-9.22815681e-01 -3.08939517e-01 8.22411180e-01 8.00474226e-01
-1.77051239e-02 3.95312071e-01 6.93531871e-01 -6.70466542e-01
-4.92814720e-01 -1.17176402e+00 -3.44197869e-01 9.04451869e-03
1.80793211e-01 -2.62045503e-01 -2.70244241e-01 1.28529742e-01] | [12.125288009643555, 2.2398176193237305] |
3e9a9946-2498-4fa0-a7a9-775c971e8aff | scene-graphs-a-survey-of-generations-and | 2104.01111 | null | https://arxiv.org/abs/2104.01111v5 | https://arxiv.org/pdf/2104.01111v5.pdf | A Comprehensive Survey of Scene Graphs: Generation and Application | Scene graph is a structured representation of a scene that can clearly express the objects, attributes, and relationships between objects in the scene. As computer vision technology continues to develop, people are no longer satisfied with simply detecting and recognizing objects in images; instead, people look forward to a higher level of understanding and reasoning about visual scenes. For example, given an image, we want to not only detect and recognize objects in the image, but also know the relationship between objects (visual relationship detection), and generate a text description (image captioning) based on the image content. Alternatively, we might want the machine to tell us what the little girl in the image is doing (Visual Question Answering (VQA)), or even remove the dog from the image and find similar images (image editing and retrieval), etc. These tasks require a higher level of understanding and reasoning for image vision tasks. The scene graph is just such a powerful tool for scene understanding. Therefore, scene graphs have attracted the attention of a large number of researchers, and related research is often cross-modal, complex, and rapidly developing. However, no relatively systematic survey of scene graphs exists at present. To this end, this survey conducts a comprehensive investigation of the current scene graph research. More specifically, we first summarized the general definition of the scene graph, then conducted a comprehensive and systematic discussion on the generation method of the scene graph (SGG) and the SGG with the aid of prior knowledge. We then investigated the main applications of scene graphs and summarized the most commonly used datasets. Finally, we provide some insights into the future development of scene graphs. We believe this will be a very helpful foundation for future research on scene graphs. | ['Alex Hauptmann', 'Xiaojiang Chen', 'Zhihui Li', 'Pengfei Xu', 'Pengzhen Ren', 'Xiaojun Chang'] | 2021-03-17 | null | null | null | null | ['visual-relationship-detection'] | ['computer-vision'] | [ 5.13086855e-01 6.22526146e-02 2.95789186e-02 -6.28814101e-01
-1.32215425e-01 -6.28742576e-01 4.93890405e-01 3.37747008e-01
6.77837953e-02 1.57528073e-01 5.04967347e-02 -2.79239655e-01
-2.52470002e-02 -9.74717438e-01 -5.69775820e-01 -5.55170357e-01
2.79363751e-01 1.78055450e-01 3.46833408e-01 -4.18449827e-02
3.28201830e-01 5.32583535e-01 -1.69479156e+00 3.96445423e-01
4.99119550e-01 8.69729519e-01 6.29010320e-01 4.68154162e-01
-4.41441506e-01 1.23988950e+00 -5.18511176e-01 -4.96029198e-01
-1.00479864e-01 -6.64531112e-01 -9.31553364e-01 9.81444001e-01
6.86237454e-01 -1.84011251e-01 -3.50378454e-01 1.41229141e+00
-1.14417732e-01 1.88824102e-01 5.19236207e-01 -1.54960132e+00
-7.59798944e-01 2.84949630e-01 -5.57905734e-01 7.38827214e-02
4.45777446e-01 -9.41727590e-03 1.00039423e+00 -9.04129148e-01
6.08610809e-01 1.20301807e+00 1.53597575e-02 3.03641587e-01
-8.39051008e-01 -2.16848478e-01 4.19669569e-01 5.07314086e-01
-1.31398284e+00 -1.69392824e-01 9.77244794e-01 -4.75678295e-01
5.17725825e-01 6.12498522e-01 7.91605592e-01 4.09549952e-01
-8.88412371e-02 9.98296380e-01 8.80384922e-01 -6.08218908e-01
1.21649794e-01 3.17722827e-01 3.18995446e-01 9.34379041e-01
2.87080169e-01 -3.23641598e-01 -3.20400000e-01 2.04887107e-01
8.92939568e-01 4.07849491e-01 -2.72350132e-01 -6.58204496e-01
-1.25349641e+00 6.98834062e-01 6.83574915e-01 3.55605304e-01
-3.24445873e-01 -6.02719635e-02 1.29919201e-01 4.01003007e-03
2.03203354e-02 3.36014807e-01 1.37088656e-01 2.67680407e-01
-5.40508509e-01 1.55185103e-01 5.60903430e-01 1.09912372e+00
9.75364685e-01 4.13386151e-02 -1.85852066e-01 6.50854468e-01
2.36370295e-01 5.03587604e-01 -7.54549727e-02 -8.98554146e-01
3.62670451e-01 1.11749446e+00 -2.01330334e-01 -1.53422034e+00
-1.61282569e-01 -5.47253005e-02 -9.40494597e-01 1.04497420e-02
2.76819378e-01 3.58842939e-01 -1.03130388e+00 1.42005992e+00
3.35641026e-01 -1.63815200e-01 -9.81359929e-02 1.10735309e+00
1.27478659e+00 8.62770259e-01 1.32016480e-01 -1.51807651e-01
1.85884273e+00 -1.03873336e+00 -7.87557185e-01 -5.68022251e-01
3.35104436e-01 -7.17212141e-01 1.17423773e+00 8.26635435e-02
-9.16357040e-01 -6.33368134e-01 -8.06131303e-01 -3.26197565e-01
-5.32611787e-01 1.88174546e-01 9.03210402e-01 3.94700229e-01
-8.17906797e-01 -4.58810739e-02 -4.84136134e-01 -7.58473933e-01
3.33205253e-01 -1.09057114e-01 -4.47729766e-01 -6.16739333e-01
-8.35303664e-01 6.53664649e-01 5.37900627e-01 2.33091325e-01
-8.24142575e-01 -1.90786749e-01 -1.19823158e+00 1.23999648e-01
6.58283412e-01 -8.07092667e-01 1.06573033e+00 -1.14172447e+00
-7.13214457e-01 1.34469020e+00 -4.03018385e-01 -1.30962849e-01
-5.66976331e-03 -3.65175912e-03 -4.56886530e-01 3.64217877e-01
2.77814567e-01 5.21605670e-01 7.45586097e-01 -1.58094585e+00
-6.99746013e-01 -4.58534092e-01 5.02835155e-01 5.60410380e-01
-1.29743842e-02 3.62545639e-01 -9.31827843e-01 -3.59585881e-01
4.95517313e-01 -5.52961469e-01 -1.90280259e-01 2.06446096e-01
-5.94072402e-01 -1.69915825e-01 1.05935788e+00 -4.90429699e-01
1.17990124e+00 -2.39593887e+00 -2.32790351e-01 1.50099680e-01
4.40902501e-01 1.19108066e-01 -1.72061801e-01 6.96568191e-01
-2.45245457e-01 4.67408001e-02 -3.37465972e-01 3.65653485e-02
-3.58524382e-01 3.21676910e-01 -5.07825494e-01 3.34871709e-01
5.75732365e-02 9.60754633e-01 -1.02437127e+00 -7.44837701e-01
6.22206807e-01 2.10789531e-01 -2.44368359e-01 2.85736233e-01
-2.36082733e-01 2.78418779e-01 -6.84224308e-01 6.79621458e-01
5.08981466e-01 -6.44133389e-01 3.87513787e-02 -3.86529714e-01
-5.02500273e-02 -3.50924246e-02 -1.15848529e+00 1.24455702e+00
-5.98011240e-02 8.82738888e-01 -1.86297283e-01 -1.23656511e+00
1.13253629e+00 3.19030927e-03 3.65586936e-01 -7.21495450e-01
8.59374553e-03 -2.72529393e-01 -4.71740402e-02 -8.15644383e-01
4.33159679e-01 -4.46225740e-02 -8.30847546e-02 3.76416564e-01
-4.54674274e-01 -5.00495255e-01 4.28490132e-01 6.12421393e-01
6.48900032e-01 -2.74970025e-01 5.92800975e-01 4.50680666e-02
6.55574143e-01 4.47600514e-01 2.37228259e-01 6.92877769e-01
-8.95576999e-02 6.34128809e-01 4.52991366e-01 -6.41768217e-01
-7.22999394e-01 -1.17831182e+00 2.58547336e-01 8.76229823e-01
8.47653866e-01 -6.53663576e-01 -5.87530434e-01 -4.46060807e-01
-2.86258519e-01 7.24556923e-01 -4.62876379e-01 -1.32872807e-02
-4.04349774e-01 -3.70164931e-01 -9.43774078e-03 4.16079968e-01
8.74389291e-01 -1.33105016e+00 -6.66732728e-01 -3.33613515e-01
-4.71922934e-01 -1.33273304e+00 -2.93606550e-01 -3.61731023e-01
-7.96709836e-01 -1.34402955e+00 -3.27660590e-01 -1.29397333e+00
1.21021152e+00 1.02885342e+00 1.19523942e+00 4.16263729e-01
-3.70298356e-01 7.10228622e-01 -5.03089130e-01 -4.88522589e-01
-1.41249493e-01 -6.70012236e-01 -2.31646031e-01 2.71057576e-01
3.85853499e-01 -1.30064338e-01 -4.63686675e-01 3.70875716e-01
-1.05809844e+00 5.21919072e-01 5.47561705e-01 4.61926192e-01
7.47006118e-01 6.23739183e-01 -8.19622427e-02 -1.11741579e+00
4.71319795e-01 -2.03208983e-01 -4.10366535e-01 6.12138391e-01
-1.06233887e-01 -1.78407431e-01 4.90926355e-01 -8.66927132e-02
-1.06961715e+00 1.60268053e-01 3.13255161e-01 -3.69340748e-01
-3.37937862e-01 5.23417890e-01 -4.55789179e-01 1.26174703e-01
4.73269284e-01 5.76385319e-01 -2.43269667e-01 -1.32320344e-01
6.86331689e-01 4.58537519e-01 6.28231764e-01 -3.37971568e-01
7.80650377e-01 7.30699241e-01 1.49606494e-02 -1.08494604e+00
-1.13892853e+00 -7.13826537e-01 -7.11023152e-01 -5.85160792e-01
9.67583179e-01 -6.72936678e-01 -4.30245072e-01 3.48284721e-01
-1.22488821e+00 -9.85851437e-02 -2.60109574e-01 2.58795649e-01
-4.76668298e-01 5.61495006e-01 -2.05206841e-01 -7.73915231e-01
9.41638872e-02 -9.30517018e-01 1.04878342e+00 4.45396006e-01
-5.10437265e-02 -9.35407102e-01 -4.90335256e-01 6.06202483e-01
9.10769999e-02 3.20888162e-01 1.25535989e+00 -2.46335819e-01
-8.97601902e-01 -1.45553276e-01 -7.61269867e-01 1.10698745e-01
5.40300488e-01 5.21064326e-02 -7.58348346e-01 4.63284552e-02
1.98774591e-01 5.93685471e-02 4.93269324e-01 3.93127352e-01
1.38180923e+00 -2.98654884e-01 -4.47794110e-01 3.17411274e-01
1.51617789e+00 5.36543846e-01 7.81079412e-01 9.55829024e-02
9.70687449e-01 1.10469091e+00 7.19253838e-01 7.84985125e-02
5.73166728e-01 5.54349959e-01 5.13971746e-01 -4.69010323e-01
-3.28267157e-01 -4.83742565e-01 1.13515981e-01 7.13380516e-01
-9.93015096e-02 -3.06361705e-01 -9.54016864e-01 4.77800339e-01
-1.82708120e+00 -9.84703481e-01 -5.58961749e-01 1.92927313e+00
2.49046966e-01 -2.58258045e-01 -9.21179280e-02 8.67010579e-02
1.03639174e+00 1.84064090e-01 -5.37384808e-01 -2.13149935e-01
-1.52622610e-01 -4.09899086e-01 1.11746751e-01 1.76997066e-01
-1.04678893e+00 1.19379961e+00 6.37489605e+00 6.02056563e-01
-1.02294588e+00 -4.27901030e-01 7.12430835e-01 5.63013792e-01
-1.70621231e-01 3.83736551e-01 -5.02645552e-01 1.04565293e-01
-1.17098782e-02 -3.99893761e-01 3.91210377e-01 9.98914540e-01
1.00117266e-01 -3.68260473e-01 -1.18174708e+00 1.53271258e+00
4.67941821e-01 -1.08314216e+00 5.32205343e-01 -1.19403675e-01
5.24341643e-01 -5.69881737e-01 -5.33916913e-02 -6.99836761e-02
1.33769736e-01 -9.66607809e-01 6.57320201e-01 4.59923029e-01
5.94813645e-01 -3.49989951e-01 4.98961776e-01 4.91256714e-01
-1.50180745e+00 8.32281262e-02 -5.59447110e-01 -3.47780645e-01
2.68443704e-01 5.59733212e-01 -7.10838437e-01 5.77128470e-01
6.78382277e-01 8.08293760e-01 -7.99410164e-01 1.10550761e+00
-5.53071856e-01 2.83661515e-01 1.26662195e-01 -1.91026181e-01
9.35961455e-02 -3.58889282e-01 3.39577436e-01 8.39950502e-01
8.37268904e-02 5.42852104e-01 2.72016883e-01 9.62389410e-01
1.10539816e-01 1.90847129e-01 -9.18890774e-01 -2.28621915e-01
2.78189719e-01 1.35623801e+00 -1.12812579e+00 -5.41369021e-01
-6.18629813e-01 9.70268726e-01 1.75380722e-01 5.28599679e-01
-5.32028973e-01 -4.22861427e-01 4.13104147e-01 2.90346920e-01
-8.82565305e-02 -3.76895696e-01 -2.37935156e-01 -1.10536122e+00
2.08994374e-02 -8.10437322e-01 6.73981965e-01 -1.45108390e+00
-1.21276128e+00 4.64840114e-01 2.66870499e-01 -1.28138101e+00
-6.58713356e-02 -6.08757436e-01 -6.43364131e-01 4.96552527e-01
-1.22604370e+00 -1.01176357e+00 -7.87978709e-01 7.63289511e-01
6.86634839e-01 2.33134508e-01 5.05941272e-01 -1.12500014e-02
-3.12055916e-01 -1.44846186e-01 -3.96741450e-01 4.67742294e-01
2.03277320e-01 -9.83195126e-01 3.73329461e-01 1.05079854e+00
6.46387279e-01 8.02652836e-01 5.92860699e-01 -5.52100658e-01
-1.40743923e+00 -1.09408414e+00 9.04005051e-01 -4.39658672e-01
4.70975846e-01 -5.18661737e-01 -9.57670987e-01 7.24090993e-01
6.24898970e-02 -1.12495221e-01 3.29835624e-01 -2.88316578e-01
-2.03553349e-01 -1.49688333e-01 -9.09753799e-01 9.31365013e-01
1.12258124e+00 -6.30673766e-01 -6.91604912e-01 4.20875758e-01
5.90093195e-01 -3.03755760e-01 -2.56755799e-01 3.37617040e-01
2.19422743e-01 -1.07415056e+00 9.90107536e-01 -5.73633552e-01
2.51057029e-01 -6.18078947e-01 -2.29558721e-01 -8.66408408e-01
-3.81802171e-01 -6.26850426e-02 2.78891861e-01 1.25374830e+00
-3.12723182e-02 -3.19807470e-01 7.40595043e-01 8.71998906e-01
-1.62401479e-02 -4.59095538e-01 -2.86618114e-01 -5.47671080e-01
-6.27096176e-01 -4.88761961e-01 4.60556835e-01 9.15731907e-01
-6.61011264e-02 5.61952591e-01 -2.51897156e-01 2.76076943e-01
6.07909560e-01 6.90988302e-01 1.01646733e+00 -1.01832211e+00
1.12699986e-01 -3.35218042e-01 -6.60862386e-01 -1.35590339e+00
-1.97261751e-01 -7.70355761e-01 8.30046274e-03 -2.19585395e+00
6.33599758e-01 -1.19531728e-01 1.30857974e-01 5.23971796e-01
-2.30946034e-01 2.06137791e-01 4.54373509e-01 2.95530409e-01
-8.72968256e-01 3.45779061e-01 1.67607236e+00 -3.03750455e-01
8.86335894e-02 -4.90046069e-02 -9.56291974e-01 9.20038402e-01
6.81130826e-01 -2.24408686e-01 -7.23567843e-01 -4.37271386e-01
2.36043528e-01 1.14469700e-01 6.24628603e-01 -6.92592561e-01
4.25897568e-01 -5.27972460e-01 3.12945902e-01 -7.47569382e-01
2.36789733e-01 -9.68563199e-01 1.00070149e-01 3.51459593e-01
-1.48432609e-02 1.67979792e-01 3.74593027e-03 4.34477836e-01
-5.71437716e-01 -1.82628781e-01 5.89948654e-01 -4.41721976e-01
-1.47399378e+00 3.29238057e-01 -4.18406516e-01 -1.45186350e-01
1.28168416e+00 -5.91038406e-01 -4.39235538e-01 -8.06027472e-01
-5.49212098e-01 3.64181578e-01 4.95533139e-01 4.88397151e-01
1.03948832e+00 -1.27187610e+00 -4.31791812e-01 2.59436011e-01
5.97909808e-01 1.88832566e-01 5.35003066e-01 3.37579042e-01
-6.00186944e-01 3.94826621e-01 -1.05788194e-01 -6.83674514e-01
-1.54511285e+00 8.15640926e-01 9.01545137e-02 2.04672337e-01
-7.06265509e-01 5.73034167e-01 1.08807313e+00 -1.07097134e-01
1.36382163e-01 -3.66581559e-01 -4.56174076e-01 -2.89855331e-01
5.93278050e-01 6.51898012e-02 -3.57510775e-01 -9.73277330e-01
-4.83392894e-01 8.35033774e-01 1.72701061e-01 2.23533615e-01
9.48235750e-01 -3.94433022e-01 -4.93851185e-01 4.58055377e-01
9.27555919e-01 -1.21722743e-01 -7.81595647e-01 -3.34961683e-01
-1.05334520e-01 -6.96636260e-01 -2.30841756e-01 -5.29879808e-01
-1.15918040e+00 1.01598203e+00 1.72405243e-01 6.27541661e-01
1.43812799e+00 5.89668453e-01 2.79669583e-01 3.73114169e-01
4.48592752e-01 -5.57416379e-01 5.13143420e-01 4.06046391e-01
9.99325633e-01 -1.16761374e+00 2.66837955e-01 -1.00452912e+00
-1.00122190e+00 1.14218640e+00 7.29721189e-01 6.98316470e-02
4.43469673e-01 -1.59721121e-01 1.46123931e-01 -8.54194760e-01
-2.93783277e-01 -4.84693199e-01 4.90621328e-01 7.53881514e-01
1.87604681e-01 1.00336649e-01 4.10288349e-02 1.75093547e-01
-2.70704716e-01 -5.00044882e-01 4.32670146e-01 7.34512389e-01
-6.46293521e-01 -9.28100586e-01 -3.96620244e-01 3.34508121e-01
-3.36390175e-02 -1.23114660e-01 -7.32862115e-01 7.52287865e-01
4.93151136e-02 1.21429944e+00 2.35940635e-01 -2.05615178e-01
4.70018029e-01 -4.30644542e-01 4.51888740e-01 -9.31664169e-01
5.21873385e-02 -2.92315006e-01 -1.92008823e-01 -4.93667871e-01
-7.56574273e-01 -4.74211693e-01 -1.27082217e+00 -9.36533436e-02
-1.88136294e-01 1.58193465e-02 6.61622822e-01 7.93427467e-01
2.83636600e-02 2.94429630e-01 5.56939840e-01 -3.72227460e-01
5.45043528e-01 -5.50110936e-01 -7.50221610e-01 7.45128095e-01
7.35550672e-02 -5.00608981e-01 -1.62493512e-01 5.66867173e-01] | [10.36331844329834, 1.5292459726333618] |
6933b067-1e5f-4a51-8bb0-aa1d5f0ab2ee | distributional-reinforcement-learning-with-2 | 2003.10903 | null | https://arxiv.org/abs/2003.10903v2 | https://arxiv.org/pdf/2003.10903v2.pdf | Distributional Reinforcement Learning with Ensembles | It is well known that ensemble methods often provide enhanced performance in reinforcement learning. In this paper, we explore this concept further by using group-aided training within the distributional reinforcement learning paradigm. Specifically, we propose an extension to categorical reinforcement learning, where distributional learning targets are implicitly based on the total information gathered by an ensemble. We empirically show that this may lead to much more robust initial learning, a stronger individual performance level, and good efficiency on a per-sample basis. | ['Karl-Olof Lindahl', 'Björn Lindenberg', 'Jonas Nordqvist'] | 2020-03-24 | null | null | null | null | ['distributional-reinforcement-learning'] | ['methodology'] | [ 1.38421506e-01 7.56280422e-02 -2.58955508e-01 -5.14694035e-01
-9.45020139e-01 -6.81326687e-01 7.46057987e-01 2.91329265e-01
-7.56697655e-01 1.26679027e+00 3.33164722e-01 -2.94420600e-01
-3.82415742e-01 -9.56435025e-01 -5.43981493e-01 -8.41940105e-01
-2.64041573e-01 4.21528071e-01 -1.46254674e-01 -4.06116635e-01
3.35829705e-01 4.09750819e-01 -1.58767796e+00 1.13477468e-01
1.15556347e+00 6.48607075e-01 -1.83471203e-01 4.39395189e-01
4.45983410e-02 8.91573906e-01 -9.04426098e-01 -2.81280249e-01
7.56619796e-02 -4.92224544e-01 -6.11149490e-01 3.32224928e-02
1.92256853e-01 -3.01379859e-01 3.76352593e-02 9.61264074e-01
6.11654758e-01 5.62349856e-01 1.01807702e+00 -1.08200157e+00
-3.93775195e-01 1.01675427e+00 -4.88758236e-01 -1.41305448e-02
3.68409008e-01 2.74206907e-01 1.13361728e+00 -7.19385087e-01
2.17617705e-01 1.27441466e+00 4.32081908e-01 5.82129598e-01
-1.70241725e+00 -7.68772781e-01 2.26615012e-01 1.85011744e-01
-9.42809582e-01 -1.72799453e-01 8.60413373e-01 -3.32062453e-01
8.43396187e-01 -1.25367835e-01 3.90653372e-01 9.53035831e-01
-3.28421593e-03 9.36210155e-01 1.55447614e+00 -6.06682539e-01
6.13059521e-01 2.47590337e-02 2.98203379e-01 3.79339606e-01
1.29298255e-01 7.11286724e-01 -2.70350128e-01 -1.56982303e-01
6.53890193e-01 -1.95502922e-01 -1.19518526e-01 -6.25903547e-01
-9.43645716e-01 1.29974425e+00 3.82834435e-01 3.80567074e-01
-5.85468411e-01 3.21074337e-01 4.84218031e-01 7.33916938e-01
5.23583055e-01 6.27989352e-01 -3.97837609e-01 -3.44988674e-01
-7.60837972e-01 6.62566960e-01 7.42917717e-01 3.76333475e-01
9.86740768e-01 4.41105872e-01 -1.74400911e-01 9.59236145e-01
1.05438165e-01 5.20778239e-01 5.10775268e-01 -1.10768354e+00
3.15584123e-01 3.92863780e-01 2.36069020e-02 -3.90827090e-01
-3.16213846e-01 -4.66057211e-01 -6.37053907e-01 6.75991535e-01
3.64141375e-01 -5.70346594e-01 -7.47116029e-01 2.10381818e+00
3.12482387e-01 2.53799349e-01 3.22202116e-01 3.67380738e-01
2.77423590e-01 2.29732588e-01 3.55171442e-01 -4.98301476e-01
7.45951355e-01 -6.51535690e-01 -6.60432637e-01 1.73036054e-01
8.69470119e-01 -2.94757754e-01 1.00884378e+00 7.40424573e-01
-8.23359311e-01 -6.89165592e-01 -1.02454650e+00 7.13753521e-01
-2.70527631e-01 -2.68915683e-01 5.93447566e-01 8.92387867e-01
-9.86259997e-01 9.35569108e-01 -3.98420572e-01 2.92738378e-02
2.73677737e-01 5.86525917e-01 -1.99896336e-01 2.71623433e-02
-1.26532936e+00 1.00861740e+00 7.07798421e-01 -5.97246706e-01
-9.61192429e-01 -6.17216706e-01 -8.50112021e-01 -9.94267464e-02
5.78571260e-01 -5.72919667e-01 1.53925765e+00 -9.06530380e-01
-1.83424604e+00 1.44376025e-01 1.33879215e-01 -6.66586280e-01
3.72809261e-01 -1.97940186e-01 -1.79067731e-01 -2.15347394e-01
-2.70223647e-01 5.50460875e-01 7.27766335e-01 -1.32041848e+00
-6.69572890e-01 -6.41092002e-01 1.42465636e-01 3.74434322e-01
-3.79717350e-01 -2.70387977e-01 6.49679542e-01 -8.15458298e-01
-5.19059539e-01 -8.20635140e-01 -5.49506009e-01 -9.07409608e-01
1.68631360e-01 -9.01047647e-01 4.92343277e-01 -2.32811615e-01
1.23276389e+00 -2.22049356e+00 3.92764956e-01 5.47239780e-01
2.23465607e-01 1.64297059e-01 -2.59952813e-01 6.25621796e-01
-2.96168655e-01 -4.91352566e-02 -4.23120767e-01 -8.62259492e-02
1.40073001e-01 6.17780983e-01 -4.38764900e-01 1.83451399e-01
2.30444357e-01 8.46586227e-01 -1.08027589e+00 -2.07174316e-01
2.73841977e-01 -5.02762198e-02 -8.28912795e-01 3.82881641e-01
-2.47826681e-01 4.55411702e-01 -3.22704703e-01 2.87616044e-01
2.91862041e-01 2.59280771e-01 3.03275436e-01 2.18405202e-01
1.04541123e-01 1.81783065e-01 -1.40514469e+00 1.34235191e+00
-3.73353958e-01 1.61505401e-01 -2.18150377e-01 -1.56789434e+00
9.53370631e-01 3.04396868e-01 6.55096471e-01 -5.99653244e-01
4.18453366e-02 2.22444534e-01 3.86514336e-01 -9.39617380e-02
5.83888590e-01 -4.65698451e-01 -1.59803927e-01 7.25753009e-01
4.69914347e-01 -2.27165148e-01 2.42332026e-01 -1.23917900e-01
8.04274559e-01 1.57787070e-01 6.19697630e-01 -3.64722848e-01
3.86175662e-01 -2.58285254e-01 2.24751189e-01 1.00062788e+00
-1.05869688e-01 -1.86680540e-01 2.42033675e-01 7.26630166e-02
-1.04174268e+00 -1.16151917e+00 -2.04762995e-01 1.44838631e+00
-3.11559349e-01 -3.63218069e-01 -6.88284099e-01 -1.15233254e+00
4.21327591e-01 1.06537592e+00 -7.48790741e-01 -2.90699989e-01
-4.97879386e-01 -6.74431980e-01 4.94982809e-01 9.42617655e-01
2.17296720e-01 -1.35833776e+00 -3.16420645e-01 3.98195744e-01
3.01239640e-01 -4.40746546e-01 -1.27999559e-01 7.42961645e-01
-1.20957136e+00 -8.16314220e-01 -7.18442380e-01 -4.12824959e-01
3.05318922e-01 -1.50832668e-01 1.15889800e+00 -7.62306824e-02
3.59292984e-01 6.04122818e-01 -4.69005436e-01 -6.50972784e-01
-5.63305378e-01 -3.86741012e-02 5.03058195e-01 -3.17016423e-01
3.89920741e-01 -6.94623470e-01 -1.46291062e-01 1.01076417e-01
-8.92928600e-01 -5.57455540e-01 4.59352016e-01 1.21235931e+00
2.61445552e-01 1.20328508e-01 1.16419280e+00 -1.01991999e+00
1.22151005e+00 -5.20716071e-01 -3.76027107e-01 8.29380155e-02
-1.03415167e+00 4.92525786e-01 6.56386256e-01 -5.23170769e-01
-8.63903761e-01 -1.84919938e-01 -3.20324361e-01 -1.91657111e-01
-4.65441614e-01 8.73656809e-01 1.14745021e-01 -1.46123275e-01
1.09672558e+00 3.32613677e-01 3.70572656e-01 -1.84070215e-01
6.70941353e-01 7.29706109e-01 1.17623068e-01 -1.04536760e+00
6.45966768e-01 -1.82616279e-01 -2.25697696e-01 -6.49412155e-01
-6.03664219e-01 -3.62052649e-01 -5.28286517e-01 -3.38827223e-01
3.19078982e-01 -7.39970744e-01 -6.35499656e-01 2.33419180e-01
-5.02262950e-01 -7.71286070e-01 -6.18891180e-01 6.75079048e-01
-1.09373653e+00 2.44765997e-01 -1.17086351e-01 -1.05747616e+00
-5.70346862e-02 -1.12292111e+00 6.87996864e-01 1.81105331e-01
-2.73042023e-01 -1.22249329e+00 5.50023913e-01 -2.18752161e-01
2.21638218e-01 3.05761900e-02 9.01156902e-01 -1.18440354e+00
1.51291311e-01 2.91046113e-01 2.27317348e-01 6.85982704e-01
2.12224707e-01 -1.35112271e-01 -8.90208483e-01 -3.59382063e-01
-1.41387507e-01 -7.79515088e-01 1.07288182e+00 3.55909467e-01
1.20365250e+00 -1.24512590e-01 1.26340106e-01 5.08844666e-02
1.42425573e+00 3.52298975e-01 4.86783832e-01 2.22484872e-01
4.56677318e-01 7.19467342e-01 7.54485786e-01 5.57920396e-01
6.17780723e-02 6.19166136e-01 1.95492864e-01 1.49827749e-01
1.72239959e-01 -1.30943865e-01 6.00497305e-01 6.55633271e-01
-2.90312320e-01 5.94920516e-02 -7.02187777e-01 4.14217949e-01
-1.99563122e+00 -1.29157293e+00 3.17735553e-01 2.16829395e+00
1.10971916e+00 -3.98047380e-02 8.16946328e-01 3.84067625e-01
4.64381486e-01 -2.66612824e-02 -5.64338505e-01 -6.34664714e-01
-5.41438870e-02 6.74133003e-01 3.61972570e-01 5.01507878e-01
-9.32021201e-01 6.84669793e-01 7.93472052e+00 1.11478245e+00
-7.81296194e-01 -1.29924834e-01 4.26344842e-01 1.73535883e-01
-5.86078942e-01 -3.00558776e-01 -7.17737675e-01 3.07024807e-01
1.14891994e+00 -4.37015623e-01 4.38488960e-01 8.50730658e-01
-1.32873133e-01 -4.78971787e-02 -1.21130455e+00 7.79110432e-01
-1.66391268e-01 -1.07056963e+00 1.83503330e-03 4.21950787e-01
9.65247691e-01 -1.42073289e-01 1.36397064e-01 9.63676035e-01
8.96464527e-01 -1.15675008e+00 2.13358447e-01 2.89881587e-01
5.25224566e-01 -1.33574581e+00 7.64919579e-01 6.15739405e-01
-7.32588470e-01 -3.73990744e-01 -2.38228753e-01 -2.29512855e-01
-2.14607790e-01 4.73217368e-01 -9.13459778e-01 8.61869633e-01
1.80662513e-01 8.30448449e-01 -3.86734724e-01 1.02546275e+00
-2.93173939e-01 9.85297143e-01 -2.44206473e-01 -4.38439935e-01
4.43846136e-01 -3.20707023e-01 3.57476830e-01 1.00676739e+00
1.40822187e-01 1.62280843e-01 4.57043648e-01 4.03483480e-01
6.19701706e-02 3.59321117e-01 -1.03692019e+00 4.73239012e-02
3.26281875e-01 1.04502857e+00 -3.29651147e-01 -5.59601724e-01
-1.80369020e-01 3.47111851e-01 6.91246688e-01 3.17679316e-01
-7.08683670e-01 -2.19498351e-01 5.22514999e-01 -3.87742907e-01
4.73842144e-01 -1.33248925e-01 -1.50048375e-01 -8.65594625e-01
-4.38447744e-01 -1.30385005e+00 5.55432320e-01 -1.39467716e-01
-1.68391764e+00 2.64553517e-01 2.64897853e-01 -1.11431742e+00
-9.16987419e-01 -8.69518995e-01 -5.36946177e-01 7.15329647e-01
-1.19975257e+00 -5.26912451e-01 1.46231502e-01 7.13923991e-01
4.27775353e-01 -4.87346739e-01 1.02994061e+00 -9.09105167e-02
-2.32518390e-01 7.85291851e-01 3.59911412e-01 5.63526861e-02
7.48555124e-01 -1.98320043e+00 -3.07197541e-01 2.57339358e-01
5.50032139e-01 4.12040502e-01 8.77713859e-01 -4.09561872e-01
-8.54520798e-01 -8.93949032e-01 1.88113973e-01 -4.09585327e-01
7.07559764e-01 1.27343521e-01 -9.73459899e-01 5.00632107e-01
3.99040073e-01 -5.12898386e-01 1.08761537e+00 7.87101269e-01
-2.36842170e-01 -3.49395014e-02 -1.09778392e+00 5.71726739e-01
6.93558991e-01 -5.70970356e-01 -1.07996225e+00 6.94010481e-02
3.98589402e-01 -1.20842196e-01 -1.30529439e+00 5.84858119e-01
3.36679965e-01 -8.50163460e-01 8.06887627e-01 -7.42190599e-01
4.24851418e-01 -4.10452373e-02 -3.09369922e-01 -2.13567090e+00
-2.84481078e-01 -3.57870013e-01 -2.70094842e-01 1.01479447e+00
3.40487570e-01 -7.25680351e-01 7.66746998e-01 1.83175266e-01
-3.06559149e-02 -8.55863571e-01 -7.83028543e-01 -9.69906151e-01
6.90223515e-01 -4.98989254e-01 5.85228682e-01 8.99091244e-01
4.91571426e-01 4.32192415e-01 -3.12111229e-01 -3.64658684e-01
7.85852730e-01 1.44149110e-01 8.46009016e-01 -1.29282010e+00
-5.25168478e-01 -7.29976594e-01 -4.29861575e-01 -9.77838695e-01
5.06790936e-01 -7.55294800e-01 2.22056076e-01 -1.00195467e+00
9.36373100e-02 -4.71304566e-01 -7.87784874e-01 4.89284396e-01
-4.96065050e-01 9.90668759e-02 7.65658841e-02 -1.44029111e-01
-6.26139641e-01 7.20402420e-01 1.20727539e+00 8.54805335e-02
-2.55211562e-01 1.10315420e-01 -8.13985527e-01 5.35001576e-01
1.11043417e+00 -4.23289150e-01 -6.21140182e-01 7.06790611e-02
-1.40256479e-01 6.15877025e-02 1.17440760e-01 -9.27789390e-01
-1.45246694e-02 -3.10827881e-01 4.30461735e-01 -4.05539036e-01
9.14029554e-02 -5.27307987e-01 -3.06859493e-01 4.83539790e-01
-6.42942607e-01 1.34855762e-01 2.41178349e-01 7.01566041e-01
-3.84070665e-01 -5.74773848e-01 6.72604859e-01 -2.67649740e-02
-6.79718018e-01 4.19939719e-02 -3.88934731e-01 -2.47001648e-03
1.18692827e+00 -1.90963909e-01 6.65980503e-02 -4.48851973e-01
-8.07969332e-01 2.12368295e-01 2.14165568e-01 8.98553357e-02
5.55176735e-01 -1.52792084e+00 -9.39984620e-01 1.25752553e-01
1.10877447e-01 -4.75301176e-01 3.83703597e-02 6.81263149e-01
3.44689131e-01 2.26030454e-01 -3.23944241e-01 -6.34292603e-01
-1.12014067e+00 5.34024715e-01 3.58418673e-01 -4.45619911e-01
-2.28717566e-01 5.92873812e-01 9.45613012e-02 -7.59545028e-01
2.64285415e-01 -5.63213825e-02 -5.83354414e-01 2.91063040e-01
5.51954031e-01 3.22830200e-01 1.01651624e-03 -2.82627016e-01
-7.85566419e-02 2.59046584e-01 -1.06747665e-01 -4.65003639e-01
1.26727605e+00 3.32291007e-01 3.04405481e-01 7.62446404e-01
8.81329477e-01 9.73104611e-02 -1.28212297e+00 -2.48270214e-01
2.45103642e-01 -2.87487984e-01 2.27236282e-02 -8.89607430e-01
-7.60962486e-01 6.78368747e-01 5.35862148e-01 5.29907048e-01
1.07052481e+00 -2.16357514e-01 1.79405101e-02 5.85828781e-01
5.07558882e-01 -1.26454926e+00 2.89184958e-01 5.92290938e-01
8.07945788e-01 -1.32124913e+00 1.04764678e-01 1.35164261e-01
-7.21400499e-01 1.11410415e+00 6.72320843e-01 -5.69260538e-01
6.77926838e-01 2.49137402e-01 -1.18470974e-01 1.24088921e-01
-9.45318401e-01 -4.22920585e-01 2.05251575e-01 8.83548558e-01
6.07019722e-01 2.10408136e-01 -2.90353030e-01 4.20870513e-01
-2.49676272e-01 -7.73781016e-02 4.06669378e-01 7.87528574e-01
-7.29586840e-01 -1.58855247e+00 -2.49518916e-01 6.52238905e-01
-4.03067142e-01 2.32498348e-03 -2.68251687e-01 7.89748967e-01
-2.38845587e-01 9.87142742e-01 6.06755428e-02 -4.60700840e-01
4.19540316e-01 3.12865734e-01 9.15696859e-01 -7.13956058e-01
-8.58277798e-01 1.94460005e-02 1.00393854e-01 -4.13427591e-01
-8.30550969e-01 -7.23569334e-01 -1.12100506e+00 -3.01210970e-01
-4.20958370e-01 6.41491055e-01 2.77258813e-01 1.14283323e+00
8.60592127e-02 6.84513748e-01 9.79380131e-01 -6.21446788e-01
-1.35816038e+00 -1.22744071e+00 -7.39250183e-01 4.22771424e-01
2.64422178e-01 -9.74324226e-01 -3.84305000e-01 -2.87818462e-01] | [4.0936360359191895, 2.455345392227173] |
63583b2a-3b75-4856-88ea-55902d3a1260 | super-pixel-cloud-detection-using | 1810.08352 | null | http://arxiv.org/abs/1810.08352v1 | http://arxiv.org/pdf/1810.08352v1.pdf | Super-pixel cloud detection using Hierarchical Fusion CNN | Cloud detection plays a very important role in the process of remote sensing
images. This paper designs a super-pixel level cloud detection method based on
convolutional neural network (CNN) and deep forest. Firstly, remote sensing
images are segmented into super-pixels through the combination of SLIC and
SEEDS. Structured forests is carried out to compute edge probability of each
pixel, based on which super-pixels are segmented more precisely. Segmented
super-pixels compose a super-pixel level remote sensing database. Though cloud
detection is essentially a binary classification problem, our database is
labeled into four categories: thick cloud, cirrus cloud, building and other
culture, to improve the generalization ability of our proposed models.
Secondly, super-pixel level database is used to train our cloud detection
models based on CNN and deep forest. Considering super-pixel level remote
sensing images contain less semantic information compared with general object
classification database, we propose a Hierarchical Fusion CNN (HFCNN). It takes
full advantage of low-level features like color and texture information and is
more applicable to cloud detection task. In test phase, every super-pixel in
remote sensing images is classified by our proposed models and then combined to
recover final binary mask by our proposed distance metric, which is used to
determine ambiguous super-pixels. Experimental results show that, compared with
conventional methods, HFCNN can achieve better precision and recall. | ['Qi Tian', 'Han Liu', 'Dan Zeng'] | 2018-10-19 | null | null | null | null | ['cloud-detection'] | ['computer-vision'] | [ 5.08382201e-01 -7.26356626e-01 5.13337031e-02 -1.56706214e-01
-2.75132388e-01 -5.23712814e-01 3.40056807e-01 -6.35030270e-02
-4.18657690e-01 5.89104474e-01 -3.88358235e-01 -4.59511250e-01
8.27093050e-02 -1.65065253e+00 -2.97521502e-01 -8.40947986e-01
1.03026638e-02 1.20549470e-01 4.20127809e-01 4.08160426e-02
2.83753097e-01 8.18950057e-01 -1.76623225e+00 4.08951610e-01
1.16225767e+00 1.08909953e+00 6.64399922e-01 5.02376020e-01
-4.22822386e-01 4.74403411e-01 -3.15572858e-01 1.65260300e-01
6.00723863e-01 -2.75130838e-01 -6.22698724e-01 4.09550339e-01
3.41009468e-01 -3.70276034e-01 1.80127293e-01 1.69791341e+00
3.68565977e-01 -1.34831995e-01 5.41385174e-01 -8.94512117e-01
-6.20442092e-01 6.28810763e-01 -7.09917545e-01 5.31350493e-01
-3.66642624e-01 1.76662013e-01 7.60925949e-01 -1.24282539e+00
3.03194195e-01 1.13231695e+00 5.12974203e-01 1.83507428e-02
-5.49715400e-01 -1.06015515e+00 3.48781854e-01 2.07792148e-01
-1.72878146e+00 1.11831993e-01 5.27291059e-01 -4.57210749e-01
3.08574319e-01 4.58813697e-01 1.09908783e+00 -4.52935435e-02
-3.18378247e-02 6.46920264e-01 1.48627663e+00 -2.44033366e-01
1.89161897e-01 2.10824132e-01 1.18124463e-01 5.21315217e-01
6.32967532e-01 5.18932603e-02 6.77919611e-02 2.33493209e-01
9.56230998e-01 5.44521272e-01 -3.27952445e-01 2.98931539e-01
-8.34249675e-01 8.33274007e-01 9.95659411e-01 5.22925079e-01
-6.25696242e-01 -2.51432121e-01 -2.73655802e-01 5.87562518e-03
4.71823633e-01 -1.22935593e-01 -2.52772391e-01 6.71433449e-01
-1.10846853e+00 2.16533035e-01 2.46323690e-01 6.37086809e-01
1.17436767e+00 1.19090855e-01 -3.48646864e-02 6.27917051e-01
3.27678949e-01 7.29367852e-01 2.19844043e-01 -5.59830129e-01
4.81905967e-01 9.23583567e-01 7.62521103e-02 -1.09391868e+00
-1.30974337e-01 -9.26221251e-01 -1.20778084e+00 5.40688336e-01
-2.90114760e-01 4.23935130e-02 -1.32273376e+00 7.23127186e-01
3.70388925e-01 1.19205542e-01 1.01585284e-01 1.21893799e+00
1.05300593e+00 7.87439525e-01 7.00759217e-02 -3.68064851e-01
1.51144147e+00 -7.48401523e-01 -3.76542538e-01 -3.01248103e-01
-1.49828017e-01 -6.18137419e-01 7.05042422e-01 4.44209665e-01
-4.81497586e-01 -7.15566218e-01 -8.89491439e-01 3.58134985e-01
-8.06928158e-01 4.26128745e-01 7.77085721e-01 6.54218435e-01
-9.59151745e-01 2.09117427e-01 -3.31105083e-01 -2.26443559e-02
6.49669945e-01 1.15665734e-01 8.27098042e-02 -1.81470618e-01
-1.08373868e+00 5.20793438e-01 7.60939956e-01 7.05770731e-01
-9.62331295e-01 -1.74867481e-01 -4.81985182e-01 2.07375467e-01
2.73534268e-01 -4.54290271e-01 7.34719694e-01 -1.13369560e+00
-7.75557637e-01 8.02439630e-01 -3.86699401e-02 -7.98572451e-02
3.32827121e-01 -1.13654425e-02 -5.24673820e-01 2.61308759e-01
1.86305910e-01 4.31945860e-01 7.71364987e-01 -1.56311345e+00
-1.24346566e+00 -4.45553929e-01 4.72472087e-02 4.03816968e-01
-9.90469605e-02 9.74536687e-02 -4.24902886e-01 -5.17250001e-01
7.87912548e-01 -6.94023967e-01 -4.62762058e-01 -1.54674232e-01
-2.89075583e-01 -9.98503640e-02 1.11575186e+00 -6.73707068e-01
1.05973816e+00 -1.78948903e+00 -4.61069286e-01 4.74865764e-01
2.21682057e-01 5.92503726e-01 -1.04292398e-02 -2.74935327e-02
-3.65187204e-03 4.19698477e-01 -8.35739851e-01 3.71371269e-01
-5.70711017e-01 6.52125850e-02 -1.21988468e-01 1.98948279e-01
4.64232087e-01 5.70294619e-01 -7.35766292e-01 -7.98775375e-01
3.32918704e-01 3.62094223e-01 -2.89240703e-02 -3.19373496e-02
-1.74209714e-01 3.60387683e-01 -7.33092129e-01 1.32213438e+00
1.49509454e+00 -8.51435587e-02 -9.78382453e-02 -8.07618052e-02
-5.16223729e-01 -2.74522692e-01 -1.35476518e+00 8.05989563e-01
-3.93064588e-01 2.94043094e-01 2.48260304e-01 -8.61095428e-01
1.18368995e+00 -1.36965245e-01 2.76101857e-01 -5.77430844e-01
7.69581571e-02 2.72255629e-01 -1.90821618e-01 -6.29562795e-01
7.45682716e-01 -1.52169868e-01 4.31864768e-01 -5.69845363e-02
-6.08776510e-01 -4.40443516e-01 -9.07421485e-02 -1.51905343e-01
3.82771671e-01 1.33617474e-02 1.06448814e-01 -3.69341634e-02
6.75103009e-01 4.73523438e-01 7.46841371e-01 7.87110567e-01
-9.14196074e-02 6.36786103e-01 -2.40929782e-01 -6.04673326e-01
-7.16080546e-01 -8.63488495e-01 -4.24019009e-01 7.87658334e-01
4.63506490e-01 3.02989572e-01 -6.94303811e-01 -4.50645924e-01
5.99330217e-02 1.80704728e-01 -4.87403244e-01 3.81574243e-01
-3.17279279e-01 -1.26317894e+00 2.04465032e-01 4.34007794e-01
1.30618227e+00 -1.27485871e+00 -4.25945580e-01 1.97715566e-01
-3.68861347e-01 -9.89239156e-01 1.08068518e-01 1.55450299e-03
-1.05405188e+00 -1.07036328e+00 -5.00180423e-01 -8.62778425e-01
6.82426155e-01 1.13791633e+00 9.46653068e-01 6.28977120e-01
-3.62812936e-01 -4.17305022e-01 -6.94581270e-01 -5.83717644e-01
2.89543942e-02 -2.17824623e-01 -4.20626223e-01 2.97496051e-01
4.58442599e-01 -5.21192670e-01 -8.03914368e-01 6.73681451e-03
-1.19718707e+00 -3.06647215e-02 1.14126503e+00 4.56295818e-01
7.73948491e-01 7.49665320e-01 8.48455075e-03 -8.19272816e-01
1.84215590e-01 -4.74258482e-01 -8.78724873e-01 2.12238878e-01
-6.70173466e-01 -6.62613094e-01 2.41758972e-01 7.26611018e-02
-1.17236960e+00 3.88256699e-01 7.58603439e-02 -4.59061623e-01
-2.23761991e-01 6.89739466e-01 -2.83465236e-01 -3.34538966e-01
3.20040882e-01 6.56888843e-01 -4.58973318e-01 -4.04347211e-01
1.43646345e-01 1.26954448e+00 5.13696074e-01 -2.09787488e-02
1.05047095e+00 7.00627446e-01 -1.53921679e-01 -7.95593202e-01
-6.11697972e-01 -7.10769236e-01 -7.44993746e-01 -3.82048756e-01
1.27498507e+00 -1.36491203e+00 -6.15615785e-01 6.94797456e-01
-1.00046980e+00 8.48269686e-02 3.50281447e-01 5.75321555e-01
2.52813399e-01 2.45844141e-01 -4.37466443e-01 -1.42092896e+00
-7.07510531e-01 -8.14214885e-01 1.24682224e+00 5.16505301e-01
1.03382039e+00 -4.26311582e-01 -4.63195592e-01 4.89991337e-01
3.24130088e-01 2.89495796e-01 4.75889951e-01 -1.28095433e-01
-1.22744501e+00 -3.18006575e-02 -7.88468659e-01 7.68108368e-01
1.05961405e-01 2.87154287e-01 -9.79430377e-01 5.41321039e-02
-3.90379503e-02 1.42215610e-01 1.45311713e+00 3.45162630e-01
1.35785353e+00 -1.78822294e-01 -3.39791983e-01 6.82116985e-01
2.12474656e+00 3.46281350e-01 7.83527195e-01 4.46144521e-01
9.77167964e-01 3.76755893e-01 8.82915139e-01 3.42312634e-01
5.00187635e-01 -8.60261247e-02 8.47367048e-01 -3.94671649e-01
7.75079876e-02 1.86046064e-01 6.92099705e-02 6.49563849e-01
-6.18491352e-01 -6.22348785e-02 -1.05755889e+00 6.22844219e-01
-1.56861925e+00 -1.15940249e+00 -6.04206562e-01 1.90842462e+00
5.64209759e-01 7.06333369e-02 -2.26654142e-01 2.85270661e-01
1.12149537e+00 6.77012503e-02 -4.51562077e-01 1.86244309e-01
-5.34176469e-01 4.68634576e-01 7.78033018e-01 2.68732011e-01
-1.43096817e+00 1.09089005e+00 5.09239149e+00 1.03828752e+00
-1.43339384e+00 1.56001255e-01 6.86699212e-01 6.23758614e-01
-2.50122011e-01 1.41938463e-01 -8.22360277e-01 5.31973004e-01
-1.59579352e-01 3.76603514e-01 3.19623083e-01 7.45105445e-01
4.21070397e-01 -5.14191449e-01 -3.82958390e-02 9.03037012e-01
-1.64177522e-01 -1.24104667e+00 3.45744491e-01 9.17374343e-02
9.29212868e-01 2.89098114e-01 -1.28615215e-01 1.03459910e-01
5.32804847e-01 -9.83010948e-01 7.43063986e-01 6.09552205e-01
7.24710047e-01 -5.72442830e-01 1.01617742e+00 5.02769351e-01
-1.81700504e+00 -4.57377762e-01 -7.70541787e-01 -1.72180831e-01
-3.92061174e-01 1.01764917e+00 -6.68143511e-01 9.62908804e-01
9.02028680e-01 5.55286646e-01 -5.52502334e-01 1.24650502e+00
-1.06126115e-01 6.23150110e-01 -2.95479894e-01 5.91317602e-02
3.52493405e-01 -7.68101454e-01 1.84536025e-01 1.13444197e+00
4.78874177e-01 5.47886610e-01 5.41462421e-01 1.02101290e+00
1.07990317e-01 9.88560915e-02 -3.58559668e-01 5.97911142e-02
7.24095821e-01 1.66229737e+00 -1.05425537e+00 -6.29801393e-01
-1.55118495e-01 6.08800709e-01 -2.49040365e-01 3.21831763e-01
-5.97612083e-01 -4.35604423e-01 3.64546508e-01 -6.83126971e-02
5.19839406e-01 -1.12724647e-01 -5.53209841e-01 -1.22335398e+00
1.82197183e-01 -6.37042463e-01 3.34009856e-01 -1.11486220e+00
-9.25145149e-01 6.33995831e-01 -2.99847960e-01 -1.65136957e+00
5.09375811e-01 -5.01620770e-01 -8.23660970e-01 1.17336667e+00
-2.10547280e+00 -1.44606531e+00 -1.14388692e+00 5.45453250e-01
4.40168232e-01 2.04171970e-01 4.67794865e-01 2.05316186e-01
-3.70424330e-01 -1.63410693e-01 6.28467053e-02 4.80652034e-01
-1.90710455e-01 -1.05380690e+00 7.71246478e-02 1.28198218e+00
-1.55255705e-01 2.75403619e-01 1.84571117e-01 -8.88649642e-01
-9.63993669e-01 -1.76046515e+00 4.32998925e-01 1.19521387e-01
8.83345008e-02 -1.84319764e-02 -8.59824121e-01 1.82178482e-01
-1.99569389e-01 1.58607066e-01 2.59363294e-01 -4.15105134e-01
-1.88614830e-01 -4.62424129e-01 -1.31369591e+00 9.95779335e-02
9.92405772e-01 -4.85594690e-01 -3.35700512e-01 6.00377977e-01
6.40077174e-01 -1.59984484e-01 -5.71061790e-01 7.74477184e-01
3.61600965e-01 -1.04879725e+00 8.55456769e-01 9.46167670e-03
5.45251191e-01 -8.22957098e-01 -1.46427840e-01 -9.53686655e-01
-5.86293936e-01 3.07364672e-01 7.18534470e-01 1.14444613e+00
1.61188662e-01 -4.64086473e-01 8.50697696e-01 -2.70119250e-01
-1.65954784e-01 -3.66454989e-01 -5.55102229e-01 -6.20478511e-01
-4.30534929e-02 -2.64946282e-01 1.04399240e+00 1.21207440e+00
-9.15837169e-01 3.51369642e-02 7.48704225e-02 8.76923859e-01
6.65016592e-01 8.31745863e-01 5.84750175e-01 -1.60315895e+00
1.50361061e-01 -5.79259455e-01 -2.72236794e-01 -6.03142023e-01
-3.35924417e-01 -7.67826974e-01 2.04593226e-01 -1.87378907e+00
5.62964559e-01 -9.43638384e-01 -3.08618039e-01 5.28398514e-01
-6.15281224e-01 7.77745724e-01 2.79394865e-01 5.63946426e-01
-2.29006454e-01 3.77350450e-01 1.50831544e+00 -3.99854094e-01
-1.86844260e-01 1.78895608e-01 -4.82093185e-01 5.06642640e-01
9.44926083e-01 -3.04435641e-01 3.04285884e-02 -5.26904106e-01
-1.28133669e-02 9.58436802e-02 6.03604376e-01 -1.35128450e+00
2.18761742e-01 -4.84708697e-01 8.33292484e-01 -1.14808500e+00
2.66914181e-02 -8.92410517e-01 2.34334767e-01 7.47896731e-01
3.15234125e-01 -2.49828547e-01 -3.15766111e-02 3.37232351e-01
-4.27787632e-01 -2.28305072e-01 7.43742228e-01 -6.36056960e-01
-1.07628179e+00 5.95100522e-01 -2.83153266e-01 -4.94931906e-01
9.20300663e-01 -6.61195636e-01 -1.32911131e-01 -3.11565083e-02
-7.28524029e-01 2.10502625e-01 2.37275392e-01 1.12499207e-01
8.17587852e-01 -1.09300101e+00 -9.46359098e-01 3.09088230e-01
2.64338374e-01 3.84359151e-01 3.45284700e-01 4.32528734e-01
-9.66619551e-01 3.40720229e-02 -2.19977707e-01 -9.07146096e-01
-1.32594585e+00 3.40047836e-01 4.36939031e-01 5.87013513e-02
-2.24135265e-01 8.77622187e-01 1.35455325e-01 -4.66143459e-01
-3.32308173e-01 -4.70104039e-01 -6.16425455e-01 1.41397461e-01
3.26153606e-01 3.18040848e-02 2.01738924e-01 -7.06870079e-01
-4.42680329e-01 8.57624054e-01 4.14453179e-01 8.77058655e-02
1.34847093e+00 -8.35689232e-02 -7.48074830e-01 -1.99277867e-02
6.38690054e-01 1.04173951e-01 -8.34109962e-01 -4.99669641e-01
-3.14387262e-01 -6.56571150e-01 4.20728385e-01 -8.41780365e-01
-1.58009613e+00 1.01154053e+00 1.07137060e+00 3.28633070e-01
1.56802309e+00 -2.54743040e-01 5.87690890e-01 2.74170846e-01
4.17997479e-01 -8.91923904e-01 -3.90854776e-01 5.39466083e-01
4.62547749e-01 -1.43460345e+00 1.99963719e-01 -8.64299953e-01
-3.18639189e-01 1.10495245e+00 7.41982043e-01 -1.63213804e-01
7.01777697e-01 4.94943857e-02 1.89649001e-01 -2.78553605e-01
-1.21935040e-01 -9.85532880e-01 -9.25865024e-02 5.70427060e-01
-1.29019946e-01 5.65196693e-01 -3.00338000e-01 3.55600089e-01
-1.51689306e-01 1.18387550e-01 4.71514016e-01 8.68776500e-01
-1.19802284e+00 -6.13512933e-01 -8.71542394e-01 6.64532781e-01
-3.34156007e-01 -5.21462858e-01 -2.94912159e-01 5.62357366e-01
9.05474186e-01 1.05281651e+00 4.45394725e-01 -7.11520970e-01
-1.40494078e-01 -4.12570566e-01 8.52574930e-02 -7.32634127e-01
-6.68229938e-01 6.43296316e-02 -3.29030275e-01 -3.44978087e-02
-8.08407247e-01 -3.99166256e-01 -1.16667676e+00 -1.26658201e-01
-7.21319497e-01 2.70726264e-01 7.55983889e-01 8.23701262e-01
-6.67103603e-02 4.33577836e-01 1.02646232e+00 -8.17578256e-01
-9.96554643e-02 -1.01407695e+00 -9.25575018e-01 3.64940912e-02
3.29460889e-01 -4.39583629e-01 -3.01084191e-01 7.89330080e-02] | [9.801094055175781, -1.7136873006820679] |
8871086e-6265-4afd-b0ef-c90ef5768e4f | improved-visual-relocalization-by-discovering | 1811.04370 | null | http://arxiv.org/abs/1811.04370v1 | http://arxiv.org/pdf/1811.04370v1.pdf | Improved Visual Relocalization by Discovering Anchor Points | We address the visual relocalization problem of predicting the location and
camera orientation or pose (6DOF) of the given input scene. We propose a method
based on how humans determine their location using the visible landmarks. We
define anchor points uniformly across the route map and propose a deep learning
architecture which predicts the most relevant anchor point present in the scene
as well as the relative offsets with respect to it. The relevant anchor point
need not be the nearest anchor point to the ground truth location, as it might
not be visible due to the pose. Hence we propose a multi task loss function,
which discovers the relevant anchor point, without needing the ground truth for
it. We validate the effectiveness of our approach by experimenting on
CambridgeLandmarks (large scale outdoor scenes) as well as 7 Scenes (indoor
scenes) using variousCNN feature extractors. Our method improves the median
error in indoor as well as outdoor localization datasets compared to the
previous best deep learning model known as PoseNet (with geometric
re-projection loss) using the same feature extractor. We improve the median
error in localization in the specific case of Street scene, by over 8m. | ['C. V. Jawahar', 'Girish Varma', 'Soham Saha'] | 2018-11-11 | null | null | null | null | ['outdoor-localization'] | ['robots'] | [-1.49457633e-01 -3.44032273e-02 7.88658932e-02 -5.09936392e-01
-7.76735723e-01 -9.11047995e-01 7.18467295e-01 4.09889668e-01
-1.00174844e+00 6.39954507e-01 2.06008270e-01 2.91686952e-02
-2.57193834e-01 -6.50206625e-01 -1.30207241e+00 -2.64254272e-01
-1.29346862e-01 5.68137705e-01 3.01366776e-01 -9.74649116e-02
2.86077142e-01 7.84052730e-01 -1.26873577e+00 -2.66174078e-01
5.04911542e-01 1.06540179e+00 2.69480377e-01 6.40131652e-01
4.43602115e-01 4.60230023e-01 -5.57314396e-01 -1.91255599e-01
5.19008279e-01 8.42045024e-02 -6.85778737e-01 5.12552559e-02
1.14788258e+00 -3.98462802e-01 -3.48997325e-01 7.89325058e-01
4.23643231e-01 4.45841730e-01 5.11679649e-01 -1.16178811e+00
-2.76295483e-01 -1.35803625e-01 -7.57512510e-01 1.72344863e-01
6.26684129e-01 -2.70149648e-01 9.82137978e-01 -9.83539045e-01
7.80611873e-01 8.57978582e-01 1.07201529e+00 -5.75364046e-02
-1.15541565e+00 -2.92343915e-01 3.26563001e-01 2.18281820e-01
-1.90366459e+00 -4.25618112e-01 7.57138610e-01 -3.12606305e-01
8.93541753e-01 8.14250186e-02 3.02195668e-01 1.10840416e+00
1.96421698e-01 3.49905014e-01 7.86667705e-01 -5.55185199e-01
3.97748798e-01 -1.66055307e-01 -1.69698924e-01 8.05813551e-01
1.58015579e-01 -1.42294779e-01 -5.98758042e-01 2.57274956e-02
6.80465043e-01 3.40308905e-01 -2.36969903e-01 -9.34541523e-01
-1.42107499e+00 6.78481281e-01 1.25462830e+00 7.99460486e-02
-3.37558866e-01 6.54763162e-01 -1.31095365e-01 4.12743539e-02
1.14136420e-01 5.60699761e-01 -3.74419034e-01 1.12292856e-01
-8.69876206e-01 2.44567335e-01 4.95959729e-01 1.07325768e+00
1.04590869e+00 -4.50405121e-01 1.42099693e-01 5.21823525e-01
4.40251499e-01 6.40522122e-01 1.87858284e-01 -1.13874662e+00
8.32599580e-01 4.75385070e-01 5.35075545e-01 -1.51961160e+00
-6.90376759e-01 -4.05778587e-01 -5.16947210e-01 2.33967572e-01
5.65995693e-01 -1.86613306e-01 -1.04330325e+00 1.70273149e+00
3.48891437e-01 1.98454648e-01 -2.90872633e-01 1.16346896e+00
5.88562250e-01 5.99935710e-01 -2.42479220e-01 5.77428102e-01
1.02981198e+00 -9.44418550e-01 -1.90846547e-01 -5.99037707e-01
5.34708321e-01 -7.33911216e-01 7.49271691e-01 1.35088623e-01
-4.72225308e-01 -6.11374021e-01 -1.19837558e+00 -3.85473639e-01
-4.56480443e-01 3.82968098e-01 5.06161809e-01 2.79525429e-01
-1.31772530e+00 4.16235656e-01 -9.02658880e-01 -8.26986670e-01
-2.99396459e-02 3.65799725e-01 -7.84432054e-01 -2.01285020e-01
-8.19496334e-01 9.38531816e-01 1.33963674e-01 3.22050333e-01
-9.68709290e-01 -5.17841220e-01 -1.09290767e+00 -9.89347622e-02
3.56211126e-01 -7.59881496e-01 9.04777765e-01 -7.05614626e-01
-1.01047623e+00 8.02178800e-01 -3.02704871e-01 -5.79429448e-01
7.18017578e-01 -5.73961318e-01 -1.43482968e-01 -4.39379811e-02
4.81211871e-01 8.47757399e-01 4.51456219e-01 -1.38376951e+00
-8.12707663e-01 -4.64350551e-01 5.23306847e-01 4.93976116e-01
1.71944559e-01 -5.12926519e-01 -5.66960335e-01 -3.32201481e-01
6.55267119e-01 -1.04711485e+00 -5.36698639e-01 3.48044634e-01
-8.14380646e-01 9.26824138e-02 6.89252555e-01 -5.09444118e-01
6.90030634e-01 -2.12528896e+00 -1.63672879e-01 4.67479706e-01
1.36080205e-01 -3.28093082e-01 1.42053114e-02 5.76292634e-01
7.47352615e-02 -1.39621034e-01 1.50535917e-02 -7.28816390e-01
7.64311701e-02 2.40197495e-01 -2.97961891e-01 9.52124000e-01
-3.83215427e-01 7.20816016e-01 -1.09277892e+00 -1.54747531e-01
6.14561558e-01 6.64559305e-01 -4.08831120e-01 7.35581219e-02
1.08563766e-01 4.00820255e-01 -2.81642944e-01 4.90143090e-01
7.62487233e-01 -1.44994006e-01 -3.16766649e-01 -1.31892279e-01
-2.25670636e-01 1.76323771e-01 -1.43636441e+00 2.30807734e+00
-7.05797076e-01 8.83254230e-01 -3.22283119e-01 -3.47864777e-01
1.06764436e+00 -1.15035526e-01 2.25797325e-01 -6.79788649e-01
-3.95792484e-01 2.10115328e-01 -5.50019801e-01 -9.59741846e-02
6.59593403e-01 3.49949092e-01 -1.66075125e-01 -4.68719862e-02
-8.76362994e-02 2.60770530e-01 -2.44765133e-01 -2.09619035e-03
1.19482529e+00 2.77297378e-01 3.20579976e-01 -1.85197696e-01
4.85018373e-01 5.77797033e-02 2.10011274e-01 8.49150062e-01
3.73176523e-02 1.06883633e+00 2.04413101e-01 -8.80073845e-01
-1.09447563e+00 -1.16780686e+00 2.66081002e-02 9.29869175e-01
5.97959757e-01 -4.88602281e-01 -4.80547637e-01 -7.93766916e-01
2.30126865e-02 4.38283473e-01 -7.23848820e-01 1.90276563e-01
-6.06843352e-01 -3.07659000e-01 3.87140572e-01 6.08669281e-01
6.76871598e-01 -5.04068196e-01 -9.51203585e-01 -1.27317131e-01
-2.57289529e-01 -1.18277740e+00 -5.39241791e-01 3.27755094e-01
-2.59641051e-01 -1.11345077e+00 -5.65757155e-01 -7.22259998e-01
1.00021923e+00 2.89384782e-01 7.28429198e-01 -2.74905652e-01
9.62550193e-02 5.13528287e-01 7.72229908e-03 2.12711701e-03
2.82161653e-01 2.46617183e-01 1.41593099e-01 1.46657839e-01
1.97772041e-01 -3.38085473e-01 -9.04959440e-01 3.11936021e-01
-1.71295881e-01 -1.26843035e-01 2.53838956e-01 4.89979118e-01
9.22845721e-01 -6.44852296e-02 -1.47599339e-01 -4.04806048e-01
-8.94038565e-03 -4.05621737e-01 -9.09931242e-01 2.00034901e-01
-2.51439065e-01 1.31919868e-02 6.37108266e-01 1.60276279e-01
-5.22164226e-01 6.07056975e-01 8.69997665e-02 -3.77710313e-01
-4.46900278e-01 2.69212931e-01 -2.09836319e-01 -4.27778661e-01
8.13554823e-01 -1.06730878e-01 -7.30171084e-01 -5.16525030e-01
4.50687021e-01 2.78182983e-01 5.96574426e-01 -2.56041825e-01
9.83331561e-01 7.70842910e-01 5.30782104e-01 -6.87374711e-01
-9.58677769e-01 -5.87285876e-01 -1.18033254e+00 -3.24519612e-02
8.14130723e-01 -1.13139033e+00 -8.02029669e-01 -2.67841443e-02
-1.48811400e+00 2.58665048e-02 -6.29696110e-03 6.43038690e-01
-5.08518755e-01 1.55502841e-01 -1.48587212e-01 -4.65096176e-01
2.35695653e-02 -9.22999024e-01 1.47024977e+00 2.81400029e-02
-2.92421579e-01 -1.05815005e+00 3.57825249e-01 -5.37731908e-02
7.67032951e-02 8.68644774e-01 4.23519731e-01 -3.80753100e-01
-9.52999294e-01 -4.22058612e-01 -1.92927182e-01 -1.74300045e-01
-1.09339803e-01 -4.56563443e-01 -1.11796856e+00 -4.00433511e-01
-2.66282111e-01 -3.45549993e-02 9.27477300e-01 5.10292113e-01
9.30989385e-01 -8.64750445e-02 -6.36557639e-01 1.13005495e+00
1.79025662e+00 5.50462585e-03 3.32324684e-01 9.06177461e-01
1.11314213e+00 3.07150185e-01 4.68640059e-01 2.63289958e-01
7.64271855e-01 1.02460897e+00 8.63703966e-01 -3.27259332e-01
-5.74021228e-02 -7.73871601e-01 1.02512017e-01 7.95925036e-02
-1.68247952e-03 -3.78829032e-01 -1.10902190e+00 6.17798865e-01
-1.94768047e+00 -4.29707795e-01 -1.40735626e-01 2.53802085e+00
3.15073058e-02 -4.29881290e-02 -2.11297408e-01 -1.85722649e-01
4.83653307e-01 3.13748449e-01 -5.12418807e-01 -1.22353524e-01
5.66721708e-02 -3.79197866e-01 1.09675288e+00 9.49336410e-01
-1.48016512e+00 8.12477529e-01 5.74209118e+00 3.60486597e-01
-1.09577584e+00 4.21225233e-03 5.63118696e-01 -1.25486270e-01
9.79313776e-02 4.80467491e-02 -1.07434833e+00 2.49466464e-01
7.62852669e-01 2.83554494e-01 4.98166353e-01 1.09402180e+00
1.57845929e-01 -3.62167269e-01 -1.29971659e+00 1.14136803e+00
4.39487994e-01 -1.15086162e+00 -3.15437645e-01 9.14898589e-02
7.51589954e-01 4.56617057e-01 4.21547368e-02 -1.09067403e-01
5.17403223e-02 -9.99017239e-01 9.59826827e-01 4.94485587e-01
6.83049262e-01 -9.55896139e-01 6.91621840e-01 3.48564774e-01
-1.19766498e+00 3.11455294e-03 -4.53072190e-01 -4.07484546e-02
2.10738908e-02 1.75975978e-01 -1.17853534e+00 4.59401429e-01
8.50960732e-01 8.26289535e-01 -1.01004136e+00 1.42500389e+00
-5.97321451e-01 -4.49179374e-02 -7.41652787e-01 2.11869538e-01
5.57237685e-01 1.01608478e-01 4.91597801e-01 1.05530488e+00
5.12959540e-01 -5.48313797e-01 4.56675321e-01 6.75595582e-01
-6.10536635e-02 -1.22098200e-01 -8.97811353e-01 8.44772875e-01
5.47302306e-01 1.06339598e+00 -9.97762263e-01 1.57781422e-01
-1.61528617e-01 1.20748281e+00 4.27098364e-01 6.07682109e-01
-8.82162392e-01 -7.30251789e-01 4.37939286e-01 5.23583293e-01
4.04962301e-01 -5.56853712e-01 -7.64071494e-02 -9.03283596e-01
4.11707580e-01 -1.63825601e-01 1.35952935e-01 -9.93125916e-01
-8.77066135e-01 6.69238210e-01 -7.42105246e-02 -1.40536916e+00
-1.99952006e-01 -4.77792442e-01 -2.02753231e-01 8.45844507e-01
-1.55730999e+00 -1.30247283e+00 -5.99071026e-01 6.48956597e-01
2.64528364e-01 1.67500481e-01 7.54534781e-01 3.96337807e-01
-6.92474768e-02 5.16671419e-01 3.27454537e-01 2.60177076e-01
7.66059577e-01 -1.41801465e+00 6.30400419e-01 1.07578373e+00
4.16854769e-01 7.30201602e-01 5.62468648e-01 -3.41373146e-01
-1.16252530e+00 -1.24914134e+00 1.21502948e+00 -1.01445699e+00
4.83578652e-01 -9.25034225e-01 -2.81020522e-01 1.01687813e+00
-1.22601300e-01 2.79724985e-01 1.28507763e-01 1.82260618e-01
-3.44264299e-01 -3.98628980e-01 -1.14651549e+00 6.54220700e-01
1.22616816e+00 -5.30105829e-01 -3.75745833e-01 6.38829887e-01
7.63772845e-01 -8.23071241e-01 -3.75988901e-01 9.74534899e-02
3.32519740e-01 -1.00861669e+00 1.23773217e+00 -2.08096132e-02
1.17837563e-01 -7.35378206e-01 -4.33370233e-01 -1.30360448e+00
-3.43277961e-01 -3.54238838e-01 -6.86530322e-02 9.01207089e-01
3.12394440e-01 -4.03969884e-01 9.62571740e-01 4.46828306e-01
-5.95910624e-02 -5.64669847e-01 -1.41545880e+00 -7.79205561e-01
-5.02280712e-01 -4.25569683e-01 6.62635922e-01 6.83438003e-01
-5.50832450e-01 4.09773797e-01 -5.47007024e-01 6.93721473e-01
5.62881708e-01 6.46752492e-02 9.82600868e-01 -9.60050464e-01
8.81316960e-02 1.70369688e-02 -1.01648772e+00 -1.38358212e+00
3.32871787e-02 -7.83270061e-01 1.40604869e-01 -1.80185711e+00
-2.22832099e-01 -4.18121010e-01 -3.35823119e-01 5.07323265e-01
3.14679474e-01 3.53732347e-01 6.47291616e-02 9.88081172e-02
-1.07965529e+00 3.46377075e-01 7.33901501e-01 -2.71008015e-02
-1.39134824e-01 4.98925224e-02 -2.11633280e-01 7.92996466e-01
4.42385346e-01 -5.33771038e-01 -3.58404785e-01 -9.39114451e-01
4.08938706e-01 -2.74206012e-01 8.94906044e-01 -1.54683936e+00
6.50834799e-01 5.85464872e-02 1.04405570e+00 -8.59386683e-01
7.26392984e-01 -1.37107205e+00 2.48692691e-01 3.74007523e-01
-3.13072294e-01 4.58832771e-01 -1.44989677e-02 8.38085175e-01
7.72524998e-02 -1.68084294e-01 3.46338004e-01 -1.13089206e-02
-1.09536612e+00 2.42640376e-01 1.58215657e-01 -2.72363305e-01
1.16112614e+00 -3.58882636e-01 -3.17438096e-01 -4.31480616e-01
-6.43579423e-01 2.79793650e-01 8.92272949e-01 4.78461891e-01
7.53880024e-01 -1.38462877e+00 -2.66844273e-01 3.00420135e-01
3.19312125e-01 9.08120498e-02 -1.22088686e-01 7.06434429e-01
-9.09799635e-01 6.41116977e-01 -1.76600348e-02 -8.56491745e-01
-9.45096076e-01 6.25189543e-01 6.53828502e-01 1.36939347e-01
-7.59594023e-01 8.56389940e-01 1.81154892e-01 -5.89809716e-01
4.03985620e-01 -4.89486784e-01 -9.08848047e-02 -2.51384825e-01
3.95778924e-01 4.57546741e-01 1.59315169e-01 -1.03111279e+00
-8.12519133e-01 1.02768123e+00 1.71940729e-01 -1.39390796e-01
1.30748582e+00 -5.05194962e-01 1.62371650e-01 4.83445972e-01
1.43690026e+00 2.10156456e-01 -1.51459420e+00 -1.40515476e-01
1.60701588e-01 -8.22959185e-01 -2.71230508e-02 -6.93911254e-01
-6.97103977e-01 7.47659326e-01 1.01602340e+00 -3.36428076e-01
5.73484659e-01 -9.11756828e-02 5.12954772e-01 8.57062519e-01
8.86998057e-01 -9.17101324e-01 -1.32592723e-01 5.77324748e-01
7.56791711e-01 -1.51040936e+00 1.44220710e-01 3.48288827e-02
-3.37445021e-01 1.00962198e+00 5.94024479e-01 -3.84957522e-01
5.91260791e-01 -1.87706962e-01 1.72046214e-01 -2.62952775e-01
-5.80884032e-02 -1.02715150e-01 3.81756812e-01 6.67164445e-01
1.54106274e-01 1.90402195e-02 3.71858209e-01 -4.64763679e-03
-3.87910992e-01 -4.11497831e-01 2.74123758e-01 8.32852364e-01
-3.10858905e-01 -5.09910583e-01 -5.30436814e-01 -8.49154666e-02
-9.88033935e-02 5.95085658e-02 -3.56843889e-01 8.90908599e-01
3.90413582e-01 7.97131240e-01 2.12719575e-01 -4.11443293e-01
5.79976201e-01 -5.26727259e-01 2.77181774e-01 -5.34628987e-01
-1.07731611e-01 -1.25516310e-01 -1.07288897e-01 -9.10426915e-01
-2.59915113e-01 -4.52212006e-01 -1.15065336e+00 -2.11249858e-01
-1.33886177e-03 -8.36291388e-02 8.08993340e-01 8.71159077e-01
3.82140815e-01 7.80772567e-02 6.51796401e-01 -1.21543181e+00
-3.20381559e-02 -4.95851398e-01 -3.27267736e-01 2.73336619e-01
8.39769006e-01 -5.82130849e-01 -2.31948122e-01 -2.23200042e-02] | [7.661238670349121, -2.1701743602752686] |
99c718f4-d624-4eae-b0f9-0af5f138c6b4 | mme-crs-multi-metric-evaluation-based-on | 2206.09403 | null | https://arxiv.org/abs/2206.09403v1 | https://arxiv.org/pdf/2206.09403v1.pdf | MME-CRS: Multi-Metric Evaluation Based on Correlation Re-Scaling for Evaluating Open-Domain Dialogue | Automatic open-domain dialogue evaluation is a crucial component of dialogue systems. Recently, learning-based evaluation metrics have achieved state-of-the-art performance in open-domain dialogue evaluation. However, these metrics, which only focus on a few qualities, are hard to evaluate dialogue comprehensively. Furthermore, these metrics lack an effective score composition approach for diverse evaluation qualities. To address the above problems, we propose a Multi-Metric Evaluation based on Correlation Re-Scaling (MME-CRS) for evaluating open-domain dialogue. Firstly, we build an evaluation metric composed of 5 groups of parallel sub-metrics called Multi-Metric Evaluation (MME) to evaluate the quality of dialogue comprehensively. Furthermore, we propose a novel score composition method called Correlation Re-Scaling (CRS) to model the relationship between sub-metrics and diverse qualities. Our approach MME-CRS ranks first on the final test data of DSTC10 track5 subtask1 Automatic Open-domain Dialogue Evaluation Challenge with a large margin, which proved the effectiveness of our proposed approach. | ['Chunyang Yuan', 'Cao Liu', 'Song Han', 'Jian Wang', 'Kaidong Yu', 'Xiaohui Hu', 'Pengfei Zhang'] | 2022-06-19 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [-3.84983450e-01 8.92704055e-02 1.01058908e-01 -5.41666389e-01
-1.07731485e+00 -6.65793002e-01 9.35003400e-01 2.38583073e-01
-4.46537495e-01 9.68228221e-01 6.98382139e-01 -8.59266892e-03
-3.50412697e-01 -5.44075429e-01 1.27342388e-01 -2.51117766e-01
1.26170143e-01 6.11501396e-01 5.75769305e-01 -1.08987033e+00
4.24090654e-01 -3.85374129e-01 -1.03621805e+00 3.33079904e-01
1.41462278e+00 8.04108679e-01 -7.10896477e-02 9.11521435e-01
-3.74958277e-01 7.04001963e-01 -1.09115934e+00 -5.94570398e-01
-1.22572385e-01 -8.60215247e-01 -1.35464978e+00 -1.82606310e-01
-1.64175302e-01 -5.23167968e-01 -1.05418719e-01 1.00243306e+00
7.62843788e-01 1.38823986e-01 7.85601974e-01 -1.14786243e+00
-4.51633096e-01 7.65632629e-01 -1.51725158e-01 -6.74817115e-02
8.88313293e-01 1.38569668e-01 1.39526653e+00 -6.57727301e-01
5.61248183e-01 1.42653310e+00 4.23154056e-01 7.02316761e-01
-9.49901104e-01 -2.60794997e-01 -3.20035070e-02 2.86501884e-01
-7.88688004e-01 -1.68581560e-01 6.00264192e-01 -3.26783657e-01
6.37551486e-01 3.18067878e-01 3.17721933e-01 9.07642245e-01
-2.29516149e-01 8.23968947e-01 1.42129374e+00 -2.81237155e-01
1.41697139e-01 2.66461313e-01 3.92207056e-01 4.86735553e-01
-4.67846692e-01 -3.43237430e-01 -4.21339244e-01 -2.55462438e-01
5.18527925e-01 -5.95039070e-01 -3.15472811e-01 -7.02200830e-02
-1.34975159e+00 1.07339537e+00 2.36155570e-01 3.94295841e-01
1.24604709e-01 -6.99278712e-01 9.49261606e-01 7.14299560e-01
4.41489697e-01 8.14249218e-01 -6.46729231e-01 -8.09035420e-01
-2.80495167e-01 3.83330017e-01 1.18407595e+00 5.93534827e-01
4.18741107e-01 -5.74474394e-01 -7.79877007e-01 1.63962781e+00
1.12895466e-01 1.86801270e-01 6.81934595e-01 -9.41152215e-01
8.05644572e-01 8.35612059e-01 2.11125419e-01 -8.57269406e-01
-6.59168422e-01 -6.86659291e-02 -8.64535630e-01 -6.06592931e-02
7.32243180e-01 -5.23977876e-01 1.19163342e-01 1.96039104e+00
3.76341254e-01 -5.15387356e-01 4.51662838e-01 1.03146124e+00
1.31063771e+00 7.25778937e-01 -2.24932700e-01 -3.60632718e-01
1.32138717e+00 -1.26221764e+00 -1.00064790e+00 4.46416974e-01
8.34365129e-01 -8.04102302e-01 1.54081750e+00 3.39445442e-01
-8.63649786e-01 -5.49595654e-01 -1.15696621e+00 -4.59600128e-02
-1.41655490e-01 -1.54776245e-01 2.80750483e-01 6.69004500e-01
-7.57651925e-01 6.13102734e-01 -1.15845650e-01 -1.13396175e-01
-3.53965461e-01 6.29467219e-02 -1.50519118e-01 3.35899055e-01
-1.86588633e+00 1.17372632e+00 2.23838732e-01 -1.88103721e-01
-6.34507895e-01 -3.74248683e-01 -4.29794818e-01 -1.23240143e-01
4.80203867e-01 -2.29850739e-01 1.58875394e+00 -4.64355856e-01
-2.20036435e+00 6.11228526e-01 3.84524971e-01 -1.77062422e-01
6.22859657e-01 -3.97182941e-01 -5.39338827e-01 2.40505338e-02
9.03436448e-03 2.04337090e-01 8.74886215e-02 -1.13682818e+00
-6.75837696e-01 -1.40099466e-01 5.88917792e-01 6.57715201e-01
-6.28047526e-01 1.30012304e-01 -2.45123252e-01 -3.00438464e-01
-2.62537241e-01 -6.18949652e-01 -6.64633811e-02 -3.82322788e-01
-3.35107535e-01 -8.48683596e-01 4.47917104e-01 -6.24404848e-01
1.80486691e+00 -1.89769483e+00 3.91995877e-01 -1.71204284e-01
5.18579900e-01 5.60863733e-01 -1.57231301e-01 5.95723569e-01
4.43408996e-01 -8.32974985e-02 -2.98141837e-01 -1.63356453e-01
4.63045359e-01 -7.27721974e-02 8.67160335e-02 -1.41899809e-01
1.92168370e-01 5.98253191e-01 -1.26699996e+00 -6.30114138e-01
4.75860536e-02 -9.44930017e-02 -4.80510324e-01 6.95652068e-01
-4.76246029e-01 4.82808769e-01 -5.29411793e-01 1.60321623e-01
4.09838110e-01 -5.28789647e-02 6.36613369e-02 -1.53401583e-01
-9.17832777e-02 5.04373312e-01 -9.41371143e-01 1.95655525e+00
-5.78407586e-01 2.98537046e-01 -4.45687436e-02 -4.51304227e-01
1.36199868e+00 2.61669308e-01 2.85389692e-01 -8.00920546e-01
2.70534784e-01 3.99692565e-01 2.96140730e-01 -4.31360573e-01
9.58598793e-01 3.63379680e-02 -4.44572568e-01 7.61790335e-01
1.99377477e-01 -2.21099034e-01 4.42599386e-01 4.20026720e-01
1.27836049e+00 -1.24560818e-01 2.78585762e-01 -2.15308174e-01
1.11732781e+00 -1.36241704e-01 3.11036795e-01 3.16706389e-01
-6.58964217e-01 5.59225500e-01 1.00101042e+00 1.08241001e-02
-9.47588027e-01 -6.52298212e-01 -2.21314818e-01 1.36075771e+00
1.73842028e-01 -7.71764994e-01 -1.00714064e+00 -9.39943671e-01
-3.66632134e-01 3.84422094e-01 -3.98977578e-01 -8.33430588e-02
-2.89041877e-01 -7.77489305e-01 1.02104175e+00 6.81347921e-02
9.80212748e-01 -8.63970876e-01 -1.05315782e-01 3.24559599e-01
-7.84750760e-01 -9.97519851e-01 -5.89747429e-01 -1.11423008e-01
-3.48736197e-01 -1.17261708e+00 -9.96123672e-01 -6.07046902e-01
-4.28743288e-02 1.20503776e-01 1.24444699e+00 5.89603484e-02
3.63920540e-01 7.46791139e-02 -9.39851284e-01 -3.50257307e-02
-8.44779134e-01 3.22808653e-01 -4.56394218e-02 -4.09834459e-02
2.57314175e-01 -2.26602003e-01 -5.35065472e-01 8.20174038e-01
-5.66935003e-01 -4.31821533e-02 3.42233092e-01 1.22695434e+00
-3.95826809e-02 -3.60995322e-01 1.18703210e+00 -7.98421085e-01
1.80382955e+00 -4.01961565e-01 -1.25075728e-01 3.88914704e-01
-7.05072701e-01 1.54612958e-01 7.86383748e-01 -2.70817727e-01
-1.11141968e+00 -8.43301594e-01 -4.46139872e-01 2.92594701e-01
1.86509460e-01 6.71105087e-01 -2.84575105e-01 9.99350399e-02
7.59387314e-01 -6.84033381e-03 2.06318364e-01 -4.22338754e-01
4.67586339e-01 1.16556323e+00 4.48338598e-01 -7.95552075e-01
3.76830369e-01 -3.90901983e-01 -3.03390503e-01 -6.31244242e-01
-9.12814677e-01 -6.06133938e-01 -4.88488466e-01 -2.86755979e-01
8.43054235e-01 -6.89976513e-01 -8.64871740e-01 6.26552105e-01
-1.30057955e+00 -2.89532393e-01 6.50492832e-02 3.71013075e-01
-4.92479205e-01 7.26731479e-01 -6.89409077e-01 -7.55722523e-01
-4.99941379e-01 -1.05709493e+00 6.89742208e-01 3.03929776e-01
-3.49046201e-01 -1.03677607e+00 6.00440562e-01 5.54167986e-01
4.99123275e-01 2.11501002e-01 7.99497426e-01 -1.10218823e+00
2.09040686e-01 1.11748636e-01 -4.22987372e-01 6.15857005e-01
5.06719351e-02 -6.95008337e-02 -7.07663536e-01 4.32856902e-02
4.02498506e-02 -9.47187245e-01 4.44530278e-01 -2.98964649e-01
6.58518374e-01 -1.61440551e-01 4.21024442e-01 1.33249080e-02
7.97427595e-01 1.30652905e-01 6.22974694e-01 5.26059031e-01
4.23103362e-01 6.96901619e-01 1.16649485e+00 8.17964315e-01
8.01480949e-01 9.75849748e-01 7.78648853e-02 2.34675631e-02
1.27173945e-01 -1.32501245e-01 4.12935674e-01 1.44978130e+00
-6.08207025e-02 -7.67329186e-02 -7.35910177e-01 2.61555761e-01
-1.95759654e+00 -6.53609633e-01 -3.76316756e-01 2.11138821e+00
1.33480763e+00 3.30503434e-01 6.10345840e-01 3.34639996e-01
7.13231862e-01 2.64808416e-01 -3.65655839e-01 -6.84451222e-01
-2.15079367e-01 -1.24617256e-01 -1.84263617e-01 6.05545819e-01
-8.39567959e-01 7.83384800e-01 5.37228298e+00 1.06808352e+00
-5.16516745e-01 1.79394186e-01 3.93714130e-01 4.55902755e-01
-3.45459729e-01 -6.32579625e-02 -7.97349215e-01 4.92067724e-01
9.48836148e-01 -4.59866464e-01 2.20899597e-01 6.82563066e-01
-6.66469857e-02 -1.06773200e-02 -9.83289301e-01 8.77022564e-01
4.47621867e-02 -7.42362738e-01 -1.37790933e-01 -5.38434014e-02
6.83043599e-01 -3.54689270e-01 -2.56530821e-01 8.54099870e-01
5.06166577e-01 -7.52783537e-01 1.48843497e-01 3.02724957e-01
6.18529856e-01 -7.33439445e-01 9.66036916e-01 3.96551698e-01
-9.00388479e-01 6.63966611e-02 -5.02325475e-01 -1.23273013e-02
1.66188076e-01 3.71321499e-01 -8.01729083e-01 8.98796737e-01
8.35025236e-02 3.56862217e-01 -5.46022534e-01 8.76136303e-01
-1.87273905e-01 3.57492924e-01 1.72529705e-02 -6.75962508e-01
4.51182902e-01 -2.53813922e-01 5.76492429e-01 1.25874257e+00
-1.92572698e-01 2.15526834e-01 2.87304580e-01 5.70454359e-01
-1.54189358e-03 6.31798565e-01 -9.82343554e-02 4.92704734e-02
5.51435888e-01 1.35569894e+00 -1.32478833e-01 -1.85351864e-01
-3.62969011e-01 7.32231200e-01 3.90129566e-01 -3.15311998e-01
-8.17771137e-01 -8.30816507e-01 5.01342475e-01 -4.38886046e-01
-3.15978289e-01 -2.58287340e-01 -9.85853970e-02 -1.13855517e+00
-2.33304761e-02 -1.25063777e+00 3.56626928e-01 -2.85841048e-01
-1.51849675e+00 1.02304387e+00 5.35408827e-03 -1.65061653e+00
-3.11903328e-01 -3.22925001e-01 -5.28421402e-01 8.08059990e-01
-1.36626267e+00 -7.74857700e-01 -5.31713903e-01 6.75220251e-01
7.87676930e-01 -4.84833628e-01 1.20678866e+00 3.06218863e-01
-5.22294760e-01 8.97727132e-01 3.32640618e-01 2.22646266e-01
1.23507154e+00 -1.48433340e+00 1.74607396e-01 2.84386128e-01
-3.29378426e-01 4.03872609e-01 6.10123336e-01 -2.57197559e-01
-9.78145540e-01 -5.26203871e-01 7.16251671e-01 -4.51848507e-01
8.99930775e-01 -1.67627856e-01 -1.07540143e+00 -2.85459161e-01
4.92059290e-01 -6.58477128e-01 9.15104210e-01 5.68523943e-01
-3.35185677e-01 -5.23914583e-02 -9.66064036e-01 4.02259052e-01
8.05520058e-01 -4.01223063e-01 -9.46020663e-01 3.72495860e-01
9.76476252e-01 -5.67731619e-01 -1.36682868e+00 5.34332871e-01
4.72298384e-01 -1.17368627e+00 5.79636276e-01 -5.10145128e-01
8.69765520e-01 -7.88199380e-02 -2.11828649e-01 -1.73995018e+00
-2.64307577e-02 -8.89631093e-01 -5.86253293e-02 1.65828669e+00
5.65220654e-01 -4.59744364e-01 3.14178050e-01 4.79763567e-01
-1.56009182e-01 -9.38980699e-01 -7.35024929e-01 -7.38817036e-01
3.03893059e-01 5.69054447e-02 7.02698112e-01 8.84028494e-01
7.45478928e-01 1.05888486e+00 -4.96701330e-01 -5.51213801e-01
3.10364127e-01 -9.55731943e-02 1.13620627e+00 -1.25419688e+00
-3.51601362e-01 -7.87771821e-01 -1.60222426e-01 -1.36261153e+00
-3.64234634e-02 -5.31854689e-01 1.14897877e-01 -1.55148125e+00
2.41088480e-01 -5.29879689e-01 -1.83417529e-01 -6.46238923e-02
-6.26677990e-01 -2.76186556e-01 1.98027179e-01 2.36115992e-01
-1.20439851e+00 1.11253476e+00 1.67087245e+00 3.73959802e-02
-4.09555346e-01 3.60846221e-02 -7.95374513e-01 4.34371024e-01
7.78733850e-01 2.00751647e-02 -5.53280115e-01 -2.34767750e-01
1.75983608e-02 5.11884511e-01 -2.47686744e-01 -9.21119452e-01
2.30546370e-01 -1.73010781e-01 -3.15847665e-01 -5.24852753e-01
2.15636671e-01 -2.37057522e-01 -6.45954251e-01 2.56664783e-01
-7.17658877e-01 1.20689822e-02 -2.70760894e-01 3.65989894e-01
-6.71119213e-01 -2.62486011e-01 7.10472226e-01 -2.55251043e-02
-5.55761635e-01 -2.95169652e-02 4.59933048e-03 7.67337859e-01
6.71202958e-01 1.39183521e-01 -7.28614032e-01 -5.49102426e-01
-4.24082756e-01 7.37925470e-01 -2.75135227e-02 6.12933457e-01
4.14087951e-01 -1.36978292e+00 -1.13644636e+00 -4.10377622e-01
3.85538489e-01 -6.50654584e-02 3.17754000e-01 7.34677017e-01
-4.31793809e-01 5.46490729e-01 -2.93949157e-01 -4.87693578e-01
-1.31408572e+00 -9.45329741e-02 2.28009596e-01 -1.04663646e+00
-2.25130141e-01 8.06154609e-01 -7.86687732e-02 -1.05932200e+00
4.05364126e-01 -7.46573433e-02 -7.01024771e-01 2.29272097e-01
6.00132823e-01 4.92804378e-01 8.10332522e-02 -5.83176076e-01
-2.03162152e-03 2.77294815e-01 -9.01940763e-02 -4.93941337e-01
1.04644847e+00 -4.01894033e-01 -9.06441584e-02 6.70718908e-01
1.12184203e+00 -6.46697357e-02 -9.07030702e-01 -5.74615479e-01
6.43281862e-02 -3.72845531e-01 -4.04434711e-01 -1.20167160e+00
-3.18810225e-01 7.83759415e-01 3.63952935e-01 5.46262980e-01
7.46805370e-01 -4.56359357e-01 9.74401891e-01 6.69304311e-01
3.69774222e-01 -1.46019113e+00 6.27669632e-01 1.23267281e+00
1.13629675e+00 -1.37877131e+00 -2.55923659e-01 -3.15764636e-01
-1.23253965e+00 1.17672658e+00 1.02677321e+00 2.23200768e-01
3.65424454e-01 -2.85520047e-01 2.68922657e-01 -1.19337570e-02
-8.51766169e-01 -2.99753785e-01 2.04445556e-01 3.57958764e-01
7.62847960e-01 2.11164743e-01 -9.85293686e-01 9.49772716e-01
-4.55671638e-01 -3.63621473e-01 4.33404297e-01 5.26711106e-01
-6.51273668e-01 -1.45107663e+00 -1.29812822e-01 3.00216109e-01
-1.75542831e-01 1.29744187e-01 -7.41236866e-01 4.87219900e-01
-4.39490318e-01 1.33584857e+00 -5.36460757e-01 -1.04579663e+00
6.35420442e-01 -4.70234565e-02 2.94474065e-01 -5.58835208e-01
-8.77863944e-01 -1.59295276e-01 6.52339756e-01 -2.17132211e-01
-5.11184812e-01 -3.02685946e-01 -1.02356708e+00 -4.55368906e-01
-6.49535120e-01 7.70786881e-01 6.02308333e-01 1.03663707e+00
-9.02100187e-03 5.64789414e-01 1.34061658e+00 -2.61500448e-01
-1.30371904e+00 -1.56908643e+00 -5.76423585e-01 6.21550381e-01
-1.59222540e-02 -7.68818080e-01 -3.27307880e-01 -7.10765243e-01] | [12.738694190979004, 8.168157577514648] |
97c573b8-a47c-466b-96f2-8cf8f077d909 | passive-motion-detection-via-mmwave | 2203.14588 | null | https://arxiv.org/abs/2203.14588v1 | https://arxiv.org/pdf/2203.14588v1.pdf | Passive Motion Detection via mmWave Communication System | In this paper, an integrated passive sensing and communication system working in 60 GHz band is elaborated, and the sensing performance is investigated in an application of hand gesture recognition. Specifically, in this integrated system, there are two radio frequency (RF) chains at the receiver and one at the transmitter. Each RF chain is connected with one phased array for analog beamforming. To facilitate simultaneous sensing and communication, the transmitter delivers one stream of information-bearing signals via two beam lobes, one is aligned with the main signal propagation path and the other is directed to the sensing target. Signals from the two lobes are received by the two RF chains at the receiver, respectively. By cross ambiguity coherent processing, the time-Doppler spectrograms of hand gestures can be obtained. Relying on the passive sensing system, a dataset of received signals, where three types of hand gestures are sensed, is collected by using Line-of-Sight (LoS) and Non-Line-of-Sight (NLoS) paths as the reference channel respectively. Then a neural network is trained by the dataset for motion detection. It is shown that the classification accuracy rate is high as long as sufficient sensing time is assured. Finally, an empirical model characterizing the relation between the classification accuracy and sensing duration is derived analytically. | ['Rui Wang', 'Yifei Sun', 'Yan Luo', 'Chao Yu', 'Jie Li'] | 2022-03-28 | null | null | null | null | ['hand-gesture-recognition', 'motion-detection', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.06870496e-01 -1.82679649e-02 -1.36882022e-01 -6.59388527e-02
-3.05439740e-01 -5.52012384e-01 2.14562565e-01 -5.53242803e-01
-5.26521206e-01 3.95852894e-01 5.88249117e-02 -1.13913506e-01
-4.70002264e-01 -8.75334501e-01 5.58573194e-02 -1.32464027e+00
-1.73980728e-01 4.13367115e-02 -1.80523291e-01 2.37134084e-01
-1.59298386e-02 5.87322950e-01 -1.19479144e+00 -2.34814659e-01
4.94540304e-01 1.40993536e+00 3.48145455e-01 7.88830101e-01
-9.87662077e-02 3.41460794e-01 -9.22026634e-01 3.26110542e-01
3.79776657e-01 -2.95379400e-01 1.53604522e-01 -5.06085694e-01
-2.45268613e-01 -5.62765241e-01 -6.04426980e-01 8.29197228e-01
7.89405882e-01 -2.42142826e-01 7.95768738e-01 -9.36943829e-01
-1.17320597e-01 6.87355399e-01 -5.58592558e-01 -5.01179621e-02
5.66915691e-01 -2.14330107e-02 5.07484972e-01 -4.86593753e-01
3.53310317e-01 8.27526987e-01 3.96927744e-01 3.47570926e-01
-6.56766593e-01 -9.89462197e-01 -3.40495884e-01 -1.82882026e-01
-1.26693630e+00 -1.91720277e-01 1.10592592e+00 -3.49199176e-01
4.03826356e-01 4.55046326e-01 7.39623606e-01 1.22411144e+00
6.29802883e-01 2.06099927e-01 8.63683641e-01 -7.17306674e-01
3.02662075e-01 -4.56497110e-02 3.88625711e-01 4.17524397e-01
3.87752026e-01 6.54170513e-01 -6.39745057e-01 -7.14512840e-02
9.14725840e-01 1.19701251e-01 -4.80615824e-01 -1.35907769e-01
-1.20912290e+00 2.38581255e-01 5.34774482e-01 9.11153316e-01
-6.44826651e-01 2.39204347e-01 -4.04243857e-01 9.04645920e-02
-1.66164935e-01 2.32656658e-01 1.12397723e-01 2.07313478e-01
-1.00943267e+00 -1.67406201e-01 9.27509010e-01 7.44129002e-01
2.05642000e-01 8.04313943e-02 -8.71187970e-02 3.42241287e-01
9.52275038e-01 1.50536513e+00 2.97039717e-01 -4.01095092e-01
6.02546573e-01 -2.63410266e-02 3.91197413e-01 -1.13481259e+00
-9.04931903e-01 -1.26555419e+00 -1.02033734e+00 -6.11536130e-02
6.53364718e-01 -7.88970113e-01 -5.71529746e-01 1.64170051e+00
1.93928257e-01 3.58025357e-02 3.53887558e-01 1.19631708e+00
6.38815582e-01 5.45013666e-01 -7.08874688e-02 -6.12771809e-01
1.43990195e+00 -3.99177253e-01 -9.45793092e-01 -4.17893767e-01
2.85204835e-02 -1.02634430e+00 3.61720145e-01 6.57968104e-01
-5.78188300e-01 -6.21590316e-01 -1.30755448e+00 7.20317900e-01
1.35190606e-01 4.97474343e-01 3.99063587e-01 1.10308707e+00
-3.15364450e-01 -1.36220813e-01 -9.17319715e-01 -1.12830095e-01
-1.14143312e-01 1.85540825e-01 2.71986961e-01 -1.04045376e-01
-1.07681525e+00 2.48955667e-01 -3.01007450e-01 6.12187803e-01
-1.72161117e-01 -4.33962047e-01 -2.89078534e-01 -1.18757270e-01
-3.57436478e-01 -6.11236930e-01 9.17794526e-01 -4.38990563e-01
-1.69809949e+00 5.34708947e-02 -1.22353114e-01 4.23720814e-02
3.41334701e-01 -1.16026647e-01 -9.02045608e-01 4.05176580e-01
1.54728368e-01 -1.66257679e-01 9.75951672e-01 -1.07168710e+00
-5.26875257e-01 -7.49929130e-01 -4.90286499e-01 1.62779659e-01
-3.57592583e-01 -2.78269291e-01 1.33389309e-01 -5.92998922e-01
1.09782255e+00 -6.96175933e-01 -5.43164387e-02 -3.14975947e-01
-5.68315566e-01 2.88974464e-01 8.89345527e-01 -5.75454652e-01
1.29126740e+00 -2.40255189e+00 -1.51877880e-01 7.82460690e-01
-1.53346032e-01 7.68828467e-02 -3.37768346e-02 7.32653201e-01
3.94191116e-01 -5.23175418e-01 1.45203527e-02 1.26487300e-01
-1.50370419e-01 -1.61242723e-01 -5.36548316e-01 6.54355407e-01
-3.73814940e-01 4.32497919e-01 -8.25250387e-01 -1.29763633e-01
1.00410230e-01 5.88056505e-01 3.06503638e-03 3.73296052e-01
4.62237000e-01 1.00387800e+00 -1.07413042e+00 7.13772297e-01
9.25991118e-01 -2.84956619e-02 3.00436288e-01 -2.69543678e-01
-2.47335851e-01 4.35765125e-02 -1.61246860e+00 1.60911584e+00
-5.69858730e-01 3.77602190e-01 4.57217008e-01 -6.60099030e-01
1.30237293e+00 4.93025303e-01 6.98027968e-01 -9.49147284e-01
1.12496957e-01 4.11946863e-01 -2.03663453e-01 -9.19413507e-01
-2.52504833e-02 -1.43587098e-01 -2.04688400e-01 7.05237985e-01
-1.74200296e-01 2.53335744e-01 -2.46259198e-01 -2.77733237e-01
1.34360516e+00 4.29345146e-02 8.41048658e-02 -1.46161970e-02
5.75125873e-01 -2.98113953e-02 2.85318047e-01 8.99759054e-01
-5.17290011e-02 2.68603880e-02 -2.58489639e-01 -3.86480032e-03
-2.31859863e-01 -1.41366768e+00 4.62429635e-02 5.85945547e-01
6.83977067e-01 2.32065916e-01 -4.59287405e-01 -1.19937332e-02
5.26553951e-02 3.49179268e-01 9.93937533e-03 8.35985839e-02
-5.84855378e-01 -4.41529483e-01 7.21819818e-01 3.19425017e-01
8.25816929e-01 -5.72618425e-01 -1.24006510e+00 4.47506189e-01
-2.30444632e-02 -9.25124645e-01 6.12439737e-02 1.46653727e-01
-6.68996871e-01 -9.53010082e-01 -7.09600687e-01 -5.71223259e-01
3.27390343e-01 5.86889446e-01 2.43348792e-01 -3.23863655e-01
-2.78352678e-01 6.54279351e-01 -2.51809686e-01 -3.96607757e-01
1.40575662e-01 -1.23929009e-01 9.27082077e-02 3.48490000e-01
4.55826640e-01 -6.33911610e-01 -8.15003753e-01 5.16637206e-01
-3.88873190e-01 -3.44256699e-01 1.06089056e+00 3.52199048e-01
4.47962172e-02 -5.65868989e-02 5.18523812e-01 3.93569954e-02
2.90709108e-01 -3.00245911e-01 -5.61853051e-01 -9.98234078e-02
-1.50788695e-01 -1.98797002e-01 4.47100222e-01 -4.32176143e-01
-1.14543831e+00 2.68749684e-01 -1.35530815e-01 9.53978449e-02
-3.26184809e-01 4.76425707e-01 -4.81895715e-01 -4.67377305e-02
6.73220277e-01 3.31607014e-01 1.91248246e-02 -3.01588595e-01
2.28608727e-01 1.15769815e+00 5.70347428e-01 9.48867500e-02
1.01051724e+00 5.14389634e-01 2.46645227e-01 -1.51874638e+00
-3.01129282e-01 -6.79802179e-01 -3.36358815e-01 -6.35345161e-01
8.37316036e-01 -7.83656836e-01 -7.13069558e-01 7.07907856e-01
-1.15154946e+00 -7.51707703e-03 2.13760331e-01 1.40887451e+00
-2.21294880e-01 1.26118571e-01 -4.03283089e-01 -1.49466491e+00
-3.43972296e-01 -7.47899055e-01 9.14868474e-01 3.95556718e-01
-3.08914959e-01 -5.71620107e-01 3.67854051e-02 1.84026942e-01
4.99135524e-01 1.48611158e-01 4.72277522e-01 -2.01729149e-01
-8.10672343e-01 -4.56584781e-01 2.22579867e-01 -3.40028703e-01
2.48724803e-01 -5.13826728e-01 -1.09645820e+00 -2.55007207e-01
3.80276382e-01 2.87958533e-01 4.78573531e-01 8.66043150e-01
3.73899043e-01 -1.33164868e-01 -1.01234782e+00 6.12958670e-01
1.49007976e+00 6.87511146e-01 4.86371934e-01 -2.38858789e-01
2.14151412e-01 4.85334575e-01 7.55010128e-01 4.50367361e-01
-2.22961128e-01 4.27136689e-01 3.35620552e-01 3.82971130e-02
9.01457947e-03 -5.41965142e-02 3.02492708e-01 4.53352094e-01
-5.08456081e-02 -6.50162339e-01 -5.15931845e-01 -1.80672571e-01
-1.30108118e+00 -8.88733685e-01 -5.32479465e-01 2.24390459e+00
1.67910494e-02 -1.44722357e-01 -8.86144787e-02 5.77153385e-01
5.99721611e-01 2.23693624e-01 -5.07471144e-01 2.73507655e-01
1.29128054e-01 3.91909838e-01 7.77504742e-01 6.67049766e-01
-9.58910108e-01 1.09743826e-01 5.23631954e+00 3.50004226e-01
-1.54038501e+00 -1.48162432e-02 -2.70639449e-01 4.02846858e-02
-1.72682986e-01 -4.50523317e-01 -5.19267380e-01 3.21871519e-01
3.59430343e-01 5.95941186e-01 1.10293090e-01 4.21231747e-01
1.94043323e-01 -3.61819983e-01 -6.39494538e-01 1.20143354e+00
-3.27970415e-01 -7.63820887e-01 -4.18285280e-01 2.67375618e-01
-7.80603066e-02 -4.33699012e-01 1.20186172e-01 -3.63601774e-01
-3.71006459e-01 -6.56299651e-01 6.70221329e-01 8.79475355e-01
8.69380176e-01 -4.19758230e-01 5.67196965e-01 7.09539890e-01
-1.55588138e+00 -3.68239939e-01 9.17431265e-02 -4.46365356e-01
2.10072175e-01 1.16194701e+00 -6.25841379e-01 8.25423062e-01
2.84518540e-01 1.61656901e-01 3.76388907e-01 6.83191776e-01
-4.71475124e-01 8.32193553e-01 -6.36164069e-01 -6.09116912e-01
-1.33046523e-01 -3.74217957e-01 8.96971524e-01 9.73226070e-01
7.74611533e-01 5.38261414e-01 -1.01888627e-01 7.95552075e-01
6.08654797e-01 -1.87659204e-01 -6.63227081e-01 8.74701142e-02
9.11912680e-01 1.25701392e+00 -7.71300018e-01 1.50856972e-01
-9.76904482e-02 7.40275323e-01 -8.06042373e-01 8.13156784e-01
-4.44000125e-01 -8.81837964e-01 4.28156942e-01 -3.74179184e-02
5.17824069e-02 -9.30878162e-01 -5.60267568e-01 -5.75396776e-01
2.09634498e-01 1.04327910e-02 -2.60731950e-03 -4.68556285e-01
-1.00015152e+00 2.87970126e-01 -5.57986200e-01 -1.47348535e+00
-4.29453075e-01 -4.67620134e-01 -3.63034785e-01 1.32524443e+00
-1.33214116e+00 -8.52058351e-01 -7.41271973e-01 5.22583544e-01
-2.83189058e-01 -1.31597042e-01 9.77416217e-01 3.72892618e-01
-1.29418820e-01 4.83320862e-01 2.47773558e-01 3.80944610e-01
5.15370369e-01 -5.37448108e-01 -3.42728734e-01 8.61457348e-01
7.18652904e-02 8.01375091e-01 7.11173892e-01 -6.90704882e-01
-1.99269032e+00 -7.58254945e-01 7.45410860e-01 2.00840488e-01
2.23472565e-01 -3.37464750e-01 -9.25795287e-02 1.27641618e-01
-2.06569239e-01 -3.29420507e-01 8.61180663e-01 -3.99800390e-01
-6.65307939e-02 -4.05975759e-01 -1.31732059e+00 1.12346813e-01
1.13383842e+00 -1.15621231e-01 -5.42796910e-01 1.05563283e-01
1.37840092e-01 -3.36481005e-01 -6.64981484e-01 1.64226174e-01
1.28595626e+00 -7.72620380e-01 9.65337455e-01 2.92823941e-01
-7.96426684e-02 -3.06223065e-01 -5.63632965e-01 -9.66025352e-01
-2.92536646e-01 -4.58681434e-01 1.08882867e-01 9.01355684e-01
2.64882267e-01 -1.17542171e+00 9.10995543e-01 -4.65981156e-01
1.63939416e-01 -4.68020022e-01 -1.08725274e+00 -8.66610527e-01
-6.53682649e-01 -5.44607818e-01 4.85526234e-01 3.75128806e-01
1.14384465e-01 5.62008917e-01 -1.08434476e-01 8.30219984e-01
1.00956726e+00 1.85252771e-01 5.59220016e-01 -1.33917809e+00
-3.97371382e-01 -8.51547271e-02 -2.78222382e-01 -1.91511464e+00
-6.05947137e-01 -5.65325618e-01 5.35704419e-02 -1.79730701e+00
-6.65358663e-01 -8.10770452e-01 -9.08056498e-02 -1.71621740e-01
5.39878368e-01 2.53091186e-01 -5.91759160e-02 4.12712276e-01
4.52201635e-01 1.03433199e-01 1.10151553e+00 -9.62379873e-02
-4.10104543e-01 6.41803086e-01 -3.69913690e-02 5.78579485e-01
6.13199472e-01 -1.91821586e-02 -4.61530238e-01 -3.28722239e-01
4.84498516e-02 7.13772416e-01 3.79835546e-01 -1.48223531e+00
5.04931033e-01 6.06746003e-02 7.10358322e-01 -8.29631865e-01
6.16495967e-01 -1.25003135e+00 4.43726420e-01 1.17853689e+00
-7.40963668e-02 -6.02420807e-01 -4.75844085e-01 7.50808895e-01
1.70635447e-01 5.08605205e-02 4.38345879e-01 3.44499618e-01
-9.75858793e-02 -6.38500527e-02 -6.91493630e-01 -7.78824747e-01
8.23856711e-01 -4.45578843e-01 -3.06796908e-01 -4.25180018e-01
-5.53876102e-01 -1.01195298e-01 -3.44863772e-01 -7.84621760e-03
6.53647959e-01 -1.31281590e+00 -2.05267936e-01 6.26612961e-01
-2.63982147e-01 -3.44365358e-01 2.56928205e-01 8.48049879e-01
-2.58617580e-01 9.23667371e-01 -1.25063136e-01 -1.08766603e+00
-1.12498009e+00 1.64455280e-01 3.66984397e-01 -1.43654889e-03
-3.49031746e-01 7.01858461e-01 -3.43428612e-01 -4.57661375e-02
4.27156687e-01 -4.15983379e-01 -3.57945055e-01 6.63219839e-02
6.78388536e-01 4.33918238e-01 -1.90183878e-01 -3.43780339e-01
-5.97926915e-01 1.23453343e+00 9.28397596e-01 -7.60805070e-01
9.67246652e-01 -2.36623168e-01 1.62263215e-01 3.27409804e-01
8.42658103e-01 6.67832553e-01 -8.48764360e-01 -2.21954107e-01
-2.54077852e-01 -4.63968068e-01 -1.54836237e-01 -1.04282582e+00
-8.52658153e-01 7.23057628e-01 1.17416465e+00 5.54295480e-01
1.38522148e+00 -9.01665986e-02 6.82850897e-01 3.44780028e-01
1.05608928e+00 -6.90243721e-01 -1.30931102e-02 7.09021464e-02
5.20258844e-01 -3.43722522e-01 -3.68305832e-01 -7.43990600e-01
2.79114664e-01 1.26309526e+00 5.39363138e-02 -1.06591038e-01
7.91851461e-01 6.29619658e-01 4.56892431e-01 -1.15636341e-01
2.27504387e-01 7.21478183e-03 1.39525263e-02 9.17346060e-01
3.74082088e-01 2.48975649e-01 -5.15577972e-01 8.60806108e-01
-4.67567712e-01 1.16471060e-01 1.32219225e-01 9.81357038e-01
-7.72495210e-01 -8.41867387e-01 -8.66394997e-01 2.96433926e-01
-6.94521219e-02 5.12234092e-01 -9.64277163e-02 6.25408232e-01
1.82617620e-01 1.49542201e+00 2.32037291e-01 -6.50734484e-01
4.98638690e-01 -3.92146111e-02 7.39231110e-01 -1.54261336e-01
-1.41077004e-02 3.79651576e-01 -7.78755546e-02 -6.70332968e-01
-3.75716120e-01 -4.17846024e-01 -1.53940213e+00 2.38494515e-01
-3.45742375e-01 2.75892675e-01 1.10740793e+00 9.55911756e-01
2.13992715e-01 6.74109638e-01 8.43471110e-01 -7.09361970e-01
-7.83083081e-01 -9.95107055e-01 -9.96014297e-01 -4.85177577e-01
5.56638718e-01 -6.87363565e-01 -4.01769012e-01 -4.93623197e-01] | [6.548395156860352, 0.8196328282356262] |
b0254bf1-a646-48df-b9b8-f90a39b4c34e | 190503042 | 1905.03042 | null | http://arxiv.org/abs/1905.03042v1 | http://arxiv.org/pdf/1905.03042v1.pdf | Rumour Detection via News Propagation Dynamics and User Representation Learning | Rumours have existed for a long time and have been known for serious
consequences. The rapid growth of social media platforms has multiplied the
negative impact of rumours; it thus becomes important to early detect them.
Many methods have been introduced to detect rumours using the content or the
social context of news. However, most existing methods ignore or do not explore
effectively the propagation pattern of news in social media, including the
sequence of interactions of social media users with news across time. In this
work, we propose a novel method for rumour detection based on deep learning.
Our method leverages the propagation process of the news by learning the users'
representation and the temporal interrelation of users' responses. Experiments
conducted on Twitter and Weibo datasets demonstrate the state-of-the-art
performance of the proposed method. | ['Xiao Luo', 'Tien Huu Do', 'Nikos Deligiannis', 'Duc Minh Nguyen'] | 2019-04-18 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-4.75275934e-01 -4.21510190e-01 -3.49853665e-01 -3.85353893e-01
9.90797728e-02 -1.86813548e-01 1.09226704e+00 5.66009641e-01
-2.52289742e-01 5.71931124e-01 6.56095147e-01 -1.53909668e-01
1.68533608e-01 -9.80546713e-01 -3.49406600e-01 -2.57293433e-01
-4.40308541e-01 2.35954285e-01 5.58285296e-01 -6.15928352e-01
6.16749704e-01 7.42463395e-02 -1.14629245e+00 4.53867167e-01
4.61625129e-01 9.01267946e-01 6.94673415e-03 4.92767394e-01
-4.34479028e-01 1.37547195e+00 -6.98012114e-01 -1.61681801e-01
-3.42498958e-01 -6.24989688e-01 -7.17913270e-01 2.27702670e-02
-2.24112853e-01 -4.20805484e-01 -9.18619514e-01 1.05738342e+00
2.29760692e-01 1.21748611e-01 2.32418254e-01 -8.69523883e-01
-7.12923586e-01 8.26985002e-01 -9.15361166e-01 8.63091707e-01
4.03894067e-01 -3.38406295e-01 9.22826469e-01 -7.18091130e-01
5.41303694e-01 1.15472472e+00 8.00950706e-01 8.87874067e-02
-7.08805382e-01 -7.30228782e-01 2.17196077e-01 3.93639177e-01
-9.38981712e-01 -1.16957836e-01 8.13227355e-01 -6.55258000e-01
4.60005075e-01 6.90175369e-02 6.96817100e-01 1.24857688e+00
5.24216235e-01 8.93892884e-01 1.15565288e+00 -9.20203179e-02
-1.35430977e-01 2.82206774e-01 4.43594158e-01 6.34907484e-01
-2.54767150e-01 -2.63840426e-03 -7.15803564e-01 -4.19430584e-01
4.69789177e-01 4.52484608e-01 -7.57441521e-02 3.64810586e-01
-8.89425993e-01 1.23321426e+00 6.21477246e-01 3.79507840e-01
-6.74572408e-01 -2.76227325e-01 6.44540906e-01 5.27002454e-01
1.23368549e+00 2.85843730e-01 -1.01363525e-01 -5.49680926e-02
-1.07111490e+00 4.18018132e-01 9.99309897e-01 3.29869688e-01
6.15983546e-01 -2.20354840e-01 6.15127906e-02 9.37935114e-01
2.40043804e-01 2.02368915e-01 6.56053424e-01 -7.98446126e-03
1.15690030e-01 5.48561394e-01 2.71664798e-01 -1.79413855e+00
-8.34182203e-01 -6.84325695e-01 -7.87652135e-01 -4.33990985e-01
3.10549259e-01 -2.74434566e-01 -4.93018955e-01 1.37406278e+00
4.17054802e-01 4.70375568e-01 -3.58098865e-01 8.23916197e-01
8.33897293e-01 9.92656708e-01 -2.90437579e-01 -3.78811359e-01
1.13642621e+00 -1.01331484e+00 -1.03810465e+00 -1.26807511e-01
5.81235290e-02 -8.40684295e-01 4.55791920e-01 2.33719006e-01
-8.59014392e-01 -1.11669667e-01 -7.83703506e-01 4.41482455e-01
-3.21008056e-01 -5.89397788e-01 7.03525841e-01 1.68134421e-01
-4.90739495e-01 8.38193715e-01 -6.55008137e-01 -4.06378955e-01
4.64059651e-01 -3.59549187e-02 3.25334519e-01 3.82142812e-02
-1.73313451e+00 8.16475630e-01 4.69125994e-02 1.51138932e-01
-9.73791182e-01 -9.35621783e-02 -3.26139987e-01 -1.36205897e-01
3.60187501e-01 -2.18997374e-01 1.51598227e+00 -1.02176058e+00
-1.30388117e+00 3.75405401e-01 -1.39132500e-01 -6.40696108e-01
7.98547924e-01 -4.88897115e-01 -8.42092037e-01 -1.15682725e-02
-2.55599599e-02 -3.23144406e-01 8.56769800e-01 -9.49373305e-01
-8.80863309e-01 -2.94095725e-01 -4.92657050e-02 -9.50426385e-02
-4.77905840e-01 5.38555562e-01 -1.08233504e-01 -5.47206581e-01
1.70979813e-01 -8.94794226e-01 -7.11005181e-02 -6.14690781e-01
-5.34796119e-01 -2.29996383e-01 9.87235963e-01 -7.04556763e-01
1.33513546e+00 -1.90796232e+00 -4.36055481e-01 -4.51615192e-02
5.73237062e-01 2.32820809e-01 3.08068663e-01 9.61078286e-01
4.40090805e-01 6.76183850e-02 1.94816142e-01 -1.99402586e-01
-2.69843102e-01 2.47624349e-02 -7.66240537e-01 8.05030525e-01
-3.41836393e-01 5.94274819e-01 -1.44313264e+00 -6.07588775e-02
-4.26917315e-01 6.02069795e-01 -2.13000193e-01 3.76264125e-01
-2.71851420e-01 6.84375346e-01 -5.57352901e-01 2.07846418e-01
5.50004840e-01 -6.25474870e-01 -3.21199894e-02 3.42146516e-01
-2.80211151e-01 8.07572722e-01 -3.91606957e-01 7.34144628e-01
-2.14288816e-01 8.77432406e-01 -9.16453972e-02 -6.70868754e-01
8.92177463e-01 4.81339663e-01 4.44037020e-01 -5.41475713e-01
2.83475697e-01 8.99696201e-02 -6.52069002e-02 -7.30556488e-01
6.56585276e-01 -3.97891372e-01 -1.26154907e-02 9.33378041e-01
-3.16170245e-01 5.13186038e-01 6.70800880e-02 3.42597455e-01
1.00598836e+00 -4.01116401e-01 3.87569457e-01 8.92116204e-02
1.76368117e-01 -7.66736940e-02 3.56043845e-01 8.37969542e-01
-1.84181362e-01 3.85502905e-01 5.49545169e-01 -4.67685819e-01
-7.36751616e-01 -6.61374629e-01 -4.71536480e-02 1.45331049e+00
1.58991903e-01 -2.01260298e-01 -3.28051925e-01 -7.22109914e-01
4.90072742e-02 5.30820489e-01 -9.44602907e-01 -7.86336735e-02
-6.08762503e-01 -1.23056293e+00 4.64771211e-01 2.14835659e-01
6.84540153e-01 -1.21251702e+00 -1.15790248e-01 5.54267764e-01
-3.52456868e-01 -1.08510804e+00 -4.31263357e-01 -4.07148242e-01
-8.37821364e-01 -8.64585996e-01 -5.63245833e-01 -4.98075098e-01
3.49088669e-01 6.75699472e-01 7.79759347e-01 2.42050692e-01
2.30702505e-01 1.75435992e-03 -5.87329328e-01 -4.42965269e-01
-6.56763852e-01 5.02841696e-02 1.34042442e-01 4.95250583e-01
1.90619260e-01 -7.60923386e-01 -6.36099994e-01 4.91062790e-01
-8.17487776e-01 -4.89884876e-02 4.12443161e-01 7.60561585e-01
-2.23794967e-01 3.06229532e-01 1.14930177e+00 -1.39823413e+00
1.04894078e+00 -1.46782756e+00 -1.71565980e-01 -3.44494253e-01
-6.81990802e-01 -3.52798045e-01 5.05873680e-01 -5.89574277e-01
-1.13506222e+00 -7.38459349e-01 6.59805238e-02 2.23255560e-01
2.36493479e-02 1.17070818e+00 6.82588816e-01 2.51069278e-01
5.75584114e-01 3.42632174e-01 -1.45798966e-01 -6.33427382e-01
4.77040894e-02 9.46618199e-01 3.24404269e-01 1.20761238e-01
4.43259716e-01 8.56631398e-01 -5.83072424e-01 -1.02888429e+00
-1.32263637e+00 -9.28554058e-01 -1.07814170e-01 -3.93702328e-01
3.07746917e-01 -6.47218049e-01 -4.86378998e-01 9.18112814e-01
-1.36814725e+00 9.98036638e-02 4.62448686e-01 4.47638571e-01
8.33978131e-02 2.74294674e-01 -1.25144875e+00 -9.63032544e-01
-4.09738272e-01 -3.88495892e-01 -5.65579087e-02 2.07122847e-01
-3.11858863e-01 -1.27848077e+00 4.20588404e-01 2.35938162e-01
9.20495927e-01 1.69175908e-01 4.72814351e-01 -1.08049905e+00
-3.25499266e-01 -5.47273159e-01 -5.59770107e-01 -4.99408171e-02
3.36789042e-01 -2.37709329e-01 -1.03142691e+00 -2.62392819e-01
2.95645386e-01 -5.53330332e-02 1.07188535e+00 1.13908030e-01
5.18392146e-01 -8.62649441e-01 -4.12695050e-01 3.08908504e-02
9.04120624e-01 -8.08103383e-02 3.45296413e-01 4.69822049e-01
6.95398450e-01 7.50705838e-01 3.21469307e-01 9.05977845e-01
5.34469545e-01 2.80756027e-01 5.81368208e-01 2.32439265e-01
3.44452322e-01 -3.39695305e-01 5.82284510e-01 1.28392577e+00
-8.16890821e-02 -3.91531080e-01 -9.96205032e-01 6.48507118e-01
-2.03597283e+00 -1.18572402e+00 -4.90408927e-01 1.96872735e+00
7.86139488e-01 4.49474692e-01 4.07544881e-01 -1.31797627e-01
1.11198449e+00 7.29013622e-01 -3.05117071e-01 -1.79009050e-01
8.52299631e-02 -5.39379179e-01 2.09847435e-01 5.32822490e-01
-1.02902973e+00 6.28924727e-01 6.31283188e+00 8.06875750e-02
-1.52839947e+00 3.39189440e-01 2.72139251e-01 -1.02357596e-01
2.13848185e-02 -1.86852217e-01 -7.83122540e-01 7.54968405e-01
1.11735785e+00 -3.57885480e-01 4.17703450e-01 5.41175067e-01
4.75104809e-01 -4.16960120e-02 -5.39561808e-01 3.71281236e-01
9.74206850e-02 -1.38788402e+00 -2.96897709e-01 -3.85308568e-03
9.49123085e-01 7.04973817e-01 1.34527475e-01 3.09771806e-01
3.14778984e-01 -6.74935699e-01 5.36501765e-01 5.74382126e-01
-6.98605999e-02 -6.82384074e-01 9.59680021e-01 7.26270914e-01
-7.79512346e-01 -1.30554616e-01 -2.02201396e-01 -5.87595344e-01
4.31318462e-01 9.80935097e-01 -1.43765616e+00 1.61160707e-01
4.59102362e-01 1.01914787e+00 -3.84146184e-01 1.24537158e+00
-4.82631832e-01 1.28945637e+00 -6.07044436e-02 -3.21472406e-01
5.16387463e-01 1.44099534e-01 8.27869177e-01 1.53827369e+00
2.33113021e-02 -1.25798166e-01 2.54868537e-01 5.65970480e-01
-3.01412851e-01 1.60403550e-01 -4.16443050e-01 -4.82006609e-01
2.42181435e-01 1.08354700e+00 -7.47152805e-01 -2.79511750e-01
-5.54163456e-01 7.74505436e-01 3.18819910e-01 2.60242730e-01
-8.49076092e-01 -9.05849561e-02 2.35695735e-01 6.69125438e-01
1.67705327e-01 -1.78491414e-01 5.08198813e-02 -1.03698575e+00
-1.58453986e-01 -6.01604640e-01 5.14499903e-01 -2.76093334e-01
-1.78150129e+00 6.46197021e-01 -2.25894406e-01 -8.75055134e-01
-1.79175109e-01 1.72546789e-01 -1.06096005e+00 5.92167556e-01
-1.72076607e+00 -7.73917615e-01 -2.27994531e-01 3.45970958e-01
4.96570706e-01 6.26530424e-02 4.03245062e-01 2.24623486e-01
-4.28306401e-01 1.94519721e-02 2.96444863e-01 4.29297179e-01
6.00820780e-01 -9.55702901e-01 6.24428332e-01 4.79543954e-01
-1.05793469e-01 7.30890632e-01 1.15500951e+00 -8.74892354e-01
-8.89139533e-01 -1.00933552e+00 1.11012816e+00 -1.80872142e-01
1.29419792e+00 -1.65199280e-01 -1.33078766e+00 7.36173868e-01
4.68492657e-01 -9.07363072e-02 6.09627783e-01 4.68509734e-01
-5.10639071e-01 1.32193670e-01 -9.01534259e-01 3.53415370e-01
4.39060360e-01 -6.13634169e-01 -7.11317956e-01 5.67255020e-01
5.70614398e-01 -9.80858952e-02 -3.49079520e-01 8.08473751e-02
6.41332686e-01 -1.01227224e+00 4.60838318e-01 -7.44075358e-01
7.96232641e-01 1.55003324e-01 3.63664806e-01 -1.54810917e+00
-4.55490172e-01 -7.51760602e-01 -5.01805007e-01 1.09859705e+00
2.73489863e-01 -9.17407632e-01 5.30020177e-01 -4.32512611e-02
1.69979721e-01 -7.39907205e-01 -3.96598071e-01 -4.21620607e-01
-2.26456434e-01 -9.51733887e-02 2.60800987e-01 1.44091821e+00
4.78925556e-01 6.63493156e-01 -8.87674332e-01 2.85967141e-01
3.70581031e-01 4.23839957e-01 5.89800835e-01 -1.38521004e+00
-2.95257986e-01 -6.20003045e-01 -2.39840761e-01 -1.22285628e+00
-1.77603856e-01 -6.34837508e-01 8.00239816e-02 -1.39857054e+00
4.86684978e-01 -2.60617763e-01 -6.24343216e-01 -5.78438640e-02
-2.43379578e-01 7.82590285e-02 -2.08231479e-01 7.24789321e-01
-8.96810889e-01 4.35144246e-01 1.17340851e+00 3.59043330e-02
-2.03988433e-01 2.83483297e-01 -4.62095946e-01 1.03820002e+00
1.01067603e+00 -8.37356031e-01 -8.28736350e-02 -1.14651322e-01
1.02209485e+00 2.16408297e-01 2.53157854e-01 -6.25320554e-01
4.65165406e-01 -1.99069723e-01 -5.44632748e-02 -7.36503601e-01
2.65444905e-01 -1.74096435e-01 -3.05839986e-01 2.66483247e-01
-4.61509436e-01 4.01566848e-02 -7.07239956e-02 1.17316473e+00
-3.97764623e-01 -2.43131872e-02 8.10660362e-01 -2.21157655e-01
-4.70816523e-01 3.16603303e-01 -8.54096770e-01 1.38629466e-01
5.76277077e-01 5.30219138e-01 -4.92544770e-01 -1.02662337e+00
-6.56294644e-01 1.53846860e-01 1.94905743e-01 6.69338942e-01
6.62353158e-01 -9.41599727e-01 -1.00669014e+00 -7.37288296e-02
-1.60122737e-01 -4.75227416e-01 2.13472590e-01 1.20767212e+00
-1.55328289e-01 1.44533426e-01 -9.00322199e-02 -1.55637309e-01
-1.08691990e+00 6.44292891e-01 1.18645437e-01 -3.55215400e-01
-6.75916493e-01 8.71664226e-01 1.56219944e-01 -2.07837835e-01
-5.97589184e-03 2.39710614e-01 -7.69159913e-01 5.40028930e-01
1.12820590e+00 5.89758575e-01 -1.57883584e-01 -7.50582933e-01
-5.27216401e-03 -3.39709699e-01 -7.86411285e-01 -3.15620229e-02
1.67339373e+00 -4.76425737e-01 -4.52709973e-01 1.04070914e+00
1.10839105e+00 2.78907835e-01 -8.47651780e-01 -7.38619804e-01
2.55290061e-01 -5.64913869e-01 3.69747788e-01 -5.23815334e-01
-9.90773141e-01 7.14061975e-01 -1.95567697e-01 1.10420513e+00
3.82863998e-01 4.33324538e-02 1.47210670e+00 1.26753420e-01
2.75629401e-01 -8.16670060e-01 2.98826754e-01 1.06254232e+00
6.49299026e-01 -1.36590457e+00 1.33117527e-01 -2.78381735e-01
-5.46141505e-01 1.06112432e+00 8.31574053e-02 -2.58455634e-01
9.64823604e-01 -9.74194780e-02 2.09248006e-01 -5.48813939e-01
-7.81268239e-01 6.30315095e-02 6.25122786e-02 -7.13065788e-02
6.07036352e-01 2.92153731e-02 -4.82150227e-01 5.97404063e-01
4.71710674e-02 -2.48862840e-02 8.89271557e-01 1.03235590e+00
-8.37897658e-01 -5.35600901e-01 -3.71163189e-01 4.76614594e-01
-9.86430168e-01 -9.45626274e-02 -4.24024194e-01 2.67046690e-01
-3.63363147e-01 1.08894289e+00 -1.86184809e-01 -5.52848399e-01
-1.81096047e-01 4.00797315e-02 -7.27849156e-02 -5.67853570e-01
-5.34794450e-01 -1.57627463e-01 2.48321399e-01 -2.18208805e-01
-3.57450217e-01 -7.30035722e-01 -1.14039564e+00 -7.53129303e-01
-5.84367275e-01 2.83820599e-01 7.01654375e-01 1.19404387e+00
2.93252438e-01 5.02964616e-01 1.06420422e+00 -6.81266904e-01
-6.76631153e-01 -1.25373030e+00 -7.16204762e-01 4.25515294e-01
7.43385077e-01 -7.43240535e-01 -5.20973623e-01 -2.87757427e-01] | [8.166282653808594, 10.145696640014648] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.