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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
e90158d1-2a36-431a-b8e4-ce4f63721534 | domain-mismatch-doesnt-always-prevent-cross | null | null | https://aclanthology.org/2022.lrec-1.94 | https://aclanthology.org/2022.lrec-1.94.pdf | Domain Mismatch Doesn’t Always Prevent Cross-lingual Transfer Learning | Cross-lingual transfer learning without labeled target language data or parallel text has been surprisingly effective in zero-shot cross-lingual classification, question answering, unsupervised machine translation, etc. However, some recent publications have claimed that domain mismatch prevents cross-lingual transfer, and their results show that unsupervised bilingual lexicon induction (UBLI) and unsupervised neural machine translation (UNMT) do not work well when the underlying monolingual corpora come from different domains (e.g., French text from Wikipedia but English text from UN proceedings). In this work, we show how a simple initialization regimen can overcome much of the effect of domain mismatch in cross-lingual transfer. We pre-train word and contextual embeddings on the concatenated domain-mismatched corpora, and use these as initializations for three tasks: MUSE UBLI, UN Parallel UNMT, and the SemEval 2017 cross-lingual word similarity task. In all cases, our results challenge the conclusions of prior work by showing that proper initialization can recover a large portion of the losses incurred by domain mismatch. | ['Noah A. Smith', 'Phillip Keung', 'Daniel Edmiston'] | null | null | null | null | lrec-2022-6 | ['word-similarity', 'unsupervised-machine-translation'] | ['natural-language-processing', 'natural-language-processing'] | [-4.38422989e-03 -1.81844726e-01 -5.00920296e-01 -3.54962617e-01
-1.44330239e+00 -9.53718066e-01 8.57831538e-01 2.01479867e-02
-8.03366661e-01 1.16144252e+00 4.25726205e-01 -5.79634190e-01
3.24913830e-01 -4.18265194e-01 -9.40760314e-01 -4.11186337e-01
4.81565267e-01 1.04433978e+00 7.14040473e-02 -4.88275409e-01
-8.24508220e-02 -9.79067087e-02 -1.11297727e+00 3.94196302e-01
1.17646849e+00 3.38701606e-01 1.03774443e-01 1.74120754e-01
-3.68161589e-01 4.20759112e-01 -2.63321728e-01 -8.16746294e-01
3.89542580e-01 -6.86861396e-01 -1.14669883e+00 -4.27773148e-01
6.77012444e-01 -7.74282636e-03 -1.74134105e-01 1.10365438e+00
7.90077507e-01 -2.34770458e-02 9.02894914e-01 -1.02481723e+00
-8.80622983e-01 8.12023580e-01 -4.24717844e-01 1.34454757e-01
2.31773734e-01 -1.02209680e-01 1.04538763e+00 -1.17630446e+00
1.01227450e+00 1.28616786e+00 8.02238524e-01 4.76307601e-01
-1.42291713e+00 -8.48516226e-01 -3.40109169e-01 2.51320153e-01
-1.27238345e+00 -7.18362808e-01 5.02385020e-01 -3.70646685e-01
1.20538223e+00 -1.20843835e-01 -2.42855493e-02 1.55711377e+00
1.39853239e-01 7.85013199e-01 1.37727261e+00 -1.00683534e+00
-2.28712901e-01 5.79041123e-01 3.15994397e-02 3.15104961e-01
2.28994697e-01 1.32189706e-01 -5.59797883e-01 -8.46782401e-02
2.48952329e-01 -6.05817258e-01 -1.69093177e-01 -4.59798217e-01
-1.26815021e+00 9.63510990e-01 -3.30851111e-03 6.74587846e-01
1.30318776e-01 -3.72533202e-01 8.52435291e-01 9.74706709e-01
8.34285498e-01 5.98380804e-01 -8.70508134e-01 7.93401618e-03
-6.49179995e-01 -1.92916289e-01 8.12657177e-01 1.17305171e+00
1.08942378e+00 -2.56496400e-01 2.36376509e-01 1.32509828e+00
-1.78740650e-01 6.51219606e-01 9.54906464e-01 -4.34419274e-01
8.25154066e-01 2.40273774e-01 -4.38740030e-02 -4.31287289e-01
-3.06550320e-02 -4.59462479e-02 -4.88506436e-01 -2.99254537e-01
7.06589520e-01 -4.13818032e-01 -6.29455090e-01 2.17427778e+00
7.82150775e-02 -1.36067644e-01 5.05352616e-01 5.33004999e-01
5.53619146e-01 5.37818551e-01 1.99352562e-01 -9.21768397e-02
1.24605918e+00 -1.00000048e+00 -6.01945579e-01 -5.01305938e-01
1.21555328e+00 -1.23263216e+00 1.31000102e+00 -2.66252756e-02
-8.60338211e-01 -5.82686305e-01 -9.92062330e-01 -3.76716733e-01
-8.29950452e-01 4.07457575e-02 4.27391917e-01 4.86287981e-01
-8.71807277e-01 4.61529851e-01 -4.56628203e-01 -1.00178695e+00
4.17991765e-02 1.55228361e-01 -7.19165146e-01 -6.77310944e-01
-1.67438841e+00 1.53432393e+00 4.50276375e-01 -4.54311609e-01
-5.94702542e-01 -6.85202181e-01 -8.62659574e-01 -4.56802636e-01
8.10053349e-02 -3.72717023e-01 1.12398124e+00 -1.39157856e+00
-1.19188666e+00 1.64742935e+00 2.67203655e-02 -4.12162751e-01
4.27421808e-01 -3.62256229e-01 -5.22824347e-01 -3.23896527e-01
6.14974320e-01 7.27602601e-01 4.08056825e-01 -8.30488920e-01
-4.32014644e-01 -4.09421325e-01 -3.03556383e-01 4.25237000e-01
-3.62033963e-01 3.47205997e-01 -3.60297441e-01 -5.26115239e-01
-4.25672352e-01 -9.82303023e-01 1.13554493e-01 -6.83267832e-01
-3.92802469e-02 -4.71634477e-01 4.11577076e-01 -8.76844406e-01
7.24380255e-01 -2.05270004e+00 2.29620382e-01 -2.29000568e-01
-5.84499419e-01 4.46709454e-01 -6.01414859e-01 7.81408608e-01
-2.44465247e-01 7.01835379e-02 -2.55645841e-01 -2.31773466e-01
8.82440135e-02 5.54181278e-01 -3.82854700e-01 5.06486595e-01
2.97552913e-01 1.05376399e+00 -9.39497709e-01 -6.31389558e-01
-9.41511393e-02 3.49228054e-01 -3.90770614e-01 8.99888724e-02
-3.46014015e-02 5.49150646e-01 8.09276253e-02 3.11859369e-01
5.34356773e-01 1.93023115e-01 5.13442516e-01 -1.15726940e-01
-7.04024583e-02 8.87166381e-01 -4.82372016e-01 2.17949700e+00
-6.58948362e-01 6.66476071e-01 -1.76754996e-01 -1.17947662e+00
7.59413779e-01 4.71725583e-01 2.07936093e-01 -1.17754400e+00
1.14343464e-01 7.64951229e-01 1.24836832e-01 -4.06233102e-01
3.33301306e-01 -6.43634617e-01 -3.92209023e-01 7.75601804e-01
6.22966409e-01 -1.81917951e-01 1.25110433e-01 -3.90800536e-02
6.95242345e-01 3.36014628e-01 2.54744679e-01 -5.41329801e-01
3.95196527e-01 3.87450933e-01 4.64411616e-01 4.17275816e-01
-2.21393079e-01 3.93103600e-01 1.55379117e-01 -2.12092191e-01
-1.33378148e+00 -1.17329967e+00 -3.07055354e-01 1.45731986e+00
-4.10179496e-02 -1.95775121e-01 -8.38205099e-01 -8.99658203e-01
-9.40796658e-02 9.62209404e-01 -4.63596195e-01 -2.65933365e-01
-8.63856256e-01 -8.44007492e-01 9.04758871e-01 2.20209092e-01
1.20187975e-01 -1.08137310e+00 1.24751292e-01 2.61205584e-01
-5.70226550e-01 -1.33478975e+00 -7.41242588e-01 4.54154313e-01
-7.63816297e-01 -8.88019323e-01 -7.18490243e-01 -1.44340038e+00
3.97817135e-01 1.07952796e-01 1.64714622e+00 -4.51052487e-01
-1.04792021e-01 3.18986893e-01 -3.49627435e-01 -1.80202216e-01
-6.75830603e-01 4.92965043e-01 4.66363698e-01 -3.37360293e-01
1.17360139e+00 -6.02757752e-01 1.36904970e-01 4.17212069e-01
-7.05912530e-01 -8.24939013e-02 5.38431346e-01 1.14137793e+00
5.06053984e-01 -5.61896026e-01 7.99275041e-01 -1.11531723e+00
7.63245106e-01 -5.48344731e-01 -3.62817764e-01 5.89698553e-01
-4.42627192e-01 3.47597808e-01 6.11512065e-01 -6.31309688e-01
-9.63686168e-01 -2.24854782e-01 -1.08835794e-01 -2.49903813e-01
-2.21176565e-01 4.12727892e-01 -2.66917765e-01 1.84995487e-01
8.97868872e-01 1.82836503e-01 -1.80109337e-01 -7.10578442e-01
6.50245547e-01 8.56740475e-01 4.97983336e-01 -9.07798827e-01
5.95665693e-01 -1.84674598e-02 -7.77494490e-01 -5.34260035e-01
-9.45074260e-01 -4.64479715e-01 -1.10658681e+00 3.29410017e-01
1.00679731e+00 -1.20349514e+00 2.52404302e-01 2.14756772e-01
-1.37929547e+00 -5.29396951e-01 -2.37086773e-01 7.75122762e-01
-5.69348097e-01 9.46526229e-02 -6.02096617e-01 -1.01607606e-01
-3.03112447e-01 -9.76561487e-01 8.07160437e-01 -2.18754321e-01
-4.13725317e-01 -1.41560698e+00 6.85024917e-01 2.52159566e-01
3.46146554e-01 -1.80430755e-01 1.17120337e+00 -1.12638032e+00
-6.76859170e-02 -8.16086121e-03 -1.32390335e-01 6.00883245e-01
2.75475413e-01 -5.99199712e-01 -8.75469625e-01 -4.91131008e-01
-1.46550406e-03 -8.89935493e-01 5.43703854e-01 -1.03921078e-01
5.93366064e-02 -1.15708731e-01 -2.20392048e-01 5.40279865e-01
1.57542038e+00 -8.18353295e-02 4.06498790e-01 3.81071895e-01
6.66639328e-01 8.49304736e-01 4.95430440e-01 -5.08611917e-01
4.66885656e-01 7.03583121e-01 -3.73122513e-01 -1.73624068e-01
-4.00329351e-01 -3.38510573e-01 8.05075765e-01 1.63076448e+00
2.55585670e-01 5.60218431e-02 -1.09365344e+00 1.08242190e+00
-1.60717440e+00 -6.64301634e-01 -1.70243122e-02 2.33576512e+00
1.35867655e+00 -1.14063002e-01 -8.43702555e-02 -5.48579037e-01
6.33654058e-01 -2.68577218e-01 -4.22397166e-01 -6.97564662e-01
-5.54575920e-01 5.97183287e-01 7.06729174e-01 6.99087381e-01
-1.04683590e+00 1.53348124e+00 6.18026781e+00 9.70397651e-01
-1.02174294e+00 8.47141623e-01 2.69470215e-01 1.48700103e-01
-3.75307441e-01 -3.60742696e-02 -7.76137531e-01 2.33497158e-01
1.18677366e+00 -4.41313893e-01 4.64875787e-01 6.59217238e-01
-5.34363627e-01 1.61339477e-01 -1.43677437e+00 9.53093827e-01
4.53642964e-01 -7.61571109e-01 -1.12047575e-01 -1.19512692e-01
1.03932595e+00 8.18130910e-01 -8.57632309e-02 7.04650521e-01
8.18934202e-01 -9.45956469e-01 3.12216997e-01 -1.82504296e-01
1.21598041e+00 -7.22687185e-01 7.92959988e-01 4.52346474e-01
-7.51084864e-01 5.72688878e-01 -6.59608126e-01 1.22239076e-01
3.93270589e-02 2.11898103e-01 -7.78704047e-01 6.26216769e-01
4.14687753e-01 6.65715158e-01 -6.02588296e-01 3.66662294e-01
-3.82619768e-01 6.62596524e-01 -2.39860401e-01 2.83609480e-01
4.22066897e-01 -3.48039687e-01 3.24279726e-01 1.49010503e+00
3.15819442e-01 -5.21586359e-01 8.43481794e-02 4.27978277e-01
-3.91739815e-01 6.89957738e-01 -1.14213049e+00 -1.28607526e-01
1.93527609e-01 7.48086512e-01 -1.97132766e-01 -4.44879711e-01
-7.70577431e-01 1.37819934e+00 8.46977353e-01 3.26042116e-01
-5.99369526e-01 -4.27153319e-01 8.11805785e-01 -1.35004848e-01
1.57234877e-01 -2.09531695e-01 -1.83628872e-01 -1.64754784e+00
-8.65873136e-03 -1.08744681e+00 3.37436646e-01 -4.80300635e-01
-1.84138370e+00 7.24221945e-01 -2.62579679e-01 -1.23551416e+00
-5.13945103e-01 -6.63861156e-01 -2.85820514e-01 1.15232778e+00
-1.54944372e+00 -1.27690411e+00 5.22336483e-01 7.15802670e-01
5.71199059e-01 -4.20643508e-01 1.15729177e+00 7.67763734e-01
-2.95422524e-01 9.64891970e-01 5.58322728e-01 4.90136802e-01
1.65087438e+00 -1.04081202e+00 5.45943201e-01 7.52726793e-01
5.59921920e-01 6.20663524e-01 5.56415319e-01 -5.46647251e-01
-1.20450222e+00 -1.17214310e+00 1.61010349e+00 -8.90249670e-01
1.01193380e+00 -7.00222731e-01 -1.00758791e+00 1.02876449e+00
8.48540783e-01 -2.87082314e-01 9.57819164e-01 3.02658945e-01
-8.16819489e-01 1.13022909e-01 -9.05798554e-01 7.09107220e-01
9.57429588e-01 -1.08057439e+00 -1.08211470e+00 6.07267082e-01
8.09757233e-01 -6.71517923e-02 -8.39034498e-01 2.21749514e-01
4.63960648e-01 -6.27904534e-01 8.37118149e-01 -1.19283736e+00
4.47985590e-01 1.18127830e-01 -3.91015232e-01 -1.72419846e+00
2.80457307e-02 -3.42225879e-01 7.05867350e-01 1.43247688e+00
7.32446432e-01 -7.80935347e-01 8.02464485e-02 9.32286531e-02
-1.34807095e-01 -8.37751180e-02 -1.04879582e+00 -1.14322329e+00
1.09550917e+00 -4.18668389e-01 2.39809796e-01 1.59105325e+00
1.87663913e-01 1.06495321e+00 -4.16545033e-01 -3.34577471e-01
6.68989182e-01 3.02487537e-02 8.22563887e-01 -9.95380580e-01
-1.65091693e-01 -2.65530556e-01 7.00857639e-02 -8.70564699e-01
6.85577393e-01 -1.46795583e+00 2.20890656e-01 -1.15271831e+00
1.95959330e-01 -5.45838952e-01 -2.91534185e-01 4.63808537e-01
-1.40985698e-01 4.53073025e-01 -4.91005927e-02 3.13948870e-01
-4.26861316e-01 4.58504945e-01 9.56398189e-01 -1.53967559e-01
7.65009969e-02 -6.07849181e-01 -6.01976991e-01 5.80881417e-01
7.51581550e-01 -8.64970922e-01 -1.35769546e-01 -1.02915275e+00
7.17232227e-02 -2.51936287e-01 -2.96810508e-01 -5.70469737e-01
6.04650155e-02 5.60618713e-02 1.83392782e-02 -1.32901773e-01
-1.11344727e-02 -6.85010195e-01 -2.06265703e-01 2.04993695e-01
-3.23430538e-01 1.35344863e-01 5.00048637e-01 1.07061327e-01
-4.21773255e-01 -2.33188123e-01 8.61946464e-01 -1.23070799e-01
-5.78515410e-01 2.79149283e-02 -3.11777562e-01 8.53655636e-01
6.65275633e-01 2.32406124e-01 -3.38229388e-01 -4.90710046e-03
-5.16115248e-01 1.20641299e-01 7.36141384e-01 8.37055922e-01
-2.20512554e-01 -1.73998713e+00 -1.06453121e+00 3.01061898e-01
4.62647527e-01 -5.24656892e-01 -1.47543922e-01 9.86067593e-01
-1.80250779e-01 6.76112354e-01 -4.11130339e-01 -4.27096605e-01
-9.47810888e-01 6.29357517e-01 1.15900680e-01 -5.42183995e-01
-1.80996060e-01 8.45925212e-01 3.36184800e-01 -1.32849729e+00
-2.95717567e-02 1.72819093e-01 1.53395861e-01 2.83238649e-01
2.88779497e-01 -8.14822093e-02 2.25164622e-01 -9.56590354e-01
-4.13564026e-01 8.63174379e-01 -3.03142130e-01 -4.45442647e-01
1.31512249e+00 -2.31689230e-01 -1.36883050e-01 7.31176615e-01
1.58586574e+00 2.86870658e-01 -4.91661072e-01 -7.86819696e-01
2.91923881e-01 -5.39259170e-04 -4.52535629e-01 -8.19160819e-01
-4.75218982e-01 1.23588932e+00 4.07505363e-01 -2.66841829e-01
8.36480856e-01 2.43877739e-01 1.02499938e+00 3.78731340e-01
6.29823208e-01 -1.37340271e+00 -3.51989627e-01 1.01293886e+00
6.67063892e-01 -1.43883312e+00 -4.48935598e-01 -2.84405239e-02
-7.33608902e-01 6.95747256e-01 5.75808287e-01 -1.61000401e-01
4.10882622e-01 1.29969642e-01 4.50889915e-01 2.06109181e-01
-7.35368073e-01 -4.34195071e-01 2.12590083e-01 7.14805901e-01
7.04697549e-01 1.10325590e-01 -5.27448535e-01 5.21466911e-01
-3.86087865e-01 -2.22350195e-01 4.27643918e-02 7.04431236e-01
-9.66768861e-02 -1.72717547e+00 -2.25449279e-01 -8.39857757e-02
-5.55336535e-01 -7.04159081e-01 -7.28140116e-01 9.12481904e-01
2.71757334e-01 7.02421546e-01 1.92423686e-01 -2.67026961e-01
2.49421313e-01 8.09105873e-01 7.55996585e-01 -7.64119625e-01
-6.12932384e-01 3.73785049e-02 3.16653848e-01 -2.71784455e-01
-3.92652154e-01 -6.14965677e-01 -7.85702169e-01 -1.44720450e-01
-2.19865471e-01 4.14038807e-01 5.61353326e-01 1.16274142e+00
1.94530770e-01 1.58359297e-02 3.16451758e-01 -3.89950097e-01
-4.77774918e-01 -1.36208892e+00 -2.38937974e-01 7.17824340e-01
-1.66967825e-03 -4.87895489e-01 -4.97514307e-01 2.69273847e-01] | [11.03689956665039, 9.970775604248047] |
bef3be3f-ec28-4d50-a8f5-a2185446e193 | 3d-meta-point-signature-learning-to-learn-3d | 2010.11159 | null | https://arxiv.org/abs/2010.11159v1 | https://arxiv.org/pdf/2010.11159v1.pdf | 3D Meta Point Signature: Learning to Learn 3D Point Signature for 3D Dense Shape Correspondence | Point signature, a representation describing the structural neighborhood of a point in 3D shapes, can be applied to establish correspondences between points in 3D shapes. Conventional methods apply a weight-sharing network, e.g., any kind of graph neural networks, across all neighborhoods to directly generate point signatures and gain the generalization ability by extensive training over a large amount of training samples from scratch. However, these methods lack the flexibility in rapidly adapting to unseen neighborhood structures and thus generalizes poorly on new point sets. In this paper, we propose a novel meta-learning based 3D point signature model, named 3Dmetapointsignature (MEPS) network, that is capable of learning robust point signatures in 3D shapes. By regarding each point signature learning process as a task, our method obtains an optimized model over the best performance on the distribution of all tasks, generating reliable signatures for new tasks, i.e., signatures of unseen point neighborhoods. Specifically, the MEPS consists of two modules: a base signature learner and a meta signature learner. During training, the base-learner is trained to perform specific signature learning tasks. In the meantime, the meta-learner is trained to update the base-learner with optimal parameters. During testing, the meta-learner that is learned with the distribution of all tasks can adaptively change parameters of the base-learner, accommodating to unseen local neighborhoods. We evaluate the MEPS model on two datasets, e.g., FAUST and TOSCA, for dense 3Dshape correspondence. Experimental results demonstrate that our method not only gains significant improvements over the baseline model and achieves state-of-the-art results, but also is capable of handling unseen 3D shapes. | ['Yi Fang', 'Xiang Li', 'Lingjing Wang', 'Hao Huang'] | 2020-10-21 | null | null | null | null | ['3d-dense-shape-correspondence'] | ['computer-vision'] | [ 4.08406258e-02 -8.74714553e-02 -5.32407276e-02 -3.51415694e-01
-7.82502055e-01 -5.58836579e-01 5.69013596e-01 -4.24025841e-02
2.93379836e-02 1.10557094e-01 -2.13082269e-01 -1.76677957e-01
-1.86484709e-01 -9.77172315e-01 -9.19967055e-01 -7.09754705e-01
-2.66491175e-01 7.86782324e-01 6.40017211e-01 -2.87078083e-01
2.71045476e-01 1.14974391e+00 -1.54565001e+00 3.15356880e-01
8.25637817e-01 8.77268791e-01 2.58951575e-01 2.72449434e-01
-2.98993558e-01 3.90482917e-02 -5.48487008e-01 -2.34347865e-01
4.87950206e-01 5.35215810e-02 -5.18442512e-01 1.47362486e-01
7.53572285e-01 -1.20937690e-01 -5.52443922e-01 9.13130105e-01
5.82122028e-01 2.76740938e-01 7.21942127e-01 -1.38345802e+00
-1.00569749e+00 3.99306118e-01 -5.48559010e-01 -7.40746558e-02
1.55067548e-01 1.45988837e-01 9.39158618e-01 -1.45967352e+00
6.63146377e-01 1.26483929e+00 7.65872657e-01 7.49747157e-01
-1.08158278e+00 -5.72101116e-01 2.84040064e-01 -7.89684281e-02
-1.53402102e+00 -2.25647181e-01 1.18400800e+00 -1.06064953e-01
8.60233784e-01 1.67927757e-01 4.45650101e-01 9.79062080e-01
-2.38331124e-01 9.22036827e-01 5.16540885e-01 -1.56179026e-01
3.23354721e-01 -1.78989083e-01 -4.71690334e-02 7.98589587e-01
2.16701567e-01 2.49426410e-01 -4.05670255e-01 -4.42060858e-01
9.46914673e-01 7.67464414e-02 -2.23753974e-01 -9.09285367e-01
-1.19719529e+00 3.67656916e-01 9.05221522e-01 3.70138168e-01
-9.32382494e-02 8.16283152e-02 2.19731688e-01 2.16735631e-01
5.42181671e-01 3.00941080e-01 -5.08239031e-01 2.13268384e-01
-5.52966475e-01 2.70148486e-01 6.61985338e-01 1.32690012e+00
9.74946618e-01 6.67998120e-02 -1.11158729e-01 1.10403335e+00
2.41831809e-01 7.35309660e-01 4.35046107e-01 -4.26373541e-01
5.83742619e-01 1.07172894e+00 -1.43474579e-01 -1.08033490e+00
-4.54354495e-01 -4.67472732e-01 -9.73489940e-01 3.37691844e-01
1.54059902e-01 1.54290810e-01 -1.46635020e+00 1.70119894e+00
7.50450611e-01 8.63761008e-01 -2.45816424e-01 7.98496664e-01
1.04413962e+00 6.54041946e-01 -3.70429426e-01 4.77933466e-01
9.11235392e-01 -7.24226415e-01 8.68922174e-02 -1.15999006e-01
5.83266854e-01 -4.30673987e-01 1.05163038e+00 -8.33478346e-02
-9.75442827e-01 -8.36418688e-01 -1.06540310e+00 8.90268683e-02
-6.86834753e-01 -6.81102201e-02 4.35635805e-01 3.26723635e-01
-1.15108919e+00 7.91730881e-01 -8.63291740e-01 -1.21060632e-01
5.77984571e-01 4.90637571e-01 -4.21452194e-01 -5.24480641e-02
-7.74279237e-01 5.84589958e-01 4.86907780e-01 2.06323359e-02
-9.02935505e-01 -1.11407673e+00 -9.78421748e-01 2.55541392e-02
2.64254928e-01 -8.74965668e-01 1.02685440e+00 -4.92054313e-01
-1.32738936e+00 1.00520182e+00 -2.52397340e-02 1.04714066e-01
4.32757407e-01 2.74323821e-01 -4.84586030e-01 -9.80065987e-02
4.34598550e-02 6.64401412e-01 1.08242226e+00 -1.77745545e+00
-7.12805808e-01 -5.02170920e-01 -6.16746508e-02 1.22513488e-01
-1.82759210e-01 -6.09701037e-01 -6.83612227e-01 -7.68404782e-01
5.63131273e-01 -1.07499063e+00 -7.29478076e-02 2.30297029e-01
-3.60239804e-01 -5.98955631e-01 1.18455458e+00 9.13801566e-02
6.51469767e-01 -2.40622640e+00 4.55876403e-02 8.58476281e-01
2.80154824e-01 4.48092014e-01 -6.79748535e-01 2.34964922e-01
-1.99908346e-01 1.69741899e-01 -2.67089635e-01 -4.63973224e-01
1.09369084e-01 3.81490856e-01 -3.61302644e-01 3.82358402e-01
3.02702010e-01 1.24105835e+00 -1.26065302e+00 -5.53598888e-02
2.83017963e-01 4.52503085e-01 -2.89145291e-01 2.20204994e-01
-2.53760964e-01 3.19773316e-01 -7.15579212e-01 9.07556891e-01
9.93847668e-01 -5.52612841e-01 -1.00046493e-01 -2.81198412e-01
1.90638959e-01 -1.21644616e-01 -1.27568328e+00 1.98205674e+00
-3.96910489e-01 -7.86703601e-02 -1.50372088e-01 -8.76244128e-01
1.42878520e+00 -2.57044155e-02 5.78402400e-01 -3.67097020e-01
-1.36225939e-01 5.57243288e-01 -1.86112270e-01 -1.62175432e-01
3.11890692e-01 1.81897074e-01 6.47901222e-02 4.46599275e-01
2.58597165e-01 -3.60471189e-01 -1.26133174e-01 -2.45318338e-02
1.14830935e+00 1.47356793e-01 -1.57968163e-01 1.13342525e-02
6.97849274e-01 -1.74083799e-01 3.00173104e-01 8.75719488e-01
2.23106388e-02 7.34869242e-01 -1.30786061e-01 -6.55738473e-01
-1.03484857e+00 -1.52276754e+00 -1.97529837e-01 1.03986633e+00
5.04211426e-01 -2.50314921e-01 -2.20576406e-01 -1.10189641e+00
6.97640479e-01 5.42432606e-01 -6.93623364e-01 -4.28520083e-01
-7.85751283e-01 -4.17505741e-01 4.35435116e-01 6.29640639e-01
6.11452878e-01 -9.53443944e-01 -7.27270395e-02 -3.49086635e-02
4.35965121e-01 -8.19572389e-01 -6.91386223e-01 -1.93757176e-01
-1.21070719e+00 -1.13522303e+00 -7.85688818e-01 -1.13996828e+00
1.12833691e+00 6.54214978e-01 1.08285487e+00 5.32635450e-01
-9.73651707e-02 5.95955193e-01 -4.37598616e-01 -3.51135612e-01
-3.75702471e-01 2.91752130e-01 3.94737758e-02 1.60693243e-01
3.92740369e-01 -9.37446475e-01 -4.72409099e-01 6.07136250e-01
-8.99224043e-01 -3.71422976e-01 7.17796862e-01 8.83658826e-01
8.29104125e-01 -1.75455660e-01 7.16154516e-01 -8.37189078e-01
4.33640301e-01 -3.96668643e-01 -5.70674419e-01 3.68375272e-01
-4.24444079e-01 8.35102201e-02 6.06569469e-01 -5.58085859e-01
-7.00135410e-01 1.64816126e-01 -8.62008780e-02 -9.24835443e-01
-2.65918761e-01 3.38723958e-01 -2.10471511e-01 -6.14285946e-01
7.86742151e-01 3.37683976e-01 6.67914823e-02 -6.07678294e-01
4.74358767e-01 4.90053862e-01 6.77360117e-01 -9.26829040e-01
1.52112401e+00 6.43136322e-01 1.95859075e-01 -7.70403147e-01
-6.88411891e-01 -6.42997324e-01 -9.51357841e-01 -2.21068516e-01
2.39551827e-01 -6.77206337e-01 -4.22341049e-01 5.95378399e-01
-9.56540704e-01 -3.07158649e-01 -4.40422058e-01 9.32225585e-02
-6.30278528e-01 3.06806475e-01 -1.94709450e-01 -3.81332695e-01
-1.96532607e-01 -8.74866962e-01 1.42379379e+00 4.86527495e-02
1.27285957e-01 -1.27458107e+00 1.79888725e-01 -1.65217474e-01
2.38022327e-01 5.98213136e-01 1.08567083e+00 -7.72947729e-01
-7.86640406e-01 -3.55262160e-01 -1.47153109e-01 1.38992757e-01
5.54125428e-01 -9.94161367e-02 -9.56747532e-01 -5.63287675e-01
-2.94346392e-01 -2.04448014e-01 7.15536952e-01 -2.86347549e-02
1.56441844e+00 -3.89733166e-02 -7.19461322e-01 8.89868617e-01
1.16498101e+00 -8.26706830e-03 5.65530300e-01 1.50297046e-01
7.44074404e-01 1.87311441e-01 3.55646819e-01 1.70315921e-01
3.56599927e-01 6.25747025e-01 4.46678042e-01 -9.42681059e-02
-3.84321302e-01 -7.55967140e-01 -1.61634330e-02 7.17987835e-01
-2.74618059e-01 4.20001820e-02 -1.04096341e+00 6.31741405e-01
-2.02052808e+00 -7.53167212e-01 2.36600280e-01 2.24785471e+00
6.51701689e-01 1.58056036e-01 8.22301731e-02 -1.01266421e-01
7.92704344e-01 2.21646428e-01 -9.13793623e-01 1.01630472e-01
-1.04533523e-01 3.16808313e-01 2.40661055e-01 2.41351679e-01
-9.77620006e-01 9.66381967e-01 5.56712341e+00 8.54932129e-01
-1.10882676e+00 -1.53269500e-01 1.62810281e-01 1.74183115e-01
-5.30882895e-01 -2.09834293e-01 -6.56739533e-01 2.97113568e-01
2.47604385e-01 -2.54711002e-01 3.75072688e-01 1.06800330e+00
-1.06630415e-01 6.70702398e-01 -1.32303905e+00 1.12300205e+00
2.31111601e-01 -1.42131901e+00 5.38733721e-01 -9.31315646e-02
9.61064398e-01 2.30640218e-01 2.02570289e-01 3.44189286e-01
2.47070521e-01 -7.67001033e-01 5.49014986e-01 4.72650290e-01
8.70683730e-01 -6.46380305e-01 5.08529961e-01 4.36224639e-01
-1.33723891e+00 4.24594879e-02 -6.59511685e-01 5.20323932e-01
1.26008153e-01 6.44626021e-01 -9.66405571e-01 7.64365733e-01
5.56174636e-01 9.68547940e-01 -6.14321530e-01 1.32390118e+00
-3.48219872e-01 1.35961711e-01 -4.16964948e-01 -1.00670062e-01
3.30381036e-01 -1.42383128e-01 8.26420784e-01 8.84288192e-01
5.39509296e-01 9.87697989e-02 3.35127264e-01 1.03542018e+00
-3.45629454e-01 -1.51418820e-01 -9.34486508e-01 2.91954160e-01
7.01336920e-01 1.29431427e+00 -5.15319824e-01 -2.17658326e-01
-2.64588207e-01 6.95569634e-01 6.89846814e-01 4.70913559e-01
-5.26573896e-01 -3.35971117e-01 6.18765473e-01 2.19714627e-01
5.63485682e-01 -4.01398450e-01 -2.91795999e-01 -1.02673137e+00
7.27359727e-02 -5.75596809e-01 4.45588440e-01 -7.22557604e-01
-1.82086384e+00 4.89691079e-01 1.04800552e-01 -1.43892348e+00
-5.61966114e-02 -8.70611072e-01 -9.39158142e-01 7.17653453e-01
-1.61174917e+00 -1.48464954e+00 -5.72562754e-01 7.48999476e-01
2.11621359e-01 -2.85641819e-01 6.21613562e-01 2.16224149e-01
-2.10410938e-01 8.10612023e-01 -4.38500308e-02 3.58909130e-01
5.24413526e-01 -1.39325678e+00 8.00982296e-01 4.09556508e-01
4.28633600e-01 6.79515839e-01 4.25087549e-02 -8.18394661e-01
-1.52118611e+00 -1.48265874e+00 5.59801280e-01 -6.10866904e-01
5.81848502e-01 -5.73697984e-01 -1.25859439e+00 6.80239558e-01
-5.18460393e-01 3.81112337e-01 4.86957967e-01 1.80505887e-01
-5.35809934e-01 -2.71186739e-01 -1.24880505e+00 7.09949672e-01
1.81826735e+00 -5.10874331e-01 -8.96527767e-01 4.04129416e-01
6.45895600e-01 -9.39977765e-01 -9.01776314e-01 8.09491932e-01
2.86380738e-01 -6.74183607e-01 1.33644831e+00 -7.84072876e-01
-5.65090813e-02 -5.77956855e-01 2.08528060e-02 -1.63223851e+00
-6.14812672e-01 -4.43941593e-01 -4.90934640e-01 9.68947172e-01
5.69797754e-01 -6.65332735e-01 1.07787848e+00 4.68329906e-01
-5.27611792e-01 -9.46478903e-01 -1.13137829e+00 -1.17843556e+00
4.08782959e-02 -4.15003359e-01 1.33000159e+00 8.87231708e-01
-3.68297249e-01 7.04313889e-02 1.72887474e-01 5.37314892e-01
7.36064732e-01 4.25749451e-01 1.08838212e+00 -1.27586377e+00
3.96007486e-02 -5.43700159e-01 -6.28530025e-01 -1.30264020e+00
4.66664106e-01 -1.48289430e+00 -9.24392119e-02 -1.33532250e+00
-2.26221547e-01 -9.90908146e-01 -3.89822751e-01 7.52506495e-01
-1.49147600e-01 1.06491670e-01 9.24355984e-02 2.61321515e-01
-3.78315568e-01 6.48450792e-01 1.49337578e+00 -5.60795605e-01
-5.23526490e-01 3.23794961e-01 -4.22290832e-01 6.20992005e-01
6.49851680e-01 -2.87331402e-01 -6.23484612e-01 -6.16800249e-01
-1.57789188e-03 -6.03382170e-01 6.10982060e-01 -8.08291852e-01
4.97973859e-01 -1.26303047e-01 5.46831310e-01 -8.58268440e-01
3.38139325e-01 -9.84535336e-01 5.83148040e-02 1.15908101e-01
7.34591782e-02 1.20172620e-01 3.11578214e-01 6.77471638e-01
-2.90466603e-02 -3.03302765e-01 5.56428194e-01 -1.87170431e-01
-9.55200016e-01 1.00343239e+00 4.94687319e-01 6.05963171e-02
9.69426215e-01 -6.46723866e-01 -3.64549190e-01 -7.25372657e-02
-6.23450398e-01 3.72203559e-01 5.74127913e-01 8.62064421e-01
1.26033163e+00 -1.85272110e+00 -6.94049835e-01 5.62947512e-01
5.52013516e-01 5.99781334e-01 1.48890197e-01 2.32191101e-01
-1.62137926e-01 -9.67822820e-02 9.62861702e-02 -1.02327597e+00
-8.36765945e-01 6.05212390e-01 5.26692688e-01 1.51769429e-01
-8.81362319e-01 7.20497370e-01 -4.88477871e-02 -9.94115889e-01
2.20039368e-01 -4.33310896e-01 7.49130771e-02 -3.51905167e-01
4.12882477e-01 3.50211352e-01 2.63130784e-01 -5.85373521e-01
-3.95213246e-01 1.02131391e+00 -5.96722923e-02 3.41227829e-01
1.45271254e+00 3.73705596e-01 -5.22666723e-02 3.26421648e-01
1.38632941e+00 -1.28493421e-02 -1.16310132e+00 -4.50709492e-01
9.67020392e-02 -8.49609017e-01 -2.49243185e-01 -6.12776101e-01
-1.13673210e+00 3.94872129e-01 3.57968956e-01 1.37142017e-01
8.23370993e-01 2.96014905e-01 7.90268660e-01 6.17922902e-01
6.49245679e-01 -8.48327816e-01 3.08453202e-01 7.18364060e-01
1.20423126e+00 -1.11856318e+00 -2.41529092e-01 -3.93870890e-01
-1.60864294e-01 1.33387089e+00 7.63448060e-01 -3.73948336e-01
1.05649960e+00 1.26243173e-03 -1.13319466e-02 -5.49919963e-01
-5.09051800e-01 -1.13103233e-01 8.65081966e-01 8.80869269e-01
-2.69333243e-01 -1.60365235e-02 4.65910673e-01 1.74153090e-01
-1.26242608e-01 -1.34187952e-01 -1.94182470e-02 8.86070728e-01
-4.21410739e-01 -1.21666598e+00 -2.42144138e-01 6.05699778e-01
5.75621188e-01 3.14255714e-01 -5.70666075e-01 9.35693920e-01
1.63074121e-01 4.61060464e-01 1.94996357e-01 -6.66030705e-01
9.82331514e-01 -5.31035438e-02 4.53170806e-01 -8.79000723e-01
-4.01354074e-01 -4.69264746e-01 -3.55872244e-01 -4.92403418e-01
-4.15320843e-01 -4.51799393e-01 -1.30300891e+00 -3.08258891e-01
-4.08172697e-01 4.11556475e-02 4.12929893e-01 6.59405529e-01
7.29746580e-01 2.11177886e-01 1.09209883e+00 -9.96918678e-01
-8.51993024e-01 -6.66155994e-01 -5.76472521e-01 8.18080723e-01
2.22839177e-01 -8.77684176e-01 -4.92327243e-01 -4.49690908e-01] | [7.945202827453613, -3.563032627105713] |
b546e193-123d-48b6-b858-d7310e3bc933 | a-flexible-schema-guided-dialogue-management | 2207.07276 | null | https://arxiv.org/abs/2207.07276v1 | https://arxiv.org/pdf/2207.07276v1.pdf | A Flexible Schema-Guided Dialogue Management Framework: From Friendly Peer to Virtual Standardized Cancer Patient | A schema-guided approach to dialogue management has been shown in recent work to be effective in creating robust customizable virtual agents capable of acting as friendly peers or task assistants. However, successful applications of these methods in open-ended, mixed-initiative domains remain elusive -- particularly within medical domains such as virtual standardized patients, where such complex interactions are commonplace -- and require more extensive and flexible dialogue management capabilities than previous systems provide. In this paper, we describe a general-purpose schema-guided dialogue management framework used to develop SOPHIE, a virtual standardized cancer patient that allows a doctor to conveniently practice for interactions with patients. We conduct a crowdsourced evaluation of conversations between medical students and SOPHIE. Our agent is judged to produce responses that are natural, emotionally appropriate, and consistent with her role as a cancer patient. Furthermore, it significantly outperforms an end-to-end neural model fine-tuned on a human standardized patient corpus, attesting to the advantages of a schema-guided approach. | ['Ehsan Hoque', 'Caleb Wohn', 'Kurtis Haut', 'Lenhart Schubert', 'Catherine Giugno', 'Benjamin Kane'] | 2022-07-15 | null | null | null | null | ['dialogue-management'] | ['natural-language-processing'] | [ 7.54891261e-02 1.08026409e+00 2.85372622e-02 -8.00205708e-01
-8.79998028e-01 -5.34787595e-01 5.05783081e-01 4.52448040e-01
-5.41702986e-01 9.82587576e-01 7.90935397e-01 -1.75451785e-01
1.52804404e-01 -2.65303463e-01 1.10304989e-01 -1.78724557e-01
9.76272225e-02 1.30990708e+00 -4.29856300e-01 -9.96614456e-01
-9.58708078e-02 -4.48471829e-02 -8.81013811e-01 5.33975005e-01
7.37144291e-01 3.07324469e-01 9.96856019e-03 1.03613400e+00
-6.56628013e-02 1.24606848e+00 -8.89365196e-01 -6.06200933e-01
-2.74887264e-01 -6.06727898e-01 -1.44142056e+00 1.51290953e-01
-1.03551663e-01 -5.10426521e-01 2.99777510e-03 5.46908796e-01
1.13133109e+00 3.73276860e-01 3.22336644e-01 -1.15882921e+00
-4.52335864e-01 6.68159723e-01 4.05419409e-01 -3.81270260e-01
1.19325399e+00 3.59080911e-01 9.89204168e-01 -1.68905914e-01
1.45446968e+00 1.22190464e+00 6.07300460e-01 1.50285649e+00
-1.10472775e+00 -1.95176393e-01 -1.88571125e-01 -6.31407678e-01
-7.25138307e-01 -8.43636990e-01 5.96147537e-01 -3.17886025e-01
1.20438886e+00 4.22102690e-01 7.40816236e-01 1.73421717e+00
1.45062014e-01 8.55805755e-01 7.28998125e-01 -1.54989958e-01
3.33865643e-01 9.23680305e-01 -2.66065467e-02 6.69712424e-01
-4.24903542e-01 -2.90609270e-01 -5.61221361e-01 -6.79812908e-01
4.95758504e-01 -4.48753923e-01 -4.82548326e-01 -2.77599931e-01
-1.56494164e+00 1.15991580e+00 2.47042328e-01 2.41548121e-01
-4.71720487e-01 -2.74029702e-01 1.10351551e+00 3.85694474e-01
5.63638806e-01 1.22225428e+00 -3.87696773e-01 -7.66787708e-01
-3.18809956e-01 4.29899633e-01 1.59637618e+00 1.04332244e+00
-2.94083238e-01 -2.70738095e-01 -4.34004843e-01 1.05775654e+00
2.74755031e-01 4.65266891e-02 4.76713032e-01 -1.06133568e+00
3.27059686e-01 6.12840652e-01 4.38277334e-01 -5.71626842e-01
-9.37660754e-01 1.32675380e-01 -6.96858943e-01 3.51239257e-02
2.65306264e-01 -6.30433142e-01 -2.50248075e-01 1.72110271e+00
6.84253871e-01 -4.25704598e-01 9.23182547e-01 8.86238039e-01
1.57405376e+00 3.02869260e-01 3.56584489e-01 -2.93764055e-01
1.56262028e+00 -1.02001703e+00 -9.55959141e-01 -1.89543605e-01
9.73550618e-01 -6.46470666e-01 1.10136437e+00 2.70279050e-01
-1.41217339e+00 -2.98499013e-03 -6.25213921e-01 -2.45688885e-01
3.39854658e-02 -4.88701969e-01 8.15792978e-01 6.17811620e-01
-1.33781219e+00 3.61558110e-01 -5.96906066e-01 -7.85836875e-01
1.22165754e-01 3.67431134e-01 -6.08611465e-01 4.44408536e-01
-1.14509690e+00 1.13328540e+00 3.88910323e-02 -2.86028564e-01
-6.44619823e-01 -4.30745453e-01 -1.09658456e+00 -2.48499036e-01
2.02004820e-01 -1.01186454e+00 2.10754919e+00 -1.06065977e+00
-1.94598925e+00 1.21619260e+00 2.47288898e-01 -3.70422304e-01
7.92610526e-01 2.88154304e-01 -2.00381771e-01 3.53017151e-01
3.30482982e-02 9.57425475e-01 4.15868424e-02 -1.04750228e+00
-3.42599928e-01 9.90656987e-02 4.97912407e-01 7.99087524e-01
-1.54035404e-01 1.38986528e-01 5.85842542e-02 -2.69337207e-01
-5.63308299e-01 -1.06033874e+00 -7.62208343e-01 -4.17958796e-02
-5.91791749e-01 -4.54902142e-01 9.14320350e-03 -4.36958492e-01
7.59046614e-01 -1.92547119e+00 2.29917929e-01 -7.82816038e-02
4.45393413e-01 2.66472638e-01 -1.10276923e-01 6.97241724e-01
6.79606348e-02 1.40139582e-02 -1.42247304e-01 -4.73352432e-01
1.02991633e-01 1.63192824e-01 2.99031347e-01 2.92608496e-02
9.57570150e-02 8.02179694e-01 -1.24985027e+00 -7.40968347e-01
-3.21337604e-03 3.97750646e-01 -5.76518834e-01 7.93220878e-01
-4.99167234e-01 7.04389036e-01 -5.90359688e-01 2.64474511e-01
8.81398097e-03 -2.96190143e-01 4.54108775e-01 4.53700900e-01
3.57378513e-01 5.06151795e-01 -5.29769659e-01 1.95962381e+00
-5.89270473e-01 2.63242781e-01 5.89487314e-01 -1.69057578e-01
8.86136532e-01 1.00863087e+00 5.88490605e-01 -4.02363986e-01
2.46381477e-01 -6.39862046e-02 -1.67895913e-01 -9.85045075e-01
7.01125443e-01 -3.89946997e-01 -4.47701573e-01 8.43540549e-01
-1.08624265e-01 -6.04815543e-01 -1.42423421e-01 7.21990943e-01
1.14434290e+00 -1.94092467e-01 5.98179281e-01 3.18517908e-04
3.84758592e-01 4.78057832e-01 2.79536873e-01 6.30550623e-01
-7.38014698e-01 3.36873502e-01 6.33722961e-01 -3.66939038e-01
-9.19309318e-01 -5.46973526e-01 8.42796713e-02 1.21673656e+00
-5.55098094e-02 -2.94553310e-01 -8.51704240e-01 -8.49928975e-01
-2.70884961e-01 9.59083378e-01 -5.10471523e-01 5.75047359e-02
-1.75353527e-01 -1.95636600e-01 7.82553017e-01 2.57555991e-01
1.55997589e-01 -1.57333243e+00 -8.35405588e-01 7.00884759e-01
-6.28626108e-01 -1.13604617e+00 -5.77474952e-01 -2.57416791e-03
-3.36463600e-01 -9.58062172e-01 -6.08638287e-01 -8.81932378e-01
4.91992414e-01 -4.72444147e-02 1.45760262e+00 1.26022995e-01
-2.82345086e-01 8.66735101e-01 -3.41920435e-01 -4.86806065e-01
-1.20419598e+00 1.57741785e-01 -1.40492916e-01 -4.38300461e-01
1.97502345e-01 -1.36251956e-01 -5.94161689e-01 3.70919943e-01
-6.59378111e-01 4.04572129e-01 3.71580534e-02 1.06889427e+00
-1.78043589e-01 -8.54553640e-01 8.39839756e-01 -1.48272407e+00
1.47582400e+00 -5.30963600e-01 1.82499915e-01 1.92089289e-01
-3.79070997e-01 -2.41805568e-01 4.67667520e-01 -2.10242346e-01
-1.28080785e+00 2.78341562e-01 -4.81875837e-01 2.30047628e-01
-4.08144534e-01 2.96753109e-01 1.54495537e-01 3.13377343e-02
1.16730237e+00 -2.07843602e-01 4.09176797e-01 7.64077753e-02
2.52396286e-01 1.21277916e+00 5.60494900e-01 -4.94284064e-01
1.95622534e-01 1.33723736e-01 -7.02397466e-01 -6.25880837e-01
-5.24064660e-01 -5.43858230e-01 -2.97121308e-03 -3.26030612e-01
1.09392822e+00 -1.02989841e+00 -1.39453948e+00 5.70383947e-03
-1.27118266e+00 -7.67521560e-01 -1.95858657e-01 3.08082581e-01
-8.04615736e-01 8.84267390e-02 -8.82703364e-01 -8.40923607e-01
-8.59256923e-01 -1.12130117e+00 1.01065445e+00 3.16676885e-01
-1.24097240e+00 -1.37701654e+00 4.65222538e-01 8.34586084e-01
4.90858644e-01 3.61582309e-01 6.17698133e-01 -1.28759813e+00
2.48912275e-01 -2.60628492e-01 1.67529523e-01 -1.54802963e-01
5.89728020e-02 -1.08413994e-01 -9.27742422e-01 -1.64968371e-01
-6.30725175e-02 -1.13891184e+00 -2.29923308e-01 -8.41181874e-02
4.05214846e-01 -4.12352920e-01 -3.13106418e-01 1.34935617e-01
7.30307698e-01 3.98064286e-01 2.39955649e-01 1.99736655e-01
2.16178522e-01 1.00834072e+00 8.66611004e-01 8.95453870e-01
9.87733066e-01 5.47477126e-01 9.92077515e-02 -3.19070727e-01
6.14016593e-01 5.87464869e-03 2.16487527e-01 4.49384481e-01
1.31993324e-01 -4.87128705e-01 -1.00651348e+00 4.85849231e-01
-1.95903027e+00 -7.99462914e-01 1.21749155e-01 1.70070708e+00
1.31497240e+00 -7.11285248e-02 2.78500915e-01 -5.59090793e-01
4.39785242e-01 2.61040982e-02 -5.12230992e-01 -1.10490036e+00
1.97667316e-01 4.26398963e-02 -1.66467473e-01 5.34855306e-01
-9.34228539e-01 9.19344962e-01 6.23919392e+00 1.70142967e-02
-8.85473192e-01 1.09858684e-01 8.64884317e-01 -3.46544415e-01
-3.57079297e-01 -6.11925900e-01 -1.93452969e-01 2.37710997e-02
1.04423630e+00 -2.95904607e-01 2.53367156e-01 1.08264291e+00
3.96645844e-01 -6.18177541e-02 -1.66113234e+00 8.73049796e-01
7.07401484e-02 -1.29588175e+00 -6.19212270e-01 -1.72264770e-01
4.67911512e-01 -1.65821701e-01 -1.45468965e-01 4.29280281e-01
9.28416610e-01 -1.17630911e+00 3.61792475e-01 2.13052064e-01
6.27277970e-01 -6.06106877e-01 1.00714195e+00 4.76227671e-01
-3.58166099e-01 2.70938128e-01 3.36250365e-02 8.75864923e-02
3.28860223e-01 -1.66356027e-01 -2.00202131e+00 1.56157851e-01
2.96874225e-01 1.32942140e-01 3.36366110e-02 5.82026541e-01
3.94004211e-02 1.56679619e-02 8.32309201e-02 -5.95687985e-01
4.16046172e-01 1.88449129e-01 4.41249818e-01 1.47387195e+00
-3.65667492e-01 5.48020840e-01 5.12244463e-01 3.20643693e-01
-1.97087735e-01 4.80422169e-01 -8.58506739e-01 -3.21383595e-01
2.78945208e-01 1.46375406e+00 -4.40458566e-01 -4.07304078e-01
-3.65392715e-02 1.11317015e+00 3.17903459e-01 -4.67718914e-02
-5.19687355e-01 -2.91725039e-01 7.13189840e-01 -3.70815009e-01
-4.58011240e-01 3.13431412e-01 -8.95980075e-02 -8.95429671e-01
-1.99516520e-01 -1.56832063e+00 4.48066413e-01 -8.67515624e-01
-1.20426464e+00 1.21129870e+00 -5.46950638e-01 -1.10343099e+00
-8.94471765e-01 -1.85011700e-01 -5.62688053e-01 6.35684669e-01
-8.59483838e-01 -1.13993669e+00 -3.38596910e-01 7.55946696e-01
6.70108795e-01 -2.87849426e-01 1.63915133e+00 -2.50685930e-01
-5.17639399e-01 6.17198288e-01 -2.76673406e-01 1.46759301e-01
1.24739981e+00 -1.36050355e+00 2.91559190e-01 -1.56721652e-01
-4.74135578e-01 4.97710377e-01 9.78498042e-01 -7.03858137e-01
-1.41585779e+00 -7.69412458e-01 1.05005527e+00 -7.02945113e-01
4.50736880e-01 -5.06812513e-01 -7.14074850e-01 6.41803503e-01
8.08368802e-01 -4.38116074e-01 1.15261221e+00 1.40188530e-01
-6.01824038e-02 4.67163116e-01 -1.78432679e+00 7.27641106e-01
8.85984838e-01 -6.12767577e-01 -5.98476946e-01 1.07305455e+00
7.65817344e-01 -8.85051966e-01 -1.11660743e+00 -2.07832932e-01
3.64113659e-01 -7.43634343e-01 5.67220867e-01 -1.03633964e+00
6.79072320e-01 4.16353703e-01 3.24654132e-01 -1.65476465e+00
6.70400113e-02 -1.35134935e+00 6.55170918e-01 1.09880316e+00
6.39875233e-01 -6.21474504e-01 7.71203756e-01 1.52961218e+00
-1.95066810e-01 -6.30985916e-01 -7.03130662e-01 1.89793110e-02
-1.85960233e-01 1.59635097e-02 4.03863311e-01 1.34904730e+00
1.30195415e+00 6.65637076e-01 -3.52998763e-01 -3.01154792e-01
-4.79976423e-02 -4.26427186e-01 9.52093959e-01 -1.19008696e+00
-4.05982792e-01 -3.47979993e-01 -7.44506493e-02 -4.70539033e-01
2.42357433e-01 -8.67713213e-01 4.73055422e-01 -1.70263696e+00
1.01525396e-01 -6.50737286e-01 3.50046933e-01 6.55948162e-01
-1.49498433e-01 -1.52099937e-01 7.76994973e-02 -1.45521209e-01
-8.13332140e-01 3.66901308e-01 1.28431630e+00 -1.37194172e-01
-6.85493112e-01 1.82325393e-01 -1.22169757e+00 5.79838395e-01
7.07417786e-01 -2.08023682e-01 -5.93226075e-01 5.51926438e-03
2.77464569e-01 9.70836043e-01 -6.66775852e-02 -4.09039617e-01
4.06873405e-01 -2.05261692e-01 -1.36159495e-01 1.10005841e-01
5.30200124e-01 -5.16970575e-01 4.88101542e-02 3.21609497e-01
-1.04171395e+00 -1.20741591e-01 1.78852439e-01 2.98305988e-01
-1.01325037e-02 -2.99308926e-01 6.92555785e-01 -4.26133394e-01
-2.24123135e-01 -8.25338438e-02 -7.62589514e-01 2.75197238e-01
1.19192994e+00 1.23839289e-01 -3.30040038e-01 -1.07453787e+00
-9.39899445e-01 7.42068172e-01 4.97208089e-01 3.61754268e-01
6.47320807e-01 -8.57759058e-01 -1.01522458e+00 -2.71535814e-01
4.60649550e-01 1.16063401e-01 2.98336536e-01 4.59752113e-01
-7.40761280e-01 3.07496279e-01 -2.69802511e-01 -4.09917295e-01
-1.68088913e+00 3.65080774e-01 3.97944450e-01 -4.12947237e-01
-7.59909570e-01 1.18606794e+00 -1.08544137e-02 -9.81572449e-01
5.25829196e-01 -8.53189975e-02 -3.72698605e-01 1.06525525e-01
5.91601551e-01 -2.18545124e-01 -7.38650337e-02 -5.04220605e-01
-2.36920908e-01 -3.21982592e-01 -3.03279907e-01 -4.96965885e-01
1.31212223e+00 -1.35064777e-02 2.60844305e-02 2.77183503e-01
8.18457544e-01 -2.49123693e-01 -6.49398744e-01 -1.67437181e-01
-8.64119455e-02 5.74887246e-02 -4.14063603e-01 -1.29976523e+00
-4.20587897e-01 4.34506297e-01 2.26014927e-01 3.65688145e-01
6.55609250e-01 -5.00092842e-02 7.92982876e-01 9.90905046e-01
3.04258108e-01 -1.26473498e+00 1.22238085e-01 3.55998874e-01
1.04083276e+00 -1.61057997e+00 -3.28579009e-01 -3.14715773e-01
-1.67414939e+00 1.07501054e+00 7.46276140e-01 4.48678136e-01
1.66508228e-01 2.05880046e-01 8.19418192e-01 -4.00915742e-01
-1.38049114e+00 3.32955629e-01 -2.41988003e-01 7.08723187e-01
7.89829612e-01 5.08170843e-01 -8.15850031e-03 4.30664062e-01
-5.18164754e-01 5.09894080e-03 1.00255525e+00 1.12070227e+00
2.03897748e-02 -9.17020321e-01 -1.83983818e-01 1.62912726e-01
-2.60424465e-01 9.37689841e-02 -1.01174510e+00 5.16607583e-01
-4.57789570e-01 1.30516493e+00 -2.91984901e-02 -9.60842445e-02
6.09943628e-01 2.62356758e-01 5.67385033e-02 -9.44984555e-01
-1.60412097e+00 -1.46174535e-01 1.21397221e+00 -6.18008554e-01
-2.80753314e-01 -5.74782073e-01 -1.39609504e+00 -2.71972418e-01
-3.34429815e-02 4.16158527e-01 8.03314328e-01 6.67776585e-01
2.60492176e-01 4.53100145e-01 8.18263590e-01 -6.85483992e-01
-6.52425945e-01 -9.03685093e-01 -1.38994426e-01 7.98954666e-01
3.40739191e-01 -5.37365675e-02 7.89436027e-02 -2.46472523e-01] | [12.75011157989502, 8.203629493713379] |
748b1e02-1882-40d2-8eea-721d0f15855b | towards-using-data-centric-approach-for | 2205.13022 | null | https://arxiv.org/abs/2205.13022v2 | https://arxiv.org/pdf/2205.13022v2.pdf | Towards Using Data-Influence Methods to Detect Noisy Samples in Source Code Corpora | Despite the recent trend of developing and applying neural source code models to software engineering tasks, the quality of such models is insufficient for real-world use. This is because there could be noise in the source code corpora used to train such models. We adapt data-influence methods to detect such noises in this paper. Data-influence methods are used in machine learning to evaluate the similarity of a target sample to the correct samples in order to determine whether or not the target sample is noisy. Our evaluation results show that data-influence methods can identify noisy samples from neural code models in classification-based tasks. This approach will contribute to the larger vision of developing better neural source code models from a data-centric perspective, which is a key driver for developing useful source code models in practice. | ['Nghi D. Q. Bui', 'Anh T. V. Dau', 'Hoang Thanh-Tung', 'Thang Nguyen-Duc'] | 2022-05-25 | null | null | null | null | ['code-classification'] | ['computer-code'] | [ 2.36253850e-02 -6.53262585e-02 -9.86575708e-02 -5.07314086e-01
-7.93230593e-01 -3.52385491e-01 1.40257418e-01 3.69410962e-01
-1.19088478e-01 1.24767281e-01 2.47570068e-01 -4.40862834e-01
1.65308279e-03 -7.17254400e-01 -8.20556760e-01 -2.89263636e-01
2.34867170e-01 -5.23370616e-02 2.13055447e-01 -1.71483949e-01
6.34549499e-01 5.96810244e-02 -1.93864930e+00 7.80763566e-01
9.80014265e-01 5.25482535e-01 3.80614698e-01 5.88888824e-01
-6.43293619e-01 9.51207519e-01 -9.18812454e-01 -3.29577506e-01
-2.43480261e-02 -6.29289091e-01 -6.34283721e-01 -3.57398838e-01
4.22447592e-01 4.90350053e-02 3.40327293e-01 1.69394577e+00
2.86054552e-01 -2.37813532e-01 3.93317163e-01 -9.91752744e-01
-7.12411880e-01 9.28783357e-01 -1.75385460e-01 3.12605977e-01
3.75518024e-01 -5.12992926e-02 9.34452474e-01 -9.83516455e-01
7.00721860e-01 1.08201730e+00 8.17106783e-01 7.30844915e-01
-1.37380636e+00 -6.38691366e-01 -6.29070774e-02 1.03219591e-01
-9.51898873e-01 -5.52901924e-01 9.35724378e-01 -8.42418134e-01
9.67482388e-01 1.98842704e-01 4.76932675e-01 1.18250906e+00
3.00636679e-01 8.40339839e-01 7.47403979e-01 -7.39910185e-01
4.47703600e-01 3.77060086e-01 6.20395720e-01 5.04495263e-01
5.23438811e-01 1.37567744e-01 -5.15048444e-01 -3.89090180e-01
3.09548914e-01 -2.81110674e-01 -3.66332918e-01 -4.01021354e-02
-7.80663729e-01 1.09392309e+00 2.56632179e-01 7.96944320e-01
-2.58789122e-01 2.24793538e-01 6.12597585e-01 6.92326128e-01
5.62103570e-01 8.66754055e-01 -6.26356125e-01 -8.22410285e-01
-1.08165407e+00 2.65092015e-01 1.05273092e+00 1.04996872e+00
6.58941209e-01 3.00998241e-01 5.01795076e-02 1.04005861e+00
4.83004063e-01 3.62819344e-01 7.81668842e-01 -8.49087656e-01
4.25870478e-01 1.19117618e+00 -2.84539580e-01 -1.07364333e+00
-5.65365776e-02 -4.58571136e-01 -3.16144735e-01 4.74470407e-01
4.14315104e-01 3.94583941e-02 -5.74179530e-01 1.40482557e+00
-1.49665317e-02 -7.98782185e-02 -7.69254491e-02 5.67163169e-01
7.18837202e-01 5.49821615e-01 -3.45410556e-01 1.67302370e-01
9.29111123e-01 -6.12744749e-01 -5.82078278e-01 -3.72484207e-01
1.06492746e+00 -8.53994250e-01 1.42272484e+00 6.62027895e-01
-6.96392655e-01 -6.91500604e-01 -1.11120987e+00 2.06370845e-01
-3.56899202e-01 3.22508067e-01 2.52500951e-01 9.95875180e-01
-9.02945995e-01 6.60816669e-01 -6.76556706e-01 -1.43789291e-01
5.54136693e-01 -1.63371256e-03 -1.01114444e-01 -7.62802064e-02
-7.38246500e-01 6.78243101e-01 3.24750215e-01 4.67568971e-02
-9.03744638e-01 -7.95615256e-01 -8.03964138e-01 2.65457351e-02
2.74126947e-01 -4.03436366e-03 1.45712996e+00 -1.56493700e+00
-8.90659273e-01 5.89363217e-01 -1.04106762e-01 -3.52602214e-01
2.20454931e-01 -1.05100907e-01 -3.79594624e-01 -5.41706443e-01
3.75022329e-02 -6.89280778e-02 8.69347692e-01 -1.30849004e+00
-6.33376837e-01 -2.21386105e-01 -1.34163618e-01 -5.33184290e-01
-3.99726421e-01 2.50289559e-01 -1.93194643e-01 -6.64485395e-01
-1.40356109e-03 -7.37766981e-01 -1.19981550e-01 -1.95362449e-01
-2.42593959e-01 -2.76153743e-01 7.47486651e-01 -5.45124173e-01
1.43415868e+00 -2.23583794e+00 -2.54971802e-01 3.84123981e-01
3.55744958e-01 4.29571420e-01 -3.73335868e-01 2.70869762e-01
-1.54546976e-01 3.97729814e-01 -3.90859991e-02 3.21648014e-03
-1.30351335e-01 -3.93247837e-03 -3.02683949e-01 1.37386709e-01
3.90945524e-01 5.07613659e-01 -9.42331851e-01 -8.10940564e-02
-1.28443003e-01 3.87986690e-01 -7.06793010e-01 4.07039255e-01
-3.99884880e-01 -8.53073522e-02 -4.81454581e-01 5.85338056e-01
3.14160496e-01 5.11526987e-02 -2.87583470e-01 2.18679264e-01
3.26165296e-02 6.26589119e-01 -1.08525026e+00 1.50863099e+00
-6.77816868e-01 1.12677777e+00 -1.04826845e-01 -7.41912961e-01
1.20941365e+00 1.93748385e-01 -6.36207089e-02 -6.74683690e-01
2.14285880e-01 4.93138701e-01 6.01568818e-01 -9.03332889e-01
5.31162143e-01 9.68997180e-02 2.08559647e-01 4.43435729e-01
2.79639941e-02 -7.52402768e-02 5.50047278e-01 -4.05187532e-02
1.29207778e+00 1.15033738e-01 4.22675647e-02 -4.76078480e-01
3.94112855e-01 7.58009925e-02 6.07543707e-01 8.36781561e-01
-1.04445025e-01 7.77666211e-01 9.19951260e-01 -3.24413985e-01
-1.11638057e+00 -4.97449458e-01 -1.36150703e-01 9.60568607e-01
-4.55032259e-01 -6.30656064e-01 -1.06043494e+00 -8.64761829e-01
-6.47708401e-02 1.04574919e+00 -7.73936689e-01 -4.75449055e-01
-3.38896751e-01 -4.34430003e-01 6.46889746e-01 4.69304711e-01
-2.39173964e-01 -1.15317523e+00 -7.17533648e-01 4.61293489e-01
-4.53642868e-02 -3.64676863e-01 -4.48867977e-01 5.20001531e-01
-9.02068436e-01 -1.41193461e+00 -2.75107831e-01 -5.80585897e-01
7.97483504e-01 1.44469440e-02 1.48574424e+00 4.97816265e-01
-4.64756275e-03 7.18404576e-02 -7.35712767e-01 -8.34497750e-01
-1.47382748e+00 1.28492936e-01 -3.64152968e-01 -4.12729681e-01
9.99636114e-01 -1.96002170e-01 2.17162743e-02 2.25724667e-01
-1.06157231e+00 -2.93704659e-01 4.37180966e-01 7.39708483e-01
9.23864394e-02 6.82063550e-02 6.80630624e-01 -1.00745344e+00
1.28868735e+00 -7.54813910e-01 -8.16506684e-01 1.69208050e-01
-8.06339562e-01 2.99574077e-01 7.73120046e-01 -5.51620603e-01
-7.19170213e-01 -2.14157581e-01 -4.41825867e-01 -1.72420859e-01
-2.27108240e-01 1.07293952e+00 7.95945898e-02 8.17642659e-02
1.42775559e+00 8.56146738e-02 9.76773128e-02 -5.62133670e-01
-2.28609636e-01 1.01560545e+00 -4.48131002e-02 -6.30318642e-01
4.26761180e-01 -3.27639580e-01 -5.84011734e-01 -7.10288942e-01
-5.62144041e-01 -5.60009122e-01 -3.85769606e-01 -2.82918364e-01
1.56310648e-01 -6.59894288e-01 9.90763120e-03 3.50099802e-01
-1.45295882e+00 -2.67656952e-01 -2.12570891e-01 2.37873301e-01
-1.12431735e-01 1.19258136e-01 -2.15023190e-01 -8.22148561e-01
-2.37412099e-02 -1.66320646e+00 7.02024162e-01 1.33792698e-01
-6.26888633e-01 -9.11035061e-01 3.03849816e-01 1.35340273e-01
6.38485074e-01 -7.64095485e-02 1.09716523e+00 -1.05805540e+00
-3.13635081e-01 -5.29376447e-01 9.61330384e-02 8.76118720e-01
6.51950464e-02 6.10667884e-01 -1.14647293e+00 4.27713543e-02
3.27725828e-01 -2.03513131e-01 4.97767776e-01 2.41003975e-01
1.11974704e+00 -1.23383500e-01 -2.07621276e-01 1.35934204e-01
1.46458244e+00 4.58271116e-01 6.21941090e-01 2.49915034e-01
5.57064950e-01 6.54929221e-01 4.57208425e-01 1.90891773e-01
6.67396188e-02 3.73570949e-01 4.41247880e-01 2.29060352e-01
-9.92348120e-02 -1.59751475e-01 6.21151686e-01 1.08825958e+00
2.82766134e-01 7.40646049e-02 -1.42839181e+00 8.64845455e-01
-1.64018118e+00 -8.32974613e-01 -7.27008879e-01 2.15429759e+00
1.13519180e+00 3.93552959e-01 -5.25285751e-02 4.77036685e-01
5.05964458e-01 -3.12685370e-01 -3.69583517e-01 -6.43815041e-01
3.08024913e-01 8.31603259e-02 -4.45031375e-02 5.11181764e-02
-7.79933453e-01 4.28929597e-01 6.24132776e+00 9.40313041e-01
-1.07627821e+00 1.05853066e-01 4.02748525e-01 1.23366453e-01
-5.13644874e-01 -1.18375402e-02 -6.80408061e-01 6.70401990e-01
1.39225459e+00 -3.24807167e-01 3.66863817e-01 1.51723993e+00
2.77736694e-01 -2.13161469e-01 -1.41997540e+00 7.58915544e-01
1.06031150e-01 -1.30323613e+00 -2.85002798e-01 -2.12504603e-02
7.35075176e-01 1.57659501e-01 -3.21504682e-01 4.91638958e-01
2.75013864e-01 -1.05016470e+00 9.48242784e-01 3.46052527e-01
2.55695403e-01 -7.65272200e-01 1.12759113e+00 6.46417379e-01
-7.88514853e-01 -6.76571354e-02 -6.53560638e-01 -1.06868804e-01
-5.36891520e-01 1.23596168e+00 -9.70614076e-01 1.06006332e-01
7.82062948e-01 5.72802365e-01 -9.71918643e-01 1.17463028e+00
-1.31727248e-01 1.03242326e+00 -7.20621869e-02 -4.87651199e-01
-7.59702176e-02 8.32702070e-02 3.31500173e-01 1.33467603e+00
4.58430141e-01 -8.28842163e-01 -2.29029715e-01 1.70848954e+00
8.67491886e-02 7.12040439e-02 -8.36640537e-01 -3.77364039e-01
4.57487226e-01 9.28959310e-01 -7.07919121e-01 8.96588434e-03
-5.88090062e-01 4.74748790e-01 1.64604887e-01 1.83717728e-01
-6.57169163e-01 -8.74822319e-01 6.05525851e-01 1.70533001e-01
1.42685786e-01 -5.63272871e-02 -6.32325888e-01 -1.11802328e+00
3.89622986e-01 -1.39566064e+00 -1.54057562e-01 -5.40634751e-01
-1.34741330e+00 8.64010751e-01 -3.29484671e-01 -1.26228666e+00
-3.65068644e-01 -8.05932283e-01 -7.53209054e-01 8.96306455e-01
-1.17202890e+00 -6.49089396e-01 -2.99001962e-01 1.20669194e-02
6.07485592e-01 -4.83584404e-01 7.43799210e-01 2.11585328e-01
-3.48980248e-01 5.44746041e-01 2.29410961e-01 5.05192220e-01
4.64476198e-01 -1.35676801e+00 7.10453808e-01 1.25361574e+00
4.27884459e-01 1.04985487e+00 9.11601126e-01 -7.82975197e-01
-1.31583011e+00 -1.06675971e+00 8.87163579e-01 -7.92963862e-01
7.12292135e-01 -4.35940117e-01 -1.35432565e+00 2.89049000e-01
-5.50196096e-02 -6.89465106e-02 7.22780764e-01 3.22508186e-01
-5.76887012e-01 1.02413855e-01 -1.04055500e+00 3.41167748e-01
6.72589719e-01 -6.83903635e-01 -7.44249463e-01 -2.23900184e-01
4.91538554e-01 -2.39711374e-01 -7.07302988e-01 -6.12198329e-03
3.52630734e-01 -1.34029031e+00 4.42089707e-01 -3.35272431e-01
7.97983885e-01 -2.99057305e-01 4.74128164e-02 -1.64301503e+00
6.54381141e-02 -1.87656596e-01 9.59547833e-02 1.40536714e+00
7.48050749e-01 -3.27712476e-01 7.03384399e-01 6.71599090e-01
-1.69091448e-01 -5.76393008e-01 -7.98807681e-01 -8.22072864e-01
2.72077203e-01 -9.96598065e-01 5.54982364e-01 9.43190098e-01
8.44798535e-02 -9.10646841e-02 2.60427147e-01 -2.92040795e-01
1.68173268e-01 -2.47465953e-01 7.21706927e-01 -1.42431414e+00
-2.29764715e-01 -7.00966775e-01 -4.57283586e-01 -5.07230520e-01
3.34327757e-01 -9.41192627e-01 5.38801849e-01 -1.24405277e+00
8.27153474e-02 -4.62061614e-01 -8.88611972e-02 3.41499239e-01
-2.53200054e-01 -1.60386458e-01 6.00986823e-04 8.92710164e-02
-9.90971103e-02 2.93882817e-01 5.84973514e-01 -7.88693801e-02
-1.81329697e-01 1.61829844e-01 -9.24806893e-01 8.12263787e-01
8.10537994e-01 -1.19620252e+00 -5.73501647e-01 -4.16125625e-01
7.32842445e-01 -4.57951814e-01 1.96196839e-01 -1.20541787e+00
3.94598469e-02 -5.53939082e-02 -4.23107855e-02 -1.29543394e-01
-4.88521725e-01 -9.92608190e-01 -6.87815547e-02 5.64462364e-01
-5.88920057e-01 1.44480258e-01 3.22211891e-01 2.84236640e-01
-4.03656840e-01 -1.17710948e+00 6.59559488e-01 -1.87792420e-01
-6.07382596e-01 -3.07945639e-01 -6.39750063e-01 2.99491376e-01
6.17715657e-01 -2.90570289e-01 -3.27904612e-01 -9.77676585e-02
-1.72941595e-01 -3.09834480e-01 6.07094109e-01 8.57464731e-01
7.20173836e-01 -1.17775238e+00 -7.41609454e-01 6.10213161e-01
6.07202291e-01 -1.15041085e-01 -3.23145956e-01 5.50351620e-01
-5.77608645e-01 2.31340323e-02 1.15127563e-01 -6.40718222e-01
-1.37016881e+00 2.68675923e-01 4.46358949e-01 4.43789549e-02
-3.33364666e-01 8.27640176e-01 -3.46306980e-01 -7.53778458e-01
5.49789011e-01 -8.20673287e-01 -1.62408903e-01 -1.37031436e-01
8.55297983e-01 2.22005337e-01 5.21640003e-01 -1.17729262e-01
-1.72246546e-01 1.06955856e-01 -1.13311239e-01 1.34526521e-01
1.40597558e+00 4.27475363e-01 -3.00612628e-01 9.55364585e-01
1.16305256e+00 3.76324728e-02 -8.61206532e-01 -8.91822502e-02
8.21673036e-01 -6.56898320e-01 1.79819897e-01 -8.95205915e-01
-1.01749873e+00 9.55583870e-01 7.10220098e-01 5.90775251e-01
9.52780783e-01 -1.84295878e-01 2.80875444e-01 5.60937881e-01
3.79719555e-01 -1.26272392e+00 1.46438375e-01 6.60806119e-01
1.01514459e+00 -1.34965968e+00 -4.45606291e-01 -7.77043328e-02
-2.65491307e-01 1.48894310e+00 7.22226322e-01 -2.07096949e-01
7.14625657e-01 7.77596176e-01 2.99532056e-01 -1.93156481e-01
-9.77237344e-01 9.27159190e-02 4.01839375e-01 8.11582446e-01
1.07496428e+00 -3.72191548e-01 -7.09712729e-02 8.61089289e-01
-1.37298033e-01 2.60680139e-01 1.08473861e+00 9.88443673e-01
-5.65105498e-01 -1.32167459e+00 -5.75800359e-01 7.65898764e-01
-4.87087965e-01 -2.56983280e-01 -5.41336775e-01 4.97490078e-01
2.26328939e-01 1.16446602e+00 -1.99639827e-01 -6.08917415e-01
5.39858401e-01 2.86844522e-01 1.73620149e-01 -1.08078563e+00
-1.10501444e+00 -9.20094028e-02 6.96996748e-02 -4.37769294e-01
-2.90953577e-01 -6.25383794e-01 -9.42274630e-01 -2.24866822e-01
-6.70687914e-01 2.27154315e-01 9.44501460e-01 8.06972444e-01
4.68577236e-01 7.78342426e-01 2.13168904e-01 -4.25521404e-01
-7.83002794e-01 -1.09281731e+00 -1.33797988e-01 4.42441821e-01
3.58859718e-01 -2.77803987e-01 -4.42304283e-01 2.77727336e-01] | [7.652684688568115, 7.801513671875] |
51f7c933-007f-4025-ace9-d2215a63e580 | compositional-networks-enable-systematic | 2008.02742 | null | https://arxiv.org/abs/2008.02742v3 | https://arxiv.org/pdf/2008.02742v3.pdf | Compositional Networks Enable Systematic Generalization for Grounded Language Understanding | Humans are remarkably flexible when understanding new sentences that include combinations of concepts they have never encountered before. Recent work has shown that while deep networks can mimic some human language abilities when presented with novel sentences, systematic variation uncovers the limitations in the language-understanding abilities of networks. We demonstrate that these limitations can be overcome by addressing the generalization challenges in the gSCAN dataset, which explicitly measures how well an agent is able to interpret novel linguistic commands grounded in vision, e.g., novel pairings of adjectives and nouns. The key principle we employ is compositionality: that the compositional structure of networks should reflect the compositional structure of the problem domain they address, while allowing other parameters to be learned end-to-end. We build a general-purpose mechanism that enables agents to generalize their language understanding to compositional domains. Crucially, our network has the same state-of-the-art performance as prior work while generalizing its knowledge when prior work does not. Our network also provides a level of interpretability that enables users to inspect what each part of networks learns. Robust grounded language understanding without dramatic failures and without corner cases is critical to building safe and fair robots; we demonstrate the significant role that compositionality can play in achieving that goal. | ['Yen-Ling Kuo', 'Boris Katz', 'Andrei Barbu'] | 2020-08-06 | null | https://aclanthology.org/2021.findings-emnlp.21 | https://aclanthology.org/2021.findings-emnlp.21.pdf | findings-emnlp-2021-11 | ['systematic-generalization'] | ['reasoning'] | [ 1.77488551e-01 3.40340346e-01 2.84184217e-01 -6.09831095e-01
-8.99338946e-02 -8.73678803e-01 8.65611613e-01 -1.47076845e-01
-5.52768290e-01 6.23676538e-01 4.55974281e-01 -3.61187845e-01
-2.99985036e-02 -7.53287613e-01 -9.67688918e-01 -2.12771475e-01
-2.77903885e-01 7.14767992e-01 7.36898482e-02 -8.76233459e-01
1.45632364e-02 4.64739591e-01 -1.45422995e+00 4.58822578e-01
9.52846885e-01 4.48309690e-01 3.75180244e-01 6.25116467e-01
4.57643904e-02 1.03546333e+00 -4.61051613e-01 -2.26408377e-01
5.58024228e-01 -4.20433134e-01 -1.08364487e+00 -2.24978924e-01
6.96634173e-01 -6.70198560e-01 -3.48852456e-01 9.49389040e-01
5.12367021e-03 2.89219856e-01 6.44115090e-01 -1.28236008e+00
-1.24278629e+00 9.85681951e-01 2.00103573e-03 3.08163226e-01
4.60935622e-01 8.84225726e-01 1.15368772e+00 -5.71588099e-01
6.26516402e-01 1.76106322e+00 7.97175109e-01 9.54083025e-01
-1.42347634e+00 -7.38598645e-01 5.96398592e-01 8.94649774e-02
-8.61686766e-01 -5.63313246e-01 3.81530821e-01 -5.03452897e-01
1.30865622e+00 -2.16599554e-01 7.88624763e-01 1.33066428e+00
1.57627270e-01 5.93755901e-01 9.77323174e-01 -1.62096128e-01
2.79092908e-01 -2.81834036e-01 2.17356563e-01 8.47507834e-01
4.70824242e-01 2.90882081e-01 -6.13690257e-01 9.78836864e-02
6.31622255e-01 7.12568462e-02 -1.95256397e-01 -4.28968221e-01
-1.44144976e+00 7.19254673e-01 9.33748424e-01 2.99580455e-01
-2.51572192e-01 5.54527700e-01 5.01791120e-01 5.92516184e-01
-1.20648608e-01 1.19790125e+00 -5.86534083e-01 1.97449028e-01
-2.55718142e-01 4.35022473e-01 8.87009025e-01 9.60162044e-01
6.79926097e-01 1.12123333e-01 2.94291884e-01 3.46486390e-01
2.47247025e-01 4.60114002e-01 5.16343415e-01 -1.31943631e+00
2.30455995e-01 6.19505882e-01 -1.33626655e-01 -7.20656633e-01
-6.48043156e-01 -3.57451618e-01 -3.55789810e-01 4.81420219e-01
4.31335062e-01 -2.66654551e-01 -8.70906830e-01 2.44469500e+00
-8.46326202e-02 -4.14955705e-01 5.97346306e-01 8.04450929e-01
6.96594954e-01 4.38769847e-01 3.01735967e-01 3.99039537e-01
1.36548817e+00 -9.55319762e-01 -1.81505352e-01 -9.24904823e-01
5.02671182e-01 -9.93773490e-02 1.57302928e+00 1.63165763e-01
-1.01269341e+00 -4.96183276e-01 -1.23245978e+00 -3.94247949e-01
-4.00110364e-01 -5.45632362e-01 1.00417340e+00 2.63424277e-01
-1.47278368e+00 4.68877673e-01 -7.10297942e-01 -8.84580493e-01
4.14939702e-01 4.40904677e-01 -4.76292431e-01 9.09535363e-02
-9.80342746e-01 1.27219713e+00 6.60762727e-01 -1.42347157e-01
-1.25406957e+00 -8.00965905e-01 -1.02824378e+00 1.27234414e-01
4.03619617e-01 -1.43533790e+00 1.74341071e+00 -1.36704159e+00
-1.28943324e+00 8.16269577e-01 -1.05815873e-01 -8.02795827e-01
2.64776886e-01 -1.45982727e-01 2.49042455e-02 2.94700503e-01
3.73912334e-01 1.33012879e+00 4.77375686e-01 -1.40567505e+00
-5.05323589e-01 -4.78517085e-01 7.78051138e-01 4.13028330e-01
-1.86481342e-01 -3.12754869e-01 1.94099888e-01 -4.13808703e-01
8.38217810e-02 -1.00621557e+00 -7.85630122e-02 3.06655467e-01
-1.87113062e-01 -3.03593129e-01 5.34409940e-01 -4.95402694e-01
1.38563469e-01 -2.07828689e+00 3.45757633e-01 -1.51212841e-01
4.58914399e-01 -1.23884030e-01 -5.57063997e-01 5.62747359e-01
6.71344846e-02 3.52159381e-01 -3.63215744e-01 -2.77009308e-01
3.48433644e-01 3.70075822e-01 -6.03359878e-01 3.92892472e-02
3.29533905e-01 1.30811822e+00 -1.09742463e+00 1.55384943e-01
-1.05961010e-01 1.31039158e-01 -8.07705343e-01 6.37548715e-02
-6.78251743e-01 4.48031366e-01 -1.35209844e-01 2.33672976e-01
2.38759190e-01 -2.02405095e-01 2.10454255e-01 -8.36961642e-02
1.11083187e-01 6.41770601e-01 -5.16303003e-01 1.99772680e+00
-5.54670513e-01 7.10759521e-01 1.87298700e-01 -8.56498599e-01
5.48516929e-01 -5.81573024e-02 -2.48118520e-01 -6.89578831e-01
8.28571394e-02 6.74059093e-02 6.30455971e-01 -5.55069149e-01
2.55274236e-01 -5.41915715e-01 -3.94512452e-02 6.95414901e-01
6.46649376e-02 -4.88512874e-01 3.20130050e-01 4.49464530e-01
1.20157754e+00 1.15197457e-01 3.36472273e-01 -6.09444559e-01
1.15535684e-01 2.60477245e-01 2.18013272e-01 1.08061004e+00
-1.85019240e-01 2.87816465e-01 4.65948045e-01 -6.96556687e-01
-1.16337752e+00 -1.42227232e+00 3.58362615e-01 1.58687878e+00
8.80329832e-02 5.29707177e-03 -6.55780196e-01 -5.83570063e-01
5.14610447e-02 1.06408989e+00 -9.08575296e-01 -5.00220537e-01
-5.33129096e-01 -1.19865902e-01 8.05697322e-01 8.19120824e-01
6.79388165e-01 -1.33264470e+00 -9.33341503e-01 -7.61090741e-02
-8.73338990e-03 -1.18036997e+00 -3.45008165e-01 2.96923578e-01
-8.00199270e-01 -1.02732682e+00 -5.21345139e-02 -1.08293033e+00
7.83453822e-01 4.79756117e-01 1.46575975e+00 2.70129591e-01
1.06600866e-01 8.30930412e-01 -2.41473079e-01 -5.18055439e-01
-6.61201894e-01 7.56421015e-02 1.27258152e-01 -6.29434109e-01
3.74985546e-01 -7.94925451e-01 -3.30100089e-01 2.68726815e-02
-1.05760264e+00 1.82749182e-01 6.38136089e-01 8.20171714e-01
-2.05970556e-01 -3.05692524e-01 7.53593862e-01 -7.29095817e-01
1.06902862e+00 -3.83690834e-01 -2.90604591e-01 1.17818676e-01
-2.12355480e-01 3.94432694e-01 9.62506533e-01 -3.67923319e-01
-9.00798619e-01 -7.10277259e-02 1.36496751e-02 1.19987711e-01
-3.89691919e-01 4.23401743e-01 -6.19944111e-02 -9.52516310e-03
1.02611697e+00 2.23846048e-01 3.53199780e-01 -6.54422864e-02
7.22237647e-01 1.26072451e-01 7.44326830e-01 -9.41373408e-01
1.01589477e+00 6.07778609e-01 -2.25769341e-01 -7.16368794e-01
-7.71426022e-01 9.69575718e-03 -7.06377089e-01 4.45521891e-01
9.31263328e-01 -1.04922616e+00 -1.12190533e+00 1.87256604e-01
-1.37699533e+00 -6.95094347e-01 -5.28170526e-01 1.51429012e-01
-6.02538824e-01 1.81428283e-01 -6.19563699e-01 -3.41460556e-01
-2.33079389e-01 -1.25463855e+00 7.46030748e-01 8.22089389e-02
-7.61981368e-01 -1.09393048e+00 -1.99909836e-01 2.45093733e-01
5.73531687e-01 4.37874682e-02 1.32147384e+00 -9.47018206e-01
-6.23139977e-01 5.03887832e-01 -2.76474893e-01 3.61675411e-01
4.81309630e-02 -3.77355576e-01 -8.18041801e-01 -4.26568657e-01
3.47337574e-02 -7.11789250e-01 1.02730238e+00 1.23009857e-04
7.27209628e-01 -5.18572271e-01 -1.14410751e-01 4.65286791e-01
1.01931393e+00 2.15179503e-01 2.72706836e-01 5.05287290e-01
4.61997360e-01 6.45819008e-01 -1.53897941e-01 -1.68149173e-01
8.96494627e-01 2.23393023e-01 5.84732473e-01 1.03157043e-01
-6.23325445e-02 -3.91075104e-01 6.01006329e-01 2.68118799e-01
3.34758580e-01 -1.27214089e-01 -1.14326048e+00 6.21630907e-01
-1.73937178e+00 -1.03273058e+00 5.43192387e-01 1.69468904e+00
9.94703770e-01 3.38102847e-01 2.94685941e-02 -4.25054282e-01
2.43391082e-01 4.57025953e-02 -9.37734962e-01 -6.19513929e-01
-1.26635283e-01 9.03028846e-02 1.07582256e-01 7.52166510e-01
-7.72842228e-01 1.18326914e+00 6.99541330e+00 4.94946204e-02
-9.06953454e-01 -1.94780737e-01 2.30187967e-01 -3.41040194e-02
-5.77081382e-01 -7.40816817e-02 -4.25706476e-01 -3.93591560e-02
6.84746683e-01 -2.25947291e-01 1.01791298e+00 6.85568690e-01
5.16400076e-02 -1.99456513e-02 -1.95454454e+00 5.77274859e-01
3.88101280e-01 -1.24007976e+00 5.00180960e-01 -1.59865946e-01
5.12498498e-01 4.01043087e-01 2.89109677e-01 5.57123125e-01
9.58970368e-01 -1.45957577e+00 1.06928778e+00 3.23854566e-01
3.79723102e-01 -2.76965410e-01 3.04020852e-01 5.38653612e-01
-8.05792511e-01 -4.72524226e-01 -3.49085927e-01 -7.52020717e-01
-1.02777608e-01 -1.80153817e-01 -1.22751462e+00 -7.98072293e-02
5.37238479e-01 6.78997397e-01 -7.51102865e-01 5.21716356e-01
-5.67135274e-01 2.29530409e-01 -3.79096478e-01 -1.50638208e-01
5.14162302e-01 1.02410078e-01 5.56601405e-01 1.00998664e+00
-2.02370901e-02 2.56028771e-01 2.30931401e-01 1.37225008e+00
-2.81691700e-01 -3.93540949e-01 -1.08945167e+00 -1.59341961e-01
5.66078424e-01 8.48258376e-01 -5.22507071e-01 -3.31056684e-01
-3.69388819e-01 8.07588577e-01 6.23034060e-01 5.78945935e-01
-3.44224840e-01 -9.79645550e-03 9.50215340e-01 -1.14352882e-01
1.86220497e-01 -6.68850362e-01 -5.98580897e-01 -1.09295821e+00
3.75071354e-02 -1.33018327e+00 1.22785181e-01 -1.15379429e+00
-1.47595441e+00 5.47195852e-01 1.60921980e-02 -6.75892115e-01
-4.25401837e-01 -9.67194676e-01 -6.94037437e-01 5.95348239e-01
-1.19488382e+00 -1.49742150e+00 -4.34035659e-01 4.19837058e-01
8.47434342e-01 -3.61543775e-01 8.31970513e-01 -4.47900593e-01
-8.91794413e-02 4.89487618e-01 -3.65355611e-01 2.87976503e-01
5.04567623e-01 -1.25752079e+00 9.18763041e-01 9.40677822e-01
1.85440153e-01 1.35395932e+00 8.82319510e-01 -5.18432081e-01
-1.42144549e+00 -8.15172970e-01 4.85808402e-01 -8.57204854e-01
7.24554956e-01 -6.08649313e-01 -8.21582258e-01 1.35343969e+00
5.20278037e-01 -3.11356723e-01 4.11917269e-01 3.58873934e-01
-9.92125452e-01 -4.25627409e-03 -1.11559606e+00 1.13349974e+00
1.57273126e+00 -7.98042655e-01 -1.34741139e+00 2.26871759e-01
1.40016258e+00 -2.53274679e-01 -3.09059292e-01 3.48976344e-01
5.66583276e-01 -1.00810170e+00 1.03624785e+00 -1.08788061e+00
6.07853770e-01 -4.63955909e-01 -2.50302702e-01 -1.74180996e+00
-4.95259762e-01 -3.29030752e-01 3.76566142e-01 7.44194865e-01
6.28228009e-01 -9.90649819e-01 3.51010472e-01 8.08167994e-01
-3.90649110e-01 -4.21163201e-01 -5.34422994e-01 -9.59697425e-01
5.64400196e-01 -3.43024313e-01 7.30484426e-01 9.02794182e-01
2.07547709e-01 7.62791812e-01 3.72024536e-01 2.52003014e-01
5.08923709e-01 -8.48957226e-02 9.08014894e-01 -1.22198379e+00
-2.94726700e-01 -7.20164776e-01 -2.88371384e-01 -1.24184799e+00
7.14498699e-01 -1.11832094e+00 7.89593607e-02 -1.58852291e+00
3.37941766e-01 -1.37778103e-01 1.31889641e-01 8.27921093e-01
6.94871768e-02 2.64976453e-02 6.17144883e-01 1.11570776e-01
-7.32091844e-01 5.74073255e-01 1.27165294e+00 -4.75565374e-01
-4.85196412e-02 -6.89394772e-01 -1.37276113e+00 1.06787074e+00
8.93223405e-01 -1.24287717e-01 -7.28106678e-01 -1.15192366e+00
4.12389547e-01 -6.78312182e-01 8.03326726e-01 -1.15623367e+00
4.92082596e-01 -2.05507994e-01 3.81670207e-01 -2.53083222e-02
2.36030236e-01 -8.64541411e-01 -1.46641761e-01 6.47253633e-01
-7.51211286e-01 4.27092016e-01 5.77888429e-01 4.83486444e-01
1.10032924e-01 -1.38712525e-01 6.65545881e-01 -6.29354000e-01
-1.09894502e+00 -7.64375478e-02 -5.01640856e-01 4.80903953e-01
7.95184493e-01 -2.22513497e-01 -7.10800111e-01 -5.02068400e-01
-5.70564449e-01 5.23978710e-01 9.42007124e-01 5.37976503e-01
5.59548616e-01 -1.02904344e+00 -6.94396853e-01 1.49984166e-01
2.36438140e-01 1.49210945e-01 -1.55958965e-01 3.57854813e-01
-6.68870807e-01 2.55932331e-01 -4.20322269e-01 -4.33391213e-01
-6.95308089e-01 5.73705792e-01 7.17402875e-01 2.11269900e-01
-6.26965523e-01 1.02299023e+00 8.16929996e-01 -8.51450205e-01
9.06325728e-02 -7.59091616e-01 3.06266174e-02 -2.93211997e-01
4.90653694e-01 -1.37977526e-01 -2.64777154e-01 -3.65619093e-01
-3.12322855e-01 3.93380076e-01 -3.16822708e-01 -2.28120208e-01
1.34163737e+00 -2.06030041e-01 -3.50460798e-01 4.68850642e-01
8.39244902e-01 -2.42899463e-01 -1.40620244e+00 -2.90268660e-01
-1.74243525e-01 -8.10712203e-03 -4.80126143e-01 -1.20163000e+00
-4.79291052e-01 8.35138857e-01 1.25918075e-01 1.35868773e-01
7.75942624e-01 2.18268812e-01 6.93543494e-01 1.25689840e+00
4.83651787e-01 -7.27304220e-01 4.18559074e-01 1.14014995e+00
1.25670135e+00 -1.37553620e+00 -2.03052819e-01 -4.96642850e-02
-6.62104905e-01 1.07797956e+00 8.88282239e-01 -2.67899275e-01
6.88294768e-02 1.33859038e-01 8.98542926e-02 -2.87457675e-01
-9.44404066e-01 -2.53051966e-01 1.09007396e-02 1.06502819e+00
2.19837874e-01 -3.26315537e-02 2.04843313e-01 4.04309422e-01
-8.30994487e-01 -3.76190931e-01 6.14881933e-01 6.82581961e-01
-6.24336302e-01 -5.48648715e-01 -2.08126664e-01 1.56307757e-01
9.47972685e-02 -3.71206552e-01 -7.84754932e-01 8.80244672e-01
-2.33156141e-02 9.44302022e-01 8.68016630e-02 -1.92485675e-01
3.66396397e-01 1.54776201e-01 5.51759303e-01 -9.39516604e-01
-5.89179814e-01 -8.90790641e-01 7.25273639e-02 -6.06849372e-01
-1.55003622e-01 -6.28132820e-01 -1.71980453e+00 -3.82768422e-01
4.49587315e-01 -7.56279230e-02 2.77682453e-01 1.23757565e+00
3.59522283e-01 5.63238800e-01 5.81533425e-02 -9.32127774e-01
-7.94996798e-01 -8.56741071e-01 -2.11225469e-02 7.49775767e-01
7.48641431e-01 -4.88046795e-01 -3.76226962e-01 1.33900732e-01] | [9.625543594360352, 7.035591125488281] |
d0d9c4a5-9cc0-4b4d-a34c-7c468eca6133 | qiuniu-a-chinese-lyrics-generation-system | null | null | https://aclanthology.org/2022.acl-demo.7 | https://aclanthology.org/2022.acl-demo.7.pdf | QiuNiu: A Chinese Lyrics Generation System with Passage-Level Input | Lyrics generation has been a very popular application of natural language generation. Previous works mainly focused on generating lyrics based on a couple of attributes or keywords, rendering very limited control over the content of the lyrics. In this paper, we demonstrate the QiuNiu, a Chinese lyrics generation system which is conditioned on passage-level text rather than a few attributes or keywords. By using the passage-level text as input, the content of generated lyrics is expected to reflect the nuances of users’ needs. The QiuNiu system supports various forms of passage-level input, such as short stories, essays, poetry. The training of it is conducted under the framework of unsupervised machine translation, due to the lack of aligned passage-level text-to-lyrics corpus. We initialize the parameters of QiuNiu with a custom pretrained Chinese GPT-2 model and adopt a two-step process to finetune the model for better alignment between passage-level text and lyrics. Additionally, a postprocess module is used to filter and rerank the generated lyrics to select the ones of highest quality. The demo video of the system is available at https://youtu.be/OCQNzahqWgM. | ['Yongzhu Chang', 'Xiaoxi Mao', 'Rongsheng Zhang', 'Le Zhang'] | null | null | null | null | acl-2022-5 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 1.09349191e-01 -1.11431777e-01 -7.46973976e-02 -1.82860643e-01
-1.02107465e+00 -7.81751394e-01 8.71670365e-01 -2.03721568e-01
-1.09109595e-01 9.67431188e-01 9.05746758e-01 -1.19012140e-01
3.46794039e-01 -7.47957170e-01 -3.61232072e-01 -5.08094430e-01
7.82200634e-01 6.31819844e-01 -1.90224707e-01 -4.76392746e-01
5.51457763e-01 -6.47259280e-02 -1.34658670e+00 4.74867672e-01
9.93583024e-01 4.51751798e-01 4.84649420e-01 8.91842484e-01
-4.98153746e-01 7.94975996e-01 -9.00211096e-01 -2.71952242e-01
8.77641812e-02 -1.17921901e+00 -6.17819846e-01 -5.44780642e-02
1.76595345e-01 -1.33028865e-01 -2.45122075e-01 8.66439581e-01
8.47164392e-01 1.96278587e-01 7.69277215e-01 -9.85113978e-01
-6.71260834e-01 8.67183805e-01 7.63323233e-02 -9.78853554e-03
7.38179564e-01 1.83229804e-01 1.15963876e+00 -1.06833184e+00
7.40164578e-01 9.85281885e-01 2.85514355e-01 7.16529787e-01
-8.93217087e-01 -5.37416756e-01 -4.73163128e-01 1.23400189e-01
-1.28680146e+00 -4.09524769e-01 9.48285282e-01 -3.91137689e-01
5.48764408e-01 4.97492850e-01 6.12235725e-01 1.17038691e+00
1.60849988e-01 8.00353467e-01 1.01673913e+00 -7.52270997e-01
1.58780888e-01 1.12470202e-01 -2.29847476e-01 1.44071087e-01
-4.18223500e-01 -3.69239688e-01 -7.01215506e-01 -8.81253108e-02
6.30911708e-01 -2.22095415e-01 -4.27186221e-01 3.92901093e-01
-1.34456527e+00 9.83359635e-01 -1.28309250e-01 4.65682954e-01
-2.83613831e-01 -1.30450577e-01 3.69908214e-01 4.04411584e-01
4.22134876e-01 9.28471148e-01 -2.06441507e-01 -6.15827799e-01
-1.21071374e+00 4.32165712e-01 8.45273733e-01 1.08493698e+00
6.48938656e-01 -1.16993079e-03 -7.08820343e-01 9.43078756e-01
1.30618408e-01 4.53852713e-01 9.21442926e-01 -6.75004125e-01
7.42067695e-01 4.40427661e-01 1.43665954e-01 -6.35845363e-01
1.61793709e-01 -1.81396425e-01 -5.06953895e-01 -2.92496532e-01
8.70389417e-02 -3.32033575e-01 -6.88102722e-01 1.54008758e+00
5.33695444e-02 -1.41414076e-01 1.61485076e-01 8.59606922e-01
1.06029427e+00 1.22317028e+00 -2.04970434e-01 -3.71649861e-01
1.29282618e+00 -1.06158769e+00 -8.13645005e-01 -1.15717798e-02
5.77408910e-01 -1.34592474e+00 1.64330530e+00 1.16504520e-01
-1.21727061e+00 -6.62302971e-01 -9.49197888e-01 -3.95844638e-01
-1.54934570e-01 5.45768023e-01 -1.18249699e-01 4.70782131e-01
-8.88326883e-01 4.16467279e-01 -1.95884243e-01 -3.27580392e-01
-2.26657793e-01 -1.73353225e-01 1.73634246e-01 8.91629606e-02
-1.54136360e+00 6.73089683e-01 5.22493064e-01 -5.95166348e-02
-5.71238399e-01 -6.24900639e-01 -7.45726764e-01 -1.40416110e-02
1.73943385e-01 -6.30020976e-01 1.36221218e+00 -9.35953379e-01
-1.96076286e+00 6.35716736e-01 -1.55847162e-01 -1.35437846e-01
6.19759500e-01 -3.52198958e-01 -2.55489528e-01 1.63713768e-01
1.66887030e-01 5.57502329e-01 7.69907355e-01 -8.18191946e-01
-6.66914105e-01 2.81558990e-01 4.88381051e-02 7.97090292e-01
-3.93231630e-01 1.92743897e-01 -5.73310971e-01 -9.56857681e-01
-2.12350756e-01 -8.54945600e-01 1.17446840e-01 -7.39989340e-01
-3.47448558e-01 -2.31834486e-01 5.68449795e-01 -7.85370290e-01
1.73520994e+00 -1.94462335e+00 6.30131662e-02 -6.45954907e-02
-3.06172043e-01 -1.96129410e-03 -2.05797881e-01 1.02449071e+00
2.34531641e-01 1.03316054e-01 -1.69134989e-01 -2.98864990e-01
1.34347126e-01 -9.05650184e-02 -5.70921481e-01 -1.12206429e-01
2.16344744e-02 9.12983656e-01 -1.06279778e+00 -5.15775084e-01
-1.70545764e-02 2.02158108e-01 -3.82492870e-01 7.31677592e-01
-4.96090651e-01 6.24725997e-01 -5.16266763e-01 2.87680626e-01
1.19083509e-01 1.03281446e-01 -7.78583735e-02 3.52733172e-02
-4.08675462e-01 8.99216890e-01 -7.20127463e-01 1.76654458e+00
-7.21552074e-01 6.58315361e-01 -6.62201285e-01 -1.17674850e-01
1.27488053e+00 7.33001769e-01 4.32385594e-01 -5.55898249e-01
1.26561880e-01 2.56782740e-01 -1.06280245e-01 -5.60899258e-01
9.61243093e-01 -1.51399717e-01 -4.13054168e-01 9.01402533e-01
-1.09783590e-01 -9.14100170e-01 5.70259631e-01 3.12227368e-01
8.44154894e-01 4.04224038e-01 3.20392430e-01 -1.09719530e-01
5.94501078e-01 1.34012282e-01 3.52105170e-01 5.81868887e-01
3.43067795e-01 1.27067089e+00 3.13988507e-01 -1.83497916e-03
-1.43696654e+00 -8.06758344e-01 1.20123662e-01 9.54658031e-01
-4.29845601e-01 -8.20843101e-01 -1.10172629e+00 -3.29267979e-01
-6.20711625e-01 1.19741905e+00 -3.36257070e-01 4.53984104e-02
-6.96944535e-01 -6.65611446e-01 7.09677756e-01 -7.36860419e-03
2.85743296e-01 -1.68624473e+00 -5.63721657e-01 4.00019765e-01
-6.98879480e-01 -8.54367793e-01 -8.56022418e-01 -3.29493642e-01
-5.20569444e-01 -7.12311506e-01 -9.39969957e-01 -8.43324244e-01
5.40158272e-01 6.85739741e-02 1.20178819e+00 7.71771297e-02
1.47770688e-01 1.39289007e-01 -9.61124003e-01 -4.16849256e-01
-8.05035174e-01 4.83028769e-01 -1.50690153e-01 -4.01720442e-02
3.95523816e-01 -5.64145386e-01 -3.69222671e-01 -5.93531393e-02
-9.73423898e-01 5.91968417e-01 4.00809646e-01 8.49822819e-01
2.96999365e-01 -2.53173321e-01 5.48228681e-01 -9.87373531e-01
1.14346099e+00 -4.70417976e-01 -1.89205095e-01 1.15276106e-01
-4.14404243e-01 -1.12128839e-01 1.02756727e+00 -4.12200511e-01
-1.13246453e+00 -1.02222122e-01 -2.96204597e-01 -1.21059522e-01
-1.39548466e-01 7.73131311e-01 -3.01116318e-01 6.62570834e-01
5.95988631e-01 5.74730515e-01 -3.16405833e-01 -4.41436082e-01
5.35771430e-01 1.04736090e+00 5.92825174e-01 -6.88375890e-01
8.88761103e-01 -1.64310932e-01 -5.99649906e-01 -7.91203797e-01
-8.26804280e-01 -4.41821009e-01 -7.40982890e-01 -4.19830143e-01
7.82538056e-01 -8.69643331e-01 -3.79702374e-02 4.29713577e-01
-1.20863700e+00 -5.57663918e-01 -3.57630134e-01 4.52750623e-01
-8.33327293e-01 1.95815444e-01 -7.50168502e-01 -5.08766174e-01
-7.54936993e-01 -8.65159512e-01 8.52844954e-01 3.87490094e-01
-5.75201154e-01 -7.85001099e-01 4.20187622e-01 2.35739276e-01
1.24790579e-01 1.80418223e-01 1.03032804e+00 -4.98894155e-01
-2.73545176e-01 -1.63714603e-01 1.24121599e-01 1.56999037e-01
4.95806187e-01 1.57454371e-01 -7.83432543e-01 -1.30470365e-01
2.42776200e-01 -4.43518281e-01 2.60107875e-01 3.63368466e-02
7.23532975e-01 -7.43656278e-01 4.44330961e-01 5.82427144e-01
1.25514627e+00 2.48728395e-01 8.59349906e-01 5.05087137e-01
7.83060431e-01 4.92752701e-01 8.10468912e-01 4.57319826e-01
4.42261994e-01 6.75275385e-01 -1.12309828e-01 8.31933096e-02
-2.54971832e-01 -6.76115572e-01 7.53639817e-01 1.63129330e+00
2.40349136e-02 -2.97970176e-01 -7.16317773e-01 6.27628088e-01
-1.69908357e+00 -1.11454391e+00 -1.56031132e-01 1.97492814e+00
1.31057823e+00 5.23214526e-02 -8.37988034e-03 -5.47550097e-02
6.17471039e-01 3.44510645e-01 -1.10462502e-01 -6.08219802e-01
-1.24660350e-01 1.51134640e-01 1.04749896e-01 5.20243049e-01
-6.13931239e-01 1.28298569e+00 5.50244093e+00 9.17159498e-01
-1.20878375e+00 -2.23079417e-03 4.56551701e-01 -2.78495133e-01
-6.72335982e-01 1.80062830e-01 -7.61007488e-01 7.31606483e-01
8.90348852e-01 -5.72240531e-01 5.40126562e-01 6.10379517e-01
7.87277281e-01 6.10192865e-02 -9.18032348e-01 8.21932852e-01
4.08796370e-01 -1.26451802e+00 2.72603929e-01 -3.55574071e-01
1.09096611e+00 -1.93074539e-01 -2.22371123e-03 3.37139606e-01
2.12537348e-01 -8.31547678e-01 1.07888877e+00 4.93062943e-01
9.96988416e-01 -7.00604916e-01 5.22069871e-01 4.90627944e-01
-1.06329155e+00 3.76030922e-01 -4.73116249e-01 -1.39760971e-01
3.60591561e-01 2.43239149e-01 -1.14962256e+00 5.08513212e-01
3.00146073e-01 6.01990759e-01 -5.28034031e-01 9.46838081e-01
-8.53782713e-01 9.62381959e-01 1.43330619e-02 -3.87718618e-01
4.54509556e-01 -4.96189326e-01 5.89938462e-01 1.47116363e+00
7.05130875e-01 3.80464554e-01 1.45610482e-01 7.93153763e-01
-2.50328109e-02 6.03598654e-01 -3.78041804e-01 -1.37519374e-01
6.16329491e-01 1.26859653e+00 -4.23471749e-01 -4.08809334e-01
-2.28065431e-01 1.21600747e+00 -9.31989774e-03 2.93013483e-01
-7.09327936e-01 -4.45432335e-01 2.18684852e-01 1.83590844e-01
6.85963184e-02 -2.86696315e-01 -3.92910987e-01 -1.18866563e+00
-8.47911760e-02 -1.25433767e+00 2.24852458e-01 -1.16350579e+00
-1.02531564e+00 9.23353672e-01 -1.59133002e-01 -1.51373315e+00
-6.95691168e-01 -2.87798364e-02 -1.07310152e+00 1.12336516e+00
-1.29651618e+00 -1.11943686e+00 -1.84218377e-01 3.66739839e-01
1.16547096e+00 -3.93269777e-01 8.52365494e-01 1.65588513e-01
-2.45111078e-01 3.76879185e-01 2.28194475e-01 1.83898538e-01
1.04582429e+00 -1.27013588e+00 6.31680608e-01 9.98916805e-01
3.13010603e-01 5.11632860e-01 1.03763866e+00 -8.48578095e-01
-9.79293168e-01 -1.21940744e+00 1.47756946e+00 -3.92808288e-01
7.54729867e-01 -4.29604590e-01 -7.22750008e-01 3.20777088e-01
6.41721070e-01 -9.65364039e-01 7.20499754e-01 -2.40271077e-01
1.68777406e-01 2.58478820e-01 -6.95437133e-01 1.10480142e+00
6.06488645e-01 -6.04430318e-01 -8.00099790e-01 3.56147915e-01
8.47718537e-01 -4.47962493e-01 -5.80927789e-01 -9.92238224e-02
3.57499033e-01 -6.07311547e-01 3.99890304e-01 -3.70211691e-01
9.14180160e-01 -4.85432416e-01 1.72392260e-02 -1.51152921e+00
-3.46305519e-01 -1.01595187e+00 1.61816239e-01 1.60535216e+00
4.52457935e-01 -4.01300825e-02 5.52459955e-01 3.54019433e-01
-2.84501851e-01 -4.49036092e-01 -4.85519052e-01 -4.06466693e-01
-1.33238062e-01 -1.98471040e-01 7.46160448e-01 8.38457823e-01
1.26372859e-01 7.98772156e-01 -6.62155986e-01 -3.12760562e-01
1.13322832e-01 2.14670792e-01 1.04585767e+00 -6.28619850e-01
-2.66031563e-01 -4.54980940e-01 1.91220418e-01 -9.53463256e-01
-8.62217098e-02 -9.23934162e-01 4.19580966e-01 -1.52293956e+00
1.01172134e-01 -3.60751122e-01 -4.79006395e-03 2.46111661e-01
-4.80468273e-01 4.94830847e-01 5.60224056e-01 5.13772011e-01
-2.27101862e-01 7.33234465e-01 1.25002146e+00 2.98365131e-02
-7.65270174e-01 1.61360413e-01 -4.32409197e-01 4.23453540e-01
1.14083552e+00 -4.88742173e-01 -5.74757695e-01 -4.03723478e-01
3.57515842e-01 8.02275464e-02 -2.22959757e-01 -9.12979662e-01
1.46041662e-01 -4.44911629e-01 1.27430841e-01 -7.80658782e-01
3.46804589e-01 -1.52424768e-01 1.76605508e-01 2.22705081e-01
-5.58880627e-01 4.74740833e-01 -1.66461647e-01 -2.20682591e-01
-4.28447753e-01 -6.96078658e-01 5.01754344e-01 -3.46631825e-01
-4.44661677e-01 1.16347305e-01 -5.33591330e-01 1.86746925e-01
5.27347445e-01 -2.56287515e-01 -1.75523087e-01 -8.37412953e-01
-2.61289716e-01 1.13493420e-01 5.96014440e-01 5.76168299e-01
6.18179619e-01 -1.59244049e+00 -1.22052193e+00 -6.30812421e-02
1.69993997e-01 -1.62559181e-01 -3.54910009e-02 3.30497026e-01
-7.34890997e-01 4.03784126e-01 -3.20230186e-01 -3.09678353e-02
-1.14373636e+00 2.68295288e-01 -1.65923208e-01 -3.62334162e-01
-5.20273864e-01 4.82641816e-01 1.08994864e-01 -3.77086967e-01
-1.22726768e-01 -5.49493060e-02 -6.33825541e-01 2.07640916e-01
6.70377314e-01 1.71043783e-01 -1.20375052e-01 -8.42360735e-01
2.30273843e-01 3.19304615e-01 9.54188630e-02 -6.63877070e-01
1.28362954e+00 -3.41487616e-01 -1.40756741e-01 7.30070949e-01
8.86553228e-01 4.09866273e-01 -1.07134390e+00 -4.45345454e-02
-3.51653732e-02 -3.90641928e-01 -3.38211060e-01 -6.94071710e-01
-4.15394306e-01 7.78979957e-01 -1.04735985e-01 4.47195433e-02
1.12797856e+00 -2.95431256e-01 1.11438537e+00 1.67602852e-01
1.09412193e-01 -1.44031107e+00 1.94321260e-01 9.58511472e-01
1.11679292e+00 -6.51938021e-01 -9.50696468e-02 -7.19362637e-03
-9.41419601e-01 1.29384971e+00 5.68143487e-01 -1.40630901e-02
7.63559192e-02 -8.84139314e-02 4.74104166e-01 3.00840318e-01
-8.01303983e-01 -5.14895916e-02 4.16546196e-01 3.19166243e-01
7.47819722e-01 8.62636343e-02 -1.02458918e+00 5.25125623e-01
-1.02909124e+00 -7.63084441e-02 8.77768219e-01 4.51968193e-01
-5.29263616e-01 -1.42023277e+00 -5.91934741e-01 2.91720688e-01
-4.95985031e-01 -6.51311576e-01 -5.47772110e-01 2.40505531e-01
9.43993479e-02 9.49170172e-01 -8.14094767e-03 -2.38303825e-01
6.59401640e-02 1.29605576e-01 1.44971326e-01 -1.01288629e+00
-8.65011513e-01 3.97304326e-01 8.59680995e-02 -1.97480358e-02
-9.20274556e-02 -7.15636075e-01 -1.05277646e+00 -2.13303670e-01
-4.87404410e-03 4.31830525e-01 5.59431911e-01 9.92171586e-01
6.70867637e-02 3.87827486e-01 9.35896277e-01 -7.12392926e-01
-2.53019571e-01 -1.29015386e+00 -2.50812918e-01 4.31241572e-01
-8.24706256e-02 1.18172370e-01 7.03821564e-03 3.36677402e-01] | [11.791081428527832, 9.017899513244629] |
51635e66-db83-4000-95bb-5c3b53fab04a | low-rank-autoregressive-tensor-completion-for-1 | 2104.14936 | null | https://arxiv.org/abs/2104.14936v1 | https://arxiv.org/pdf/2104.14936v1.pdf | Low-Rank Autoregressive Tensor Completion for Spatiotemporal Traffic Data Imputation | Spatiotemporal traffic time series (e.g., traffic volume/speed) collected from sensing systems are often incomplete with considerable corruption and large amounts of missing values, preventing users from harnessing the full power of the data. Missing data imputation has been a long-standing research topic and critical application for real-world intelligent transportation systems. A widely applied imputation method is low-rank matrix/tensor completion; however, the low-rank assumption only preserves the global structure while ignores the strong local consistency in spatiotemporal data. In this paper, we propose a low-rank autoregressive tensor completion (LATC) framework by introducing \textit{temporal variation} as a new regularization term into the completion of a third-order (sensor $\times$ time of day $\times$ day) tensor. The third-order tensor structure allows us to better capture the global consistency of traffic data, such as the inherent seasonality and day-to-day similarity. To achieve local consistency, we design the temporal variation by imposing an AR($p$) model for each time series with coefficients as learnable parameters. Different from previous spatial and temporal regularization schemes, the minimization of temporal variation can better characterize temporal generative mechanisms beyond local smoothness, allowing us to deal with more challenging scenarios such "blackout" missing. To solve the optimization problem in LATC, we introduce an alternating minimization scheme that estimates the low-rank tensor and autoregressive coefficients iteratively. We conduct extensive numerical experiments on several real-world traffic data sets, and our results demonstrate the effectiveness of LATC in diverse missing scenarios. | ['Lijun Sun', 'Nicolas Saunier', 'MengYing Lei', 'Xinyu Chen'] | 2021-04-30 | null | null | null | null | ['traffic-data-imputation'] | ['time-series'] | [-1.46922573e-01 -7.41856873e-01 -2.63956130e-01 -4.30384070e-01
-7.56591916e-01 -2.59097278e-01 3.29150230e-01 -5.78868151e-01
-5.55297993e-02 6.70401275e-01 3.96481484e-01 -2.13266194e-01
-6.64080739e-01 -6.46827757e-01 -8.73262107e-01 -9.40944791e-01
-2.10895866e-01 1.14536770e-01 -1.63988695e-01 -2.86489576e-01
-1.36674210e-01 3.58165771e-01 -1.31938624e+00 -3.41993757e-02
1.19089711e+00 9.87768233e-01 1.31326556e-01 3.11886910e-02
-5.49259968e-02 8.19674313e-01 -7.29211494e-02 -1.91871151e-01
4.04505551e-01 -9.96965449e-03 -1.81798413e-01 2.95000732e-01
1.86036542e-01 -1.44318461e-01 -8.21421683e-01 8.43941689e-01
-6.20160140e-02 3.09996456e-01 2.62438595e-01 -1.65954816e+00
-4.33178991e-01 2.64252812e-01 -8.58108461e-01 2.07505241e-01
5.62381148e-02 3.78429621e-01 7.36317933e-01 -9.93875384e-01
2.72771239e-01 1.07097614e+00 8.19317222e-01 6.60552979e-02
-1.37511563e+00 -8.68542910e-01 4.02576208e-01 2.26343319e-01
-1.66162491e+00 -4.78613168e-01 1.32491398e+00 -8.16731572e-01
4.29202080e-01 4.63419408e-01 4.41084474e-01 1.18443227e+00
1.17279887e-01 5.71248412e-01 1.07584679e+00 3.81500900e-01
2.19114665e-02 -4.25880611e-01 2.62256026e-01 3.81224006e-01
2.04615891e-01 1.75140828e-01 -4.82256263e-01 -2.99635053e-01
6.04392290e-01 6.75803423e-01 2.12779604e-02 -1.75091967e-01
-1.32456040e+00 5.80116212e-01 3.04232836e-01 6.09996468e-02
-7.20194757e-01 4.07029569e-01 1.48987234e-01 6.43930882e-02
7.46052802e-01 -4.14452434e-01 -2.62391478e-01 -3.29091370e-01
-8.83532763e-01 2.18178958e-01 1.15973227e-01 9.80636239e-01
1.10582209e+00 5.34206629e-01 -1.88491806e-01 1.03489089e+00
2.98688263e-01 1.12726498e+00 -4.55213189e-02 -1.26171219e+00
9.65141714e-01 6.19513869e-01 3.28694373e-01 -1.46739089e+00
-3.11502248e-01 -4.44957942e-01 -1.47743928e+00 -3.80274981e-01
5.16658723e-01 -3.84352297e-01 -5.75421751e-01 2.03542709e+00
3.02512348e-01 7.38924861e-01 -3.94315183e-01 1.07994592e+00
1.58041850e-01 7.20575988e-01 -1.22579010e-02 -4.01632041e-01
1.10481906e+00 -2.49439374e-01 -9.00236785e-01 -1.39687791e-01
4.28889990e-01 -6.41122878e-01 7.95113146e-01 6.27523437e-02
-6.26987994e-01 -3.50955606e-01 -4.66562241e-01 8.36348459e-02
-1.19676441e-01 1.32469967e-01 6.93925440e-01 2.09029466e-01
-5.04095137e-01 1.64888635e-01 -1.00243831e+00 -2.01390591e-02
7.52384439e-02 -7.30168400e-03 -2.73032486e-01 -4.41416651e-01
-1.21318591e+00 2.81434447e-01 -4.57943916e-01 7.52612412e-01
-7.91293859e-01 -9.93123174e-01 -7.69401550e-01 -2.92142123e-01
5.56164205e-01 -4.53604162e-01 3.71484756e-01 -2.44082779e-01
-8.74196827e-01 2.83818901e-01 -7.33853817e-01 -2.14244470e-01
4.06714708e-01 -7.64288828e-02 -9.14783061e-01 -4.07103390e-01
5.19395590e-01 3.25201713e-02 9.40812469e-01 -1.18854237e+00
-2.88500965e-01 -4.48960394e-01 -1.95506498e-01 -4.06602323e-01
-7.07468614e-02 -2.41517186e-01 -4.53484833e-01 -1.04656315e+00
5.32329023e-01 -1.13097274e+00 -3.57845634e-01 -1.63163587e-01
-2.43360817e-01 3.61233093e-02 9.04629827e-01 -9.13487732e-01
1.53864777e+00 -2.34442139e+00 -1.57630332e-02 4.09876943e-01
1.77470416e-01 -2.95753866e-01 -2.91769236e-01 5.54311097e-01
-7.15734661e-02 -8.09528753e-02 -6.20451748e-01 -4.02764618e-01
-5.15446719e-03 6.79794550e-01 -5.44620931e-01 6.07261181e-01
1.63522899e-01 6.99839413e-01 -8.12267721e-01 -2.08072409e-01
2.37622291e-01 6.83183432e-01 -3.39867800e-01 -1.73422828e-01
-3.84502142e-04 8.59196246e-01 -7.11165071e-01 6.46050692e-01
8.68395627e-01 -1.46840751e-01 -2.64403194e-01 -3.29307288e-01
-3.90468180e-01 -7.54371136e-02 -1.39582479e+00 1.56338501e+00
-4.52826858e-01 5.14919221e-01 4.02411819e-01 -1.10179472e+00
8.68870437e-01 1.50986880e-01 1.23912895e+00 -1.04124033e+00
-2.53014952e-01 1.25168353e-01 -2.62510598e-01 -7.54801273e-01
4.69000429e-01 6.33337349e-02 -2.03657135e-01 1.97217241e-01
-6.42693639e-01 4.39025551e-01 2.11640164e-01 2.20041052e-01
1.12389243e+00 -2.74142027e-02 -6.74846888e-01 -1.32809058e-01
4.23938036e-01 5.67187145e-02 1.25194919e+00 4.16381985e-01
-1.28380969e-01 4.53227282e-01 4.57801700e-01 -5.42551279e-01
-9.04806137e-01 -1.00713456e+00 -1.84921563e-01 8.40720773e-01
-5.96337654e-02 -1.98780388e-01 -2.65575707e-01 -2.49095216e-01
2.93662131e-01 4.08709288e-01 -6.03621423e-01 8.96381512e-02
-8.79468977e-01 -1.09389067e+00 3.10752094e-01 2.91510850e-01
4.59996432e-01 -4.03762043e-01 1.25024378e-01 3.68979901e-01
-9.47233856e-01 -1.33946383e+00 -8.70551944e-01 -4.83765453e-01
-8.73813450e-01 -8.36874306e-01 -3.88536483e-01 -1.08526677e-01
7.23499835e-01 8.34880710e-01 8.39898050e-01 -1.77944019e-01
-1.36573493e-01 4.17179465e-01 -1.40376553e-01 2.95218527e-02
4.39989001e-01 -3.65293473e-01 2.66254276e-01 8.91819179e-01
2.63889700e-01 -9.14455593e-01 -6.24395728e-01 7.15923667e-01
-9.10668671e-01 -8.49350542e-02 4.66687530e-01 8.28234017e-01
7.95105040e-01 1.94676831e-01 4.28255975e-01 -4.47040915e-01
3.85356694e-01 -9.70089555e-01 -7.52240539e-01 -3.46247433e-03
-3.92017961e-01 -4.71385866e-02 3.83863091e-01 -5.19516885e-01
-9.48739767e-01 8.49859864e-02 1.88968271e-01 -9.03823614e-01
2.64356971e-01 8.88861656e-01 -2.20028073e-01 9.66561437e-02
1.30907476e-01 4.01072562e-01 1.28393993e-01 -6.71710432e-01
2.57076561e-01 1.76420316e-01 4.76496339e-01 -8.21007669e-01
1.08655560e+00 8.62078309e-01 1.54665679e-01 -1.11003602e+00
-7.11550593e-01 -6.58832371e-01 -4.41145003e-01 -4.05811906e-01
5.19005537e-01 -1.11985803e+00 -9.68379974e-01 4.38053608e-01
-9.08741415e-01 -1.94643199e-01 -3.08834285e-01 7.31620073e-01
-2.73976505e-01 3.39348823e-01 -4.12772477e-01 -9.53555644e-01
2.96000868e-01 -9.34647501e-01 9.37108099e-01 -5.00783563e-01
2.24476039e-01 -8.49377334e-01 1.79827437e-01 6.65770173e-01
5.20144105e-01 4.10520762e-01 6.40391767e-01 2.87273556e-01
-8.84362936e-01 -2.22428039e-01 -2.87882686e-01 2.16386959e-01
1.16093755e-01 -6.61901310e-02 -5.54373741e-01 -4.28053252e-02
1.91127926e-01 5.27333379e-01 8.71372879e-01 6.45711482e-01
1.28737640e+00 -6.21675611e-01 -1.72144398e-01 6.84857547e-01
1.27258933e+00 -1.01852760e-01 6.48194909e-01 -1.73131928e-01
1.10858631e+00 7.71141052e-01 5.12098670e-01 7.57972181e-01
9.02113318e-01 7.34727204e-01 3.45313638e-01 1.91981122e-02
1.11501418e-01 -2.53709465e-01 5.14454961e-01 1.23712552e+00
-2.43094191e-01 9.54131037e-02 -1.03553021e+00 7.35874712e-01
-2.17728782e+00 -1.32835984e+00 -7.62935638e-01 2.40757251e+00
2.93569148e-01 -2.33926147e-01 2.64310449e-01 -6.67999387e-02
6.23120070e-01 4.89980072e-01 -6.20622098e-01 2.14745596e-01
-3.85694414e-01 -2.65145272e-01 8.01900983e-01 5.59893191e-01
-8.63346934e-01 4.91957724e-01 5.87661648e+00 7.58975506e-01
-9.76406872e-01 3.22665572e-01 3.11749756e-01 -1.46613404e-01
-5.74371934e-01 1.35465458e-01 -5.08614421e-01 9.07474339e-01
8.98070514e-01 1.09229304e-01 7.17760801e-01 3.34690809e-01
1.13421047e+00 1.14098690e-01 -4.96347725e-01 1.11794269e+00
-3.28804523e-01 -1.24742031e+00 -1.45385787e-01 4.74733055e-01
7.34944522e-01 3.35587680e-01 1.50328204e-01 3.42697054e-01
4.55813080e-01 -7.39262640e-01 3.73367429e-01 1.05792356e+00
5.89205801e-01 -5.29966295e-01 2.91448593e-01 3.22408348e-01
-1.63625669e+00 -1.29826933e-01 -1.95177376e-01 -1.81656793e-01
6.40731394e-01 1.22539449e+00 1.27017915e-01 6.25025094e-01
8.42227340e-01 1.22489870e+00 -3.67703438e-01 7.35241294e-01
2.73743451e-01 7.84620404e-01 -5.76221108e-01 5.63801110e-01
2.06800759e-01 -1.09869432e+00 7.92454660e-01 8.23731720e-01
4.90140051e-01 3.44240367e-01 4.56996173e-01 8.65130723e-01
1.68087736e-01 -2.85452783e-01 -7.60075808e-01 1.22564003e-01
4.94190454e-01 1.02324259e+00 8.43033120e-02 -6.97672218e-02
-6.67017162e-01 5.11536598e-01 -2.11573504e-02 9.57426667e-01
-1.00290763e+00 1.88366830e-01 1.15628040e+00 3.29213440e-01
1.23878241e-01 -1.03596783e+00 -4.23542112e-01 -1.47550368e+00
7.03338265e-01 -5.98433614e-01 2.16525778e-01 -5.96460998e-01
-1.67819929e+00 1.51630417e-01 -2.33685765e-02 -1.66949713e+00
-1.13244861e-01 -5.56937940e-02 -5.45784175e-01 8.60663593e-01
-1.40903580e+00 -1.33232260e+00 -3.22501272e-01 1.08873272e+00
3.10756803e-01 1.10026971e-01 1.63217753e-01 7.83027828e-01
-1.21594298e+00 1.29704177e-01 4.15216088e-01 2.02119842e-01
3.78638417e-01 -6.86728716e-01 7.15521425e-02 1.14331627e+00
-1.69952184e-01 8.32867086e-01 6.28867149e-01 -8.15712631e-01
-1.95080113e+00 -1.43116105e+00 8.69462848e-01 -4.23241556e-01
1.19786453e+00 -3.42396617e-01 -9.86466527e-01 8.64264250e-01
-3.36117625e-01 3.78331751e-01 5.26524365e-01 3.09804291e-01
-5.54602623e-01 -8.18082273e-01 -8.49836946e-01 5.39626837e-01
1.14060700e+00 -7.26789773e-01 -1.86083242e-02 3.90831053e-01
6.11912072e-01 1.67051256e-02 -1.05729282e+00 5.87930679e-01
5.27773738e-01 -4.43892896e-01 1.04384065e+00 -5.03544331e-01
-1.07623473e-01 -7.05100656e-01 -5.38612843e-01 -8.90643060e-01
-3.59226942e-01 -7.84657598e-01 -2.03551844e-01 1.49037695e+00
3.02216530e-01 -7.13659286e-01 6.68932021e-01 1.14588571e+00
-1.87441438e-01 -3.15845072e-01 -1.29654169e+00 -9.01608765e-01
-1.21331856e-01 -1.14474571e+00 7.28984773e-01 1.15620756e+00
-4.67535049e-01 1.27764475e-02 -1.10025489e+00 4.52103317e-01
1.14140701e+00 -2.43083909e-02 9.82025266e-01 -1.17610335e+00
5.99841848e-02 -1.66863784e-01 -1.83327988e-01 -1.23497355e+00
2.36520901e-01 -5.57979822e-01 2.99521107e-02 -1.06066442e+00
5.83644286e-02 -8.87266517e-01 -2.91399390e-01 3.61581862e-01
1.89755455e-01 -2.57438887e-02 -8.01993236e-02 5.41729271e-01
-3.74434948e-01 9.37041163e-01 1.04314125e+00 -3.87871534e-01
-1.20864943e-01 -2.97479983e-02 -3.30591470e-01 5.47971010e-01
5.03757954e-01 -4.98168319e-01 -4.28261817e-01 -8.61168683e-01
2.87911773e-01 4.24791425e-01 6.97238922e-01 -8.61295342e-01
3.92551959e-01 -8.46111357e-01 -3.48771289e-02 -8.59305680e-01
5.37662566e-01 -1.19893956e+00 7.50672042e-01 -2.80058440e-02
1.31493375e-01 3.54974151e-01 -9.72605348e-02 9.70386565e-01
-3.39357525e-01 6.07984781e-01 1.96988538e-01 2.86234885e-01
-6.66776478e-01 8.64782810e-01 -4.58816111e-01 2.22261790e-02
6.05324745e-01 -1.32287920e-01 -3.21882218e-01 -3.74924481e-01
-8.14135373e-01 7.12613523e-01 2.53269732e-01 5.57492673e-01
6.06614470e-01 -1.74404931e+00 -8.96522999e-01 3.01396966e-01
2.23436415e-01 -2.35575005e-01 9.56633925e-01 1.46473467e+00
1.57666460e-01 2.97050029e-01 2.44030446e-01 -9.12206531e-01
-5.37714660e-01 5.74744523e-01 1.70557305e-01 -9.70261320e-02
-6.03726566e-01 3.85670736e-02 4.66088429e-02 -4.81751859e-01
-1.13342211e-01 -2.43425846e-01 2.10052967e-01 2.17239019e-02
2.69281149e-01 7.63277233e-01 -8.11149478e-02 -1.16779852e+00
-4.30556953e-01 8.94688487e-01 4.69372034e-01 -5.51943183e-02
1.37807631e+00 -6.47186518e-01 -2.73804843e-01 6.94860816e-01
1.28853595e+00 1.03690527e-01 -1.47676063e+00 -5.10654211e-01
-6.34686127e-02 -6.05328023e-01 -3.59211676e-02 -2.50575125e-01
-1.47111845e+00 6.95340455e-01 4.12913322e-01 1.56921029e-01
1.10961938e+00 -3.50043565e-01 9.95579422e-01 1.52250677e-01
5.84110796e-01 -9.81904864e-01 -2.41325170e-01 6.30337179e-01
9.40915883e-01 -1.30311525e+00 -1.80366069e-01 -6.03152096e-01
-5.96539915e-01 4.70493108e-01 2.39469528e-01 -2.30482399e-01
1.08076441e+00 -5.21935150e-02 -1.30149081e-01 -3.23880106e-01
-7.15781450e-01 -3.69595021e-01 3.60557139e-01 4.23404008e-01
1.35666698e-01 4.46766317e-01 -3.04014117e-01 3.71269614e-01
7.03881960e-03 -1.54168963e-01 2.94184357e-01 5.90669930e-01
9.12904665e-02 -9.80566025e-01 -5.64951956e-01 4.50618327e-01
-1.35848075e-01 1.40207857e-01 1.88985214e-01 5.80565095e-01
1.41050324e-01 1.38240612e+00 5.27293496e-02 -4.86956209e-01
4.76452231e-01 -1.51191875e-01 -2.47402862e-01 6.49805665e-02
-1.62598304e-02 2.41623446e-01 -7.73060769e-02 -9.65224504e-01
-4.32401001e-01 -1.09737754e+00 -8.45676959e-01 -7.86677063e-01
1.46207005e-01 2.20849633e-01 6.64577067e-01 1.02494478e+00
6.95596039e-01 3.06531698e-01 1.10493886e+00 -5.39938211e-01
-2.96253897e-02 -7.25950181e-01 -6.87331796e-01 8.55051279e-01
7.10058451e-01 -9.54378963e-01 -4.84971762e-01 1.16800822e-01] | [6.586612224578857, 2.1495306491851807] |
840a105d-a3b3-4b0c-a8cd-bc7b4a937bfb | markerless-camera-to-robot-pose-estimation | 2302.14332 | null | https://arxiv.org/abs/2302.14332v2 | https://arxiv.org/pdf/2302.14332v2.pdf | Markerless Camera-to-Robot Pose Estimation via Self-supervised Sim-to-Real Transfer | Solving the camera-to-robot pose is a fundamental requirement for vision-based robot control, and is a process that takes considerable effort and cares to make accurate. Traditional approaches require modification of the robot via markers, and subsequent deep learning approaches enabled markerless feature extraction. Mainstream deep learning methods only use synthetic data and rely on Domain Randomization to fill the sim-to-real gap, because acquiring the 3D annotation is labor-intensive. In this work, we go beyond the limitation of 3D annotations for real-world data. We propose an end-to-end pose estimation framework that is capable of online camera-to-robot calibration and a self-supervised training method to scale the training to unlabeled real-world data. Our framework combines deep learning and geometric vision for solving the robot pose, and the pipeline is fully differentiable. To train the Camera-to-Robot Pose Estimation Network (CtRNet), we leverage foreground segmentation and differentiable rendering for image-level self-supervision. The pose prediction is visualized through a renderer and the image loss with the input image is back-propagated to train the neural network. Our experimental results on two public real datasets confirm the effectiveness of our approach over existing works. We also integrate our framework into a visual servoing system to demonstrate the promise of real-time precise robot pose estimation for automation tasks. | ['Michael C. Yip', 'Florian Richter', 'Jingpei Lu'] | 2023-02-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lu_Markerless_Camera-to-Robot_Pose_Estimation_via_Self-Supervised_Sim-to-Real_Transfer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lu_Markerless_Camera-to-Robot_Pose_Estimation_via_Self-Supervised_Sim-to-Real_Transfer_CVPR_2023_paper.pdf | cvpr-2023-1 | ['foreground-segmentation'] | ['computer-vision'] | [ 2.13171303e-01 2.28944331e-01 -1.15421988e-01 -4.46792811e-01
-8.83806705e-01 -5.98727286e-01 3.94096941e-01 -4.67059344e-01
-4.55223888e-01 1.81683257e-01 -5.51325917e-01 -4.58387315e-01
3.26397717e-01 -3.64317000e-01 -1.38303041e+00 -3.06514144e-01
2.71985054e-01 8.30920041e-01 3.97993773e-01 -1.33534819e-01
2.54947841e-01 5.81393301e-01 -1.27446508e+00 -3.00796211e-01
8.17515790e-01 1.07530642e+00 5.46662271e-01 7.51736879e-01
2.36773700e-01 5.64579248e-01 -3.46315145e-01 -6.04949035e-02
7.27745295e-01 1.62372842e-01 -6.56107128e-01 4.84574467e-01
6.68506503e-01 -6.63838506e-01 -2.18791187e-01 1.10458124e+00
2.54715860e-01 -1.48176420e-02 5.10890424e-01 -1.52788627e+00
-2.34935015e-01 -1.78930938e-01 -7.75589108e-01 -4.83051509e-01
3.19288999e-01 5.88212311e-01 5.95120609e-01 -8.90214264e-01
5.86812139e-01 1.33335996e+00 9.83006477e-01 3.56608272e-01
-1.06619728e+00 -5.92039704e-01 1.87682554e-01 -9.82906595e-02
-1.22608662e+00 -2.24883527e-01 7.59468615e-01 -7.11580515e-01
1.01760423e+00 -3.53023320e-01 6.41199350e-01 1.04516482e+00
2.08609834e-01 5.14448285e-01 6.50716841e-01 -2.59045273e-01
3.13691758e-02 -2.67878413e-01 -3.30723077e-01 1.15892935e+00
7.85189420e-02 1.99757785e-01 -1.26630127e-01 3.43408555e-01
1.37467611e+00 2.56962419e-01 -1.02503113e-01 -1.10777962e+00
-1.43564928e+00 6.19991302e-01 6.34192765e-01 -5.50711215e-01
-1.71658307e-01 6.17168248e-01 1.92330942e-01 1.03446819e-01
1.95173159e-01 6.04043901e-01 -8.27795565e-01 -1.60854772e-01
-5.43615639e-01 5.67686595e-02 6.61105335e-01 1.33941460e+00
1.04937208e+00 -6.36879951e-02 3.50526571e-01 5.06544590e-01
3.94150943e-01 6.82208300e-01 3.16763580e-01 -1.69515860e+00
5.47057986e-01 7.50570238e-01 3.93976867e-01 -7.97965765e-01
-5.14298737e-01 -9.25186351e-02 -5.54573655e-01 7.99360991e-01
5.05699575e-01 -1.84841678e-01 -1.02432966e+00 1.35034883e+00
5.75252593e-01 -8.69430788e-03 -1.29419714e-01 1.18591154e+00
2.38286257e-01 3.80760252e-01 -2.94883192e-01 3.35003704e-01
1.02551579e+00 -1.43249917e+00 -2.75147557e-01 -7.44921267e-01
6.10840738e-01 -6.27224565e-01 1.43191612e+00 3.82990986e-01
-7.31008470e-01 -6.77309453e-01 -1.06540406e+00 -3.78070921e-01
-9.45892632e-02 5.48502624e-01 4.73736286e-01 1.20362163e-01
-8.10549915e-01 5.80014408e-01 -1.16809237e+00 -3.63966405e-01
3.51048768e-01 5.55681467e-01 -6.38674617e-01 -2.11048052e-02
-6.45459116e-01 1.06902778e+00 4.44936186e-01 1.81955129e-01
-1.10228801e+00 -5.14102519e-01 -1.09769809e+00 -3.90152514e-01
7.63951659e-01 -8.92643034e-01 1.57612669e+00 -9.02150214e-01
-1.88093913e+00 9.60724115e-01 2.17084274e-01 -4.47054505e-01
9.33445156e-01 -6.36646032e-01 4.57707822e-01 7.93320611e-02
2.11735100e-01 1.06198251e+00 1.13070536e+00 -1.37782407e+00
-7.38783538e-01 -2.95273870e-01 1.73502192e-01 2.00231478e-01
1.77246794e-01 -4.84409958e-01 -8.81313443e-01 -1.84952661e-01
4.18742299e-01 -1.31623518e+00 -4.70749229e-01 5.40928721e-01
-4.74693239e-01 2.40304638e-02 9.97612953e-01 -5.58742225e-01
2.05461398e-01 -2.02264237e+00 2.41533611e-02 -8.61277580e-02
9.52154994e-02 8.27223510e-02 -1.19883351e-01 3.56903742e-03
9.95620340e-02 -4.20046151e-01 -7.09300488e-02 -5.78773081e-01
7.66074210e-02 2.29527712e-01 -3.51272076e-01 8.13648880e-01
4.72197384e-01 9.82984424e-01 -8.60799611e-01 -4.90233868e-01
7.04277337e-01 4.45300698e-01 -7.09693491e-01 4.96040523e-01
-6.16807222e-01 7.09032297e-01 -3.11814904e-01 6.08970940e-01
5.35066843e-01 -3.84327412e-01 -3.70523743e-02 -3.91895056e-01
-7.71178231e-02 1.48335159e-01 -9.54822838e-01 2.07809472e+00
-5.65963984e-01 5.43909013e-01 1.60430491e-01 -7.68932343e-01
1.05135405e+00 -2.54017353e-01 4.03587073e-01 -3.64694446e-01
3.95268410e-01 1.28097892e-01 -3.34726751e-01 -4.93269980e-01
5.52470803e-01 3.49267751e-01 -1.06180549e-01 -1.32338544e-02
-6.95081279e-02 -9.05083954e-01 -2.17788726e-01 -1.04861110e-01
9.59055781e-01 9.75276291e-01 7.34940842e-02 2.89250284e-01
1.72270253e-01 5.29567838e-01 5.33324957e-01 3.20578873e-01
-2.10736096e-01 7.67221570e-01 4.20308650e-01 -4.00537878e-01
-1.43572247e+00 -9.01540756e-01 2.69082963e-01 9.48619187e-01
6.46394253e-01 3.15290540e-02 -7.98535287e-01 -7.73376048e-01
1.67777687e-01 5.00472009e-01 -4.39807653e-01 -1.85609415e-01
-6.13810778e-01 -4.96208034e-02 2.79560238e-01 6.51516795e-01
5.01015902e-01 -7.84116328e-01 -1.04719448e+00 1.42805511e-02
3.15603539e-02 -1.49996436e+00 -5.37062287e-01 3.86624277e-01
-8.78339767e-01 -1.39424491e+00 -4.89649236e-01 -1.04516363e+00
9.30360615e-01 3.47794414e-01 7.04056144e-01 -2.24231392e-01
-3.32860917e-01 5.67043245e-01 -2.02907901e-02 -3.40874344e-01
-4.51417148e-01 1.49532348e-01 9.06430930e-02 -5.07588625e-01
-1.98364660e-01 -3.59357297e-01 -6.78347826e-01 4.94873673e-01
-3.41086268e-01 3.22138906e-01 7.96613932e-01 8.72550249e-01
8.41105700e-01 -4.16834652e-01 1.71304797e-03 -4.99468178e-01
1.17195502e-01 1.02545448e-01 -1.27580285e+00 -1.47394538e-01
-6.52399778e-01 9.12691653e-02 4.72521603e-01 -6.25075340e-01
-8.08731258e-01 1.13597929e+00 6.57524243e-02 -1.13568878e+00
-1.13538772e-01 1.26647055e-01 -2.26640657e-01 -2.99859643e-01
5.52440405e-01 -1.76356196e-01 3.91037196e-01 -2.47471660e-01
6.43486023e-01 3.84803742e-01 1.10463476e+00 -4.89319742e-01
9.89800453e-01 4.74856526e-01 2.22763103e-02 -3.27585757e-01
-8.56404006e-01 -4.05011028e-01 -1.03436172e+00 -1.80333346e-01
9.44696665e-01 -1.09886694e+00 -9.88927484e-01 4.83828694e-01
-1.42803574e+00 -8.73865843e-01 -2.73750097e-01 4.22997922e-01
-9.90880132e-01 3.29534054e-01 -5.07848024e-01 -6.19881809e-01
-1.92601845e-01 -1.51928258e+00 1.66495466e+00 -5.77894598e-02
-7.86186680e-02 -5.93266428e-01 -1.45388290e-01 3.91323507e-01
-1.22637905e-01 3.47582370e-01 3.59066457e-01 -1.35904014e-01
-9.39548671e-01 -4.41727072e-01 -3.49319547e-01 3.03281695e-01
7.46012153e-03 5.51663712e-02 -9.03022349e-01 -3.08843732e-01
-1.56101182e-01 -8.09824586e-01 5.93057394e-01 2.00271189e-01
1.16858923e+00 -7.90855810e-02 -4.31216270e-01 1.02540493e+00
1.16848838e+00 -1.97060630e-01 3.63269627e-01 7.41101146e-01
1.23733699e+00 4.94869828e-01 1.13096821e+00 2.77763575e-01
5.81404865e-01 6.65155530e-01 1.11408496e+00 -2.45945007e-01
5.46721332e-02 -5.09417713e-01 4.33361590e-01 5.14812946e-01
2.47789979e-01 2.67788798e-01 -9.14131403e-01 2.95089930e-01
-1.82688200e+00 -4.50200915e-01 -1.02260016e-01 2.10824633e+00
8.79958272e-01 2.58890390e-01 -7.05200508e-02 -2.49283388e-01
6.21803105e-01 -1.66333839e-01 -9.79602218e-01 4.46943492e-02
3.86322439e-01 -3.72961849e-01 9.55359042e-01 6.33050740e-01
-1.33540964e+00 1.34524465e+00 5.31159401e+00 3.42777997e-01
-1.40551734e+00 -1.71413019e-01 3.87099594e-01 9.78863537e-02
2.32848078e-01 5.68909608e-02 -6.54948950e-01 6.72191307e-02
4.83417481e-01 3.28323573e-01 4.86896366e-01 1.38725555e+00
1.39815837e-01 -1.31612688e-01 -1.45969272e+00 1.23252010e+00
-6.53251335e-02 -1.01255596e+00 -3.82927179e-01 1.46556785e-02
6.74996793e-01 3.53990495e-01 1.04013033e-01 1.24350362e-01
6.62107170e-01 -8.36975455e-01 9.75450158e-01 3.34465683e-01
1.07127213e+00 -5.02696693e-01 4.37138915e-01 6.71305239e-01
-1.02451921e+00 -1.05871037e-01 -3.61986041e-01 3.24842408e-02
1.84418529e-01 2.79416919e-01 -1.41610825e+00 2.17554837e-01
6.96473479e-01 8.24947476e-01 -4.96134967e-01 8.29717934e-01
-4.36670363e-01 -6.98042512e-02 -4.75953162e-01 2.18217418e-01
1.62157163e-01 -1.82250038e-01 3.41287881e-01 8.24808896e-01
3.00858259e-01 -4.50173676e-01 4.93867189e-01 9.56237793e-01
-7.38525689e-02 -4.67964739e-01 -6.28381193e-01 2.01771930e-01
4.94510859e-01 1.40982616e+00 -6.62567258e-01 -6.75937682e-02
-1.08287767e-01 1.40498185e+00 4.94342893e-01 1.49840653e-01
-9.61225271e-01 -3.77053857e-01 5.94367564e-01 3.49194300e-03
4.37297404e-01 -7.30667889e-01 -3.79022479e-01 -9.77591395e-01
1.30954191e-01 -6.78505242e-01 -2.54171789e-01 -1.05404925e+00
-9.98172045e-01 2.05398113e-01 -2.22488537e-01 -1.51551616e+00
-5.58401823e-01 -9.78626549e-01 -2.03191772e-01 6.43769145e-01
-1.56843328e+00 -1.37452877e+00 -6.76342607e-01 4.17717874e-01
6.95615768e-01 7.60032237e-02 5.94779670e-01 2.57866252e-02
-3.85255367e-01 4.86221641e-01 -1.94082752e-01 2.95112222e-01
9.32551026e-01 -1.15215707e+00 7.15892375e-01 5.92292726e-01
-1.32023320e-01 3.88376653e-01 5.58296621e-01 -6.55866027e-01
-1.84534514e+00 -1.43719709e+00 2.29225129e-01 -8.42124283e-01
6.99109793e-01 -4.03162777e-01 -5.42129934e-01 9.15666699e-01
-2.04893202e-01 3.17983299e-01 -2.56524414e-01 -4.54542726e-01
-2.45837554e-01 -8.22172239e-02 -9.83996987e-01 6.30741656e-01
1.11254442e+00 -4.28096384e-01 -5.05040050e-01 4.38121587e-01
1.14783478e+00 -1.13923991e+00 -6.11513674e-01 3.67523819e-01
5.73266506e-01 -6.73313797e-01 1.10784566e+00 -2.00566903e-01
4.85493690e-01 -7.22288966e-01 6.98635355e-03 -1.15887809e+00
5.63169606e-02 -7.37511158e-01 1.24606319e-01 8.63433719e-01
3.39130551e-01 -4.09628510e-01 1.01965868e+00 6.72137976e-01
-4.11895752e-01 -6.78547978e-01 -7.23785937e-01 -6.97216570e-01
-1.63871899e-01 -5.75112641e-01 3.44810873e-01 8.09677780e-01
-4.67235029e-01 4.76161659e-01 -2.56671369e-01 5.51053822e-01
5.72354376e-01 8.59694704e-02 1.52174389e+00 -1.18857813e+00
-2.22809151e-01 -1.80290252e-01 -3.99442911e-01 -1.47531378e+00
4.08414662e-01 -4.87936825e-01 6.02171957e-01 -1.52031255e+00
-2.17286631e-01 -4.66496497e-01 4.53686565e-01 6.18884206e-01
5.84651567e-02 1.66377500e-01 2.22230658e-01 4.08937722e-01
-7.93801844e-01 5.96519530e-01 1.40051854e+00 -2.98272669e-01
-3.72659564e-01 1.49357021e-02 -1.11331493e-01 1.08023071e+00
6.96343541e-01 -3.05643737e-01 -3.72828990e-01 -7.52980053e-01
9.59092900e-02 -3.36333690e-03 7.63301849e-01 -1.00391221e+00
3.68623078e-01 -1.94071308e-01 5.77948749e-01 -7.06173360e-01
6.52147710e-01 -1.28075254e+00 -2.13461936e-01 5.27388871e-01
-2.18847662e-01 1.30887568e-01 7.22817779e-02 6.40042424e-01
1.76774025e-01 -5.77793121e-02 7.02857077e-01 -1.07054867e-01
-9.09268260e-01 5.27253866e-01 6.89811409e-02 1.09168841e-02
1.13301611e+00 -3.13037783e-01 -7.29142502e-02 -3.12030584e-01
-2.23938942e-01 4.20992285e-01 1.03748941e+00 4.34474230e-01
6.99454486e-01 -1.06967270e+00 -1.81640491e-01 3.02466393e-01
2.31567249e-01 9.95507240e-01 -3.21302235e-01 8.00318837e-01
-1.02000725e+00 1.09388888e-01 8.84129200e-03 -1.21953189e+00
-9.38846350e-01 6.53577566e-01 5.47485769e-01 1.19822145e-01
-7.92558193e-01 7.57370234e-01 2.16866955e-01 -1.02161574e+00
6.09443247e-01 -7.14644611e-01 3.20434153e-01 -4.98220354e-01
-5.49281389e-02 2.09191054e-01 -5.89807816e-02 -4.98007387e-01
-2.32852101e-01 8.09367955e-01 1.67462602e-01 -1.40607461e-01
1.22582603e+00 -2.86714375e-01 5.51474504e-02 4.01643008e-01
1.27845168e+00 -4.13708299e-01 -2.31463122e+00 2.83347145e-02
-5.65595180e-02 -3.17280889e-01 -6.85743615e-02 -6.04692101e-01
-1.05911517e+00 1.00281858e+00 6.61342084e-01 -4.72839028e-01
5.99099576e-01 -3.38346213e-02 7.94206917e-01 9.92525637e-01
5.34506500e-01 -1.35507059e+00 4.04898316e-01 8.21569264e-01
9.05650139e-01 -1.68483460e+00 -2.65598968e-02 -4.82952476e-01
-6.76920295e-01 1.27015972e+00 1.08009422e+00 -3.81229013e-01
1.77455381e-01 5.10780156e-01 3.71868163e-01 5.14620394e-02
-3.14716101e-01 3.05453502e-02 1.22037150e-01 7.65341461e-01
-3.05294860e-02 -2.36767232e-01 6.97973788e-01 3.56522053e-01
-3.14736307e-01 4.41253856e-02 2.28933379e-01 1.01083243e+00
-4.30280298e-01 -8.39475751e-01 -4.92163002e-01 3.30421552e-02
8.78658742e-02 1.97988808e-01 -4.06910300e-01 8.79902959e-01
5.68975434e-02 5.90193510e-01 1.91719905e-01 -4.78015751e-01
5.64671695e-01 -2.14937776e-01 6.12113237e-01 -5.45501351e-01
-1.81341365e-01 1.65557310e-01 -1.88052818e-01 -1.00393319e+00
-1.39691681e-01 -4.25912976e-01 -1.74628651e+00 3.66212688e-02
-3.91528785e-01 -5.46577990e-01 8.68625581e-01 8.42601776e-01
4.12593931e-01 6.78700030e-01 6.29466951e-01 -1.76565623e+00
-8.23595703e-01 -7.07380295e-01 -1.27075702e-01 2.68756896e-01
6.59412026e-01 -8.48774731e-01 -1.96378365e-01 4.13268834e-01] | [7.531944751739502, -2.586904287338257] |
0beacacc-5e73-46b7-8e68-19db52fdfab3 | temporally-layered-architecture-for-adaptive | 2301.00723 | null | https://arxiv.org/abs/2301.00723v2 | https://arxiv.org/pdf/2301.00723v2.pdf | Temporally Layered Architecture for Adaptive, Distributed and Continuous Control | We present temporally layered architecture (TLA), a biologically inspired system for temporally adaptive distributed control. TLA layers a fast and a slow controller together to achieve temporal abstraction that allows each layer to focus on a different time-scale. Our design is biologically inspired and draws on the architecture of the human brain which executes actions at different timescales depending on the environment's demands. Such distributed control design is widespread across biological systems because it increases survivability and accuracy in certain and uncertain environments. We demonstrate that TLA can provide many advantages over existing approaches, including persistent exploration, adaptive control, explainable temporal behavior, compute efficiency and distributed control. We present two different algorithms for training TLA: (a) Closed-loop control, where the fast controller is trained over a pre-trained slow controller, allowing better exploration for the fast controller and closed-loop control where the fast controller decides whether to "act-or-not" at each timestep; and (b) Partially open loop control, where the slow controller is trained over a pre-trained fast controller, allowing for open loop-control where the slow controller picks a temporally extended action or defers the next n-actions to the fast controller. We evaluated our method on a suite of continuous control tasks and demonstrate the advantages of TLA over several strong baselines. | ['Terrence Sejnowski', 'Hava Siegelmann', 'Tauhidur Rahman', 'Francesca Walsh', 'Joshua Russell', 'Devdhar Patel'] | 2022-12-25 | null | null | null | null | ['continuous-control'] | ['playing-games'] | [-4.88591231e-02 -8.27382654e-02 4.43381965e-02 3.56344655e-02
1.38695557e-02 -6.58857107e-01 8.16980243e-01 2.44539119e-02
-4.02679443e-01 9.07715321e-01 2.79834390e-01 -1.75269663e-01
-4.07187343e-01 -6.54141366e-01 -6.77105367e-01 -1.14861298e+00
-7.31425166e-01 6.73585594e-01 6.58806443e-01 -3.33039790e-01
2.17104658e-01 6.72221065e-01 -1.40093529e+00 2.51711190e-01
3.06005627e-01 8.33670497e-01 2.29095995e-01 1.03031981e+00
3.47020924e-01 9.72358763e-01 -3.51318449e-01 6.56871319e-01
3.48557442e-01 -6.20755255e-01 -6.89904869e-01 -2.88928390e-01
-3.35708052e-01 8.68486054e-03 -1.31265074e-01 3.27425689e-01
5.29768288e-01 5.12396336e-01 5.64907968e-01 -1.18250227e+00
-2.98634261e-01 6.06862843e-01 -1.44822866e-01 3.52418154e-01
5.65725230e-02 8.71545494e-01 5.40510237e-01 -3.26088786e-01
6.95232809e-01 1.50197530e+00 5.26524663e-01 9.91957605e-01
-1.67338037e+00 -6.47758722e-01 6.26343250e-01 -2.37202391e-01
-1.10033882e+00 -7.11082995e-01 1.67350605e-01 -3.59553695e-01
1.40032089e+00 8.78787637e-02 1.02918434e+00 1.07619607e+00
8.27559769e-01 5.42685866e-01 1.04657722e+00 -4.42456156e-02
1.08321011e+00 -7.40500152e-01 -4.03268546e-01 6.08225644e-01
-2.24427268e-01 8.48743081e-01 -6.93616033e-01 -2.64951646e-01
9.96112525e-01 -1.37786657e-01 -1.37459919e-01 -4.54685569e-01
-1.57123816e+00 4.33017790e-01 5.93158424e-01 2.41273358e-01
-7.21656084e-01 8.29674959e-01 3.62914771e-01 5.82775474e-01
-9.22928974e-02 9.95903194e-01 -7.82453299e-01 -1.63933277e-01
-6.27471805e-01 6.63599432e-01 7.83637822e-01 6.25544667e-01
5.06149709e-01 1.88490644e-01 -3.60786140e-01 3.97995561e-01
2.76252404e-02 8.76389295e-02 6.09840095e-01 -1.66028523e+00
-2.20878735e-01 2.82768428e-01 4.33433086e-01 -5.72435021e-01
-5.94068706e-01 -4.47637290e-02 -8.88353050e-01 8.74154449e-01
2.45469615e-01 -4.87842470e-01 -1.11150372e+00 2.13015294e+00
4.47474331e-01 2.95401454e-01 6.60265014e-02 7.46029913e-01
-4.13648844e-01 9.78090644e-01 1.83882967e-01 -3.82060319e-01
8.75206172e-01 -1.02434325e+00 -4.21165586e-01 -3.20760041e-01
1.94335788e-01 5.80251310e-03 8.01471412e-01 4.86512810e-01
-1.19449234e+00 -3.17290097e-01 -9.38162506e-01 3.14918190e-01
-4.83169436e-01 -3.91898334e-01 7.11040735e-01 -1.80299953e-01
-1.57866549e+00 8.84400249e-01 -1.44352806e+00 -6.29345000e-01
1.79046050e-01 6.05152488e-01 -2.49875158e-01 4.97313648e-01
-9.47981536e-01 8.15457463e-01 6.19825840e-01 -9.74205285e-02
-1.53654063e+00 -7.06918120e-01 -4.85494077e-01 1.22902259e-01
3.13416421e-01 -1.09609628e+00 1.54423320e+00 -6.88830256e-01
-2.06639051e+00 3.82181674e-01 4.66480963e-02 -8.53175044e-01
4.18187410e-01 1.92685556e-02 9.50075537e-02 -7.34610930e-02
5.26044965e-02 1.26144469e+00 6.83297276e-01 -9.48594391e-01
-5.51374972e-01 -1.45424992e-01 1.16202869e-01 2.40321353e-01
-2.86689103e-02 -2.36889496e-01 -2.19213888e-01 -6.37366652e-01
-1.31669894e-01 -1.23217452e+00 -5.47733307e-01 4.98214602e-01
9.07654762e-02 -1.70186892e-01 9.57525194e-01 6.81041628e-02
1.22076845e+00 -1.92342639e+00 7.44914293e-01 -1.62403032e-01
1.34306177e-01 -6.51447102e-02 -2.90417522e-01 7.90413439e-01
-8.30849260e-02 7.70112947e-02 -3.98050815e-01 1.95440874e-02
-1.46435365e-01 4.07550424e-01 -2.95626879e-01 2.44166508e-01
1.06219605e-01 6.45589054e-01 -1.12234926e+00 -3.43064785e-01
-2.05439385e-02 3.05323172e-02 -7.03280509e-01 3.13162059e-01
-8.79432142e-01 8.51802170e-01 -6.72075033e-01 5.53970456e-01
-1.48588657e-01 -2.92744756e-01 1.78766221e-01 4.33281898e-01
-7.73220181e-01 1.82192907e-01 -1.02471793e+00 1.53421867e+00
-3.49963516e-01 4.67511207e-01 5.69427133e-01 -5.31136155e-01
6.94943130e-01 5.78328073e-01 6.61317825e-01 -6.12681985e-01
1.37706891e-01 2.22350992e-02 1.82365954e-01 -2.00528651e-01
-1.11453943e-01 -5.55798374e-02 -6.50535375e-02 6.82709694e-01
-2.04257667e-01 -5.90384007e-01 2.89297134e-01 2.02644914e-01
1.85284674e+00 5.81426024e-01 3.80728602e-01 -5.72287858e-01
2.90580451e-01 7.01183826e-02 9.13744509e-01 9.19071913e-01
-6.58284128e-01 1.29392043e-01 5.21299005e-01 -7.69510210e-01
-1.03547788e+00 -9.54585314e-01 3.43863547e-01 1.57725060e+00
1.60306320e-02 -2.76910752e-01 -5.17401159e-01 -1.39851019e-01
2.05468491e-01 6.27348244e-01 -9.75465477e-01 -4.01241779e-01
-8.00971985e-01 -1.75650641e-01 2.85277158e-01 6.29190147e-01
5.23080528e-01 -1.69051635e+00 -1.39216435e+00 4.67599422e-01
3.30272615e-01 -3.07972789e-01 -7.80079722e-01 7.72748172e-01
-8.81851435e-01 -6.26513183e-01 -3.91516209e-01 -6.06167972e-01
3.71015459e-01 -9.27158371e-02 8.87199223e-01 1.82943270e-01
-9.21301246e-02 2.91814685e-01 1.33580729e-01 -3.43106031e-01
-2.40190580e-01 -1.66903466e-01 4.49620038e-01 -2.70895958e-01
-4.30047095e-01 -9.81668293e-01 -6.94450319e-01 4.80193287e-01
-7.44021058e-01 3.25261205e-01 4.06207561e-01 8.23888123e-01
6.38661146e-01 2.53349682e-03 6.19980156e-01 -3.82069051e-01
6.26837492e-01 -4.88223374e-01 -8.50428581e-01 1.99675392e-02
-4.78041142e-01 5.33792675e-01 6.46200061e-01 -9.83808160e-01
-8.23285341e-01 4.37829584e-01 4.03644830e-01 -6.73600018e-01
6.43735379e-02 2.41259083e-01 7.52960742e-02 -5.07753938e-02
7.79339492e-01 3.10614973e-01 1.60473496e-01 -5.46331219e-02
4.07229066e-01 2.16266103e-02 6.42413914e-01 -9.77964699e-01
3.76242071e-01 4.66238737e-01 -2.28030588e-02 -4.27025974e-01
-1.68011501e-01 7.80344233e-02 -5.81876636e-01 -4.30162936e-01
8.85828555e-01 -5.00464797e-01 -1.02315414e+00 5.70635438e-01
-9.46068883e-01 -1.32874930e+00 -3.70606869e-01 1.29027873e-01
-1.05126727e+00 -5.98590612e-01 -6.86627507e-01 -8.22310746e-01
-2.45667920e-01 -1.00391138e+00 8.85806382e-01 4.45387214e-01
-6.33342564e-01 -8.49233866e-01 5.14615715e-01 -5.32964349e-01
8.59892368e-01 4.88366574e-01 9.68608618e-01 -4.39492643e-01
-5.12006640e-01 2.89673507e-01 4.10894960e-01 -2.79587924e-01
8.44219923e-02 1.90786049e-01 -5.18681169e-01 -5.81926525e-01
-1.17425710e-01 -5.56795597e-01 7.06017971e-01 5.38871050e-01
9.63999450e-01 -4.39204276e-01 -7.82217503e-01 4.16961342e-01
1.02442908e+00 6.34420991e-01 3.54133785e-01 4.64329034e-01
3.95179354e-02 3.54081035e-01 4.56106395e-01 5.01491845e-01
1.95601597e-01 3.81731838e-01 6.22376025e-01 2.68953443e-01
1.63152125e-02 -5.65971108e-03 5.74265599e-01 4.03883189e-01
4.40414138e-02 -3.39723110e-01 -1.00972414e+00 7.78194845e-01
-2.21555448e+00 -1.04077578e+00 7.36893654e-01 2.13849092e+00
1.12045634e+00 3.30799192e-01 3.14458609e-01 -2.70964772e-01
5.11824548e-01 1.67202786e-01 -1.19229615e+00 -5.85252166e-01
1.80498123e-01 7.81380683e-02 2.56646425e-01 4.79773730e-01
-9.72353220e-01 1.21002495e+00 7.25670958e+00 3.52389753e-01
-1.39448941e+00 -2.36276567e-01 5.88982165e-01 -6.07463479e-01
6.35270849e-02 2.08054349e-01 -4.13959771e-01 4.90497857e-01
1.18945146e+00 -3.53533715e-01 8.65868270e-01 5.37285566e-01
5.88133693e-01 -1.76743582e-01 -1.50863063e+00 2.75155813e-01
-6.46693707e-01 -1.53592682e+00 -1.06352836e-01 1.25653103e-01
8.01125050e-01 1.21703707e-01 -1.23479620e-01 5.23441255e-01
1.11746538e+00 -9.42091107e-01 9.16971684e-01 8.30167174e-01
2.86357850e-01 -5.33990443e-01 -1.77396595e-01 7.95784593e-01
-1.35593593e+00 -5.24740934e-01 -1.23835668e-01 -4.34106082e-01
9.63478312e-02 -2.07248591e-02 -4.12231177e-01 -2.33780265e-01
8.53759348e-01 4.65258181e-01 -8.31438377e-02 8.96022558e-01
-4.35502268e-02 3.73965681e-01 -5.72597504e-01 -3.18062574e-01
4.40288126e-01 -8.91019180e-02 7.86908805e-01 9.73003268e-01
2.70658024e-02 5.91643155e-01 4.83300567e-01 9.25525367e-01
2.83521742e-01 -6.10331416e-01 -7.28776991e-01 -2.54056752e-01
7.31970906e-01 8.99511993e-01 -8.86811793e-01 -3.66268069e-01
1.46197185e-01 7.06560552e-01 4.73135531e-01 4.41073716e-01
-7.38692105e-01 -2.58527339e-01 9.79901135e-01 8.86074081e-02
5.37811816e-01 -5.98014593e-01 -6.83999211e-02 -7.96513200e-01
-7.19900072e-01 -8.44430625e-01 5.72824180e-01 -1.02786660e+00
-8.79525423e-01 4.68762577e-01 2.27048658e-02 -8.10058594e-01
-6.48374617e-01 -2.88261235e-01 -8.15886855e-01 5.54143608e-01
-7.23437369e-01 -7.71120071e-01 7.16662928e-02 6.68858230e-01
7.33686745e-01 -7.70593137e-02 8.47240806e-01 -3.94817829e-01
-5.36839068e-01 1.41595732e-02 -3.31300385e-02 -2.69602925e-01
4.71685469e-01 -1.17928195e+00 3.04228902e-01 7.23681331e-01
-5.49597979e-01 6.45435035e-01 9.17933702e-01 -6.99488521e-01
-1.55053556e+00 -1.02359700e+00 4.90734965e-01 -2.38454297e-01
6.81121707e-01 -3.79171461e-01 -8.93441200e-01 7.12058842e-01
2.40004078e-01 7.86223784e-02 6.14151582e-02 -1.84036002e-01
-6.70971349e-02 -1.53766260e-01 -1.19720459e+00 1.04718900e+00
8.99811029e-01 -2.34624803e-01 -5.25009155e-01 2.32891038e-01
9.37515318e-01 -3.78063232e-01 -6.25340462e-01 2.12827384e-01
7.86688268e-01 -1.03490388e+00 7.02900290e-01 -7.24108875e-01
1.62209749e-01 -6.50577962e-01 5.79390712e-02 -1.41646850e+00
-7.87288725e-01 -1.09649611e+00 -3.21807981e-01 5.86897850e-01
2.44688526e-01 -7.76075125e-01 4.45476264e-01 6.81974649e-01
-1.76697403e-01 -1.01017129e+00 -9.83348370e-01 -8.58805239e-01
3.66614521e-01 6.42020777e-02 4.86820370e-01 6.04448199e-01
4.06520739e-02 3.40090357e-02 -5.93087077e-02 3.27071309e-01
9.98616815e-02 1.46843657e-01 4.22726035e-01 -8.12438846e-01
-3.81443173e-01 -7.97826529e-01 -6.37927353e-02 -9.01071906e-01
-3.65469768e-03 -3.97408992e-01 5.82280874e-01 -1.15374863e+00
-1.70231655e-01 -2.89752901e-01 -3.39089513e-01 1.06538248e+00
2.13719651e-01 -3.44474971e-01 6.81721121e-02 3.55920255e-01
-6.04101002e-01 7.29364693e-01 1.17356336e+00 -8.36288407e-02
-6.88626051e-01 -2.78980851e-01 -3.53616774e-01 6.06278241e-01
7.09712327e-01 -5.09871364e-01 -6.15639269e-01 -4.78960723e-01
3.07888705e-02 5.47099411e-01 2.40230426e-01 -1.39271975e+00
7.46675611e-01 -9.14668024e-01 3.45024079e-01 1.17116403e-02
3.43615294e-01 -4.39031631e-01 2.99204767e-01 1.09537184e+00
-6.85597301e-01 4.48272139e-01 4.63256836e-01 7.77024329e-01
2.15013623e-01 6.04991078e-01 1.12191868e+00 -3.92887652e-01
-5.75206041e-01 3.94145310e-01 -1.08681965e+00 -2.44294032e-01
1.28274405e+00 -1.66346982e-01 -1.45043328e-01 -3.56741071e-01
-9.93150949e-01 8.60974312e-01 6.36023402e-01 3.62124920e-01
4.31729406e-01 -1.10189044e+00 -4.36438739e-01 1.55741408e-01
-1.32693738e-01 -4.70609367e-02 -1.88343853e-01 6.47625029e-01
-5.21501899e-01 3.32269788e-01 -4.50426519e-01 -5.90178013e-01
-7.37833321e-01 7.80548334e-01 8.31479669e-01 -1.23443827e-01
-6.64851069e-01 8.73008788e-01 3.19622993e-01 -3.17325890e-01
2.10967362e-01 -4.57998663e-01 6.77195936e-02 -4.17631775e-01
5.42745411e-01 2.88661152e-01 -3.48850697e-01 -8.23388621e-02
-5.50885499e-01 4.23876166e-01 1.44309834e-01 -6.77519500e-01
1.42621636e+00 -5.15974090e-02 -3.85629624e-01 7.34412372e-01
4.20487463e-01 -4.94880170e-01 -1.84099209e+00 1.39297619e-01
-6.95192218e-02 1.05070487e-01 9.09552947e-02 -1.05476296e+00
-7.63333380e-01 5.07417500e-01 3.31303418e-01 1.41753241e-01
1.26736116e+00 -2.03883171e-01 4.94724184e-01 6.33492231e-01
7.61451662e-01 -9.97385859e-01 3.65674675e-01 9.73124325e-01
1.26327312e+00 -5.48470795e-01 -1.40666008e-01 4.51656312e-01
-6.61877573e-01 1.01174796e+00 9.44722116e-01 -4.99943823e-01
6.44956648e-01 6.00463450e-01 -1.35739177e-01 -1.79593906e-01
-2.01706147e+00 8.71210992e-02 -2.73388356e-01 5.22930264e-01
1.15647338e-01 -1.48118541e-01 -7.42240669e-03 1.08670525e-01
2.22765431e-02 1.33732885e-01 2.09418118e-01 1.30624926e+00
-6.03348434e-01 -8.51703644e-01 -9.32435766e-02 1.44769862e-01
1.32999986e-01 2.65777022e-01 -5.30065715e-01 7.98693597e-01
5.84896952e-02 6.89574182e-01 2.47992709e-01 -2.98518032e-01
-1.05705336e-01 1.47481441e-01 3.48359168e-01 -6.63132608e-01
-7.89374352e-01 1.38382912e-01 -2.37463385e-01 -1.11301672e+00
3.34754866e-03 -7.81845033e-01 -1.75712371e+00 -3.10003817e-01
2.09857434e-01 9.18733254e-02 3.56606364e-01 8.81188512e-01
8.09102833e-01 6.65314078e-01 4.60883528e-01 -1.34732425e+00
-5.09711802e-01 -7.78101742e-01 -2.12392166e-01 -1.62961915e-01
6.96857810e-01 -6.01816714e-01 -3.20131212e-01 1.85899600e-01] | [4.161577224731445, 1.6920766830444336] |
337a2e33-f408-49c5-b5bf-f246b4d37d17 | can-foundation-models-perform-zero-shot-task | 2204.11134 | null | https://arxiv.org/abs/2204.11134v1 | https://arxiv.org/pdf/2204.11134v1.pdf | Can Foundation Models Perform Zero-Shot Task Specification For Robot Manipulation? | Task specification is at the core of programming autonomous robots. A low-effort modality for task specification is critical for engagement of non-expert end-users and ultimate adoption of personalized robot agents. A widely studied approach to task specification is through goals, using either compact state vectors or goal images from the same robot scene. The former is hard to interpret for non-experts and necessitates detailed state estimation and scene understanding. The latter requires the generation of desired goal image, which often requires a human to complete the task, defeating the purpose of having autonomous robots. In this work, we explore alternate and more general forms of goal specification that are expected to be easier for humans to specify and use such as images obtained from the internet, hand sketches that provide a visual description of the desired task, or simple language descriptions. As a preliminary step towards this, we investigate the capabilities of large scale pre-trained models (foundation models) for zero-shot goal specification, and find promising results in a collection of simulated robot manipulation tasks and real-world datasets. | ['Aravind Rajeswaran', 'Vikash Kumar', 'Abhinav Gupta', 'Scott Niekum', 'Yuchen Cui'] | 2022-04-23 | null | null | null | null | ['robot-manipulation'] | ['robots'] | [ 3.75187725e-01 3.82817715e-01 5.87170795e-02 -4.85554963e-01
-3.65950972e-01 -7.76037872e-01 8.03709865e-01 4.77175750e-02
-3.59294266e-01 5.11088967e-01 6.77012801e-02 -1.72898322e-01
-2.23556146e-01 -3.67397815e-01 -5.11335254e-01 -2.91391551e-01
-1.68313924e-02 8.61897945e-01 1.57929197e-01 -4.28355515e-01
4.57903266e-01 4.63113248e-01 -1.67419839e+00 1.21933617e-01
5.97391307e-01 6.62710547e-01 9.06964719e-01 7.28442192e-01
-1.25063270e-01 6.67904258e-01 -2.99567580e-01 2.21217964e-02
2.56658971e-01 -1.54688656e-01 -8.05817008e-01 5.49146652e-01
1.31378144e-01 -4.18185443e-01 1.88720990e-02 1.04295444e+00
4.11274321e-02 3.36362094e-01 8.96670938e-01 -1.66134167e+00
-3.86416346e-01 6.07753456e-01 2.19914541e-01 -5.30009687e-01
6.00684285e-01 6.86166704e-01 9.09169495e-01 -5.29175758e-01
7.70289361e-01 1.28564942e+00 2.00113684e-01 7.74454594e-01
-1.36555862e+00 -2.30768502e-01 2.62779832e-01 -6.58831820e-02
-1.11261845e+00 -6.45780325e-01 8.64618063e-01 -7.46887922e-01
9.30732608e-01 -1.36722714e-01 5.49099326e-01 1.16991878e+00
2.17565507e-01 6.66027069e-01 8.20609152e-01 -2.57951856e-01
5.54493427e-01 2.98324764e-01 1.03328094e-01 7.20694602e-01
2.31625125e-01 8.77863243e-02 -3.91015023e-01 -1.68919131e-01
1.09863651e+00 -2.85512321e-02 -2.41090879e-01 -1.34043324e+00
-1.42612910e+00 5.63858151e-01 3.12106848e-01 2.26815775e-01
-7.74201155e-01 4.78652924e-01 4.16689485e-01 4.43152070e-01
-1.55506328e-01 8.55648398e-01 -3.48998666e-01 -2.86239564e-01
-3.97014707e-01 5.26667535e-01 9.88733590e-01 1.44195366e+00
6.59498334e-01 -6.93829134e-02 -1.17681399e-01 4.59050149e-01
3.20536941e-01 2.40708992e-01 2.18329921e-01 -1.36411250e+00
4.15550828e-01 6.28506243e-01 8.74008536e-01 -6.58726394e-01
-5.87370396e-01 -2.74470504e-02 -3.08226854e-01 7.06714928e-01
6.31710589e-01 -1.26417503e-01 -1.04477751e+00 1.59577906e+00
5.66339940e-02 -3.58174652e-01 3.63030434e-01 1.09826541e+00
4.43873942e-01 6.17767990e-01 3.69610339e-01 1.51909804e-02
1.36816287e+00 -1.05720448e+00 -5.15365899e-01 -7.78431237e-01
5.98919034e-01 -3.15121084e-01 1.37624359e+00 4.69988406e-01
-8.65744710e-01 -5.16475379e-01 -8.60566258e-01 -4.10689563e-02
-2.56036341e-01 2.35192239e-01 7.04123139e-01 2.41229042e-01
-9.99128580e-01 5.33708811e-01 -8.22741091e-01 -6.52770758e-01
1.13748312e-01 3.53492767e-01 -5.93681574e-01 -2.80690610e-01
-5.24156928e-01 1.18586528e+00 5.56464851e-01 -8.54375511e-02
-1.34204328e+00 -1.65491477e-01 -1.19744802e+00 2.16742262e-01
6.56361341e-01 -6.92773759e-01 1.63185632e+00 -9.22049224e-01
-1.73950779e+00 6.88856244e-01 1.80553913e-01 -2.12075397e-01
5.04317403e-01 -2.17834115e-01 4.26248580e-01 -1.31036118e-01
1.41270801e-01 9.46505904e-01 1.00823236e+00 -1.47401738e+00
-5.75335443e-01 -2.04707101e-01 5.71141660e-01 4.84533817e-01
1.06950156e-01 -1.37555152e-01 -3.89826089e-01 -8.54969099e-02
1.20558916e-02 -1.24389052e+00 -5.66695035e-01 2.90330499e-01
-2.58936554e-01 -3.89276713e-01 4.88458186e-01 -4.50980067e-01
3.54089886e-01 -2.08927798e+00 4.84784633e-01 -4.16820347e-02
-2.04280261e-02 -1.09154120e-01 -3.71817857e-01 5.42365193e-01
1.23437695e-01 -1.37760997e-01 -1.27158120e-01 -4.10355836e-01
4.94861692e-01 2.10229054e-01 -8.40882212e-02 3.54320437e-01
2.99830198e-01 8.02619398e-01 -1.23610103e+00 -3.21985930e-01
4.30630535e-01 1.15694918e-01 -4.56021935e-01 5.80622077e-01
-7.41846025e-01 5.18380165e-01 -6.11522496e-01 2.55747467e-01
-2.24348474e-02 -1.60472929e-01 2.71099508e-01 -5.74124195e-02
-1.02327660e-01 3.90916407e-01 -1.16545415e+00 1.97587478e+00
-6.45414233e-01 3.97055179e-01 2.32742578e-01 -8.85007620e-01
8.89865339e-01 4.69152182e-01 2.70851284e-01 -2.08514929e-01
1.51242688e-01 1.49428695e-01 -4.07265536e-02 -6.40473127e-01
4.96146262e-01 -5.00493832e-02 -4.81645674e-01 6.20304286e-01
5.55810444e-02 -9.05755401e-01 4.29569989e-01 -1.62707805e-03
1.06470823e+00 6.69479430e-01 4.83305603e-01 -2.36985147e-01
8.69278014e-02 6.07616127e-01 1.20979667e-01 9.40783203e-01
-2.01028585e-01 3.69037390e-01 4.18570340e-01 -4.86110926e-01
-1.21195602e+00 -8.64080131e-01 5.12890339e-01 1.14451170e+00
2.96823353e-01 -2.43213668e-01 -4.52288210e-01 -5.70210516e-01
-3.08521658e-01 1.06243479e+00 -3.49266350e-01 -7.54816905e-02
-5.64101338e-01 1.83873430e-01 -4.18513790e-02 2.99094647e-01
2.44312793e-01 -1.52207041e+00 -1.56343484e+00 3.99670511e-01
-1.78695604e-01 -1.17801511e+00 -2.91469604e-01 3.78031671e-01
-7.16375768e-01 -9.17502701e-01 -7.40111828e-01 -9.14968312e-01
9.92479086e-01 1.89997926e-01 1.00641894e+00 -1.16870448e-01
-1.59529716e-01 7.82684088e-01 -3.37728560e-01 -5.10666847e-01
-5.76401591e-01 -9.41634849e-02 -3.70752974e-03 -4.03756648e-01
-1.85267359e-01 -4.12227124e-01 -1.41716868e-01 1.58492073e-01
-6.09113216e-01 6.21229887e-01 7.43039727e-01 7.38963962e-01
2.67700404e-01 5.43400599e-03 2.10557595e-01 -5.17579615e-01
1.01930487e+00 -2.67963916e-01 -7.26702750e-01 2.89454848e-01
-3.31336617e-01 4.59007800e-01 6.32320642e-01 -7.52578974e-01
-8.96038890e-01 7.25294352e-01 3.23052824e-01 -3.36288214e-01
-4.29880738e-01 5.52198529e-01 9.80169252e-02 1.65310532e-01
8.08353066e-01 2.63376147e-01 6.84574023e-02 -2.38254026e-01
5.66392720e-01 4.84950244e-01 5.38520038e-01 -7.30202019e-01
6.08555019e-01 1.34099022e-01 -1.56234384e-01 -7.46527612e-01
-3.39587599e-01 -4.06029135e-01 -5.59074819e-01 -1.67831674e-01
7.03426480e-01 -6.17792428e-01 -6.55174553e-01 1.47457480e-01
-1.40328932e+00 -9.91892993e-01 -1.41103014e-01 2.05830783e-01
-1.27753341e+00 1.72104891e-02 -1.18867993e-01 -9.83195543e-01
-3.51978168e-02 -1.48915982e+00 1.27696180e+00 1.97369397e-01
-6.80158794e-01 -5.95701993e-01 -3.21872115e-01 -6.12878762e-02
4.54935342e-01 3.76198925e-02 9.77676332e-01 -5.52084327e-01
-6.04286969e-01 -2.93024838e-01 -1.34936869e-01 -6.87044263e-02
2.34159589e-01 -3.85302991e-01 -5.44066072e-01 -1.22661471e-01
-7.27293417e-02 -6.40018463e-01 9.72078741e-02 2.25983664e-01
7.25627780e-01 -1.47117361e-01 -4.01987463e-01 -2.12787976e-03
1.26206577e+00 3.97223473e-01 4.06147301e-01 1.56653747e-01
3.93531680e-01 1.00050783e+00 1.04369187e+00 4.83961910e-01
5.93267977e-01 9.03883815e-01 5.13812363e-01 2.84918427e-01
1.06428839e-01 -1.24362856e-01 4.13468689e-01 -6.38492852e-02
-2.82049663e-02 -1.54105335e-01 -1.09803045e+00 7.57538438e-01
-2.15820074e+00 -7.28125751e-01 2.62997955e-01 1.97598171e+00
5.22006512e-01 1.59388602e-01 5.96200228e-02 -2.58636713e-01
5.18181086e-01 -2.14026719e-01 -6.93943799e-01 -3.98748398e-01
7.49567211e-01 -2.26543233e-01 1.71983972e-01 4.23655599e-01
-9.56833482e-01 1.21119058e+00 5.99950552e+00 1.90283477e-01
-9.71628606e-01 -1.00141779e-01 1.58440500e-01 1.72979042e-01
-4.86948416e-02 1.98412493e-01 -4.84942883e-01 1.29938290e-01
6.91416919e-01 -2.14373410e-01 7.22779870e-01 1.25629556e+00
3.19305897e-01 -3.63692373e-01 -1.75861084e+00 1.10250878e+00
-1.68311045e-01 -9.43494499e-01 -2.62943745e-01 -1.89224452e-01
2.29010135e-01 4.98977304e-03 -1.94339737e-01 6.24177337e-01
7.90657163e-01 -9.90885794e-01 1.10119557e+00 4.71648216e-01
6.18088245e-01 -2.58669496e-01 3.83311659e-01 8.71199548e-01
-9.42620695e-01 -3.93885761e-01 -2.69414872e-01 -2.90291578e-01
4.32132751e-01 -2.53466338e-01 -1.17675138e+00 1.38971373e-01
4.15421009e-01 3.60914737e-01 -2.20730916e-01 8.35258424e-01
-3.60244989e-01 -1.72925130e-01 -3.76150131e-01 -4.21378225e-01
3.09638917e-01 3.12689580e-02 5.89579701e-01 8.20030630e-01
2.66890436e-01 6.90162107e-02 6.48587942e-01 1.04767120e+00
4.51019466e-01 -2.20571667e-01 -1.00725698e+00 -3.79562855e-01
2.60981381e-01 1.15883112e+00 -6.91292644e-01 -2.93196589e-01
-1.36365697e-01 1.11027598e+00 5.29194772e-01 3.80573362e-01
-6.03200257e-01 -2.91576177e-01 7.00485229e-01 -5.34653775e-02
-2.06457730e-02 -8.02853286e-01 -1.64174631e-01 -9.17639017e-01
-5.24420850e-02 -9.69534278e-01 -1.28530562e-01 -1.26560032e+00
-8.05570006e-01 4.05725986e-01 3.43806595e-01 -1.19438744e+00
-8.54058623e-01 -7.75467753e-01 -4.12431389e-01 7.56673634e-01
-1.13857698e+00 -1.26985776e+00 -6.00100577e-01 2.92436868e-01
9.42829788e-01 -3.27437259e-02 9.48296726e-01 -3.19947064e-01
-5.29423133e-02 -2.18840376e-01 -5.38358688e-01 -2.85108358e-01
4.09890413e-01 -1.00842452e+00 5.26516676e-01 5.32856345e-01
7.77962580e-02 6.00612581e-01 1.15168381e+00 -8.06796908e-01
-1.77268612e+00 -7.71784723e-01 6.03613794e-01 -5.22004843e-01
4.31167901e-01 -4.84229118e-01 -5.11419058e-01 8.18278253e-01
-3.41607220e-02 -2.69890577e-01 -6.66653439e-02 -9.60744098e-02
7.14812279e-02 3.46824318e-01 -1.11999404e+00 1.05936086e+00
1.10153067e+00 -3.62616479e-01 -7.17901766e-01 3.08862150e-01
6.17949307e-01 -3.00151616e-01 -4.37236935e-01 1.91754624e-01
6.19545937e-01 -5.81819952e-01 8.31096470e-01 -5.41309059e-01
4.03481841e-01 -4.32354122e-01 -6.17019236e-02 -1.38521504e+00
-4.57013905e-01 -5.22175491e-01 2.68825412e-01 7.41394162e-01
4.40408051e-01 -2.85699010e-01 7.23962545e-01 1.29725671e+00
-2.77047813e-01 -2.58326888e-01 -4.41508621e-01 -8.95248652e-01
-4.57258821e-01 -5.55551291e-01 3.72408122e-01 5.96846998e-01
4.10855144e-01 4.52233583e-01 -2.79518396e-01 2.10841209e-01
4.28037196e-01 1.46962449e-01 1.22799385e+00 -1.24503183e+00
-1.53331205e-01 -4.10345078e-01 -2.42616042e-01 -1.16247535e+00
3.33752066e-01 -6.32807970e-01 8.37926745e-01 -1.99898040e+00
-6.19703643e-02 -5.98649383e-01 3.37128907e-01 6.83010638e-01
1.53709605e-01 -2.90565848e-01 3.87298822e-01 1.89410821e-01
-7.13392735e-01 5.61588824e-01 1.28937197e+00 -1.85063884e-01
-5.31164408e-01 2.77671241e-03 -5.59677303e-01 7.56139815e-01
8.80198896e-01 -4.54230905e-01 -8.46727312e-01 -5.35518765e-01
2.13982627e-01 4.15241659e-01 4.22513872e-01 -1.08561313e+00
3.50191981e-01 -7.11293995e-01 -1.58888936e-01 -1.24444380e-01
7.42027700e-01 -1.17860401e+00 3.96946520e-01 5.61432242e-01
-4.47353154e-01 6.97735045e-03 3.36938538e-02 3.91538948e-01
1.22441180e-01 -6.95856273e-01 5.25284171e-01 -6.23960376e-01
-1.21481454e+00 4.66257744e-02 -5.48480868e-01 -5.21555126e-01
1.31608248e+00 -2.23342389e-01 -9.77227092e-02 -7.67687023e-01
-7.74357498e-01 4.60685849e-01 7.56976128e-01 5.94895840e-01
6.68521523e-01 -8.91305566e-01 -4.72656369e-01 -8.59122127e-02
4.82093334e-01 1.94331288e-01 -2.36709267e-01 3.67070585e-01
-4.48621660e-01 4.24332529e-01 -4.70976114e-01 -5.62349856e-01
-1.04489577e+00 5.17397523e-01 1.69028789e-01 -1.70242578e-01
-6.99117899e-01 5.70221901e-01 5.02220869e-01 -5.88807642e-01
2.10660532e-01 -3.62840682e-01 -2.97576219e-01 -4.31541324e-01
2.68451512e-01 -1.58652782e-01 -5.18612921e-01 -3.95038575e-01
-1.89211100e-01 1.82310894e-01 7.09598139e-02 -4.35108215e-01
1.34074712e+00 -1.77434966e-01 1.16737194e-01 5.44678032e-01
4.69489366e-01 -7.36502647e-01 -1.63900685e+00 -1.19753920e-01
2.65634477e-01 -2.08594278e-01 -2.27127671e-01 -7.27368593e-01
-2.49763697e-01 7.36438036e-01 3.05407017e-01 1.77444711e-01
6.50473714e-01 1.45681828e-01 2.34800041e-01 1.05509186e+00
1.10319293e+00 -9.93884683e-01 5.32295644e-01 6.67401314e-01
1.37194049e+00 -1.30523372e+00 -2.11804926e-01 -3.38934571e-01
-1.03568566e+00 1.06029916e+00 8.29490840e-01 -1.20444372e-01
1.42597988e-01 -8.05978291e-03 -4.11509648e-02 -4.16673183e-01
-8.71279240e-01 -3.47744226e-01 1.64243728e-01 8.51701736e-01
9.16168168e-02 7.36992806e-02 1.73388794e-02 5.47692895e-01
-8.82769376e-02 8.32792670e-02 6.85023427e-01 1.28971696e+00
-5.26476920e-01 -9.61933970e-01 -1.19428314e-01 4.31382000e-01
1.01447478e-01 1.85525492e-01 -3.68635297e-01 5.25295854e-01
-2.66770989e-01 9.64266896e-01 -4.22531702e-02 -4.58089411e-02
4.50512707e-01 3.49506289e-01 6.77289605e-01 -1.22470295e+00
-2.33496547e-01 -2.72937566e-01 4.90411490e-01 -6.35663033e-01
-1.11911774e-01 -5.56323349e-01 -1.27541053e+00 3.26346755e-01
-1.34947583e-01 -1.84011981e-01 1.20480263e+00 1.01878130e+00
2.30783060e-01 3.60813200e-01 1.24212354e-01 -1.57091630e+00
-7.52396703e-01 -9.58968341e-01 -1.93079874e-01 5.72754920e-01
2.04829246e-01 -8.46927762e-01 3.05568129e-02 4.53568190e-01] | [4.498474597930908, 0.8354924917221069] |
524295b0-0654-4e8b-bf8c-101ab75a115d | the-interpreter-understands-your-meaning-end | 2305.09652 | null | https://arxiv.org/abs/2305.09652v1 | https://arxiv.org/pdf/2305.09652v1.pdf | The Interpreter Understands Your Meaning: End-to-end Spoken Language Understanding Aided by Speech Translation | End-to-end spoken language understanding (SLU) remains elusive even with current large pretrained language models on text and speech, especially in multilingual cases. Machine translation has been established as a powerful pretraining objective on text as it enables the model to capture high-level semantics of the input utterance and associations between different languages, which is desired for speech models that work on lower-level acoustic frames. Motivated particularly by the task of cross-lingual SLU, we demonstrate that the task of speech translation (ST) is a good means of pretraining speech models for end-to-end SLU on both monolingual and cross-lingual scenarios. By introducing ST, our models give higher performance over current baselines on monolingual and multilingual intent classification as well as spoken question answering using SLURP, MINDS-14, and NMSQA benchmarks. To verify the effectiveness of our methods, we also release two new benchmark datasets from both synthetic and real sources, for the tasks of abstractive summarization from speech and low-resource or zero-shot transfer from English to French. We further show the value of preserving knowledge from the pretraining task, and explore Bayesian transfer learning on pretrained speech models based on continual learning regularizers for that. | ['Philip N. Garner', 'Mutian He'] | 2023-05-16 | null | null | null | null | ['abstractive-text-summarization', 'spoken-language-understanding', 'intent-classification', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 2.36888409e-01 2.13469312e-01 -1.91509724e-01 -6.46998227e-01
-1.42965662e+00 -5.29157400e-01 6.03852510e-01 -1.16386972e-01
-5.01884580e-01 6.72455072e-01 8.38413060e-01 -6.13151014e-01
4.98827726e-01 -2.83498973e-01 -1.13310659e+00 -8.43078494e-02
9.18694884e-02 6.61997259e-01 -3.21390666e-02 -4.07539010e-01
-4.23960924e-01 -2.93047518e-01 -9.68909979e-01 6.09406233e-01
1.07111597e+00 6.70423210e-01 5.25186479e-01 7.19278991e-01
-2.30803251e-01 7.80540466e-01 -5.24685085e-01 -4.43538576e-01
-2.30491921e-01 -7.10367560e-01 -1.13414812e+00 8.31952170e-02
5.70390165e-01 -2.98006505e-01 -4.06521767e-01 6.73831999e-01
6.64668739e-01 2.45242372e-01 8.43139887e-01 -7.09162295e-01
-8.38135958e-01 1.18870103e+00 -2.42400859e-02 2.44872600e-01
5.70498466e-01 1.52948782e-01 1.36969602e+00 -1.21525621e+00
4.69198704e-01 1.66968715e+00 5.13995230e-01 7.69313097e-01
-1.30388486e+00 -4.83931094e-01 1.91644669e-01 2.98630059e-01
-1.02174473e+00 -1.14664209e+00 5.41409850e-01 4.52500805e-02
1.42478955e+00 2.85650432e-01 8.33431706e-02 1.62942421e+00
1.62845757e-02 1.44774246e+00 8.57689738e-01 -6.03061736e-01
1.23473674e-01 8.05136785e-02 2.49439687e-01 4.59038854e-01
-5.14507115e-01 -1.10551059e-01 -9.54463840e-01 1.03713162e-01
2.53590252e-02 -6.94812059e-01 -6.39368951e-01 1.09111369e-01
-1.37679815e+00 8.16918135e-01 1.19675197e-01 9.47206020e-02
-2.03651503e-01 -1.10041033e-02 7.45155573e-01 6.02795243e-01
7.90924370e-01 2.95205355e-01 -7.41191030e-01 -2.02031583e-01
-8.08182597e-01 -2.22734362e-01 1.01028144e+00 9.88078535e-01
5.33510685e-01 4.60130095e-01 -2.77090609e-01 1.37175477e+00
2.46298119e-01 9.52751815e-01 8.23136747e-01 -7.15579927e-01
7.82526314e-01 -1.27531998e-02 -3.81899655e-01 -1.35984957e-01
-8.91578570e-02 -3.47251683e-01 -7.64851868e-01 -6.38056934e-01
1.26052037e-01 -2.70085424e-01 -9.11447227e-01 2.06888986e+00
-6.44363761e-02 3.15625429e-01 9.49971676e-01 4.69856322e-01
1.09577239e+00 1.33092499e+00 1.00706639e-02 -4.17290747e-01
1.33102179e+00 -1.38474572e+00 -9.19982851e-01 -6.89395905e-01
8.47520351e-01 -8.01733375e-01 1.65600860e+00 3.95668447e-02
-1.17107177e+00 -6.64167523e-01 -8.61241162e-01 -3.66736054e-01
-2.09981099e-01 1.86076313e-01 4.97157872e-02 3.71840417e-01
-1.06813025e+00 1.12279341e-01 -8.38365912e-01 -6.00498617e-01
4.87622339e-03 -1.04843333e-01 -1.90430656e-01 -2.60181814e-01
-1.66495514e+00 1.06253016e+00 4.14725721e-01 -1.89510673e-01
-1.05186880e+00 -7.78226018e-01 -1.18943608e+00 1.16738878e-01
2.76759118e-01 -6.36211395e-01 1.67350316e+00 -8.21386516e-01
-1.87101352e+00 7.92942464e-01 -4.63167131e-01 -8.63780260e-01
1.01740226e-01 -3.81737828e-01 -3.24575812e-01 -8.02746266e-02
6.83789402e-02 7.43902981e-01 7.56260157e-01 -9.50864673e-01
-5.09276271e-01 -1.78348035e-01 -3.79654109e-01 6.42328918e-01
-2.86059022e-01 1.70870528e-01 -1.10472813e-01 -5.53226411e-01
-1.89806476e-01 -6.86881125e-01 3.15809458e-01 -8.49007905e-01
-3.81453514e-01 -6.22362256e-01 9.05723631e-01 -1.20988333e+00
9.58666921e-01 -2.07643437e+00 4.02976990e-01 -4.43691313e-01
-6.48772597e-01 3.21020424e-01 -4.58284199e-01 6.66581273e-01
1.73684567e-01 -1.03037432e-01 -4.06111330e-01 -9.42434072e-01
1.52520150e-01 6.38651729e-01 -8.04381371e-01 -3.37726660e-02
2.98413575e-01 1.16397810e+00 -8.25449347e-01 -1.39331132e-01
8.43374729e-02 3.38074476e-01 -4.45697993e-01 6.45016670e-01
-4.25876617e-01 6.41028523e-01 2.56950874e-02 3.15396994e-01
-9.35639665e-02 1.38300121e-01 5.79327494e-02 3.76999862e-02
1.85777739e-01 1.18488240e+00 -5.25049448e-01 2.11877441e+00
-8.90541434e-01 3.84437740e-01 1.81392491e-01 -1.02918601e+00
6.77701652e-01 7.88497567e-01 -1.38987422e-01 -8.12268734e-01
6.79288991e-03 4.49408233e-01 7.23826215e-02 -4.52961802e-01
4.71171856e-01 -3.48094642e-01 -3.10131341e-01 5.78497708e-01
6.94764912e-01 -6.74338877e-01 -2.09117718e-02 2.59308577e-01
8.83346319e-01 -9.76568386e-02 3.71898741e-01 -2.99772054e-01
5.85181773e-01 -2.10168049e-01 1.26856759e-01 7.26433754e-01
1.02761455e-01 6.66154385e-01 4.09219004e-02 1.13192879e-01
-7.80747652e-01 -1.14325452e+00 -7.55534619e-02 1.68118739e+00
-2.92124361e-01 -3.03886861e-01 -9.67250288e-01 -8.20107639e-01
-1.73343271e-01 1.45433974e+00 -1.23059876e-01 -3.01295489e-01
-6.53601587e-01 -4.83742207e-01 7.41052151e-01 3.73653919e-01
2.18665317e-01 -1.22938240e+00 2.35012397e-01 3.82401437e-01
-7.54578292e-01 -1.62667966e+00 -9.49979007e-01 2.08173946e-01
-5.17879725e-01 -5.49286306e-01 -7.36110449e-01 -1.10623789e+00
1.34567708e-01 1.38804510e-01 1.30494380e+00 -4.49632764e-01
4.26977515e-01 6.36931062e-01 -5.01321554e-01 -4.82402593e-01
-1.16524208e+00 5.37291408e-01 4.14992630e-01 1.20253347e-01
1.51199535e-01 -4.15122628e-01 3.98259014e-02 1.50889114e-01
-7.01499164e-01 1.86174110e-01 4.71358031e-01 1.02009404e+00
4.31640357e-01 -7.50873983e-01 1.15504682e+00 -6.11415863e-01
8.68870437e-01 -3.86194766e-01 -1.25509143e-01 5.12994468e-01
5.40191568e-02 2.67738849e-01 7.26406336e-01 -3.43775928e-01
-1.10178185e+00 -3.17362785e-01 -7.01397121e-01 -2.37543866e-01
-1.81950107e-01 7.62256622e-01 -4.01575029e-01 6.60491943e-01
5.93668997e-01 6.95677340e-01 1.42457202e-01 -5.01501739e-01
7.24207819e-01 1.11307383e+00 7.92952478e-01 -7.21433043e-01
3.31924796e-01 -4.01436277e-02 -8.75039518e-01 -1.36707115e+00
-1.24094641e+00 -4.80765045e-01 -5.33787131e-01 1.25111133e-01
9.80939031e-01 -1.18839490e+00 -1.93390608e-01 4.10693377e-01
-1.62908113e+00 -6.33790851e-01 -3.41599941e-01 5.50965190e-01
-7.71402538e-01 4.07307982e-01 -7.98645258e-01 -6.14164770e-01
-6.96197987e-01 -1.27328742e+00 1.07564652e+00 -2.82406628e-01
-4.60230827e-01 -1.28556883e+00 2.31535912e-01 5.94016373e-01
5.30793607e-01 -6.40404284e-01 1.17969835e+00 -1.14317524e+00
-2.30375931e-01 1.95725352e-01 2.30341598e-01 9.91245151e-01
1.69359863e-01 -5.48755050e-01 -1.09978092e+00 -4.67963606e-01
1.77592143e-01 -9.59136128e-01 9.03563797e-01 2.59364754e-01
5.94049633e-01 -4.77937162e-01 1.46889687e-01 3.81593645e-01
6.58844292e-01 -1.33419126e-01 2.32992098e-01 -2.68744886e-01
6.44261122e-01 8.09673309e-01 1.77589551e-01 -2.22851068e-01
9.31614935e-01 5.52869797e-01 3.51235196e-02 -3.69804539e-02
-5.42667031e-01 -4.67869967e-01 1.09246433e+00 2.09460711e+00
5.91427445e-01 -4.13223088e-01 -1.00662255e+00 7.16490269e-01
-1.82201397e+00 -5.20760715e-01 2.50897497e-01 2.12656546e+00
1.22472620e+00 -1.83693692e-02 -1.25851765e-01 -3.94531935e-01
3.68642718e-01 3.46309394e-01 -3.71377796e-01 -5.23452520e-01
-2.56748676e-01 1.90380663e-01 -1.75262745e-02 8.54678929e-01
-8.97214651e-01 1.39912391e+00 5.79958248e+00 1.00656724e+00
-1.16294098e+00 5.78069806e-01 6.21662140e-01 1.83604777e-01
-2.95534074e-01 -2.07964703e-01 -9.50092733e-01 3.32319856e-01
1.57598352e+00 -2.62224734e-01 4.40945596e-01 5.81441283e-01
2.87604630e-01 1.80310145e-01 -1.54905736e+00 7.53774345e-01
4.81876761e-01 -1.10233593e+00 4.00289804e-01 -4.80042219e-01
8.62734437e-01 7.46026933e-01 -7.12463111e-02 9.12426353e-01
4.63105798e-01 -1.03427219e+00 7.12438762e-01 -2.52270289e-02
8.58722806e-01 -5.76364696e-01 6.57872736e-01 7.73226142e-01
-8.31125200e-01 3.79996836e-01 -3.62753630e-01 8.03747699e-02
6.00472748e-01 -7.15179145e-02 -1.30186951e+00 6.30568266e-01
1.63597137e-01 7.16219068e-01 -1.52523518e-01 2.67675906e-01
-5.11842251e-01 1.30198908e+00 -3.96859765e-01 -2.06444170e-02
4.03557360e-01 7.19411671e-02 7.61854649e-01 1.49544799e+00
2.76422530e-01 -1.55313790e-01 2.62524635e-01 5.91323614e-01
-5.09054959e-01 4.89033073e-01 -6.56132162e-01 -1.41896293e-01
3.50102127e-01 6.27645433e-01 -8.04389864e-02 -3.91070694e-01
-6.84305131e-01 1.17610610e+00 4.54890579e-01 5.19589067e-01
-4.62604463e-01 8.64968151e-02 6.02996469e-01 -6.97751939e-02
9.19332132e-02 -5.05188584e-01 8.17358643e-02 -1.48589051e+00
-1.28728434e-01 -1.26184928e+00 2.34802604e-01 -6.46252155e-01
-1.41145778e+00 8.90281320e-01 -2.77062118e-01 -7.84545958e-01
-7.45107234e-01 -4.16358888e-01 -6.91232502e-01 9.78524566e-01
-1.82560027e+00 -1.57626891e+00 4.71729189e-01 5.17876804e-01
1.43545401e+00 -4.28772271e-01 1.08666873e+00 5.79243079e-02
-4.50570911e-01 6.10116124e-01 2.85405129e-01 1.91101402e-01
9.47401226e-01 -1.11886525e+00 7.76263475e-01 9.97414887e-01
6.89657032e-01 1.94278985e-01 6.78867996e-01 -3.63385230e-01
-1.44628298e+00 -1.18514752e+00 1.01366425e+00 -7.03268528e-01
9.07582343e-01 -7.40101755e-01 -1.32351446e+00 1.01015341e+00
8.78019214e-01 -3.41702789e-01 6.04907632e-01 3.69734168e-01
-1.22891180e-01 5.20584397e-02 -4.17838752e-01 6.27846718e-01
7.78532922e-01 -1.02170062e+00 -1.16764092e+00 4.63937551e-01
1.28224576e+00 -3.67702037e-01 -6.19797826e-01 4.81239825e-01
1.45652562e-01 -5.15904248e-01 8.43683422e-01 -7.70621777e-01
3.20880532e-01 3.13726127e-01 -4.63347912e-01 -1.96412838e+00
2.77863562e-01 -7.77602196e-01 5.54920249e-02 1.40322185e+00
8.04999530e-01 -5.28757334e-01 1.74400657e-01 -4.17770408e-02
-7.61061072e-01 -4.73528445e-01 -1.26464725e+00 -9.13027644e-01
4.00249511e-01 -7.44134963e-01 2.46382341e-01 9.10167158e-01
1.43235952e-01 1.28294969e+00 -3.93721104e-01 1.60106927e-01
4.19025719e-01 -3.44316483e-01 8.40482116e-01 -6.31540477e-01
-4.92973924e-01 -3.26062113e-01 1.62232131e-01 -1.51764393e+00
1.05618072e+00 -1.32454216e+00 5.29455125e-01 -1.52074373e+00
-5.51104248e-02 -3.77032980e-02 -4.07347493e-02 4.09414649e-01
-3.62320483e-01 -1.37793526e-01 9.73813981e-02 9.05157104e-02
-6.01249874e-01 1.19355786e+00 1.06188035e+00 -2.10336819e-01
-1.24218151e-01 6.08431511e-02 -3.29429328e-01 6.99290633e-01
5.20478606e-01 -3.26831669e-01 -5.46802342e-01 -7.84437895e-01
-3.68501186e-01 3.76694858e-01 -2.03370646e-01 -7.20736623e-01
1.29834145e-01 1.98863357e-01 -4.68732715e-01 -5.11473894e-01
5.28013825e-01 -3.48083615e-01 -6.67439699e-01 1.47636011e-01
-5.93696475e-01 -2.24533379e-01 3.02780569e-01 2.98944205e-01
-6.26592219e-01 -2.02541500e-01 7.27154374e-01 -6.88099116e-02
-6.98219001e-01 4.22007926e-02 -4.73958105e-01 6.80496216e-01
4.08268899e-01 2.50933200e-01 -2.63536602e-01 -8.86646152e-01
-7.12661266e-01 5.22513151e-01 -4.50967960e-02 7.97432721e-01
5.37501335e-01 -1.08307099e+00 -1.32251096e+00 4.57726121e-01
1.94725037e-01 2.35400274e-02 9.80888903e-02 8.85778248e-01
-3.13614216e-03 6.60198569e-01 2.93622404e-01 -7.96565473e-01
-1.14276576e+00 1.60966575e-01 3.38695735e-01 -2.48158097e-01
-3.50232184e-01 1.03909385e+00 5.67051709e-01 -1.05180681e+00
4.53810543e-01 -6.06071293e-01 9.44760814e-02 -2.21408401e-02
4.56346065e-01 4.09916490e-02 2.15780661e-01 -8.18723440e-01
-1.90912142e-01 1.52714234e-02 -2.72415251e-01 -5.49871564e-01
1.13662124e+00 -5.03349721e-01 7.46878907e-02 8.83431852e-01
1.20654690e+00 8.11212659e-02 -1.05264711e+00 -6.49837732e-01
1.23450220e-01 3.95426065e-01 -5.81729300e-02 -8.28189850e-01
-2.33229265e-01 1.27431905e+00 -2.62043756e-02 1.32406401e-02
5.35932183e-01 4.16468740e-01 1.21691871e+00 6.78233385e-01
1.67458609e-01 -1.07517052e+00 2.70893902e-01 1.37323678e+00
1.33802783e+00 -1.40490890e+00 -7.54868388e-01 -8.90889317e-02
-9.93509591e-01 7.09672570e-01 3.64050448e-01 2.49108419e-01
4.50366795e-01 1.37845352e-01 3.39855880e-01 3.29725564e-01
-9.79735374e-01 -1.36859208e-01 5.00092506e-01 3.06125969e-01
5.14544010e-01 2.38925010e-01 5.99664487e-02 7.52531648e-01
-5.42892337e-01 -4.92470652e-01 4.62621152e-01 5.61706245e-01
-5.44608176e-01 -1.09740758e+00 -1.46258503e-01 1.96121752e-01
-4.68367934e-01 -6.02753401e-01 -2.99207479e-01 5.27013659e-01
-5.74735761e-01 1.02836740e+00 7.72309527e-02 -1.18854843e-01
4.22280192e-01 7.55839288e-01 2.64416128e-01 -1.12159884e+00
-3.50977212e-01 1.95122063e-01 5.85422158e-01 -2.16662198e-01
-1.41577467e-01 -5.85766554e-01 -1.22912717e+00 1.44565374e-01
-2.68142551e-01 3.70550007e-01 6.70950472e-01 1.40474164e+00
2.33698651e-01 5.95267355e-01 5.27099729e-01 -7.14097023e-01
-9.84872878e-01 -1.54289639e+00 -2.74177909e-01 3.08911830e-01
4.30774301e-01 -6.26894534e-02 -5.01950622e-01 2.13959754e-01] | [14.088396072387695, 7.029539585113525] |
4d2666ed-8e43-43f8-b8a6-f660b456c705 | the-effect-of-spectrogram-reconstruction-on | 2010.09969 | null | https://arxiv.org/abs/2010.09969v1 | https://arxiv.org/pdf/2010.09969v1.pdf | The Effect of Spectrogram Reconstruction on Automatic Music Transcription: An Alternative Approach to Improve Transcription Accuracy | Most of the state-of-the-art automatic music transcription (AMT) models break down the main transcription task into sub-tasks such as onset prediction and offset prediction and train them with onset and offset labels. These predictions are then concatenated together and used as the input to train another model with the pitch labels to obtain the final transcription. We attempt to use only the pitch labels (together with spectrogram reconstruction loss) and explore how far this model can go without introducing supervised sub-tasks. In this paper, we do not aim at achieving state-of-the-art transcription accuracy, instead, we explore the effect that spectrogram reconstruction has on our AMT model. Our proposed model consists of two U-nets: the first U-net transcribes the spectrogram into a posteriorgram, and a second U-net transforms the posteriorgram back into a spectrogram. A reconstruction loss is applied between the original spectrogram and the reconstructed spectrogram to constrain the second U-net to focus only on reconstruction. We train our model on three different datasets: MAPS, MAESTRO, and MusicNet. Our experiments show that adding the reconstruction loss can generally improve the note-level transcription accuracy when compared to the same model without the reconstruction part. Moreover, it can also boost the frame-level precision to be higher than the state-of-the-art models. The feature maps learned by our U-net contain gridlike structures (not present in the baseline model) which implies that with the presence of the reconstruction loss, the model is probably trying to count along both the time and frequency axis, resulting in a higher note-level transcription accuracy. | ['Dorien Herremans', 'Emmanouil Benetos', 'Yin-Jyun Luo', 'Kin Wai Cheuk'] | 2020-10-20 | null | null | null | null | ['music-transcription'] | ['music'] | [ 2.79851526e-01 1.72853574e-01 9.56321973e-03 -1.03388734e-01
-9.36208963e-01 -5.94240606e-01 4.31367964e-01 8.82632360e-02
-2.98009187e-01 4.34156120e-01 3.84917885e-01 -9.45236757e-02
1.17776692e-01 -4.96784419e-01 -8.30867767e-01 -8.22746277e-01
1.01368390e-01 2.95750290e-01 1.62250146e-01 -4.02889550e-02
6.63163289e-02 6.56167194e-02 -1.49278915e+00 5.10018408e-01
8.46325517e-01 1.08570755e+00 1.80724174e-01 5.93407869e-01
-3.02468017e-02 6.82389498e-01 -6.27396941e-01 -2.99322486e-01
3.22287321e-01 -7.98894167e-01 -5.72188079e-01 -9.35171619e-02
3.07255685e-01 -1.23913594e-01 -1.27564490e-01 8.42802167e-01
5.68143547e-01 1.80820435e-01 5.94519794e-01 -7.17961907e-01
-2.43469447e-01 1.13488543e+00 -4.08982396e-01 -2.77578831e-01
3.65886629e-01 -1.69853645e-03 1.26557815e+00 -8.29803467e-01
4.97115642e-01 9.52746451e-01 9.25451100e-01 1.08596891e-01
-1.42856860e+00 -6.85519040e-01 8.32977071e-02 1.85717195e-01
-1.37030339e+00 -5.20125687e-01 9.55596387e-01 -3.83103371e-01
1.03024924e+00 2.48157084e-01 5.81999123e-01 1.07447588e+00
-2.47508846e-02 6.06526613e-01 8.36480498e-01 -6.14615798e-01
1.00256354e-01 -1.85909957e-01 -9.69614908e-02 3.50163907e-01
-5.19170523e-01 7.54314661e-02 -6.98637664e-01 1.10505395e-01
7.62772024e-01 -4.18801069e-01 -5.09585619e-01 4.30432782e-02
-1.14792454e+00 4.65320498e-01 4.59008902e-01 6.64541781e-01
-4.85098958e-01 1.74341202e-01 4.35775071e-01 3.54697913e-01
3.79408985e-01 6.20851219e-01 -3.46261173e-01 -3.53933007e-01
-1.39205718e+00 2.76405990e-01 6.66722715e-01 4.65649158e-01
6.33273005e-01 1.06185369e-01 -5.61197877e-01 1.12923694e+00
-4.43307571e-02 6.89401031e-02 5.93227863e-01 -9.87560630e-01
6.33874416e-01 1.55921847e-01 4.97774929e-02 -7.56727874e-01
-3.72815549e-01 -7.60707140e-01 -7.24363804e-01 -8.05909559e-02
6.60275400e-01 -1.88851446e-01 -7.78894246e-01 1.86015737e+00
-6.85481206e-02 4.45892066e-01 -1.72130167e-01 1.12020552e+00
3.64475489e-01 9.89171267e-01 -2.90780425e-01 -4.29298729e-01
1.07232928e+00 -1.18810689e+00 -9.49518323e-01 3.38893421e-02
5.44655800e-01 -1.10634756e+00 1.21024656e+00 6.54575169e-01
-1.25449562e+00 -1.02317357e+00 -1.11073720e+00 -1.97786286e-01
1.25900805e-01 6.65993273e-01 1.58188120e-01 2.39709124e-01
-8.83202076e-01 1.21028185e+00 -8.12009752e-01 -9.42916796e-02
-3.49005312e-02 2.28808716e-01 -1.22657433e-01 2.73926556e-01
-1.24591947e+00 7.27843225e-01 4.50978845e-01 2.13699546e-02
-6.23723388e-01 -6.19376779e-01 -6.12684608e-01 4.39947009e-01
3.29434037e-01 -4.97369409e-01 1.32830322e+00 -9.74740326e-01
-1.93547118e+00 3.47842485e-01 -1.85343832e-01 -4.97415274e-01
5.17138362e-01 -2.68026322e-01 -2.15253383e-01 -1.16163716e-01
-1.04576901e-01 6.89953327e-01 8.72974455e-01 -1.15828133e+00
-5.64495206e-01 2.82885153e-02 -1.95630535e-01 2.50791788e-01
-1.93869755e-01 -2.92306453e-01 -6.21600151e-01 -9.90524888e-01
2.32197568e-01 -1.21310818e+00 1.95021451e-01 -4.37996298e-01
-4.49996620e-01 -1.34847462e-01 4.80545908e-01 -8.28063607e-01
1.65880692e+00 -2.42921162e+00 4.38853145e-01 7.31728747e-02
-3.95842373e-01 1.49120614e-01 -2.86080599e-01 6.43014610e-01
-2.79157519e-01 4.45646979e-02 -2.70790249e-01 -7.83798099e-01
8.92559905e-03 9.95197669e-02 -5.96862257e-01 2.05402404e-01
7.50489384e-02 6.10238075e-01 -8.05291235e-01 -2.33173985e-02
9.54808891e-02 5.79990268e-01 -7.98069239e-01 1.04550585e-01
-3.43569875e-01 6.95902944e-01 2.09063023e-01 8.30171034e-02
3.45566124e-01 2.37167135e-01 2.76892006e-01 -2.50859112e-01
-3.36967528e-01 1.06760871e+00 -1.17102754e+00 1.94633400e+00
-6.31628454e-01 5.77149510e-01 -1.80412680e-01 -6.88152432e-01
9.01671171e-01 7.51310050e-01 4.50051785e-01 -4.86933678e-01
1.07154414e-01 4.07215387e-01 3.57825577e-01 -1.91582412e-01
4.59012032e-01 -3.77022237e-01 4.97630537e-02 4.77362037e-01
3.48973066e-01 -1.23493023e-01 1.80041209e-01 -3.00566643e-01
7.96101093e-01 5.43167353e-01 -6.93857074e-02 7.57957846e-02
4.61544871e-01 -4.28970277e-01 6.32567823e-01 5.98342538e-01
3.11539054e-01 1.00501716e+00 7.28230774e-01 -1.81000724e-01
-1.10044885e+00 -8.64238501e-01 6.73785657e-02 1.17570066e+00
-3.87548298e-01 -8.91826332e-01 -9.25124347e-01 -5.20212352e-01
-2.31414661e-01 8.69494677e-01 -4.53068137e-01 -1.57430500e-01
-6.40209556e-01 -4.03122187e-01 7.69685090e-01 4.85329092e-01
3.80847961e-01 -1.07839668e+00 -2.88084298e-01 4.70764339e-01
-5.75253665e-01 -9.12638783e-01 -7.81326413e-01 4.48898047e-01
-8.77957046e-01 -4.23515707e-01 -6.40455604e-01 -5.40221691e-01
1.52215078e-01 -2.90111750e-01 8.79524410e-01 -1.16999187e-01
4.01932567e-01 -3.10933053e-01 -5.46948910e-01 -3.52361888e-01
-4.71482217e-01 4.81193781e-01 7.30120465e-02 4.16711181e-01
-1.18629113e-01 -1.03698456e+00 -2.93522209e-01 1.90368563e-01
-7.24495888e-01 2.44465232e-01 4.75320339e-01 7.54760742e-01
7.14829862e-01 -7.83333834e-03 6.50696635e-01 -7.05292642e-01
4.98718321e-01 5.31502776e-02 -4.05932426e-01 -3.88643071e-02
-3.78425181e-01 1.54904798e-01 9.33761299e-01 -6.22794867e-01
-7.59382725e-01 2.25729108e-01 -5.29427171e-01 -8.50960732e-01
1.02274388e-01 6.44267023e-01 -7.40707889e-02 4.90667105e-01
6.32783115e-01 2.02784196e-01 -2.68735886e-01 -8.03737462e-01
2.89865017e-01 6.18860364e-01 6.78095698e-01 -4.82712567e-01
6.30346775e-01 -2.84555592e-02 -2.18351871e-01 -6.42216921e-01
-1.09237683e+00 -3.11559647e-01 -6.79190099e-01 -1.53876513e-01
7.85594165e-01 -6.56732202e-01 -5.18526495e-01 3.41956705e-01
-1.35231304e+00 -4.14917082e-01 -4.76264238e-01 6.37012541e-01
-7.46690214e-01 9.34490114e-02 -8.32696497e-01 -8.07064593e-01
-1.97782233e-01 -1.14834654e+00 1.09891760e+00 -1.09420672e-01
-4.18586880e-01 -6.30350411e-01 1.78757817e-01 1.53085813e-01
2.90882647e-01 -1.24241024e-01 1.02858388e+00 -4.25223887e-01
-3.60924989e-01 6.79698512e-02 1.67865872e-01 4.62987751e-01
7.01878294e-02 2.25917026e-02 -1.52100420e+00 -3.60977314e-02
2.24996656e-01 -1.23762183e-01 1.23448944e+00 4.93424147e-01
1.28405702e+00 -4.11550820e-01 8.89860168e-02 7.51042426e-01
1.10243225e+00 1.70325145e-01 8.53456974e-01 -4.28160158e-04
6.62362754e-01 5.35928011e-01 5.16098499e-01 2.24069968e-01
7.86732435e-02 1.04890561e+00 1.93816707e-01 2.71610357e-02
-3.21350873e-01 -6.35759592e-01 5.84410429e-01 1.34902692e+00
-4.53131616e-01 -2.97067314e-01 -7.53952920e-01 4.36950862e-01
-2.05286407e+00 -9.89017129e-01 -4.72269505e-02 2.46789598e+00
1.05077636e+00 2.51039177e-01 2.20952928e-01 6.77114069e-01
4.53963459e-01 1.94084466e-01 -2.06541568e-01 -5.68875074e-01
1.52563557e-01 3.47626150e-01 1.64108962e-01 5.90721965e-01
-9.86819029e-01 8.34657907e-01 5.84819174e+00 1.09365439e+00
-1.59058702e+00 1.09971896e-01 3.98596883e-01 -2.11852893e-01
-1.21973485e-01 9.92331952e-02 -4.90793020e-01 4.40037996e-01
1.10599387e+00 3.58841360e-01 7.13648558e-01 4.29249018e-01
3.66095304e-01 6.57686666e-02 -1.35195112e+00 9.03489649e-01
-2.02815339e-01 -1.13846147e+00 -1.58842234e-03 5.50956652e-02
6.06289446e-01 -1.64315030e-01 -8.36932436e-02 5.72972894e-01
-3.01072806e-01 -1.02980161e+00 1.13388145e+00 4.85009074e-01
8.72760892e-01 -8.96961689e-01 8.19892049e-01 5.01697600e-01
-1.27339184e+00 -5.43246791e-02 -1.89315438e-01 -3.48395407e-01
3.61248165e-01 7.46435702e-01 -1.00777769e+00 8.62892568e-01
4.59820628e-01 7.02143729e-01 -1.73620865e-01 9.39811885e-01
-5.23749352e-01 1.07070410e+00 -3.18117768e-01 4.75689441e-01
2.39505261e-01 -4.18831289e-01 5.85931778e-01 1.33686566e+00
4.61118847e-01 -1.87286943e-01 2.56586283e-01 8.61327410e-01
-1.76230818e-01 1.54410616e-01 -1.59815967e-01 -1.64434060e-01
4.39175755e-01 1.04652405e+00 -6.21246219e-01 -2.59394795e-01
-1.22869365e-01 1.04620159e+00 3.46225560e-01 3.88678759e-01
-9.78247464e-01 -3.29841226e-01 4.60365087e-01 3.33043575e-01
4.78198737e-01 -7.96652883e-02 -5.05730629e-01 -9.99780118e-01
-4.66676801e-02 -8.88192475e-01 -9.05554090e-03 -7.52181470e-01
-1.02752757e+00 6.45339727e-01 -2.98500389e-01 -1.32419074e+00
-6.76938891e-01 -1.18277460e-01 -6.06398106e-01 1.01407778e+00
-1.16457510e+00 -9.55075204e-01 1.55677319e-01 1.92464814e-01
5.20289004e-01 2.40217105e-01 9.76133883e-01 5.14295638e-01
-4.24582601e-01 6.47306144e-01 -1.74777910e-01 1.86513394e-01
1.03237128e+00 -1.29051304e+00 3.49100769e-01 6.53512001e-01
5.87879419e-01 4.95035321e-01 7.65446484e-01 -3.88343841e-01
-7.71438181e-01 -1.05230772e+00 1.18815768e+00 -2.94735700e-01
4.62418288e-01 -6.43755734e-01 -1.07437468e+00 6.11247182e-01
1.37776911e-01 -2.02001825e-01 4.72244203e-01 3.89214128e-01
-4.07701254e-01 -2.21116066e-01 -5.30895710e-01 5.34228683e-01
8.26772988e-01 -7.08873153e-01 -6.14196479e-01 -1.45936981e-01
8.17621529e-01 -5.09370506e-01 -7.06053138e-01 3.94863129e-01
8.40951085e-01 -1.05692756e+00 6.38859391e-01 -1.28455520e-01
5.88312626e-01 -4.13950920e-01 -5.14814854e-02 -1.59772778e+00
-3.75104457e-01 -6.95326567e-01 -6.00672886e-02 1.44118071e+00
5.77368498e-01 -2.55629361e-01 2.80459702e-01 -1.47581488e-01
-4.85884488e-01 -8.55922520e-01 -9.61423993e-01 -8.56865466e-01
-4.70837131e-02 -6.60364807e-01 4.40925747e-01 8.82552505e-01
8.31082836e-02 6.65176868e-01 -5.58840692e-01 -5.66969067e-02
1.39567465e-01 1.70053720e-01 6.39406443e-01 -1.08755791e+00
-7.73063421e-01 -5.44141471e-01 5.34123331e-02 -1.15361953e+00
-5.65936714e-02 -9.26508605e-01 3.26786399e-01 -1.27245605e+00
-1.67860880e-01 -3.65960628e-01 -3.94106686e-01 8.06782246e-01
-8.03992599e-02 3.59018683e-01 7.03446746e-01 3.48822474e-01
-1.61248118e-01 6.10291004e-01 1.12742853e+00 1.30400164e-02
-7.07742393e-01 2.72894111e-02 -3.25646907e-01 7.26910353e-01
6.00111246e-01 -6.55242443e-01 -3.08221549e-01 -3.82778168e-01
7.45468736e-02 4.23039526e-01 1.88340638e-02 -1.30693889e+00
1.66504085e-01 2.07492337e-01 3.27699691e-01 -4.97796118e-01
5.43081105e-01 -5.24162650e-01 4.24271166e-01 2.56329507e-01
-4.93362248e-01 -2.65116215e-01 1.61394730e-01 1.47660345e-01
-4.34611142e-01 -2.70728558e-01 7.42190957e-01 1.56438634e-01
1.52279302e-01 -9.32724625e-02 -2.47673288e-01 -3.86311293e-01
2.98418075e-01 -9.63613242e-02 8.36583376e-02 -5.23882687e-01
-9.67027724e-01 -2.25535586e-01 3.71493161e-01 3.44877362e-01
-1.24440563e-03 -1.33089435e+00 -7.48032749e-01 2.96510726e-01
-1.50739476e-01 -5.50052859e-02 1.60205990e-01 1.05080640e+00
-1.17151119e-01 4.30092484e-01 4.88662720e-03 -6.63306117e-01
-9.39641118e-01 4.21952009e-01 4.04223651e-01 -6.39313400e-01
-5.90035081e-01 6.99726224e-01 2.72491395e-01 -4.77320284e-01
5.32388806e-01 -6.88427925e-01 -8.88102129e-02 3.12069952e-01
3.18930298e-01 1.37242705e-01 3.45415890e-01 -4.97127324e-01
-1.32996991e-01 6.04094982e-01 2.18713194e-01 -5.96535802e-01
1.33994114e+00 9.05621126e-02 5.02337702e-03 9.97048914e-01
9.67157364e-01 4.02170748e-01 -1.22008502e+00 7.15084597e-02
-1.10334292e-01 -1.41210064e-01 1.32664323e-01 -1.04238391e+00
-9.86325145e-01 1.04262900e+00 2.46003151e-01 1.61214024e-01
1.32316434e+00 -2.57613480e-01 8.11471701e-01 8.66915658e-02
2.37220764e-01 -1.08830547e+00 2.86193304e-02 8.30326855e-01
1.02786720e+00 -5.73772848e-01 -3.43551695e-01 -3.79942030e-01
-5.73258102e-01 1.08748078e+00 1.49528384e-01 -3.33639115e-01
3.39218765e-01 2.99058646e-01 7.27861300e-02 3.37715447e-01
-7.98722565e-01 -2.68201113e-01 5.70395112e-01 2.41503298e-01
7.94778407e-01 -5.42099737e-02 -3.85711223e-01 1.05524945e+00
-7.46421814e-01 7.33429864e-02 4.06679064e-01 3.78432631e-01
-3.24352413e-01 -1.36651158e+00 -4.56367880e-01 1.67864993e-01
-6.26990914e-01 -1.05381489e-01 -4.79853153e-01 4.41267461e-01
4.93193775e-01 9.16822672e-01 1.22607596e-01 -8.00136387e-01
4.98518765e-01 6.04179680e-01 5.20624876e-01 -7.17782557e-01
-9.89802420e-01 7.40235806e-01 1.10822663e-01 -4.03353184e-01
-2.48439848e-01 -4.56507385e-01 -1.29987061e+00 3.69671583e-02
-4.57681715e-01 2.47054011e-01 7.10497737e-01 8.71765137e-01
1.61570236e-01 9.63682711e-01 4.85775739e-01 -1.21337748e+00
-4.58986282e-01 -1.35244548e+00 -4.89511132e-01 3.13112676e-01
4.28999633e-01 -3.71670425e-01 -3.41866732e-01 1.69934139e-01] | [15.790068626403809, 5.46782922744751] |
8dc3e803-2024-46d0-9e1c-39a709ab3562 | deep-transfer-learning-for-texture | 2004.01614 | null | https://arxiv.org/abs/2004.01614v1 | https://arxiv.org/pdf/2004.01614v1.pdf | Deep Transfer Learning for Texture Classification in Colorectal Cancer Histology | Microscopic examination of tissues or histopathology is one of the diagnostic procedures for detecting colorectal cancer. The pathologist involved in such an examination usually identifies tissue type based on texture analysis, especially focusing on tumour-stroma ratio. In this work, we automate the task of tissue classification within colorectal cancer histology samples using deep transfer learning. We use discriminative fine-tuning with one-cycle-policy and apply structure-preserving colour normalization to boost our results. We also provide visual explanations of the deep neural network's decision on texture classification. With achieving state-of-the-art test accuracy of 96.2% we also embark on using deployment friendly architecture called SqueezeNet for memory-limited hardware. | ['Srinath Jayachandran', 'Ashlin Ghosh'] | 2020-04-03 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 1.36002406e-01 -8.50081630e-03 -1.84332073e-01 -3.09122711e-01
-8.01430404e-01 -4.42953050e-01 1.45600975e-01 4.60656762e-01
-6.57256722e-01 4.34989184e-01 -3.13383371e-01 -1.03355229e+00
1.20326795e-01 -8.96071196e-01 -4.47624356e-01 -1.11636972e+00
-2.37494543e-01 3.43440861e-01 6.80463389e-03 4.35897969e-02
2.13244304e-01 7.89407492e-01 -7.51775026e-01 8.71543825e-01
2.22998396e-01 1.01792133e+00 9.37874392e-02 1.34639740e+00
-2.09923998e-01 6.41730011e-01 -4.31863636e-01 -1.69307977e-01
8.60942006e-02 -9.04264301e-02 -9.74549592e-01 2.16429122e-02
2.02449948e-01 -1.14135355e-01 -2.31907666e-01 9.22670901e-01
4.82480347e-01 -4.68670279e-01 7.70863533e-01 -5.95684171e-01
-2.55514652e-01 2.21808299e-01 -7.04206645e-01 6.12414300e-01
-3.56495529e-01 2.78161585e-01 5.94370544e-01 -5.76616406e-01
5.07036984e-01 5.92465997e-01 6.59924924e-01 4.14712399e-01
-1.27326655e+00 -2.68113792e-01 -3.38046730e-01 2.48682410e-01
-1.39731705e+00 -4.56019163e-01 4.81778771e-01 -4.89239365e-01
9.50239301e-01 6.29800737e-01 6.87178314e-01 6.47428095e-01
1.02297544e+00 4.41620827e-01 1.39807189e+00 -6.80337012e-01
2.76480407e-01 2.94473320e-01 -1.67600915e-01 9.06925559e-01
1.99048266e-01 -1.99964836e-01 6.06096722e-02 -9.93101597e-02
8.29614043e-01 -7.79591501e-03 -5.54982759e-02 -1.53580353e-01
-1.05747700e+00 7.68941820e-01 7.30638862e-01 4.75549966e-01
2.92718355e-02 4.46063846e-01 7.56993532e-01 4.14373964e-01
4.21660602e-01 8.28477666e-02 -4.15220380e-01 3.04489225e-01
-8.66178632e-01 -5.15880644e-01 5.79763055e-01 5.12419999e-01
4.10714775e-01 -3.97683084e-01 -3.36503893e-01 6.71649218e-01
1.70363098e-01 4.12856638e-02 7.57775664e-01 -3.27592760e-01
-3.66896451e-01 6.34041548e-01 -3.73197764e-01 -4.08425748e-01
-9.68718648e-01 -6.47701800e-01 -1.25875580e+00 5.58015883e-01
5.92455149e-01 -1.45897632e-02 -1.25865710e+00 9.56891239e-01
2.18463451e-01 -1.79482117e-01 -4.47869338e-02 8.93593073e-01
6.57775939e-01 1.90593690e-01 2.78597593e-01 2.26293892e-01
1.86944282e+00 -7.78986871e-01 -2.91761011e-01 -6.59777820e-02
1.15228486e+00 -7.92118073e-01 8.21902931e-01 3.97053212e-01
-5.52086592e-01 -3.22266757e-01 -1.07851219e+00 -2.24391744e-01
-5.53505301e-01 8.79716992e-01 8.24714005e-01 9.58327293e-01
-1.29099369e+00 4.83653575e-01 -1.11672401e+00 -6.90887094e-01
5.61273098e-01 7.11826682e-01 -7.66436517e-01 1.21802613e-01
-4.04739112e-01 6.05833888e-01 2.48484105e-01 3.64495724e-01
-6.76094353e-01 -4.96116430e-01 -4.05940711e-01 -3.76048908e-02
-2.55971193e-01 -7.58053839e-01 1.19270647e+00 -1.15915251e+00
-1.49729502e+00 1.61390853e+00 4.26430069e-02 -3.01529437e-01
5.31160235e-01 6.58687353e-01 -3.95744979e-01 1.69929206e-01
-2.95031309e-01 5.86279213e-01 4.08378839e-01 -7.25730300e-01
-7.16120601e-01 -3.10957164e-01 -2.03795463e-01 -1.79580867e-01
-2.89892048e-01 -1.17581308e-01 -6.27821684e-01 -5.16842484e-01
-1.32152900e-01 -1.07502794e+00 -4.57525402e-01 6.60528660e-01
-3.74068081e-01 1.99029967e-01 8.18624854e-01 -6.78168476e-01
8.93130064e-01 -2.31125593e+00 -2.77152389e-01 4.93173331e-01
2.62487501e-01 4.98470180e-02 -5.98013075e-03 4.05440778e-02
-1.47083089e-01 1.33730188e-01 1.73309416e-01 -1.40865132e-01
-1.52464941e-01 -5.49840555e-02 3.34418952e-01 1.06448293e+00
3.31916869e-01 9.26904082e-01 -4.91066724e-01 -7.22938299e-01
1.44138500e-01 5.80770731e-01 -3.09939206e-01 7.24733695e-02
2.05649048e-01 2.21231014e-01 -2.81386048e-01 9.23322141e-01
6.50295258e-01 -5.74233353e-01 6.02572441e-01 -5.55440426e-01
9.70413536e-02 -1.63682282e-01 -5.41024566e-01 1.59784794e+00
-4.79505837e-01 9.48734403e-01 3.73185188e-01 -8.84349585e-01
6.36138678e-01 1.93544164e-01 2.97007650e-01 -6.94253266e-01
2.85187632e-01 1.53182343e-01 3.36796254e-01 -5.31249344e-01
1.26649588e-01 -2.37965375e-01 1.08476140e-01 2.37146050e-01
-1.57211870e-01 2.48834670e-01 8.22506249e-02 -1.42419636e-01
1.32102847e+00 -9.61795673e-02 5.29484093e-01 -7.22955644e-01
4.65805322e-01 1.92826867e-01 1.35897383e-01 4.16994870e-01
-4.95670199e-01 3.27571332e-01 8.96985650e-01 -8.67459238e-01
-1.07617593e+00 -6.52863085e-01 -5.13586283e-01 1.13955057e+00
-3.68003726e-01 2.31984705e-01 -5.96273780e-01 -8.18627775e-01
-3.21708359e-02 2.66543794e-02 -1.36587262e+00 -4.19133529e-03
-5.33276320e-01 -1.12822831e+00 5.55996954e-01 5.27315915e-01
2.02265516e-01 -8.91740918e-01 -9.21967387e-01 1.65428802e-01
4.10402268e-01 -5.79269946e-01 -2.71840036e-01 8.74658525e-01
-7.65238404e-01 -1.25930333e+00 -8.54128778e-01 -1.18422282e+00
1.16398978e+00 1.51543379e-01 1.02155185e+00 4.92069811e-01
-9.33426857e-01 -5.99569641e-02 -2.50261366e-01 -2.86588371e-01
-5.60172677e-01 2.42234424e-01 -6.39672995e-01 -6.03415556e-02
4.69281077e-01 -6.47764876e-02 -9.83820558e-01 4.48286757e-02
-8.27390015e-01 2.28526935e-01 1.19747138e+00 1.02036417e+00
6.88286304e-01 3.64053436e-02 -1.20972797e-01 -9.58864629e-01
3.13117236e-01 -2.23541200e-01 -4.10301596e-01 3.35298359e-01
-2.71262914e-01 2.04719529e-01 5.37091315e-01 -3.36363286e-01
-7.96766579e-01 2.08840087e-01 -1.22737408e-01 -1.46071732e-01
-2.68076062e-01 5.02264917e-01 3.52898687e-01 -5.91091156e-01
7.08033621e-01 -1.77366175e-02 8.28027204e-02 -4.08301391e-02
-3.00571531e-01 7.65643597e-01 3.11958849e-01 -2.83994347e-01
2.41615459e-01 7.44875729e-01 1.88744366e-01 -5.38547099e-01
-1.47964314e-01 -5.76002598e-01 -4.52901453e-01 -3.39182079e-01
8.02259564e-01 -7.61091769e-01 -1.06184220e+00 3.20629776e-01
-7.75603473e-01 -8.43278706e-01 1.67446166e-01 3.00304204e-01
-2.66347378e-01 -8.89646336e-02 -1.23697484e+00 -3.70707184e-01
-6.42055929e-01 -1.21319664e+00 1.13653028e+00 1.88102245e-01
-1.24159284e-01 -1.12759018e+00 2.54570633e-01 -7.57818809e-03
7.40133107e-01 3.41787338e-01 1.05146074e+00 -4.53615576e-01
-3.53667319e-01 -2.99380869e-01 -5.88503122e-01 -2.57955432e-01
1.31720454e-01 4.43318218e-01 -1.32490671e+00 -6.13778949e-01
-4.86759216e-01 -9.53044295e-02 9.39184427e-01 6.56633377e-01
1.48911107e+00 7.03307763e-02 -8.42648387e-01 9.66722608e-01
1.69909203e+00 1.31256478e-02 6.26106501e-01 7.74981499e-01
4.33870584e-01 7.07993507e-01 2.47023895e-01 2.83471048e-01
2.11769622e-02 3.18683803e-01 4.63318884e-01 -7.78452635e-01
-3.39540452e-01 2.75946021e-01 -3.39107327e-02 1.97925404e-01
9.26469490e-02 -1.74295917e-01 -1.05105650e+00 5.67311406e-01
-1.33755517e+00 -3.98613870e-01 -8.06501061e-02 1.84105253e+00
7.62004256e-01 2.46256977e-01 -4.27476838e-02 1.23461477e-01
5.61763048e-01 -5.04490793e-01 -3.07674706e-01 -6.09052837e-01
1.20747589e-01 2.55640805e-01 1.02090836e+00 3.30066055e-01
-1.34121001e+00 5.18103302e-01 6.28621864e+00 1.13707900e+00
-1.80721855e+00 -2.44357020e-01 1.51216114e+00 -5.76474667e-02
1.82290271e-01 -3.62515688e-01 -3.44725937e-01 2.01868340e-01
8.50810885e-01 2.50146091e-01 1.06421253e-02 6.66392028e-01
1.08052239e-01 -5.49180627e-01 -1.10439491e+00 7.12911963e-01
-3.03228587e-01 -1.42891598e+00 -3.17846239e-01 4.43720847e-01
3.39021385e-01 -2.30543911e-02 1.40331402e-01 1.81207687e-01
8.25913250e-02 -1.28033090e+00 2.95828938e-01 3.40405941e-01
1.37792134e+00 -7.61573136e-01 1.23773682e+00 -1.62892163e-01
-9.79587495e-01 7.37296045e-02 -2.78968185e-01 3.74004632e-01
-5.84671319e-01 6.85180426e-01 -1.43043649e+00 3.99723239e-02
8.29276145e-01 2.35970497e-01 -8.37681890e-01 1.13156295e+00
3.37142825e-01 6.52075171e-01 -1.14558905e-01 -2.08486646e-01
2.00134173e-01 2.16092274e-01 -9.18198749e-02 2.00864792e+00
4.75025535e-01 -2.03184381e-01 -6.34514093e-02 3.59032810e-01
5.77885024e-02 2.40505293e-01 -7.37693086e-02 9.67911780e-02
-8.87778848e-02 1.78140712e+00 -1.57100987e+00 -2.59903967e-01
-1.67868018e-01 1.02620006e+00 3.71214956e-01 5.72206192e-02
-5.65901756e-01 -3.46174181e-01 4.00138259e-01 1.39687166e-01
3.08418334e-01 1.09262757e-01 -5.94189107e-01 -5.96083403e-01
-3.04978549e-01 -7.14684367e-01 4.42746937e-01 -4.35991466e-01
-9.49595511e-01 7.36551046e-01 -6.05265260e-01 -1.21099281e+00
2.18819112e-01 -1.30539072e+00 -6.94081128e-01 7.95986593e-01
-1.62567055e+00 -1.43316293e+00 -4.35941964e-01 2.90226102e-01
1.79896995e-01 1.65406074e-02 1.28970063e+00 1.75363719e-01
-6.65139318e-01 7.21766531e-01 1.55790657e-01 3.65572244e-01
9.83708441e-01 -1.58781266e+00 1.47108272e-01 3.80541503e-01
-4.73976135e-01 3.48282099e-01 7.09442437e-01 -2.45690137e-01
-1.38539553e+00 -1.00752521e+00 4.67864990e-01 1.41295508e-01
7.75136769e-01 -3.82681042e-01 -6.05891645e-01 3.75465542e-01
3.62227947e-01 5.51752150e-01 1.24674702e+00 -1.40844792e-01
-1.23370647e-01 -2.12247401e-01 -1.37638736e+00 5.28031170e-01
2.17424065e-01 -4.48641419e-01 4.00232673e-01 4.97675598e-01
-5.30283749e-02 -6.54796600e-01 -9.73956466e-01 1.61487371e-01
7.01479733e-01 -8.81223142e-01 6.17910266e-01 -3.72296900e-01
2.85870045e-01 -3.71282309e-01 1.47273615e-01 -1.19593251e+00
-9.10943329e-01 -1.42578229e-01 5.66115856e-01 4.06017870e-01
4.55307066e-01 -4.11985338e-01 1.43545961e+00 2.20522240e-01
-3.00329123e-02 -8.27391982e-01 -8.74570072e-01 -1.64305627e-01
2.58287281e-01 4.31925096e-02 -1.31877158e-02 8.66283059e-01
2.01966077e-01 -2.37128168e-01 1.66398764e-01 4.57246900e-02
3.11641335e-01 -2.67062876e-02 5.64802408e-01 -6.71575785e-01
-5.18272161e-01 -7.76274383e-01 -7.62779713e-01 -3.10363322e-01
-3.67507875e-01 -8.72144997e-01 -1.17578790e-01 -1.24137723e+00
4.96488452e-01 -3.01668614e-01 -6.47879958e-01 7.97028959e-01
8.98573399e-02 8.33225787e-01 -3.33643973e-01 -9.91738960e-02
-4.88915354e-01 -3.74936134e-01 1.19117045e+00 -4.70337868e-01
2.46937707e-01 -3.22222084e-01 -6.12297535e-01 2.93915957e-01
1.01189315e+00 -4.30025071e-01 2.72257119e-01 -2.37712666e-01
1.39547393e-01 5.96939623e-02 4.40839916e-01 -1.00399542e+00
3.15177888e-01 -8.83528441e-02 1.02520537e+00 -2.73002356e-01
2.54665557e-02 -8.68438482e-01 2.58281529e-01 9.90720630e-01
-3.69492799e-01 -1.64072603e-01 7.00592160e-01 1.90218836e-01
-1.37033328e-01 9.05149132e-02 9.34001267e-01 -1.06163330e-01
-4.24943030e-01 5.13929836e-02 -8.02249312e-01 -8.50138068e-01
1.20738721e+00 -3.82797450e-01 -5.59035063e-01 2.22518623e-01
-7.13778257e-01 -1.00008927e-01 6.14362538e-01 -1.92828760e-01
3.84171277e-01 -1.04786634e+00 -9.94387746e-01 3.64603043e-01
4.31006432e-01 -2.62684107e-01 5.92218757e-01 1.18220365e+00
-1.30430377e+00 5.37359715e-01 -4.71399486e-01 -6.29691482e-01
-1.64152479e+00 4.42520440e-01 7.93550193e-01 -7.91825712e-01
-3.47411156e-01 1.08357930e+00 2.71729589e-01 -1.36447370e-01
-6.04969896e-02 -4.09699082e-01 -1.31373212e-01 -3.64833534e-01
5.48530936e-01 4.04638089e-02 7.51141310e-01 -1.65254980e-01
-4.84681606e-01 2.72994399e-01 -5.85938275e-01 1.95624158e-01
1.11749506e+00 3.03700894e-01 -2.72886366e-01 1.25680894e-01
1.55958927e+00 -8.65034163e-02 -1.03091192e+00 1.05273947e-01
-1.47884294e-01 -1.65828720e-01 5.45640647e-01 -1.03529453e+00
-1.20691276e+00 7.72461653e-01 1.19542694e+00 1.47099242e-01
1.29584348e+00 -2.23299280e-01 2.30903417e-01 3.23244035e-01
4.36899178e-02 -9.28804278e-01 -3.25717568e-01 1.13997877e-01
6.51658595e-01 -1.47821283e+00 1.31991506e-01 -1.74898401e-01
-2.17962220e-01 1.68309844e+00 4.46276456e-01 -3.08178872e-01
5.94997883e-01 8.62568080e-01 6.20335162e-01 -1.94135249e-01
-9.55356836e-01 1.37078315e-02 2.62707889e-01 4.56097364e-01
9.72194016e-01 4.56808507e-01 -1.24015354e-01 1.01981781e-01
5.54567836e-02 6.31959885e-02 2.74555117e-01 1.08307719e+00
-4.07497644e-01 -1.03582764e+00 -4.48617727e-01 7.01509595e-01
-8.87681365e-01 -1.30468264e-01 -4.38989222e-01 7.51342118e-01
-1.18489817e-01 4.97362375e-01 3.77536088e-01 -2.94696510e-01
-1.61792427e-01 -4.26676244e-01 6.15421295e-01 -2.21638113e-01
-8.24893773e-01 5.41838527e-01 1.15437351e-01 -2.60661036e-01
-1.82411402e-01 -3.28612208e-01 -1.07616615e+00 -4.03603375e-01
3.89452055e-02 -8.36790279e-02 9.44310784e-01 5.70785284e-01
1.47529542e-01 9.01404977e-01 3.25274050e-01 -7.06898093e-01
1.67129487e-02 -1.05944407e+00 -6.13871217e-01 5.04417084e-02
5.28951287e-01 1.23337079e-02 -1.03598043e-01 2.18171671e-01] | [15.141414642333984, -2.9891250133514404] |
ab685a6a-d746-4844-adc7-b13f27f3d043 | q-shed-distributed-optimization-at-the-edge | 2305.10852 | null | https://arxiv.org/abs/2305.10852v1 | https://arxiv.org/pdf/2305.10852v1.pdf | Q-SHED: Distributed Optimization at the Edge via Hessian Eigenvectors Quantization | Edge networks call for communication efficient (low overhead) and robust distributed optimization (DO) algorithms. These are, in fact, desirable qualities for DO frameworks, such as federated edge learning techniques, in the presence of data and system heterogeneity, and in scenarios where internode communication is the main bottleneck. Although computationally demanding, Newton-type (NT) methods have been recently advocated as enablers of robust convergence rates in challenging DO problems where edge devices have sufficient computational power. Along these lines, in this work we propose Q-SHED, an original NT algorithm for DO featuring a novel bit-allocation scheme based on incremental Hessian eigenvectors quantization. The proposed technique is integrated with the recent SHED algorithm, from which it inherits appealing features like the small number of required Hessian computations, while being bandwidth-versatile at a bit-resolution level. Our empirical evaluation against competing approaches shows that Q-SHED can reduce by up to 60% the number of communication rounds required for convergence. | ['Subhrakanti Dey', 'Luca Schenato', 'Michele Rossi', 'Nicolò Dal Fabbro'] | 2023-05-18 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-2.88805604e-01 9.78830084e-02 -4.70678300e-01 8.81327614e-02
-6.35905266e-01 -3.52559537e-01 3.37683707e-01 1.65938482e-01
-2.77370691e-01 9.67176855e-01 1.00812934e-01 -4.90758121e-01
-7.57170022e-01 -6.39546394e-01 -4.14744794e-01 -9.30261672e-01
-3.04553062e-01 2.91271240e-01 -1.34895325e-01 -3.25615443e-02
4.54591475e-02 6.16613984e-01 -1.22691286e+00 -5.06443262e-01
8.94382179e-01 1.52244937e+00 -1.65547967e-01 7.35082984e-01
-9.31043625e-02 7.98832953e-01 -2.41377190e-01 -6.15375817e-01
5.68864584e-01 -4.48717237e-01 -5.08010805e-01 1.34948939e-01
9.04452428e-02 -3.07158828e-01 -3.10061306e-01 8.92461300e-01
8.43598962e-01 1.97616499e-02 4.05965030e-01 -1.30046558e+00
3.14161777e-02 6.68929875e-01 -5.56045055e-01 3.33898813e-01
2.93194279e-02 6.51042461e-02 1.18006277e+00 -7.29861677e-01
7.26500034e-01 5.71471989e-01 8.97538185e-01 3.15512002e-01
-1.39359581e+00 -4.68372881e-01 -9.64009687e-02 2.66258627e-01
-1.48545694e+00 -9.18512404e-01 8.56547236e-01 4.22679633e-02
7.55574882e-01 4.35762912e-01 7.56063640e-01 3.85655761e-01
2.02772275e-01 8.24663579e-01 8.83547723e-01 -3.65774661e-01
7.55952775e-01 2.86075383e-01 -4.12836671e-01 7.06406474e-01
5.96937358e-01 -6.01396039e-02 -9.77640510e-01 -2.33754620e-01
5.20193458e-01 -1.63975000e-01 -2.95898914e-01 -7.27716267e-01
-8.33905578e-01 7.23737121e-01 4.30634409e-01 2.12059975e-01
-8.39135170e-01 4.88496929e-01 6.63166761e-01 5.68892896e-01
6.40291274e-01 -8.57669488e-02 -4.35104519e-01 -5.89316785e-01
-1.34177411e+00 -1.71971530e-01 1.09053075e+00 8.89978528e-01
7.17797160e-01 1.90193921e-01 -4.26335521e-02 1.84203133e-01
2.94206768e-01 3.50934684e-01 -9.33625326e-02 -1.05865085e+00
3.82418960e-01 4.03643250e-01 6.62406385e-02 -1.05324483e+00
-5.40538371e-01 -1.01535630e+00 -1.29094136e+00 1.10378064e-01
1.60915807e-01 -6.31188393e-01 -4.78244759e-02 1.50355685e+00
8.13541114e-01 1.17090344e-01 2.61696894e-02 9.79152381e-01
3.13337982e-01 3.92519027e-01 -3.06525290e-01 -6.22361541e-01
8.35057020e-01 -8.05864155e-01 -7.42171824e-01 4.20614243e-01
9.05565858e-01 -5.56436956e-01 3.67343396e-01 5.26501536e-01
-1.33227384e+00 -8.85853171e-02 -8.63271832e-01 2.62107939e-01
2.98989881e-02 7.51138572e-03 1.06874704e+00 1.55061686e+00
-1.57165170e+00 5.77466667e-01 -7.61303186e-01 -3.30134720e-01
6.62062764e-01 7.18710840e-01 5.41215427e-02 1.02141820e-01
-4.65027869e-01 2.00940028e-01 4.11160011e-03 1.07417561e-01
-5.82975984e-01 -8.69764030e-01 -2.62699127e-01 4.56559032e-01
5.44937193e-01 -9.93887782e-01 8.25913012e-01 -8.02287638e-01
-1.92143691e+00 3.00476104e-01 -3.48122814e-03 -7.73604393e-01
8.44481945e-01 -1.44086117e-02 -1.95292711e-01 3.30664188e-01
-2.83597112e-01 2.41037667e-01 8.25643182e-01 -7.32780874e-01
-7.16583550e-01 -3.91159683e-01 1.90477949e-02 1.28665820e-01
-7.70887196e-01 -3.40515912e-01 -3.02488089e-01 -3.41252804e-01
-2.31257780e-03 -7.77057230e-01 -5.50878465e-01 3.72814298e-01
-3.38573188e-01 -1.66230261e-01 7.46859670e-01 -1.73523903e-01
1.35449839e+00 -2.04132056e+00 1.33067533e-01 6.27874076e-01
7.31947899e-01 2.44857017e-02 3.17621410e-01 6.96505666e-01
6.90256476e-01 4.21601795e-02 3.13620359e-01 -4.12714988e-01
1.97473258e-01 -4.01576087e-02 2.17824697e-01 7.89410949e-01
-4.19410616e-01 7.25798965e-01 -7.43260980e-01 -4.21844333e-01
8.40302110e-02 4.32816416e-01 -9.14876103e-01 -1.69430241e-01
1.35123178e-01 3.66353482e-01 -5.73476136e-01 5.68555534e-01
5.14896333e-01 -5.64265251e-01 4.17121232e-01 -2.90729433e-01
-6.52424395e-02 2.54456662e-02 -1.59618568e+00 1.63624215e+00
-7.05222011e-01 6.66414082e-01 7.99257636e-01 -1.09634054e+00
7.96275496e-01 5.33533514e-01 1.06402171e+00 -6.04877651e-01
2.06123278e-01 4.76318479e-01 -2.64466256e-01 -9.23501998e-02
4.77226019e-01 1.33766115e-01 2.85533935e-01 5.51483095e-01
-1.63607553e-01 2.39057422e-01 2.27000788e-01 4.05723482e-01
1.34007633e+00 -2.65025705e-01 9.67900231e-02 -5.91290951e-01
4.42095280e-01 -2.75620341e-01 6.55430555e-01 7.02293813e-01
-3.53377521e-01 -1.95215687e-01 3.98673773e-01 -3.34684581e-01
-1.05369735e+00 -7.87034631e-01 -6.52211308e-02 7.51331091e-01
3.23946923e-01 -6.82037175e-01 -5.21730483e-01 -4.33436573e-01
2.60201916e-02 1.75370112e-01 -2.44947542e-02 3.86711918e-02
-3.31838876e-02 -7.27570355e-01 3.48599643e-01 2.26553515e-01
6.05179667e-01 -2.51259089e-01 -8.17772686e-01 5.57723701e-01
3.12922597e-01 -1.00028896e+00 -3.81091207e-01 2.61213273e-01
-1.01326895e+00 -7.05472052e-01 -6.39691651e-01 -3.72200549e-01
5.12198746e-01 2.90760726e-01 9.68789577e-01 -8.67706537e-03
-2.77349174e-01 7.34182179e-01 -2.77276367e-01 1.73938461e-02
6.35076761e-02 3.02121907e-01 2.48478755e-01 3.77215862e-01
-9.27793514e-03 -8.62024963e-01 -9.08977747e-01 1.51803866e-01
-6.91021502e-01 -4.40642498e-02 5.06358027e-01 9.61344421e-01
5.11021376e-01 2.84047425e-01 8.62896025e-01 -9.14418221e-01
5.81955552e-01 -5.86748481e-01 -6.91743970e-01 -5.60815185e-02
-1.08215296e+00 -4.39727642e-02 9.01366711e-01 1.64406389e-01
-8.75400245e-01 -8.31503514e-03 1.01869896e-01 -3.01833749e-01
3.96281719e-01 4.38790262e-01 1.16970859e-01 -6.72607005e-01
6.62376583e-01 1.28274456e-01 9.06943064e-03 -1.90918818e-01
5.67060411e-01 8.00807238e-01 3.60452570e-02 -4.58954781e-01
6.92589402e-01 8.87769103e-01 5.57337761e-01 -8.86846423e-01
-3.57402503e-01 -6.00686848e-01 -1.97773233e-01 -4.01514709e-01
1.45668134e-01 -1.08038116e+00 -1.15980983e+00 -7.55512193e-02
-6.88327372e-01 -2.62990803e-01 -5.36824286e-01 5.45563817e-01
-6.96369886e-01 3.22066545e-01 -5.73406398e-01 -1.13478458e+00
-7.05778480e-01 -8.01452875e-01 7.12717891e-01 4.28084314e-01
1.29675299e-01 -1.18364394e+00 -2.80302674e-01 2.15613291e-01
1.03204155e+00 1.70551807e-01 3.78077745e-01 -2.91173726e-01
-8.73040855e-01 -1.03126362e-01 -2.77427077e-01 -1.78926200e-01
8.25537592e-02 -1.84699893e-01 -5.92473924e-01 -7.95636714e-01
-2.50103414e-01 -9.73467305e-02 4.30361152e-01 4.76345599e-01
9.12926018e-01 -3.97414654e-01 -3.18402916e-01 9.25066590e-01
1.75559127e+00 -3.07895243e-01 1.66674718e-01 -5.28436489e-02
3.76478553e-01 1.89438358e-01 2.21498489e-01 1.17180014e+00
4.70948517e-01 6.30942881e-01 4.80388254e-01 -2.79983938e-01
-1.46196947e-01 1.80356398e-01 3.14787716e-01 1.07706881e+00
-1.88715719e-02 -3.58254254e-01 -4.93300289e-01 6.85093164e-01
-1.81285870e+00 -8.57703686e-01 -1.53405711e-01 2.24565101e+00
5.16489267e-01 2.90177226e-01 3.79572570e-01 4.05693531e-01
4.10360932e-01 9.55372751e-02 -7.93480456e-01 -5.60999513e-01
-5.27344681e-02 4.72871996e-02 9.19256270e-01 1.53958634e-01
-6.45619690e-01 4.57737744e-01 5.70791245e+00 9.64279711e-01
-1.10194111e+00 2.87258685e-01 6.79462373e-01 -3.59257907e-01
-4.37698066e-01 -2.65407786e-02 -6.41627312e-01 4.75728124e-01
1.35526037e+00 -5.15082359e-01 5.29792964e-01 9.10390019e-01
5.02574384e-01 -1.77282885e-01 -8.04934919e-01 1.41117084e+00
-2.34115854e-01 -1.74496758e+00 -5.17664909e-01 3.93666744e-01
1.01542330e+00 2.85800576e-01 -8.41065049e-02 -1.33257151e-01
1.84300378e-01 -6.47522449e-01 4.35536206e-01 3.25804234e-01
7.48547852e-01 -9.31654632e-01 6.96818829e-01 1.89389780e-01
-1.31181490e+00 -3.15800130e-01 -2.03783587e-01 -7.21685737e-02
3.93075407e-01 1.38052368e+00 -2.79020280e-01 8.21744740e-01
7.42503524e-01 6.19771481e-01 -1.27859404e-02 1.35435534e+00
2.59401351e-01 7.22881734e-01 -8.58568549e-01 -3.71289968e-01
2.54679829e-01 -3.94784570e-01 8.05792212e-01 9.46930468e-01
3.92272860e-01 -2.22558156e-01 1.82392135e-01 5.10576606e-01
-3.28252643e-01 4.46579069e-01 -4.75718349e-01 2.41246913e-02
7.36436725e-01 1.44057453e+00 -8.27639699e-01 -2.47347459e-01
-5.18243372e-01 9.26087260e-01 2.26174891e-02 2.82404661e-01
-5.18038869e-01 -4.75038677e-01 6.12344265e-01 -6.11928031e-02
5.13314188e-01 -4.31079775e-01 -4.56877142e-01 -8.33399773e-01
2.89516598e-01 -5.01063883e-01 2.86115527e-01 -1.62484914e-01
-1.06272125e+00 2.84243137e-01 -6.90914869e-01 -1.08716094e+00
-1.96537778e-01 -2.44042650e-01 -4.15030330e-01 3.52024704e-01
-1.85036564e+00 -9.11560476e-01 -2.12387130e-01 7.07798958e-01
3.14600527e-01 -1.38379276e-01 5.48213780e-01 7.63440490e-01
-5.99556029e-01 7.94227123e-01 7.77597308e-01 -4.37809527e-01
2.35295400e-01 -1.13493431e+00 -1.63012311e-01 6.79010272e-01
3.12843978e-01 2.57869691e-01 7.74668157e-01 -3.08840156e-01
-2.20341325e+00 -8.31982136e-01 5.31112373e-01 3.61913323e-01
7.30263531e-01 -3.10263485e-01 -1.81641385e-01 5.42548150e-02
1.12015903e-01 4.68626291e-01 8.94774616e-01 1.51990294e-01
1.66512296e-01 -6.23113453e-01 -1.16096592e+00 6.61372125e-01
1.06128371e+00 -4.89890844e-01 5.57500422e-01 5.51294923e-01
2.34456122e-01 -2.89507836e-01 -1.11553895e+00 -1.44865081e-01
3.84315282e-01 -1.29889059e+00 5.41225851e-01 -1.32366776e-01
-2.39279598e-01 -2.44767517e-01 -1.86696827e-01 -1.09959638e+00
-1.22680686e-01 -1.73517823e+00 -6.71773970e-01 1.21153700e+00
3.04365873e-01 -6.17122233e-01 1.44916725e+00 2.97890753e-01
2.16941647e-02 -9.02775228e-01 -1.40259349e+00 -1.01240027e+00
-4.26193237e-01 -4.11933929e-01 4.00166154e-01 5.76180816e-01
2.71222830e-01 2.07961723e-01 -4.29008782e-01 7.72721320e-02
7.90966272e-01 -1.51408568e-03 9.64969456e-01 -1.32536530e+00
-5.23602068e-01 -4.37175393e-01 -6.55810237e-01 -1.34047556e+00
-3.20203364e-01 -8.67206633e-01 -4.63859856e-01 -1.27023304e+00
-1.84772000e-01 -8.77902448e-01 -2.46215165e-01 6.66075274e-02
2.68869102e-01 2.22571671e-01 2.09007412e-01 2.78017163e-01
-9.53419328e-01 7.94532359e-01 7.90171087e-01 8.27290192e-02
-4.22305405e-01 2.05482796e-01 -5.14836550e-01 2.25625336e-01
7.48152494e-01 -3.87421399e-01 -5.02172410e-01 -2.36680239e-01
6.67570531e-01 3.02047759e-01 -3.81569844e-03 -1.18991792e+00
8.76661003e-01 1.64412096e-01 -5.31957000e-02 -2.24789366e-01
3.13099056e-01 -1.34421253e+00 3.75360340e-01 7.17635632e-01
1.98300347e-01 7.75697976e-02 -6.97279051e-02 7.94019759e-01
-2.03152690e-02 2.72011936e-01 4.68428046e-01 3.75257522e-01
-5.79233587e-01 5.10022104e-01 -2.64021009e-01 -6.31973520e-02
1.24048555e+00 -4.97924894e-01 -7.37561509e-02 -6.83461726e-01
-4.78889048e-01 5.13918579e-01 2.63310879e-01 -2.96161860e-01
1.69625878e-01 -1.05624366e+00 -6.21067822e-01 -1.52932614e-01
-3.43141437e-01 -2.78037786e-01 3.70439619e-01 1.62597907e+00
-5.13060927e-01 5.58230937e-01 5.55069000e-02 -6.59269691e-01
-8.65658641e-01 3.26764792e-01 2.86403924e-01 -2.75457799e-01
-8.03378642e-01 8.00919354e-01 -6.16179526e-01 -4.18538265e-02
3.21771771e-01 2.70828038e-01 4.86597300e-01 7.97223579e-03
2.57984757e-01 8.30674469e-01 3.65963846e-01 -1.15842506e-01
-2.29853883e-01 3.48412275e-01 1.65093228e-01 1.29693434e-01
1.43038285e+00 -7.55994678e-01 -1.79910678e-02 -5.10487743e-02
9.69272554e-01 4.10846680e-01 -1.46078932e+00 -3.29410553e-01
1.85688972e-01 -3.27713251e-01 8.13666463e-01 -3.84331882e-01
-1.55889022e+00 4.02520388e-01 7.98409760e-01 5.23559153e-01
1.42012870e+00 -2.95593590e-01 9.57501769e-01 2.66949981e-01
8.92674923e-01 -1.17760909e+00 -2.73445755e-01 -1.03306256e-01
-5.01610599e-02 -8.97508442e-01 1.31906718e-01 -5.36426723e-01
-5.68139516e-02 1.11101878e+00 3.10635865e-02 6.75972849e-02
8.07501674e-01 4.63392764e-01 -1.98002070e-01 -2.61830151e-01
-9.00837958e-01 -2.87734896e-01 -2.82553583e-01 4.70332950e-01
1.37619436e-01 4.65905946e-03 -7.23916829e-01 1.77901983e-01
5.98908626e-02 1.15680315e-01 5.33750534e-01 8.49357605e-01
-2.77276337e-01 -1.03188157e+00 -6.93011051e-03 6.02455854e-01
-5.52834809e-01 -5.50888851e-02 9.68092680e-02 4.26477700e-01
-1.58669069e-01 1.07010508e+00 3.85739654e-02 -1.59032643e-01
-1.18657120e-01 -3.93187851e-01 7.49616250e-02 -9.70796049e-02
-6.97412133e-01 -2.82107089e-02 1.66947097e-01 -8.96206856e-01
-5.18304467e-01 -4.73932028e-01 -1.01668155e+00 -6.89140618e-01
-6.09651744e-01 4.55077261e-01 9.60868061e-01 8.64147604e-01
9.73378301e-01 3.56835514e-01 1.08967423e+00 -5.92321932e-01
-5.95901668e-01 -2.90985644e-01 -9.38092053e-01 -1.88822508e-01
3.28646123e-01 -2.24577457e-01 -5.25731444e-01 -3.46623242e-01] | [6.186124324798584, 5.065540790557861] |
d80726a6-a488-41d1-9fa2-28812eccb022 | deep-masking-generative-network-a-unified | 2010.04324 | null | https://arxiv.org/abs/2010.04324v2 | https://arxiv.org/pdf/2010.04324v2.pdf | Deep-Masking Generative Network: A Unified Framework for Background Restoration from Superimposed Images | Restoring the clean background from the superimposed images containing a noisy layer is the common crux of a classical category of tasks on image restoration such as image reflection removal, image deraining and image dehazing. These tasks are typically formulated and tackled individually due to the diverse and complicated appearance patterns of noise layers within the image. In this work we present the Deep-Masking Generative Network (DMGN), which is a unified framework for background restoration from the superimposed images and is able to cope with different types of noise. Our proposed DMGN follows a coarse-to-fine generative process: a coarse background image and a noise image are first generated in parallel, then the noise image is further leveraged to refine the background image to achieve a higher-quality background image. In particular, we design the novel Residual Deep-Masking Cell as the core operating unit for our DMGN to enhance the effective information and suppress the negative information during image generation via learning a gating mask to control the information flow. By iteratively employing this Residual Deep-Masking Cell, our proposed DMGN is able to generate both high-quality background image and noisy image progressively. Furthermore, we propose a two-pronged strategy to effectively leverage the generated noise image as contrasting cues to facilitate the refinement of the background image. Extensive experiments across three typical tasks for image background restoration, including image reflection removal, image rain steak removal and image dehazing, show that our DMGN consistently outperforms state-of-the-art methods specifically designed for each single task. | ['Fanglin Chen', 'Guangming Lu', 'David Zhang', 'Zihui Jia', 'Wenjie Pei', 'Xin Feng'] | 2020-10-09 | null | null | null | null | ['reflection-removal'] | ['computer-vision'] | [ 9.80995357e-01 -4.20304596e-01 4.64068919e-01 -6.78151771e-02
-6.00817919e-01 -2.29066595e-01 5.92471659e-01 -5.44510663e-01
-1.16291717e-01 6.80020392e-01 1.57287866e-01 -3.15166265e-01
1.60122424e-01 -9.28188384e-01 -7.46590257e-01 -1.42747509e+00
4.25766677e-01 -3.29713047e-01 2.50201523e-01 -4.80909228e-01
8.94001797e-02 4.67952371e-01 -1.72645330e+00 3.81149232e-01
1.19773972e+00 8.79529297e-01 7.08913863e-01 8.17985713e-01
-1.58919588e-01 1.02896500e+00 -7.74710536e-01 -9.09660459e-02
3.59552264e-01 -7.83671379e-01 -1.35355800e-01 5.47984362e-01
4.43002760e-01 -2.51456499e-01 -5.60196519e-01 1.42167318e+00
7.59677708e-01 2.35670015e-01 2.75777340e-01 -7.99645722e-01
-7.55890548e-01 2.94872463e-01 -7.91757703e-01 4.92946833e-01
9.79290158e-02 5.18305361e-01 2.87775517e-01 -7.78282225e-01
3.23680580e-01 1.20145619e+00 3.91199082e-01 8.54577184e-01
-1.50643051e+00 -6.33557439e-01 2.84384489e-01 -1.02834202e-01
-1.28919435e+00 -5.62604725e-01 8.59518647e-01 -2.54301727e-01
2.97338217e-01 5.95101595e-01 5.01274407e-01 9.81661737e-01
2.85536461e-02 5.20210266e-01 1.35456157e+00 -3.75961244e-01
8.28165859e-02 -1.68431923e-01 -9.47097242e-02 3.82019907e-01
4.05171573e-01 1.71135768e-01 -3.40921253e-01 1.80750564e-01
9.11790609e-01 1.35019928e-01 -8.45025003e-01 3.07519794e-01
-8.18003476e-01 3.99371058e-01 5.10202765e-01 3.32399666e-01
-5.45242906e-01 9.24437791e-02 -1.72438726e-01 1.34434521e-01
5.12246370e-01 1.60539940e-01 -5.66413738e-02 4.23832625e-01
-1.06111610e+00 3.38596493e-01 6.19015872e-01 5.48614681e-01
8.47893834e-01 4.20543075e-01 -5.90792894e-01 8.61248732e-01
1.55288294e-01 6.50446296e-01 1.72910094e-01 -7.61641920e-01
3.32205325e-01 5.14572561e-01 2.53283054e-01 -1.05321991e+00
-8.28119218e-02 -6.20709181e-01 -1.55107915e+00 5.37453413e-01
7.34884813e-02 -1.06830239e-01 -1.40114403e+00 1.62775731e+00
3.48518610e-01 4.48626071e-01 1.49179280e-01 1.03798473e+00
1.00233281e+00 8.07808340e-01 -8.80171806e-02 -3.28259170e-01
1.25999713e+00 -9.12242532e-01 -9.44282711e-01 -6.24526203e-01
-2.71949470e-01 -8.28855038e-01 9.02430058e-01 4.59542304e-01
-1.22137427e+00 -9.15007949e-01 -1.00848067e+00 -9.05200988e-02
4.76241522e-02 1.55960843e-01 1.61004856e-01 7.35867321e-01
-1.16826189e+00 4.79269236e-01 -6.36068285e-01 5.11643663e-02
6.42477810e-01 1.08249836e-01 -1.77019909e-01 -5.02915442e-01
-9.79641616e-01 6.66227996e-01 2.10890070e-01 7.48393059e-01
-1.24360764e+00 -5.24260163e-01 -8.45986605e-01 -2.15583742e-02
3.99165899e-01 -9.47257400e-01 5.18957436e-01 -1.03049636e+00
-1.56125164e+00 9.28711534e-01 -2.45710120e-01 -2.75016189e-01
5.15045285e-01 -1.91269904e-01 -5.25042474e-01 3.50327580e-03
-4.27793339e-02 1.48561016e-01 1.38370860e+00 -1.86304522e+00
-4.78215009e-01 -1.25178888e-01 -1.70368865e-01 3.71969491e-01
2.20240459e-01 -8.61893967e-02 -7.87641525e-01 -1.03258955e+00
1.29640907e-01 -4.43702996e-01 -5.55865586e-01 -3.88977885e-01
-5.29105723e-01 6.32081091e-01 8.50065768e-01 -8.97855580e-01
1.34975636e+00 -2.21508551e+00 3.27107519e-01 1.46029042e-02
4.92347211e-01 5.79051316e-01 -3.27683628e-01 -1.58772111e-01
-1.05005339e-01 9.40809920e-02 -6.85837209e-01 -3.70785415e-01
-3.35583389e-01 2.84203976e-01 -4.08216178e-01 4.82214928e-01
5.24600267e-01 8.13904881e-01 -8.29671621e-01 1.41328964e-02
3.76480997e-01 8.57819974e-01 -3.91160488e-01 4.98241484e-01
-2.12697655e-01 9.59752262e-01 -8.34956467e-02 5.59246719e-01
1.29689050e+00 -1.72472775e-01 8.85040388e-02 -3.86743337e-01
-1.04544677e-01 -1.02644414e-01 -1.16545129e+00 1.26943517e+00
-4.21662182e-01 4.30756718e-01 7.40285814e-01 -7.90547431e-01
9.75143433e-01 -6.31019026e-02 -4.57506999e-02 -9.11595047e-01
1.31038696e-01 1.38894305e-01 -4.85549355e-03 -4.82319921e-01
4.29738820e-01 -3.38419765e-01 2.99159348e-01 1.38268232e-01
-1.29807830e-01 -2.02106446e-01 1.82845965e-01 1.58683024e-02
1.10940456e+00 7.97505584e-03 -5.51449433e-02 -8.69517177e-02
9.63569164e-01 -3.77366602e-01 6.34764433e-01 1.08621323e+00
4.76015247e-02 1.11323988e+00 1.77529395e-01 -2.72429973e-01
-8.71148944e-01 -9.70253289e-01 1.02840669e-01 8.31896007e-01
4.99771923e-01 -8.01010355e-02 -9.09350097e-01 -3.57857704e-01
-4.29625869e-01 5.39037168e-01 -5.49061239e-01 -9.36044008e-02
-8.19107234e-01 -1.37996531e+00 3.91861707e-01 -9.08528827e-03
9.63001192e-01 -1.31099856e+00 -2.09240004e-01 1.73516199e-01
-5.08084595e-01 -1.08302224e+00 -3.78192991e-01 3.01802605e-01
-3.85200202e-01 -1.03620350e+00 -6.69275939e-01 -7.57923961e-01
8.76007497e-01 6.67496741e-01 1.25586021e+00 5.72351098e-01
-5.19623339e-01 -1.85325947e-02 -1.48134604e-01 1.73083767e-02
-6.32026255e-01 -4.76157099e-01 -3.35577935e-01 5.43632567e-01
-1.50376379e-01 -8.13587189e-01 -9.10066962e-01 2.09057003e-01
-1.57227886e+00 2.38142863e-01 7.32092202e-01 1.05012202e+00
7.26201952e-01 8.23183835e-01 1.01371273e-01 -8.81249785e-01
6.26145542e-01 -3.17824006e-01 -7.54491210e-01 7.25619271e-02
6.30468735e-03 -1.58492118e-01 6.41525447e-01 -5.18271863e-01
-1.54157150e+00 -2.18485549e-01 -1.83327392e-01 -3.69530976e-01
-2.24633276e-01 1.00863330e-01 -7.82986343e-01 -2.66991049e-01
5.73721409e-01 7.27423668e-01 -1.34869188e-01 -5.64253986e-01
4.46175814e-01 2.79175252e-01 1.04478550e+00 -3.73216540e-01
1.22365594e+00 7.81897306e-01 1.68731697e-02 -9.27144229e-01
-8.13166142e-01 -2.89058894e-01 -3.26668471e-01 -1.08911529e-01
9.94031012e-01 -1.11520398e+00 -6.02856338e-01 1.14412189e+00
-1.01859140e+00 -8.43102813e-01 -1.85471907e-01 -1.90264717e-01
-1.83642998e-01 4.20906514e-01 -5.77168584e-01 -9.38199818e-01
-3.93704921e-01 -1.22125483e+00 1.08048213e+00 5.84375083e-01
5.95954835e-01 -8.19218338e-01 -2.46227369e-01 3.27844828e-01
8.34860444e-01 3.35016817e-01 8.38941693e-01 1.58719048e-01
-9.70703244e-01 2.12128922e-01 -5.18501937e-01 7.80397415e-01
4.37753260e-01 -2.96426684e-01 -1.14441657e+00 -3.11464459e-01
5.35461843e-01 9.96816456e-02 1.44392896e+00 5.51925480e-01
1.12344468e+00 -3.52564365e-01 -1.15411833e-01 9.87293184e-01
1.62947059e+00 9.92137492e-02 1.32826877e+00 4.24342066e-01
8.60130191e-01 3.29443395e-01 2.74406224e-01 2.25738108e-01
-1.17209069e-02 4.83870953e-01 5.02827585e-01 -7.89204061e-01
-7.00251937e-01 1.19886197e-01 4.12780911e-01 5.35277605e-01
1.17533822e-02 -5.53669691e-01 -5.20166099e-01 3.37090164e-01
-1.66054368e+00 -1.15753448e+00 -1.63997099e-01 2.10139418e+00
9.56386447e-01 6.91931741e-03 -4.31802243e-01 6.59237895e-03
9.78583574e-01 4.77643490e-01 -4.08763885e-01 2.79202163e-01
-7.05160737e-01 3.95021588e-01 3.96694869e-01 6.83339298e-01
-1.23986435e+00 9.68180358e-01 5.88323259e+00 9.17921960e-01
-1.15080762e+00 -6.99169710e-02 9.57044542e-01 7.48355389e-02
-3.05782288e-01 -1.73024908e-01 -8.11029255e-01 7.83763945e-01
3.59117329e-01 2.15507850e-01 7.94355392e-01 2.05139473e-01
4.19598669e-01 -3.52917403e-01 -3.73838425e-01 1.01529384e+00
1.75163433e-01 -1.29208577e+00 2.91872293e-01 -1.55645192e-01
9.48705077e-01 -2.36465111e-01 1.89867318e-01 5.03927097e-02
3.58774126e-01 -1.23239374e+00 5.01699448e-01 8.26434374e-01
7.06765234e-01 -5.64197063e-01 5.74496567e-01 2.65837193e-01
-9.56244349e-01 -5.42048290e-02 -3.14877361e-01 1.65496796e-01
2.72557676e-01 9.51682270e-01 -1.12383530e-01 7.61989892e-01
8.21341038e-01 4.86181378e-01 -5.30316114e-01 9.55908835e-01
-5.45654714e-01 5.42197168e-01 -2.80117802e-02 8.23618650e-01
-8.58253539e-02 -5.97003877e-01 7.89138377e-01 1.24542332e+00
2.65923351e-01 2.91659623e-01 2.00799391e-01 1.19426763e+00
-1.09170668e-01 -5.42299569e-01 -2.56093919e-01 2.74300218e-01
4.86833118e-02 1.46326733e+00 -7.47157991e-01 -3.58698428e-01
-2.70662904e-01 1.28702641e+00 -9.57222134e-02 7.85432160e-01
-8.36599946e-01 -2.89547682e-01 9.46676254e-01 2.20573097e-01
6.35630846e-01 1.12337314e-01 -1.76869631e-01 -1.32630420e+00
1.78402692e-01 -1.26498318e+00 1.03429621e-02 -9.49499965e-01
-1.42612886e+00 9.53360021e-01 -4.25116599e-01 -1.00890326e+00
4.03859615e-01 -4.23962146e-01 -8.77753019e-01 1.35734177e+00
-2.05516434e+00 -1.00465250e+00 -9.65614617e-01 7.78257966e-01
4.11805242e-01 9.77561474e-02 3.78974468e-01 5.61791480e-01
-8.64751637e-01 5.66893071e-02 6.77189380e-02 4.97691371e-02
5.37715197e-01 -1.16768909e+00 5.68168998e-01 1.61403120e+00
-3.74172747e-01 6.13241553e-01 7.26351202e-01 -9.69355524e-01
-1.22940934e+00 -1.48770535e+00 9.82800797e-02 -6.94915876e-02
3.69683921e-01 -4.30981517e-01 -1.17046642e+00 2.45675370e-01
1.22294441e-01 2.53448337e-01 2.64005035e-01 -6.54244006e-01
-1.48669183e-01 -2.53933311e-01 -1.03977060e+00 7.84099698e-01
1.03631449e+00 -2.79455632e-01 -1.52645752e-01 7.63828233e-02
7.78959572e-01 -6.50651991e-01 -1.36002034e-01 4.08402801e-01
-5.60220256e-02 -1.19949615e+00 1.14405346e+00 -5.73303439e-02
4.47352409e-01 -1.01600742e+00 -2.40750998e-01 -1.23108029e+00
-4.75916028e-01 -1.14519358e+00 -3.22164316e-03 1.41094983e+00
1.50139518e-02 -4.77755994e-01 5.20257056e-01 1.69071898e-01
-1.13741964e-01 -2.43669137e-01 -5.66960692e-01 -3.37700516e-01
-3.69524419e-01 -2.07489341e-01 5.16757309e-01 6.64265752e-01
-1.10518706e+00 1.47496849e-01 -7.23408759e-01 5.41091681e-01
9.19164062e-01 1.09270215e-03 9.39038634e-01 -7.73108065e-01
-3.34129214e-01 -3.57483119e-01 -2.11306676e-01 -1.12695026e+00
-2.18199551e-01 -6.25990629e-01 3.65866065e-01 -1.63602912e+00
1.80988312e-01 -1.56981438e-01 -2.47625232e-01 1.86891228e-01
-8.92922819e-01 7.01244593e-01 2.16932967e-01 1.29056737e-01
-4.69911098e-01 6.34540141e-01 1.40678048e+00 -1.56928197e-01
-3.11453074e-01 7.10687041e-02 -1.14186871e+00 6.73320115e-01
5.04020810e-01 -3.44338089e-01 -1.37243941e-01 -4.19995159e-01
5.40537434e-03 -1.83800921e-01 7.55621195e-01 -8.93635511e-01
5.54130040e-02 -1.18570678e-01 6.94269955e-01 -4.21065599e-01
2.27353528e-01 -6.38194025e-01 3.14657003e-01 1.67734176e-01
1.94520459e-01 -3.83435726e-01 3.82603943e-01 7.79137731e-01
-3.86248797e-01 2.09183082e-01 1.14729035e+00 -2.87447751e-01
-6.43575311e-01 2.11030096e-01 -4.99008834e-01 -6.16252981e-03
5.96772254e-01 -1.69107243e-01 -5.18188894e-01 -3.60767126e-01
-7.59159923e-01 -2.65669879e-02 5.11716843e-01 1.97301760e-01
6.98889136e-01 -1.14287639e+00 -9.07906890e-01 5.08571625e-01
-3.65864396e-01 2.30550200e-01 7.84233510e-01 9.06940043e-01
-4.99549568e-01 -2.84990847e-01 -9.08661038e-02 -5.01033902e-01
-1.24533927e+00 6.06182039e-01 5.95229208e-01 -5.10829985e-01
-9.60804701e-01 8.73797417e-01 9.08148825e-01 -9.27701667e-02
1.68688774e-01 -1.83813661e-01 -1.13367058e-01 -2.62715161e-01
1.06930554e+00 3.80603850e-01 9.10426080e-02 -5.85732877e-01
2.63584312e-02 4.46712732e-01 -4.56751846e-02 9.99880135e-02
1.42044210e+00 -5.88355541e-01 -5.63388586e-01 -5.99244535e-02
6.81945682e-01 2.03854114e-01 -1.60614765e+00 -2.94752240e-01
-4.45768297e-01 -7.12099493e-01 2.26974681e-01 -9.41332161e-01
-1.47370684e+00 6.48307264e-01 6.26444161e-01 2.09864482e-01
1.98071086e+00 -4.48000878e-01 6.27425075e-01 -2.04163492e-01
-1.41705632e-01 -5.67312777e-01 2.73030668e-01 3.82544070e-01
9.13225353e-01 -1.08348107e+00 3.62176411e-02 -5.21255851e-01
-5.05929410e-01 9.04918432e-01 5.82003295e-01 -7.87676424e-02
3.34314704e-01 6.13067985e-01 3.81457865e-01 -2.31990561e-01
-4.39650685e-01 -6.49788022e-01 2.59792268e-01 7.59023368e-01
9.35523883e-02 -2.79797405e-01 1.41910508e-01 6.96565509e-01
2.01191917e-01 -1.69304669e-01 5.35792291e-01 6.82511091e-01
-4.33871686e-01 -9.58128989e-01 -7.67203867e-01 4.72089425e-02
-6.28410518e-01 -4.73149180e-01 -1.98153317e-01 5.06790340e-01
3.30790371e-01 1.24444830e+00 -1.32702172e-01 -1.24779999e-01
2.44887367e-01 -3.90420020e-01 4.36387688e-01 -5.12632191e-01
-6.69498503e-01 4.63805169e-01 -3.44380826e-01 -5.70057154e-01
-6.17987692e-01 -5.79004660e-02 -7.16543019e-01 -1.78394273e-01
-3.67945582e-01 -2.67495871e-01 3.08268249e-01 7.20587492e-01
3.61678153e-01 9.92351770e-01 5.60670137e-01 -1.27842569e+00
9.95450094e-02 -8.42142701e-01 -6.76623821e-01 6.39778435e-01
8.13034356e-01 -4.02137250e-01 -7.34255552e-01 2.44639948e-01] | [11.022128105163574, -2.5310730934143066] |
e211a354-e059-48f5-9e74-b672736a218d | signal-processing-on-large-networks-with | 2303.17065 | null | https://arxiv.org/abs/2303.17065v1 | https://arxiv.org/pdf/2303.17065v1.pdf | Signal processing on large networks with group symmetries | Current methods of graph signal processing rely heavily on the specific structure of the underlying network: the shift operator and the graph Fourier transform are both derived directly from a specific graph. In many cases, the network is subject to error or natural changes over time. This motivated a new perspective on GSP, where the signal processing framework is developed for an entire class of graphs with similar structures. This approach can be formalized via the theory of graph limits, where graphs are considered as random samples from a distribution represented by a graphon. When the network under consideration has underlying symmetries, they may be modeled as samples from Cayley graphons. In Cayley graphons, vertices are sampled from a group, and the link probability between two vertices is determined by a function of the two corresponding group elements. Infinite groups such as the 1-dimensional torus can be used to model networks with an underlying spatial reality. Cayley graphons on finite groups give rise to a Stochastic Block Model, where the link probabilities between blocks form a (edge-weighted) Cayley graph. This manuscript summarizes some work on graph signal processing on large networks, in particular samples of Cayley graphons. | ['Nauzer Kalyaniwalla', 'Jeannette Janssen', 'Mahya Ghandehari', 'Kathryn Beck'] | 2023-03-29 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 2.74522692e-01 4.53453153e-01 1.25925526e-01 1.58201963e-01
3.13243240e-01 -6.80543184e-01 7.03958929e-01 5.33914268e-02
6.76432401e-02 3.90839607e-01 -4.86490019e-02 -7.67907724e-02
-3.60339105e-01 -1.25921214e+00 -6.94603264e-01 -9.94907618e-01
-7.71167934e-01 3.01706284e-01 2.59944469e-01 -3.69862169e-01
-1.97346240e-01 6.04846299e-01 -9.83036101e-01 -2.36789867e-01
4.39310580e-01 5.15768528e-01 -1.37787178e-01 1.02509737e+00
6.37910366e-02 4.39017981e-01 -4.18793201e-01 -1.75170094e-01
5.45492053e-01 -8.64547729e-01 -3.01083535e-01 4.46831673e-01
1.27365753e-01 3.55929434e-01 -1.08458459e+00 1.48992372e+00
2.17964992e-01 4.81726304e-02 7.45902717e-01 -1.59635210e+00
-4.45666611e-01 1.03172708e+00 -4.12811577e-01 7.91420862e-02
3.99931252e-01 -2.84333199e-01 1.04908049e+00 -3.83180290e-01
8.32592607e-01 1.23927438e+00 6.83934391e-01 -5.10119535e-02
-2.03242278e+00 -3.00254464e-01 -4.22824621e-02 2.89093703e-01
-1.59529769e+00 -1.35711119e-01 1.09927309e+00 -4.66127336e-01
4.30369526e-01 4.05749142e-01 1.19647276e+00 8.73118460e-01
6.24903560e-01 9.19937938e-02 7.15125442e-01 -7.04144657e-01
5.88634729e-01 -3.69035900e-01 3.37683380e-01 7.14139998e-01
7.43793070e-01 2.61298902e-02 -3.11414421e-01 -2.20223725e-01
7.21773446e-01 -2.22503692e-01 -5.35548151e-01 -9.91536319e-01
-1.13545477e+00 8.68093073e-01 5.78687966e-01 5.54794133e-01
-4.23274755e-01 3.71277899e-01 2.94164956e-01 7.61617899e-01
1.59888700e-01 2.57250201e-02 4.14202154e-01 5.45479655e-01
-8.17031324e-01 1.66963369e-01 1.23205149e+00 1.24088573e+00
7.83964455e-01 2.16091052e-01 2.59907633e-01 1.95964366e-01
2.31467605e-01 7.63868153e-01 -1.83479667e-01 -6.29289091e-01
1.18292637e-01 3.58825564e-01 -4.63865608e-01 -1.58867586e+00
-6.27840638e-01 -5.83680749e-01 -1.29735112e+00 4.98139560e-02
6.62615120e-01 8.50649364e-03 -5.98917782e-01 1.86543441e+00
4.46479246e-02 2.28908762e-01 -1.61457121e-01 4.25963312e-01
3.27849537e-01 6.18868232e-01 -4.88454223e-01 -3.42231810e-01
1.16107738e+00 -1.35761812e-01 -6.51110709e-01 9.34518352e-02
4.45405781e-01 -3.40424180e-01 4.27810073e-01 5.64988375e-01
-9.66512501e-01 -2.60083050e-01 -1.14866233e+00 4.58222330e-01
-2.35988617e-01 -4.08365339e-01 -4.00297455e-02 6.10148907e-01
-1.50654304e+00 7.29619861e-01 -5.81618488e-01 -7.02410996e-01
-2.05687046e-01 2.32283041e-01 -3.01540524e-01 -2.18002900e-01
-1.10656750e+00 4.05762136e-01 2.27614298e-01 4.27440882e-01
-5.62097669e-01 -2.34283864e-01 -8.21701288e-01 4.60154451e-02
4.19562638e-01 -5.33487439e-01 5.53916931e-01 -7.61293352e-01
-9.73516107e-01 5.92131376e-01 2.49790817e-01 -5.51939666e-01
4.78705823e-01 7.56757021e-01 -7.55041718e-01 4.98783648e-01
3.50691974e-02 -9.02322456e-02 1.25962639e+00 -8.62723768e-01
1.80906236e-01 -2.18971089e-01 4.41211462e-02 -2.24458218e-01
2.17019185e-01 -2.98155487e-01 -1.73969083e-02 -6.40135050e-01
5.99298835e-01 -1.30582666e+00 -1.61992490e-01 -3.42442244e-01
-8.46931636e-01 2.13207796e-01 6.65093839e-01 -3.76612008e-01
1.32396388e+00 -2.24320483e+00 5.08195281e-01 1.00640249e+00
6.54561579e-01 -3.78310233e-01 -2.33337075e-01 1.08637428e+00
-5.20777583e-01 2.15353712e-01 -7.39835501e-02 4.07390684e-01
-2.49118749e-02 -5.21602668e-02 -1.26292214e-01 9.67158079e-01
-1.76963493e-01 6.46926999e-01 -9.49957967e-01 -1.03347078e-02
5.71545493e-03 3.15753609e-01 -6.12046897e-01 -3.58858436e-01
-2.34059170e-02 1.03294767e-01 -5.85354269e-01 -7.33739287e-02
7.03759789e-01 -3.27400476e-01 5.26710570e-01 -2.99269170e-01
5.04955947e-02 -7.81555027e-02 -1.48185456e+00 1.21687460e+00
-1.00739770e-01 6.84589446e-01 4.76714820e-01 -1.34496367e+00
9.85651493e-01 2.47655988e-01 4.73109365e-01 -2.28942875e-02
1.44759119e-01 -9.69191492e-02 6.91602826e-01 -7.68119320e-02
1.95544839e-01 3.36267985e-02 -1.90658927e-01 5.11968255e-01
7.64113590e-02 -3.89971614e-01 5.61102331e-01 8.73855174e-01
1.86044836e+00 -6.17407918e-01 6.02936089e-01 -7.94498920e-01
4.60880369e-01 -2.29919121e-01 2.32556894e-01 7.90335834e-01
3.69558409e-02 4.02792990e-01 1.17422271e+00 -1.05426885e-01
-1.09779775e+00 -1.41971946e+00 -6.02609701e-02 1.16673172e-01
-1.01803513e-02 -7.54326701e-01 -8.68850887e-01 -6.81431666e-02
4.86980332e-03 3.10000002e-01 -5.71441352e-01 -4.56888974e-01
-4.05360281e-01 -6.76809192e-01 2.86108106e-01 -2.89937586e-01
4.11543757e-01 -6.91201091e-01 6.49052998e-03 5.04716039e-01
1.10435657e-01 -1.23207653e+00 -5.57804346e-01 3.18181925e-02
-8.92583013e-01 -1.44836533e+00 -3.63680601e-01 -7.49791741e-01
8.20836544e-01 4.30551112e-01 9.61810708e-01 -2.41343416e-02
-3.07791919e-01 9.06964183e-01 -3.05057466e-01 -9.72041711e-02
-7.12836742e-01 -1.51110679e-01 2.89243311e-01 6.80703700e-01
-4.29950282e-02 -1.01320255e+00 -1.61582023e-01 3.23256046e-01
-9.37704623e-01 -1.62284523e-01 5.07404618e-02 4.52510059e-01
3.63350034e-01 5.64815581e-01 3.10779333e-01 -8.36323261e-01
6.82288170e-01 -5.63786447e-01 -8.77054334e-01 1.69403911e-01
-1.40766636e-01 2.34369904e-01 9.85282362e-01 -3.55195254e-01
-2.00455874e-01 -2.36636996e-02 7.70177722e-01 -2.16221482e-01
4.27050561e-01 7.26289511e-01 -2.80211985e-01 -4.10985291e-01
6.58587217e-01 2.20864162e-01 2.48277888e-01 -2.69805025e-02
4.69104916e-01 1.26188532e-01 2.21213102e-01 -3.82534802e-01
1.00881863e+00 4.47164029e-01 8.70036185e-01 -1.71779537e+00
5.29129915e-02 4.25244644e-02 -7.58196831e-01 -6.33525193e-01
6.56978726e-01 -6.52566373e-01 -6.85466647e-01 6.36104226e-01
-9.28258598e-01 -3.33418369e-01 -4.91831809e-01 5.48492849e-01
-5.66199005e-01 4.47863340e-01 -6.13635242e-01 -5.71426094e-01
2.43151858e-01 -6.32588208e-01 5.85695028e-01 -2.43115872e-01
-1.30559936e-01 -1.36453903e+00 1.12819299e-01 -5.52640676e-01
3.91383827e-01 5.07444084e-01 1.14656126e+00 -4.66819674e-01
-8.67212534e-01 -5.88863671e-01 2.43201658e-01 2.11429462e-01
8.74235295e-03 2.71500111e-01 -1.91447273e-01 -4.40987736e-01
2.54662097e-01 3.89834762e-01 3.93758804e-01 7.84264088e-01
5.77779889e-01 -1.06984332e-01 -4.31741774e-01 2.56963283e-01
1.36026442e+00 9.78346393e-02 3.40272725e-01 -2.91440040e-01
6.42776370e-01 4.62503284e-01 -2.93111295e-01 1.67331368e-01
-2.09026530e-01 6.41095936e-01 2.35678896e-01 4.83537108e-01
-5.57372831e-02 -3.62074822e-01 2.48114645e-01 1.29194748e+00
-9.33965221e-02 -5.26281297e-01 -8.36893737e-01 3.14007923e-02
-1.51299524e+00 -1.04198563e+00 -7.33274162e-01 2.30275583e+00
-6.74586594e-02 2.33011201e-01 1.92361683e-01 3.03863168e-01
1.45389163e+00 5.33403456e-01 -3.20671290e-01 4.06889059e-02
-2.51013696e-01 1.49606662e-02 6.72971308e-01 7.29064167e-01
-5.73163331e-01 3.17563713e-01 6.84681082e+00 5.05262077e-01
-7.72791266e-01 -1.97640911e-01 6.58514723e-02 8.31372589e-02
-2.24665061e-01 2.58897841e-01 -3.07270676e-01 5.34069479e-01
9.20296371e-01 -9.88787711e-01 8.03812921e-01 2.94546664e-01
3.43107820e-01 -5.80591373e-02 -1.19166446e+00 9.89574730e-01
-3.41800414e-02 -9.57661271e-01 2.31250338e-02 7.14447677e-01
5.58983743e-01 -1.36081353e-01 -2.18545988e-01 -1.47522524e-01
4.30271536e-01 -6.31879389e-01 4.99182165e-01 4.83171135e-01
6.81922257e-01 -3.85923862e-01 1.89601690e-01 2.39373639e-01
-1.60601723e+00 3.15223396e-01 -4.03780520e-01 -1.09528556e-01
3.89786601e-01 1.03866148e+00 -6.13061249e-01 7.72040486e-01
8.83095190e-02 9.35297728e-01 -2.29687586e-01 9.62798297e-01
-2.16821343e-01 7.64131606e-01 -6.29383564e-01 -1.76753893e-01
7.45828599e-02 -1.12351036e+00 1.17373288e+00 6.21506512e-01
6.13731205e-01 3.07981282e-01 1.16379581e-01 7.62693882e-01
-2.47205663e-02 1.92204174e-02 -1.26447284e+00 -1.86683461e-01
3.76607388e-01 1.21896541e+00 -1.42755282e+00 -7.27981851e-02
-5.89318037e-01 4.99730110e-01 -2.25226544e-02 6.92345321e-01
-5.02168000e-01 -6.03233457e-01 5.23407519e-01 6.31564915e-01
3.43529463e-01 -5.07604539e-01 3.57641816e-01 -1.19608688e+00
-6.00895621e-02 -6.27628565e-01 3.69033255e-02 -7.61123836e-01
-1.09808564e+00 5.64358175e-01 2.15533882e-01 -1.31768227e+00
-1.06181651e-01 -5.66532731e-01 -7.81740904e-01 4.79124933e-01
-3.60338598e-01 -4.42767203e-01 -2.23583728e-01 8.29947829e-01
-3.49132627e-01 -1.59146443e-01 5.43629229e-01 8.09838437e-03
-2.57350266e-01 -1.87975556e-01 3.18362653e-01 1.73022434e-01
3.22178602e-02 -1.26028883e+00 5.96550822e-01 1.13279843e+00
4.18442428e-01 4.21117485e-01 9.46409583e-01 -8.35036516e-01
-1.70520020e+00 -9.23084080e-01 5.63605011e-01 2.53081396e-02
1.21346366e+00 -1.00075555e+00 -6.74303651e-01 9.34463561e-01
-2.29933448e-02 4.41030651e-01 1.91624433e-01 -7.80461431e-02
-8.08267370e-02 -2.17421338e-01 -9.81682777e-01 7.82798052e-01
1.52731228e+00 -6.22003198e-01 -1.11245818e-01 4.38533694e-01
2.54099607e-01 6.05957322e-02 -6.67366505e-01 -3.97039950e-02
1.63981214e-01 -9.98433471e-01 6.94294870e-01 -4.67544675e-01
-1.13923013e-01 -5.13316631e-01 -8.89688581e-02 -1.89408755e+00
-6.08155608e-01 -1.11961484e+00 1.90183982e-01 7.09629536e-01
7.81536698e-02 -8.17342103e-01 6.82734430e-01 1.79674625e-01
2.18267024e-01 8.14336091e-02 -1.08918953e+00 -1.14088118e+00
-2.39431322e-01 -4.49458778e-01 4.59627897e-01 7.81594455e-01
2.94176012e-01 6.12664640e-01 -1.14971921e-01 3.12628716e-01
1.05375767e+00 -1.72641292e-01 5.85229337e-01 -1.54549360e+00
-3.96321625e-01 -3.95440578e-01 -1.25558782e+00 -8.17425251e-01
1.01742782e-01 -1.46589708e+00 -2.23826811e-01 -1.28413332e+00
-2.13504657e-01 -1.83072478e-01 2.94774890e-01 -4.53375638e-01
6.46142304e-01 1.78861432e-02 2.19825521e-01 -1.75635010e-01
-1.55347317e-01 3.78105074e-01 1.06739545e+00 1.25948079e-02
-1.92868620e-01 2.68127799e-01 -2.49741614e-01 6.16070867e-01
5.42219520e-01 -5.26361525e-01 -7.06264377e-01 1.84425086e-01
7.29895830e-01 5.09875000e-01 3.29736143e-01 -1.11326647e+00
4.59695727e-01 3.48936617e-01 -2.33144388e-01 -2.11978361e-01
4.75152098e-02 -9.45785403e-01 9.78196383e-01 5.99954844e-01
-9.91845354e-02 2.32862532e-01 -3.23205143e-01 1.09246874e+00
-2.21777007e-01 -1.95015505e-01 7.64031589e-01 1.58443314e-03
-1.39448032e-01 4.44954067e-01 -6.49965346e-01 2.42240638e-01
1.25557482e+00 -2.03225926e-01 -1.40357152e-01 -8.57352018e-01
-1.28174245e+00 -5.47588468e-02 4.96873468e-01 -7.41528049e-02
3.02024722e-01 -1.61478412e+00 -6.99038148e-01 4.69618320e-01
-1.87582523e-01 -3.14649075e-01 1.40496850e-01 8.91391993e-01
-6.44510150e-01 2.41928577e-01 -8.01794752e-02 -7.20115185e-01
-1.22483575e+00 5.87418795e-01 4.24142987e-01 -6.98984638e-02
-8.00170958e-01 3.59654486e-01 3.38263005e-01 -1.88817963e-01
-1.62576482e-01 -5.10567009e-01 1.45656422e-01 1.27153352e-01
3.18750888e-01 5.56789160e-01 -3.90650518e-02 -7.67444432e-01
-1.66091427e-01 6.49039507e-01 2.45144024e-01 -1.18290722e-01
1.19436145e+00 -8.63404423e-02 -5.23136318e-01 6.37890160e-01
1.22598279e+00 3.88056934e-01 -7.62469292e-01 -2.96692461e-01
-1.47458822e-01 -2.40550339e-01 -1.75171286e-01 1.63551226e-01
-1.17121935e+00 4.15245891e-01 -1.76001161e-01 1.37113583e+00
9.24776256e-01 2.20604047e-01 -8.09838846e-02 2.62873828e-01
7.14437246e-01 -6.22808576e-01 -1.55518457e-01 5.29086947e-01
1.08269310e+00 -3.58319014e-01 2.18235441e-02 -9.20151889e-01
1.50119483e-01 1.52755082e+00 -2.41079733e-01 -8.96520197e-01
1.37503576e+00 2.63369203e-01 -6.71608567e-01 -3.94931525e-01
-4.58169490e-01 9.74104330e-02 1.96088612e-01 6.90517843e-01
7.92849138e-02 3.17266494e-01 -2.30276152e-01 4.74757329e-02
-5.53006053e-01 -3.12681764e-01 1.34121573e+00 4.80814427e-01
-3.72475415e-01 -1.00454640e+00 -4.44395661e-01 7.69212306e-01
6.74988627e-02 -6.98976889e-02 -3.43645841e-01 1.02492642e+00
-4.56003785e-01 8.70650053e-01 1.86030582e-01 -4.09154981e-01
4.07140076e-01 -1.28270313e-01 6.95110381e-01 -7.55280316e-01
2.57385701e-01 -1.59919634e-01 3.55275571e-02 -4.92482424e-01
-4.44379300e-01 -1.09074152e+00 -8.28973413e-01 -7.52964377e-01
-1.37863815e-01 3.62329930e-01 3.78038585e-01 5.68129420e-01
1.14287458e-01 5.06372809e-01 6.24631107e-01 -7.70124733e-01
-3.36102605e-01 -7.61099398e-01 -1.80046046e+00 4.04091217e-02
3.95821005e-01 -5.33719182e-01 -8.98079216e-01 -2.19873190e-01] | [6.971334934234619, 5.176416873931885] |
23192025-26ef-47a0-9a7c-bf01520e6c96 | text-based-person-search-in-full-images-via | 2109.12965 | null | https://arxiv.org/abs/2109.12965v1 | https://arxiv.org/pdf/2109.12965v1.pdf | Text-based Person Search in Full Images via Semantic-Driven Proposal Generation | Finding target persons in full scene images with a query of text description has important practical applications in intelligent video surveillance.However, different from the real-world scenarios where the bounding boxes are not available, existing text-based person retrieval methods mainly focus on the cross modal matching between the query text descriptions and the gallery of cropped pedestrian images. To close the gap, we study the problem of text-based person search in full images by proposing a new end-to-end learning framework which jointly optimize the pedestrian detection, identification and visual-semantic feature embedding tasks. To take full advantage of the query text, the semantic features are leveraged to instruct the Region Proposal Network to pay more attention to the text-described proposals. Besides, a cross-scale visual-semantic embedding mechanism is utilized to improve the performance. To validate the proposed method, we collect and annotate two large-scale benchmark datasets based on the widely adopted image-based person search datasets CUHK-SYSU and PRW. Comprehensive experiments are conducted on the two datasets and compared with the baseline methods, our method achieves the state-of-the-art performance. | ['Yanning Zhang', 'Kai Niu', 'Qian Zhang', 'Liying Gao', 'Yitao Gao', 'Duo Long', 'Shizhou Zhang'] | 2021-09-27 | null | null | null | null | ['nlp-based-person-retrival', 'person-search'] | ['computer-vision', 'computer-vision'] | [-1.67955175e-01 -5.23611069e-01 -7.39900023e-02 -4.96855438e-01
-8.73613894e-01 -3.53311598e-01 7.49113798e-01 -2.06163734e-01
-9.03905272e-01 4.95199561e-01 4.63054627e-01 1.78688779e-01
-3.35258767e-02 -6.20304942e-01 -4.95374978e-01 -6.89666212e-01
3.61092120e-01 4.17543739e-01 5.69488108e-01 1.87357329e-02
-3.23284231e-03 8.83875489e-02 -1.42399037e+00 2.98534751e-01
5.24560750e-01 1.19682479e+00 2.28736609e-01 3.04072171e-01
1.83152229e-01 4.25395697e-01 -5.54780960e-01 -8.63352478e-01
4.24619317e-01 9.46960449e-02 -4.99707699e-01 3.67712706e-01
7.79628158e-01 -6.17417693e-01 -1.09268260e+00 1.07959843e+00
9.43160355e-01 3.92975837e-01 5.89771271e-01 -1.31432128e+00
-8.55844080e-01 -1.13489171e-02 -6.96511805e-01 3.64027411e-01
5.47634184e-01 4.33870852e-01 9.41770852e-01 -1.09424865e+00
5.19813001e-01 1.75612211e+00 4.26621497e-01 5.55370688e-01
-6.57766223e-01 -6.70730591e-01 4.33710068e-01 6.35100484e-01
-1.69142711e+00 -5.02318561e-01 6.60654902e-01 -2.90151685e-01
6.19260609e-01 1.97262585e-01 7.08220363e-01 1.21307003e+00
-1.94513991e-01 1.31865823e+00 6.82166696e-01 -8.12439546e-02
-2.33798012e-01 4.38493252e-01 1.98418841e-01 8.78988683e-01
4.74066734e-01 2.68370658e-01 -3.83429825e-01 -2.25819901e-01
4.50673699e-01 5.45103371e-01 -3.62980425e-01 -5.32543480e-01
-1.21954250e+00 7.40055799e-01 7.26650596e-01 4.81251515e-02
-1.89516544e-01 7.81782717e-02 7.43511915e-01 -2.20110789e-01
2.60498166e-01 -2.48487711e-01 -1.02931015e-01 3.66276801e-01
-8.82826090e-01 5.26066184e-01 3.70788693e-01 1.19308889e+00
4.58417386e-01 -4.49168414e-01 -8.74301493e-01 9.73511696e-01
6.54992759e-01 9.18172777e-01 2.16820121e-01 -4.61657584e-01
7.97786653e-01 7.63513029e-01 4.90736157e-01 -1.23187840e+00
7.87273515e-03 -4.12796080e-01 -7.10299551e-01 -4.70144153e-01
4.42411274e-01 1.60544559e-01 -7.76349366e-01 1.32409132e+00
5.64807117e-01 -3.97417024e-02 6.92230090e-02 1.54194546e+00
1.27995002e+00 6.27849579e-01 3.01443666e-01 1.84241340e-01
1.96197534e+00 -1.46958244e+00 -4.81783211e-01 -4.66052562e-01
1.53105900e-01 -6.62290037e-01 7.45465457e-01 -2.39292338e-01
-8.27550232e-01 -7.08699107e-01 -7.06683278e-01 -2.56177872e-01
-6.86707616e-01 7.73947716e-01 1.89406484e-01 5.43560624e-01
-7.93983340e-01 -3.10205430e-01 -3.48175913e-01 -8.43861639e-01
6.70485854e-01 1.59979060e-01 -5.13380051e-01 -6.03537560e-01
-1.30541170e+00 6.42318547e-01 5.01960874e-01 4.49777454e-01
-9.43303585e-01 -2.08969846e-01 -7.90942729e-01 4.75133024e-02
5.79714179e-01 -9.97773826e-01 8.28068852e-01 -5.86600006e-01
-6.87525511e-01 1.26163709e+00 -2.77729064e-01 -2.77080476e-01
8.49065065e-01 -1.52939543e-01 -2.40978584e-01 3.12951028e-01
6.31348848e-01 8.18806052e-01 6.57375455e-01 -1.02297258e+00
-1.00041950e+00 -6.42251730e-01 1.13273501e-01 2.64739901e-01
-5.96066535e-01 3.92149627e-01 -1.34300327e+00 -5.97908795e-01
-3.01940858e-01 -7.17817843e-01 -8.12510699e-02 4.48234558e-01
-5.47145307e-01 -6.18966460e-01 8.62067401e-01 -8.03748965e-01
9.49406683e-01 -1.90003586e+00 -4.45085168e-02 -1.19010247e-01
1.80940211e-01 3.30806792e-01 -4.83472738e-03 2.64131486e-01
3.93276185e-01 -4.09521133e-01 1.86766416e-01 -5.11696994e-01
3.26155394e-01 -1.06589779e-01 -1.05683118e-01 6.49541795e-01
-8.14117584e-03 1.18357348e+00 -7.77418554e-01 -1.07709777e+00
3.13069016e-01 3.79078001e-01 -2.26544306e-01 2.29169279e-01
1.34325206e-01 1.72711790e-01 -9.85399544e-01 9.69681382e-01
6.88754022e-01 -2.76491761e-01 -4.30581987e-01 -4.68113214e-01
7.00886622e-02 -4.36272234e-01 -1.07698715e+00 1.71492577e+00
2.69758999e-01 4.55820441e-01 -2.16058102e-02 -1.09425867e+00
6.08028710e-01 8.24396536e-02 3.45315248e-01 -7.59823263e-01
8.07505175e-02 -4.42526750e-02 -5.77480614e-01 -7.75961697e-01
5.84728658e-01 4.63202864e-01 -3.11401933e-01 1.31901190e-01
-1.01567402e-01 6.98800743e-01 3.24781746e-01 3.20331603e-01
7.74153829e-01 1.02435581e-01 1.90436035e-01 -5.86275272e-02
9.78718936e-01 1.34076223e-01 4.74047363e-01 9.37551022e-01
-7.54877150e-01 4.16418821e-01 2.75181490e-03 -7.68028200e-01
-1.06460035e+00 -9.77515817e-01 -2.09302157e-02 1.30832028e+00
5.94421804e-01 -2.41707325e-01 -7.76375353e-01 -9.00507092e-01
4.91282977e-02 9.29799750e-02 -6.28321648e-01 1.45863313e-02
-3.62702668e-01 -6.47473395e-01 6.67748749e-01 4.16949570e-01
1.23007727e+00 -9.18386102e-01 -6.60025835e-01 -1.27380475e-01
-5.49036801e-01 -1.61929333e+00 -1.19122684e+00 -7.58939266e-01
-9.97916535e-02 -1.29675770e+00 -1.36268389e+00 -1.09077907e+00
7.66790509e-01 7.11078465e-01 8.26399744e-01 8.88133273e-02
-7.48716891e-01 8.62318277e-01 -3.34147871e-01 -1.89715460e-01
4.18050587e-01 -5.78413308e-02 -1.66847017e-02 2.46659353e-01
7.83779681e-01 3.66713345e-01 -1.17032814e+00 6.10062897e-01
-7.32834339e-01 -4.30246331e-02 6.26136065e-01 9.48722064e-01
4.04548466e-01 3.27472165e-02 2.47231036e-01 4.18107817e-03
3.68160784e-01 -3.08162063e-01 -5.81207931e-01 7.69861519e-01
-1.54664040e-01 -2.27360576e-01 1.47440210e-01 -4.01481211e-01
-9.88744557e-01 6.12813458e-02 1.79739594e-01 -4.97248411e-01
-2.19713032e-01 -4.94839437e-02 -4.51218575e-01 -2.42268443e-02
1.90403476e-01 7.11284697e-01 -4.12025422e-01 -2.00663581e-01
2.26181552e-01 8.72502685e-01 6.66813314e-01 -5.46659350e-01
8.83944690e-01 6.33438587e-01 -4.07502741e-01 -6.38128579e-01
-9.10146117e-01 -9.94237423e-01 -5.86681128e-01 -4.29950058e-01
1.19541264e+00 -1.27232277e+00 -8.27663958e-01 3.11572373e-01
-1.39837861e+00 3.88182938e-01 2.77520120e-01 3.04195315e-01
-2.53097922e-01 5.96076727e-01 -3.58189970e-01 -1.05887902e+00
-6.36393905e-01 -1.29959512e+00 1.77165675e+00 4.19088662e-01
5.98639786e-01 -7.04419136e-01 -3.54883134e-01 7.19599605e-01
1.73332512e-01 -3.89853641e-02 4.34820652e-01 -7.90545046e-01
-9.67083514e-01 -5.92023432e-01 -8.51960838e-01 -1.19646348e-01
-2.02990294e-01 -6.72409415e-01 -8.73044491e-01 -5.98164737e-01
-3.97306174e-01 -2.72816867e-01 1.15606821e+00 2.35947669e-01
9.82101381e-01 -3.33410144e-01 -8.52130055e-01 4.66785789e-01
1.26705229e+00 -2.07229614e-01 3.19776028e-01 5.77049375e-01
5.72903872e-01 6.63092792e-01 8.07131767e-01 4.99955088e-01
7.50750899e-01 9.84466076e-01 1.83598071e-01 -7.74956271e-02
-2.17470359e-02 -5.40228903e-01 1.67402253e-01 -7.02391490e-02
1.92891240e-01 -4.10385132e-01 -6.63495243e-01 6.64581656e-01
-2.21130872e+00 -1.23272324e+00 2.76454210e-01 2.03374863e+00
3.27836275e-01 -1.40267342e-01 2.74871320e-01 -3.72456074e-01
1.04653037e+00 3.48840356e-01 -5.15926540e-01 7.15544283e-01
-2.21970394e-01 -6.74154818e-01 4.78293806e-01 -5.97130449e-04
-1.72446072e+00 1.15810406e+00 5.23208666e+00 1.09832609e+00
-6.30512953e-01 3.32860321e-01 6.39072835e-01 -1.26482993e-01
3.12818319e-01 -3.30443114e-01 -1.26303542e+00 6.41919792e-01
1.55701995e-01 -5.78112006e-02 8.64054114e-02 9.39451098e-01
3.86192977e-01 4.84516174e-02 -1.07255900e+00 1.54103601e+00
4.30243313e-01 -1.05467200e+00 2.76310265e-01 -1.28757255e-02
4.87948269e-01 -2.55482823e-01 1.07190885e-01 4.28936988e-01
-4.63389046e-02 -8.37064087e-01 7.37654805e-01 7.72489071e-01
6.47974014e-01 -6.19298518e-01 8.57574642e-01 3.54458839e-01
-1.68882430e+00 -4.14457172e-01 -8.12052608e-01 4.76738721e-01
3.02240700e-01 3.47983837e-02 -4.07239377e-01 6.42973542e-01
1.23611414e+00 8.51131797e-01 -9.52182233e-01 1.34042573e+00
1.11123249e-01 7.77884945e-02 -2.31418028e-01 -2.19278082e-01
4.22094852e-01 -7.12040961e-02 5.58870256e-01 1.38605118e+00
1.17851354e-01 1.72664925e-01 6.49869561e-01 1.03164232e+00
3.60199325e-02 1.93822816e-01 -6.12185895e-01 2.51651227e-01
3.68772000e-01 1.20831954e+00 -4.36327159e-01 -5.11117697e-01
-6.19622648e-01 1.32568169e+00 1.73748598e-01 6.34985089e-01
-9.70873892e-01 -2.76593536e-01 5.18757522e-01 2.08806153e-02
4.82967407e-01 -2.03376915e-02 4.94845450e-01 -1.30550206e+00
3.84569228e-01 -6.57347322e-01 6.36021674e-01 -9.29703176e-01
-1.65635514e+00 3.85387361e-01 2.37014756e-01 -1.07108998e+00
1.19980805e-01 -6.43730938e-01 -5.52951097e-01 9.51491714e-01
-1.53315270e+00 -1.89951384e+00 -7.01189101e-01 8.58291268e-01
8.09157491e-01 -5.78367293e-01 3.18497270e-01 6.68735504e-01
-7.34674096e-01 9.83591080e-01 2.10522432e-02 7.69815803e-01
9.05265391e-01 -6.43763006e-01 2.42392868e-01 9.37746942e-01
-2.18062788e-01 6.57482862e-01 4.25336331e-01 -6.68266118e-01
-1.47340262e+00 -1.41650808e+00 6.53228164e-01 -6.44118249e-01
4.40752357e-01 -4.12031591e-01 -4.67612237e-01 4.82341349e-01
1.78037465e-01 3.25122654e-01 3.02593380e-01 -4.79586840e-01
-2.39115193e-01 -2.46711537e-01 -1.08482504e+00 6.03788912e-01
1.25472927e+00 -4.06583309e-01 -5.88647604e-01 5.84740818e-01
7.55032539e-01 -2.38126650e-01 -4.57663476e-01 2.01221526e-01
6.95493877e-01 -7.28440702e-01 1.63034797e+00 -7.15578079e-01
1.88540127e-02 -4.44245875e-01 -3.45000029e-01 -7.12828159e-01
-1.41005367e-01 7.77978590e-03 2.76531965e-01 1.22572076e+00
-1.63349628e-01 -5.87417006e-01 7.74581075e-01 6.10857785e-01
2.11311251e-01 -5.28291583e-01 -1.08756423e+00 -5.71648359e-01
-2.93536037e-01 -1.77548663e-03 4.71573353e-01 3.67560774e-01
-4.09162790e-01 1.97832942e-01 -7.04838097e-01 3.64453942e-01
1.21198130e+00 4.77132089e-02 8.95252287e-01 -1.00523293e+00
-6.33316673e-03 -3.47705901e-01 -7.43449688e-01 -1.48937619e+00
2.87969798e-01 -7.40969956e-01 8.07833299e-02 -1.62815297e+00
9.90742683e-01 -1.64461538e-01 -2.62038022e-01 1.83436438e-01
-5.32581687e-01 2.38107860e-01 2.62659788e-01 4.04980659e-01
-1.41445673e+00 8.39960575e-01 1.02751124e+00 -6.88305914e-01
3.43538553e-01 8.76608025e-03 -5.03526092e-01 4.32570189e-01
1.87263995e-01 -2.38544121e-01 -2.20684752e-01 -6.29791737e-01
-2.90479839e-01 3.02048754e-02 1.12248385e+00 -9.24765468e-01
7.59669006e-01 -2.52135973e-02 7.44805872e-01 -1.03817725e+00
5.81736982e-01 -8.99691522e-01 -3.73199373e-01 3.82746577e-01
-4.63696718e-01 4.14390191e-02 -1.33091778e-01 1.09182096e+00
-1.84836313e-01 -6.50193691e-02 5.32979250e-01 -2.94589609e-01
-1.03049552e+00 8.79627585e-01 8.38578213e-03 3.09409183e-02
1.08595514e+00 -3.33640188e-01 -4.99636710e-01 -2.35305771e-01
-3.00209314e-01 8.47775936e-01 3.33395451e-01 6.39379442e-01
9.45522308e-01 -1.51727903e+00 -8.68661344e-01 -2.21337080e-02
6.68491781e-01 -2.75878638e-01 4.92621779e-01 7.24571347e-01
-1.30621061e-01 1.00151920e+00 1.02047034e-01 -7.90845692e-01
-1.42031574e+00 8.70472848e-01 5.76149702e-01 -3.89134884e-02
-5.04425406e-01 6.85587168e-01 5.90195715e-01 -3.34827363e-01
5.47723949e-01 2.88938463e-01 -2.89333075e-01 -2.60982931e-01
7.45379090e-01 2.45493144e-01 -5.49738109e-01 -1.03533602e+00
-5.40198147e-01 8.00180078e-01 -4.79416884e-02 -1.78289637e-02
7.90813744e-01 -4.65648741e-01 4.99790050e-02 -2.33809173e-01
1.28197050e+00 -4.20037150e-01 -1.11522877e+00 -7.67440200e-01
-1.96351945e-01 -8.05071414e-01 -2.84166843e-01 -5.01690626e-01
-1.03414774e+00 8.19674015e-01 1.09944057e+00 -2.96932489e-01
8.41537297e-01 2.41259754e-01 9.86907959e-01 7.44782329e-01
4.97741908e-01 -1.18113196e+00 3.77931744e-01 1.06172092e-01
7.35220611e-01 -1.77531099e+00 9.90308896e-02 -2.75335759e-01
-5.95374584e-01 8.25680315e-01 7.63153851e-01 3.02936044e-02
4.24790531e-01 -5.42641342e-01 -2.22012803e-01 -1.15838215e-01
-4.74648029e-01 -5.84754586e-01 6.21052742e-01 7.20334709e-01
-1.16404578e-01 -9.01854560e-02 -1.10604391e-01 7.58717835e-01
2.62542307e-01 -1.45729631e-01 -2.37222776e-01 6.25826895e-01
-6.77859366e-01 -7.15988517e-01 -7.60667801e-01 3.52260500e-01
-2.84896463e-01 -4.43174206e-02 -3.17598611e-01 6.76783860e-01
1.43713012e-01 1.19736207e+00 -2.09272906e-01 -1.02454007e-01
4.10286248e-01 1.73090789e-02 1.60160944e-01 -1.39920503e-01
-2.59220511e-01 3.27263144e-03 -8.46103579e-02 -4.63232070e-01
-4.91083950e-01 -6.83081031e-01 -7.19919026e-01 -1.28680736e-01
-2.65435785e-01 8.76563787e-02 3.06939602e-01 9.69742179e-01
2.05714986e-01 2.21641272e-01 3.22965950e-01 -7.67677188e-01
-6.11552238e-01 -8.90388846e-01 -2.76180446e-01 6.44241095e-01
2.08448902e-01 -8.69764984e-01 1.27033750e-02 4.82076406e-02] | [14.674172401428223, 0.8195810914039612] |
d0ef18c8-7a38-43b1-8856-cc546f5727d2 | physically-plausible-spectral-reconstruction | 2001.00558 | null | https://arxiv.org/abs/2001.00558v1 | https://arxiv.org/pdf/2001.00558v1.pdf | Physically Plausible Spectral Reconstruction from RGB Images | Recently Convolutional Neural Networks (CNN) have been used to reconstruct hyperspectral information from RGB images. Moreover, this spectral reconstruction problem (SR) can often be solved with good (low) error. However, these methods are not physically plausible: that is when the recovered spectra are reintegrated with the underlying camera sensitivities, the resulting predicted RGB is not the same as the actual RGB, and sometimes this discrepancy can be large. The problem is further compounded by exposure change. Indeed, most learning-based SR models train for a fixed exposure setting and we show that this can result in poor performance when exposure varies. In this paper we show how CNN learning can be extended so that physical plausibility is enforced and the problem resulting from changing exposures is mitigated. Our SR solution improves the state-of-the-art spectral recovery performance under varying exposure conditions while simultaneously ensuring physical plausibility (the recovered spectra reintegrate to the input RGBs exactly). | ['Yi-Tun Lin', 'Graham D. Finlayson'] | 2020-01-02 | null | null | null | null | ['spectral-reconstruction'] | ['computer-vision'] | [ 7.88967073e-01 -1.52357727e-01 1.38066754e-01 -6.76429644e-02
-7.99856663e-01 -8.59808266e-01 2.02405334e-01 -7.52123743e-02
-2.13314354e-01 1.04684281e+00 -5.17616905e-02 -2.14534998e-01
-4.33338374e-01 -9.36839044e-01 -1.06254518e+00 -1.00231516e+00
3.32172304e-01 -5.13730347e-02 -3.20645273e-02 -7.59568512e-02
-1.89234599e-01 7.32881188e-01 -1.57970822e+00 -1.34321362e-01
9.66637135e-01 1.00847471e+00 2.63615638e-01 7.64574528e-01
1.42212883e-01 5.88940203e-01 -4.61021781e-01 -1.16249874e-01
8.36345732e-01 -4.37290430e-01 -8.05086195e-01 3.45412821e-01
7.10356355e-01 -4.51339334e-01 -5.28541386e-01 1.36319470e+00
3.49329114e-01 2.65209436e-01 3.40298444e-01 -1.03812265e+00
-6.89702094e-01 2.88509399e-01 -6.23629332e-01 -4.27849114e-01
-5.88724278e-02 2.37162963e-01 7.79997706e-01 -5.79716980e-01
2.87520707e-01 6.24293864e-01 1.07091689e+00 1.65431008e-01
-1.49500573e+00 -5.13088286e-01 -8.32595900e-02 6.99727908e-02
-1.49976957e+00 -2.63704211e-01 6.69326484e-01 -2.93970257e-01
6.44704401e-01 3.09519738e-01 7.54947424e-01 8.15790355e-01
-1.23317935e-01 4.58025970e-02 1.22005475e+00 -4.69099462e-01
5.59906587e-02 -1.25592351e-01 -1.69841960e-01 3.76273692e-01
2.68334240e-01 3.47658277e-01 -4.97487277e-01 9.52180289e-03
7.10970819e-01 4.36796695e-02 -9.15532708e-01 -7.06720501e-02
-9.59257007e-01 6.01249635e-01 8.83729875e-01 -2.15019986e-01
-5.42367816e-01 3.07986230e-01 -1.04957908e-01 3.21949661e-01
4.00773793e-01 6.77407265e-01 -6.56670392e-01 4.01798487e-01
-1.01347744e+00 1.34191155e-01 3.80901635e-01 5.42930305e-01
1.16163540e+00 1.04415059e-01 2.76398510e-01 7.73235738e-01
8.14920962e-02 7.45378673e-01 -4.02598530e-02 -1.16107440e+00
1.17322572e-01 1.67676494e-01 4.43943053e-01 -8.11208904e-01
-3.87544870e-01 -5.78444719e-01 -7.92450547e-01 5.66645980e-01
5.61244905e-01 -1.10012189e-01 -1.05903912e+00 1.75442445e+00
1.16148721e-02 4.18389469e-01 2.78834492e-01 9.42589104e-01
6.39037013e-01 7.39000797e-01 -3.67580564e-03 -2.00628027e-01
8.54862154e-01 -5.87717831e-01 -6.29044652e-01 -5.23057997e-01
1.32445797e-01 -6.50509357e-01 1.12965059e+00 4.79142100e-01
-9.25510466e-01 -3.04422975e-01 -1.20341849e+00 8.27715024e-02
-3.03232223e-01 3.93551728e-03 7.55257249e-01 5.67932069e-01
-1.09249794e+00 1.23460698e+00 -6.33656263e-01 -3.29931229e-01
2.73428142e-01 2.61187792e-01 -3.26214343e-01 -1.06670000e-01
-1.19701552e+00 9.64417338e-01 5.16399086e-01 3.83906394e-01
-7.36391366e-01 -1.06205809e+00 -4.90712851e-01 2.99417917e-02
5.86594820e-01 -4.53654349e-01 1.18932629e+00 -1.39225566e+00
-1.35172594e+00 5.99308014e-01 1.02219135e-01 -2.14175627e-01
5.16334772e-01 -3.71692814e-02 -5.48999131e-01 1.33350998e-01
-2.41666749e-01 3.44004810e-01 8.16235423e-01 -1.71651912e+00
-1.94665447e-01 -1.80764064e-01 2.25639656e-01 1.65490508e-01
-1.42648041e-01 -2.41158754e-01 -2.88041443e-01 -2.76365131e-01
4.42930013e-01 -1.16663420e+00 -1.42628968e-01 6.16375744e-01
-4.75768566e-01 6.90345168e-01 4.30027813e-01 -6.67891026e-01
5.68507612e-01 -2.04156828e+00 9.26230550e-02 1.29670352e-01
6.60874648e-03 2.44472742e-01 -2.17812136e-01 3.13220203e-01
-4.83155608e-01 2.77038485e-01 -7.59632826e-01 -2.28821233e-01
-2.27931127e-01 2.48749718e-01 -4.44479525e-01 6.34118021e-01
1.41680539e-01 6.03036284e-01 -7.39508748e-01 2.00682089e-01
3.37546676e-01 6.27461970e-01 -9.83326361e-02 3.60076040e-01
-2.96169668e-01 4.76187021e-01 2.14794233e-01 3.91083390e-01
1.11373436e+00 -3.77909362e-01 2.09433675e-01 -5.84679365e-01
-1.92397058e-01 7.51023591e-02 -1.33704329e+00 1.57020164e+00
-6.84990942e-01 6.73413336e-01 2.57888496e-01 -6.99154973e-01
5.51475346e-01 2.02831462e-01 4.76437986e-01 -4.12042826e-01
-3.35202366e-01 2.38926008e-01 -1.73354551e-01 -4.83093649e-01
5.69560945e-01 -6.81727052e-01 6.33353114e-01 4.01910573e-01
-2.28937641e-01 -5.24380565e-01 -6.43823445e-01 -1.22818314e-01
5.69441020e-01 3.64651322e-01 7.39855543e-02 -1.29542917e-01
1.70223311e-01 1.50492981e-01 5.49696922e-01 7.28523433e-01
1.72031894e-02 1.12327135e+00 3.50681603e-01 -1.91736728e-01
-1.12816429e+00 -1.02040803e+00 -2.64430583e-01 5.39121449e-01
1.85661942e-01 2.50850022e-01 -5.69462299e-01 -2.03303128e-01
-4.52403128e-02 6.18948877e-01 -5.84118366e-01 -3.34997684e-01
-2.57376041e-02 -1.18807185e+00 6.64715588e-01 4.98348713e-01
7.80799389e-01 -6.94383740e-01 -4.78264600e-01 1.46736667e-01
-1.74170822e-01 -1.15724647e+00 1.62125707e-01 3.57119918e-01
-6.98686719e-01 -1.23835981e+00 -3.39811623e-01 -4.72703278e-02
4.88632977e-01 6.36618733e-01 9.51249242e-01 2.82121867e-01
-2.77638674e-01 8.56615603e-02 -2.86153376e-01 -2.94032067e-01
-4.03582543e-01 -3.20767254e-01 -4.24262658e-02 4.00701948e-02
-1.64489985e-01 -6.30730510e-01 -6.68854833e-01 2.77737137e-02
-1.50585413e+00 2.17739955e-01 4.73613255e-02 7.95696855e-01
6.14845514e-01 6.24576151e-01 4.00044501e-01 -8.36396694e-01
5.88830337e-02 -3.57263923e-01 -8.56324792e-01 4.40859884e-01
-7.87228584e-01 -1.55097051e-02 6.67734325e-01 -3.91645640e-01
-1.29328251e+00 2.75203437e-01 7.38565773e-02 -5.40830851e-01
-2.39117682e-01 8.00432026e-01 -1.90868765e-01 -5.79185545e-01
8.91122222e-01 6.03996404e-02 -2.22178757e-01 -4.41186249e-01
2.79409617e-01 5.17159820e-01 8.55862200e-01 -4.93826330e-01
9.81775224e-01 6.91204548e-01 3.63789201e-01 -8.26568305e-01
-1.15490162e+00 -5.39980419e-02 -5.21543860e-01 -5.00305474e-01
7.19701827e-01 -1.03039908e+00 -3.48155051e-01 8.73724937e-01
-9.92749870e-01 -7.82590210e-01 -2.51378059e-01 3.96717072e-01
-2.59819835e-01 5.53894460e-01 -3.64599258e-01 -1.06013596e+00
1.86283735e-03 -1.06795454e+00 9.08630908e-01 2.88351536e-01
4.66940343e-01 -9.61398363e-01 -2.13574618e-01 1.73092216e-01
5.33465505e-01 6.32396758e-01 9.81055915e-01 3.39042157e-01
-5.96478462e-01 9.25039202e-02 -7.38756120e-01 3.74916285e-01
4.58154291e-01 3.81387234e-01 -1.47703314e+00 -2.46182576e-01
-1.61625370e-02 -3.75423163e-01 1.09468114e+00 4.31877762e-01
1.32725835e+00 -8.99871141e-02 1.98235586e-01 1.06833363e+00
2.03662133e+00 -2.46137977e-01 9.11823332e-01 4.93428409e-01
8.20744336e-01 7.29637623e-01 1.55366570e-01 3.25981975e-01
-4.53505851e-02 4.65124369e-01 1.01168656e+00 -4.98590887e-01
-3.02754641e-01 7.07914457e-02 2.69501340e-02 -5.42185754e-02
-2.07839966e-01 -4.85483527e-01 -8.12239051e-01 4.34959531e-01
-1.87623000e+00 -9.33013499e-01 -6.12972200e-01 2.64598727e+00
8.74100804e-01 -2.14240715e-01 -2.11461127e-01 1.93608180e-01
6.73895001e-01 2.44920284e-01 -8.99728656e-01 2.03356951e-01
-6.00277662e-01 2.90513903e-01 1.22703850e+00 6.51684105e-01
-8.31819475e-01 7.34784186e-01 6.81757975e+00 2.41112530e-01
-1.41318321e+00 7.14250952e-02 3.40820193e-01 -3.81217934e-02
-4.94588435e-01 1.58693269e-01 -6.64057508e-02 -5.33779105e-03
7.75311053e-01 2.74400264e-01 9.77292538e-01 2.53091395e-01
2.33677685e-01 -3.78604800e-01 -6.52057350e-01 1.07460380e+00
-8.45590308e-02 -1.26699758e+00 -4.65229005e-01 1.83498114e-02
7.69699156e-01 2.09341943e-01 8.10534135e-02 -2.58585066e-01
3.65218699e-01 -1.17331636e+00 7.32623458e-01 8.37542951e-01
1.17909729e+00 -7.22508669e-01 4.81733203e-01 -2.70260591e-03
-9.26182389e-01 -1.64898038e-01 -6.72222197e-01 -6.96574152e-02
1.59852114e-02 9.22433197e-01 -5.19958377e-01 9.77676630e-01
7.18235433e-01 6.22562587e-01 -5.43953121e-01 1.05449355e+00
-4.72630322e-01 5.25652528e-01 -3.43021095e-01 6.43923640e-01
9.10895020e-02 -5.03915131e-01 3.35969776e-01 7.04017043e-01
5.26209831e-01 3.34618926e-01 -1.38603508e-01 1.22925687e+00
-1.42916664e-01 -3.75590980e-01 -5.64826965e-01 -8.57827603e-04
2.43891641e-01 1.17167711e+00 -5.41016102e-01 -6.68939576e-02
-4.21682715e-01 1.09146416e+00 1.56121910e-01 7.29486346e-01
-7.64765263e-01 -3.01752090e-02 9.31069970e-01 4.28451896e-02
1.17005549e-01 -7.57229552e-02 -2.81148762e-01 -1.15400684e+00
-9.20220278e-04 -6.42499745e-01 1.60853654e-01 -1.65811455e+00
-1.40972376e+00 2.92380929e-01 -1.22983001e-01 -1.11878455e+00
3.19607407e-01 -8.19479704e-01 -4.09313738e-01 1.12562215e+00
-2.05835581e+00 -1.12396348e+00 -8.19547236e-01 4.49389279e-01
8.73595029e-02 4.76193786e-01 7.54506350e-01 8.27702582e-02
-6.38527632e-01 8.96827057e-02 3.45432878e-01 -3.99680555e-01
8.21506262e-01 -1.35490668e+00 -7.23403543e-02 1.09408164e+00
-2.07783639e-01 2.31645525e-01 8.49134028e-01 -5.68170965e-01
-1.50705194e+00 -1.41152954e+00 2.69437939e-01 -1.11009374e-01
8.66492450e-01 2.20166147e-01 -1.30693817e+00 4.63187069e-01
6.13715649e-02 2.31471896e-01 5.68550169e-01 -1.81976512e-01
-7.20323205e-01 -3.36444318e-01 -1.23413992e+00 5.38221598e-01
9.81132448e-01 -8.74517560e-01 5.62647991e-02 7.26484239e-01
8.33741546e-01 -4.54398960e-01 -8.49017262e-01 4.46877182e-01
4.61214662e-01 -1.17712235e+00 1.07128179e+00 -5.04822195e-01
5.09149253e-01 -5.87703407e-01 -4.35043722e-01 -1.37039840e+00
-1.63381532e-01 -3.44883353e-01 2.58192986e-01 9.00718451e-01
5.05186737e-01 -5.99657953e-01 4.23749804e-01 1.02164590e+00
-1.94668651e-01 2.51690461e-03 -6.17444873e-01 -1.05373287e+00
8.64695534e-02 -5.85420012e-01 9.98597920e-01 1.19624078e+00
-5.46054244e-01 -2.33850405e-01 -5.00387907e-01 1.00794554e+00
7.23917723e-01 2.15124190e-01 3.24464768e-01 -1.27808690e+00
-4.50004727e-01 -4.19708163e-01 1.39126763e-01 -4.81712818e-01
2.38618076e-01 -5.94182372e-01 4.14835334e-01 -1.47097576e+00
2.38041237e-01 -5.47729313e-01 -2.68729389e-01 7.94360101e-01
-2.98446000e-01 3.01766306e-01 1.78019509e-01 2.93650717e-01
3.06131035e-01 2.82711446e-01 1.10297537e+00 -1.77986890e-01
-2.53840536e-01 -2.63859540e-01 -5.48462629e-01 4.76816982e-01
9.68394041e-01 -5.93076050e-01 -3.36759448e-01 -5.87998271e-01
8.19717169e-01 2.19843224e-01 9.03798938e-01 -1.02937424e+00
8.14519823e-02 -4.71891284e-01 3.77217472e-01 -2.61829793e-01
3.47282141e-01 -9.68507111e-01 9.25214946e-01 1.33242533e-01
-2.60036379e-01 -4.80149984e-01 4.16259259e-01 6.09955132e-01
2.68219203e-01 -4.80552584e-01 1.08349979e+00 -8.95154327e-02
-4.25763577e-01 3.92533720e-01 -8.68911743e-02 -5.33072591e-01
4.23971534e-01 -1.96833059e-01 -3.71151626e-01 -5.24310112e-01
-4.74044651e-01 -3.11086297e-01 9.97161031e-01 -2.01834857e-01
2.90362567e-01 -1.09517145e+00 -5.65986574e-01 -1.81832492e-01
1.97985634e-01 3.50382209e-01 3.83087188e-01 6.64501190e-01
-6.47281647e-01 -2.22194940e-01 9.62680280e-02 -3.70711178e-01
-8.56476009e-01 4.43253666e-01 9.67977405e-01 1.77023426e-01
-6.29404187e-01 8.25064361e-01 -4.94756959e-02 -5.13029099e-01
-1.03150696e-01 -1.38982475e-01 2.56569773e-01 -1.53642878e-01
3.70719433e-01 1.69318601e-01 3.83353025e-01 -5.19642353e-01
-1.98888534e-04 6.32113695e-01 6.43197834e-01 -1.35684818e-01
1.63950896e+00 -1.31090909e-01 -4.77702200e-01 3.00825149e-01
9.34522212e-01 8.58502463e-02 -1.81514323e+00 -2.94936836e-01
-5.57441413e-01 -6.59329176e-01 5.75191081e-01 -1.09213555e+00
-1.60799968e+00 8.22999239e-01 6.06412053e-01 3.91749948e-01
1.50617671e+00 -4.14121836e-01 6.70743883e-01 5.33929229e-01
1.20531820e-01 -9.58388686e-01 -2.47571245e-01 3.13934863e-01
8.44433784e-01 -1.38802910e+00 2.98752099e-01 -4.81399328e-01
-4.77016330e-01 1.38268113e+00 3.47984105e-01 4.08027433e-02
5.68521798e-01 -3.93424369e-02 1.15898639e-01 -2.31653914e-01
-6.68973699e-02 -3.06932658e-01 1.08138219e-01 7.13529944e-01
1.12354510e-01 1.51174024e-01 4.52052921e-01 2.41462782e-01
-4.52630366e-06 -2.07615584e-01 1.00716901e+00 5.14876246e-01
-3.56654376e-01 -8.31086516e-01 -7.04730451e-01 2.99895674e-01
-6.96033686e-02 -3.80849600e-01 -3.80765557e-01 6.37440205e-01
1.83870971e-01 1.07718277e+00 -6.49846867e-02 -3.19996893e-01
1.41068518e-01 1.14170341e-02 4.68587667e-01 -5.00582159e-01
-3.11980993e-01 1.21606275e-01 7.50442743e-02 -5.58601856e-01
-7.91402459e-01 -5.65898597e-01 -1.14582157e+00 -3.80936563e-01
-6.30755067e-01 -3.10472518e-01 8.38093281e-01 9.13649440e-01
8.01177099e-02 5.99420130e-01 6.81497097e-01 -8.69171560e-01
-4.39410239e-01 -6.92582130e-01 -8.62260163e-01 2.38942400e-01
7.50434697e-01 -5.84374189e-01 -7.42599308e-01 1.10044055e-01] | [10.22118091583252, -2.252643346786499] |
383bcfca-5585-4fb0-8bd2-43781944c147 | efficient-deep-learning-a-survey-on-making | 2106.08962 | null | https://arxiv.org/abs/2106.08962v2 | https://arxiv.org/pdf/2106.08962v2.pdf | Efficient Deep Learning: A Survey on Making Deep Learning Models Smaller, Faster, and Better | Deep Learning has revolutionized the fields of computer vision, natural language understanding, speech recognition, information retrieval and more. However, with the progressive improvements in deep learning models, their number of parameters, latency, resources required to train, etc. have all have increased significantly. Consequently, it has become important to pay attention to these footprint metrics of a model as well, not just its quality. We present and motivate the problem of efficiency in deep learning, followed by a thorough survey of the five core areas of model efficiency (spanning modeling techniques, infrastructure, and hardware) and the seminal work there. We also present an experiment-based guide along with code, for practitioners to optimize their model training and deployment. We believe this is the first comprehensive survey in the efficient deep learning space that covers the landscape of model efficiency from modeling techniques to hardware support. Our hope is that this survey would provide the reader with the mental model and the necessary understanding of the field to apply generic efficiency techniques to immediately get significant improvements, and also equip them with ideas for further research and experimentation to achieve additional gains. | ['Gaurav Menghani'] | 2021-06-16 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [-1.34780973e-01 -1.09665170e-01 -4.43063676e-01 -6.03859663e-01
-6.05776966e-01 -5.48039734e-01 3.14428449e-01 -6.85515702e-02
-6.62241101e-01 1.86571658e-01 -9.38306972e-02 -7.32820809e-01
-8.38796869e-02 -5.33302367e-01 -4.58527625e-01 -4.74655360e-01
-1.64871946e-01 5.51854551e-01 2.11441070e-01 -2.51086596e-02
2.59690851e-01 9.23465312e-01 -1.47947013e+00 9.71889123e-02
2.14349970e-01 8.78787220e-01 3.26851338e-01 8.41186285e-01
-1.17657810e-01 1.09360504e+00 -5.63387215e-01 -4.04797047e-01
-2.64869947e-02 7.78427720e-02 -1.22807574e+00 6.76556351e-03
4.57666397e-01 -5.33276141e-01 -5.14554679e-01 5.68587422e-01
5.95326722e-01 1.41168991e-02 1.57309651e-01 -1.23367178e+00
-3.77658941e-02 4.06543881e-01 -4.27200794e-01 5.63819528e-01
-4.82957959e-01 1.79604173e-01 8.11712146e-01 -7.90496349e-01
3.28081071e-01 1.11415648e+00 7.25170910e-01 7.55350530e-01
-6.46159947e-01 -6.23623371e-01 2.87650079e-01 4.96828169e-01
-1.18865156e+00 -9.14371908e-01 4.53814358e-01 -1.46742553e-01
1.52501440e+00 2.10601151e-01 7.87565470e-01 8.87185037e-01
2.66161598e-02 1.08182037e+00 6.03461623e-01 -6.71471179e-01
2.19935745e-01 2.16133729e-01 4.93790656e-01 9.65253651e-01
1.37836859e-01 1.82163924e-01 -5.45714796e-01 2.02748831e-02
7.92802811e-01 -3.66562083e-02 2.20099688e-01 -4.29997087e-01
-7.44181156e-01 7.76742458e-01 2.90976316e-01 3.47758263e-01
-1.13590494e-01 5.83095193e-01 6.07683003e-01 2.86652714e-01
1.58111170e-01 4.03987169e-01 -8.54351044e-01 -6.39179945e-01
-9.85243022e-01 4.17893350e-01 9.66838956e-01 1.10833752e+00
6.65065527e-01 3.98319632e-01 5.49406290e-01 9.40362632e-01
2.77087569e-01 2.08985731e-01 4.03344780e-01 -1.26142192e+00
2.04223529e-01 3.52493316e-01 -1.68821841e-01 -7.72781730e-01
-6.19559526e-01 -7.76956677e-01 -7.20511734e-01 1.46893054e-01
1.07634038e-01 -3.69561106e-01 -7.61782706e-01 1.57227457e+00
5.47759086e-02 2.44835749e-01 -2.52878100e-01 5.11524081e-01
7.43646741e-01 5.32956898e-01 -1.30307665e-02 9.01968107e-02
1.36481380e+00 -1.22126806e+00 -3.80715072e-01 -6.45043850e-01
1.15278077e+00 -9.10200655e-01 9.62543488e-01 4.11395848e-01
-1.42965961e+00 -4.43930566e-01 -1.05533147e+00 -2.46739700e-01
-2.45808810e-01 -5.86569682e-02 1.23840201e+00 7.91402340e-01
-1.21581876e+00 5.09772480e-01 -1.54505944e+00 -5.83848178e-01
5.57710469e-01 6.63565278e-01 -5.24261314e-03 -1.24152869e-01
-7.72457302e-01 1.16706800e+00 1.66540831e-01 -1.03861108e-01
-8.64203632e-01 -7.43305504e-01 -6.42396390e-01 2.18051985e-01
1.01889811e-01 -6.33960485e-01 1.93573308e+00 -6.65780008e-01
-1.17555523e+00 7.84013867e-01 -2.84793109e-01 -6.89724803e-01
1.11500256e-01 -6.16268888e-02 -1.92363858e-01 -1.80623829e-01
-5.03487110e-01 9.36640680e-01 2.68110931e-01 -9.13521290e-01
-9.31829154e-01 -3.95944476e-01 3.46165478e-01 1.19178005e-01
-6.40116930e-01 3.91296446e-01 -7.59269953e-01 -2.42391855e-01
-1.28355116e-01 -1.00665665e+00 -4.94756401e-01 4.54206727e-02
8.94443914e-02 -2.36381501e-01 7.44136631e-01 -2.57259935e-01
1.33696187e+00 -2.00447679e+00 -2.87090570e-01 -6.48284052e-03
5.27729332e-01 4.13605779e-01 -2.26516932e-01 5.44878304e-01
1.40863866e-01 1.23269312e-01 1.33653805e-01 -4.52509314e-01
-6.02124371e-02 4.44266289e-01 -1.65874273e-01 3.05380076e-01
2.23362595e-02 9.13083136e-01 -6.97844505e-01 -3.68123114e-01
4.41871822e-01 7.68988967e-01 -6.98849916e-01 -8.12414940e-03
1.14528283e-01 -1.27270281e-01 -4.10205990e-01 3.70190352e-01
6.02760613e-01 -5.80762148e-01 5.18579222e-02 -1.24612324e-01
-2.08860114e-01 6.17822766e-01 -9.91993368e-01 1.58925796e+00
-7.05697238e-01 1.06290352e+00 2.95916349e-01 -1.12509203e+00
6.30961120e-01 3.39849174e-01 5.18023431e-01 -6.66118979e-01
1.25865072e-01 2.49547467e-01 4.53887992e-02 -2.64082849e-01
4.50260073e-01 1.44732818e-01 1.90131351e-01 4.80471253e-01
4.81526591e-02 -4.97635268e-02 5.63869923e-02 2.26281956e-01
1.06781089e+00 -3.42890203e-01 5.84093481e-02 -2.29836583e-01
1.43660843e-01 1.60076663e-01 9.29322243e-02 8.30222607e-01
-2.83699602e-01 9.44886804e-02 1.77000165e-01 -1.02339041e+00
-1.30332470e+00 -6.68452322e-01 -1.78841963e-01 1.22973132e+00
-2.73799628e-01 -6.14854217e-01 -9.92634833e-01 -2.62172133e-01
-3.83130789e-01 4.00440276e-01 -3.06059122e-01 -1.19568273e-01
-9.11144435e-01 -9.22605693e-01 6.23316228e-01 9.36216474e-01
5.40604472e-01 -9.05842304e-01 -1.07340765e+00 1.74947888e-01
5.12172747e-03 -1.12553537e+00 -9.95859802e-02 6.09118760e-01
-1.42133963e+00 -7.75997937e-01 -4.13259983e-01 -1.09026659e+00
4.09153283e-01 6.00099623e-01 1.39220965e+00 7.24570572e-01
-3.76578540e-01 2.96250612e-01 -2.40392953e-01 -6.09501719e-01
-2.09493473e-01 5.01369655e-01 1.45269639e-03 -8.01279426e-01
7.01000631e-01 -5.18099546e-01 -6.05446458e-01 2.38000173e-02
-7.94398367e-01 2.47295678e-01 6.01167023e-01 6.88181698e-01
1.43313572e-01 1.16623335e-01 1.28256649e-01 -8.04880381e-01
4.54866618e-01 -1.54969275e-01 -4.79340017e-01 1.17337376e-01
-9.12112355e-01 7.08702486e-03 3.62265170e-01 -7.30836615e-02
-6.76633656e-01 -8.31867531e-02 -5.92363834e-01 -1.24542810e-01
-4.37017232e-02 4.81851041e-01 1.76822804e-02 -3.46276313e-01
7.85617828e-01 1.86236963e-01 -2.09986180e-01 -6.88444495e-01
2.31794104e-01 7.79581368e-01 3.32727343e-01 -3.93750548e-01
3.84345621e-01 3.44055325e-01 -1.05443306e-01 -1.09566832e+00
-6.47525191e-01 -5.43151200e-01 -5.30753613e-01 7.07073286e-02
3.24838787e-01 -8.84247363e-01 -6.23451710e-01 4.52584893e-01
-1.30110371e+00 -5.25057912e-01 -3.14284079e-02 3.36942554e-01
-5.56424856e-01 1.74169078e-01 -9.20400381e-01 -6.44403279e-01
-4.21093792e-01 -1.40757239e+00 8.93109858e-01 1.14807457e-01
-1.78755775e-01 -1.23340988e+00 -7.09780604e-02 4.50617343e-01
6.86336637e-01 -4.29093003e-01 1.04704809e+00 -5.10127664e-01
-8.52239788e-01 -4.83450055e-01 -2.98007071e-01 3.31160009e-01
-2.68597901e-01 1.34753168e-01 -1.19420505e+00 -5.31311691e-01
-1.38863727e-01 -5.21426320e-01 7.62956321e-01 6.01805389e-01
1.48565376e+00 -3.31774116e-01 -4.99438703e-01 7.33042598e-01
1.39142525e+00 4.03576642e-01 4.39779073e-01 5.23238719e-01
6.14353538e-01 3.92496109e-01 1.14270173e-01 4.31298584e-01
4.26919043e-01 7.65706241e-01 3.79221886e-01 -5.54809213e-01
-1.10637575e-01 1.85584962e-01 6.13609292e-02 1.24476063e+00
-1.28300041e-01 -3.22764009e-01 -1.21928954e+00 4.26679313e-01
-1.65809166e+00 -9.09482956e-01 1.92565471e-01 2.04336977e+00
5.57160378e-01 2.41099209e-01 1.67606190e-01 2.81475961e-01
2.52964169e-01 9.68749076e-02 -6.12444639e-01 -6.37289822e-01
3.87761891e-01 4.14405167e-01 6.26402140e-01 6.68841064e-01
-8.98028255e-01 1.17252433e+00 7.95592070e+00 8.43577921e-01
-1.27188885e+00 -6.83696941e-03 7.90120721e-01 -2.91854531e-01
-1.16244294e-01 -1.32449925e-01 -9.37168062e-01 -6.77811503e-02
1.44312656e+00 -2.56612957e-01 6.78711832e-01 1.18469930e+00
3.06134850e-01 2.49605313e-01 -1.11388898e+00 1.14337134e+00
-2.14410245e-01 -1.58408928e+00 -1.38698652e-01 2.04786777e-01
4.61035728e-01 7.21382082e-01 1.28100321e-01 4.96222794e-01
4.21033412e-01 -1.13214493e+00 3.68734896e-01 4.56950292e-02
4.86715794e-01 -1.05302501e+00 6.40643418e-01 4.86338615e-01
-1.13254547e+00 -1.23688849e-02 -6.48627996e-01 -3.77731472e-01
-9.38100964e-02 2.86313891e-01 -9.03292596e-01 1.06402840e-02
7.86108732e-01 4.93274510e-01 -4.88873124e-01 1.16866851e+00
3.91350031e-01 6.56362891e-01 -3.26127559e-01 -2.62203097e-01
4.81253177e-01 4.32788253e-01 -1.30821645e-01 1.40261710e+00
9.12224408e-03 -5.31868190e-02 2.11492460e-02 2.42757469e-01
-2.19760701e-01 1.04553429e-02 -3.09291333e-01 -1.89338505e-01
7.83274889e-01 1.18232250e+00 -7.45360911e-01 -5.85237801e-01
-7.96723962e-01 5.83337843e-01 4.06261712e-01 2.05117613e-01
-8.82997155e-01 -2.85846859e-01 8.75471830e-01 7.75420815e-02
-4.88251969e-02 -6.05775893e-01 -5.96966624e-01 -9.55526650e-01
-2.44817406e-01 -1.13575363e+00 1.54230043e-01 -5.97175837e-01
-6.17245197e-01 7.65590429e-01 5.47231883e-02 -6.71427071e-01
-3.11805457e-01 -7.73594558e-01 -3.42993259e-01 6.61205113e-01
-1.62985539e+00 -7.74457812e-01 -3.93494695e-01 2.53955483e-01
8.73752117e-01 -1.42736599e-01 9.50778604e-01 5.15278876e-01
-6.23669028e-01 8.96511018e-01 2.17075929e-01 1.26064256e-01
1.72605455e-01 -8.05220783e-01 1.03004551e+00 5.66102445e-01
4.31439400e-01 7.91821897e-01 6.96334183e-01 2.22952291e-02
-1.46472573e+00 -8.62207413e-01 9.89872456e-01 -5.41494668e-01
5.72138071e-01 -2.68451869e-01 -5.81552505e-01 1.01329267e+00
2.13150725e-01 -2.47464150e-01 7.01298475e-01 4.50203806e-01
-2.01710656e-01 -2.61331350e-01 -7.65575230e-01 6.94295526e-01
9.68160570e-01 -3.41729999e-01 5.47511270e-03 3.85154903e-01
7.22458541e-01 -5.40092349e-01 -6.57408357e-01 3.92922610e-01
7.41789401e-01 -1.09267974e+00 8.46687555e-01 -7.44349778e-01
1.66249946e-01 2.54680395e-01 -1.24858513e-01 -7.73813307e-01
-4.26050723e-01 -3.82071912e-01 -3.26600581e-01 5.70587277e-01
4.94998753e-01 -2.72201955e-01 1.48087323e+00 8.15181911e-01
-4.21633989e-01 -1.22529948e+00 -6.89877748e-01 -7.01623797e-01
2.52881318e-01 -9.55332696e-01 7.17963099e-01 7.64394820e-01
-1.97924435e-01 4.28498179e-01 -2.40126699e-01 -1.03764333e-01
4.35882568e-01 -2.72097319e-01 9.60591912e-01 -1.17548430e+00
-3.11896086e-01 -6.35516465e-01 -5.32932758e-01 -1.59706426e+00
-3.31145555e-01 -6.63348198e-01 -3.09559733e-01 -1.63156974e+00
3.59811544e-01 -5.79564035e-01 -2.28508443e-01 5.91889799e-01
3.84532481e-01 2.37521023e-01 2.39367068e-01 3.23522866e-01
-6.45488977e-01 1.71315894e-01 1.11850214e+00 1.64738577e-02
-5.85899688e-02 1.27187476e-01 -9.02226865e-01 9.20692205e-01
8.73340189e-01 -4.71188962e-01 -6.60883963e-01 -1.11668754e+00
1.39309719e-01 -1.62661165e-01 2.34331310e-01 -1.24002743e+00
3.92943174e-01 -2.07820624e-01 3.70968729e-01 -4.55664963e-01
5.80325842e-01 -9.64676559e-01 -1.54312521e-01 6.70030713e-01
-4.37696308e-01 5.60914278e-01 4.58536386e-01 -6.06289171e-02
-1.46573648e-01 -3.02494824e-01 1.06813145e+00 -3.16590607e-01
-9.82050240e-01 5.80158174e-01 -5.26408136e-01 -1.06570639e-01
8.53802919e-01 -3.12855154e-01 -1.79522723e-01 -4.11705881e-01
-5.92124462e-01 1.00025699e-01 5.24630547e-01 3.62521678e-01
4.04207617e-01 -9.11282480e-01 -3.75322670e-01 1.72577769e-01
-6.31153136e-02 3.08571421e-02 3.08809519e-01 6.00694835e-01
-8.55595112e-01 8.93434942e-01 -6.36977181e-02 -5.01695395e-01
-1.46077442e+00 4.47382212e-01 5.72432339e-01 -2.80090660e-01
-4.96611327e-01 1.15567422e+00 -4.61274851e-03 -3.11591506e-01
8.18518817e-01 -3.34608108e-01 6.57056496e-02 -4.31141883e-01
7.52911389e-01 5.30206740e-01 4.45605755e-01 -1.96346790e-01
-4.99648303e-01 3.66440803e-01 -4.24624324e-01 1.27162738e-02
1.28839242e+00 4.25597616e-02 1.16267607e-01 2.58376688e-01
1.23706830e+00 -4.09053355e-01 -1.13663876e+00 -1.37669146e-01
8.89229104e-02 -1.63776934e-01 5.10312557e-01 -6.93686724e-01
-1.29555631e+00 1.38866627e+00 8.00657988e-01 9.80543345e-02
1.20180833e+00 -5.67004038e-03 9.88774240e-01 7.15549111e-01
4.96065915e-01 -9.55867350e-01 1.19766772e-01 8.12056184e-01
4.07924831e-01 -9.22317982e-01 -5.88362776e-02 -7.42902905e-02
-2.02964649e-01 1.17874777e+00 6.04621530e-01 -1.03426483e-02
9.53436196e-01 7.96960294e-01 2.80953556e-01 -2.73901165e-01
-1.16910636e+00 2.96713207e-02 -2.33882323e-01 6.56998754e-01
9.19375598e-01 -1.45878270e-01 1.94555879e-01 5.67834005e-02
-5.33786833e-01 1.77849397e-01 2.34179944e-01 9.84710813e-01
-8.73959780e-01 -1.39639425e+00 4.91858460e-02 4.92499083e-01
-3.88359040e-01 -2.36469388e-01 -1.54032290e-01 8.94181192e-01
7.37679610e-03 9.94183660e-01 3.30783367e-01 -4.48881030e-01
3.20815891e-01 -9.18883383e-02 6.91606343e-01 -5.84512949e-01
-5.15275359e-01 -4.11885343e-02 1.67625099e-01 -5.16348481e-01
-1.20436892e-01 -2.40277186e-01 -1.18234193e+00 -1.20963979e+00
-3.92359763e-01 1.72796384e-01 1.11864829e+00 8.89007449e-01
5.02692103e-01 4.17466730e-01 4.41361725e-01 -8.28252494e-01
-6.10144973e-01 -8.92063439e-01 -3.84601861e-01 -3.31632078e-01
3.36972445e-01 -2.97588319e-01 -1.60733894e-01 -4.46045436e-02] | [8.69515609741211, 2.934115409851074] |
ea4c264c-724b-4886-a1ad-33edfcf4ac79 | robust-model-predictive-control-with-1 | 2008.04980 | null | https://arxiv.org/abs/2008.04980v4 | https://arxiv.org/pdf/2008.04980v4.pdf | Robust Output Feedback MPC with Reduced Conservatism under Ellipsoidal Uncertainty | Robust design of autonomous systems under uncertainty is an important yet challenging problem. This work proposes a robust controller that consists of a state estimator and a tube based predictive control law. The class of linear systems under ellipsoidal uncertainty is considered. In contrast to existing approaches based on polytopic sets, the constraint tightening is directly computed from the ellipsoidal sets of disturbances without over-approximation, thus leading to less conservative bounds. Conditions to guarantee robust constraint satisfaction and robust stability are presented. Further, by avoiding the usage of Minkowski sum in set computation, the proposed approach can also scale up to high-dimensional systems. The results are illustrated by examples. | ['Katherine Driggs-Campbell', 'Junyi Geng', 'Tianchen Ji'] | 2020-08-11 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [ 1.32111341e-01 6.03433669e-01 -6.50200248e-02 2.04928681e-01
-3.26059759e-01 -8.11162293e-01 4.99280065e-01 2.36483693e-01
-2.67037928e-01 1.20581031e+00 -4.28764313e-01 -1.15987405e-01
-6.14096880e-01 -6.68867767e-01 -7.34908164e-01 -8.46134782e-01
-2.64594518e-02 3.83875370e-01 1.68399379e-01 -2.82363325e-01
2.03904778e-01 6.34603918e-01 -1.42349708e+00 -8.00315619e-01
1.05477488e+00 1.21130693e+00 -1.12525500e-01 2.66284078e-01
5.63568234e-01 -9.19544175e-02 -4.99933600e-01 1.26326561e-01
6.67239428e-01 -1.48320660e-01 -1.42446399e-01 6.03267670e-01
-5.60668763e-03 -4.96255420e-02 6.45254627e-02 1.34796691e+00
4.04697895e-01 5.21165192e-01 8.25055599e-01 -1.23451030e+00
-3.25522572e-01 1.95539504e-01 -1.42134637e-01 -3.90537739e-01
1.76770329e-01 -1.90264151e-01 3.92941743e-01 -9.26043808e-01
4.74671334e-01 1.18029559e+00 4.56831098e-01 3.04119527e-01
-1.04731381e+00 -2.45114639e-01 2.46769264e-01 -9.50012654e-02
-1.71792114e+00 -3.74399722e-01 4.19130355e-01 -5.45399964e-01
4.79569525e-01 4.07609731e-01 5.72286189e-01 2.85689026e-01
3.99780512e-01 -1.14658298e-02 9.54278708e-01 -3.03520709e-01
6.12164199e-01 3.39909762e-01 -1.60453320e-01 4.43106115e-01
1.16055989e+00 2.00014815e-01 4.14933503e-01 -1.32969424e-01
7.67726541e-01 -1.68614417e-01 -5.44873059e-01 -9.81428325e-01
-9.67699111e-01 8.09398115e-01 1.51641205e-01 -1.59470588e-02
-1.69274509e-01 -1.16279304e-01 2.83108860e-01 1.40007824e-01
2.57571518e-01 6.45865023e-01 -1.32186517e-01 2.25032777e-01
-4.51234341e-01 4.40405935e-01 1.07471371e+00 1.50043929e+00
1.90471396e-01 5.83128929e-01 1.23396754e-01 4.17153269e-01
5.27320147e-01 7.96981037e-01 -1.07032862e-02 -8.60537767e-01
2.39731357e-01 7.07257032e-01 7.28609979e-01 -9.88183081e-01
-3.82542044e-01 -2.94531912e-01 -6.69631124e-01 7.18852282e-01
5.19244492e-01 -4.85547513e-01 -7.33852983e-01 1.38111985e+00
5.73792040e-01 -2.25512668e-01 2.66450375e-01 9.65997040e-01
-1.07299671e-01 6.28404438e-01 -7.14458942e-01 -8.99101257e-01
9.98976707e-01 -4.20255423e-01 -1.18006599e+00 6.50701523e-02
1.27525449e-01 -4.27175194e-01 3.18164676e-01 6.30091906e-01
-1.06801295e+00 -1.42485067e-01 -1.34490621e+00 4.77697998e-01
-4.90696758e-01 -7.69912824e-03 -2.36510679e-01 6.03498757e-01
-8.02527964e-01 4.79906201e-01 -4.88681555e-01 -4.00008112e-01
-2.50363559e-01 4.82024789e-01 -2.87945658e-01 2.48777106e-01
-1.06452644e+00 1.46893883e+00 8.70365977e-01 6.23330534e-01
-4.61969644e-01 -4.03519690e-01 -1.15840888e+00 -1.58142745e-01
1.00380743e+00 -3.20293009e-01 9.28528488e-01 -2.24932611e-01
-2.11143541e+00 -4.64539975e-02 1.43657446e-01 -4.80107784e-01
7.28685737e-01 -1.59267217e-01 -2.98745930e-02 7.02901334e-02
-3.60768914e-01 -2.87153810e-01 1.04031122e+00 -1.33841908e+00
-2.89380848e-01 -1.81273893e-01 1.08675987e-01 3.19405735e-01
-1.24912322e-01 -4.30155620e-02 -1.55488253e-01 -3.10011595e-01
2.66021430e-01 -1.18125427e+00 -7.55396545e-01 1.42984524e-01
-5.23969352e-01 4.61427541e-03 9.09337521e-01 -1.37156188e-01
1.27887917e+00 -1.69302988e+00 5.83153725e-01 6.08351767e-01
-9.10527334e-02 3.32929254e-01 5.44646323e-01 6.38869762e-01
1.94158047e-01 -4.28451635e-02 -4.80477095e-01 7.28347450e-02
2.39383742e-01 1.86614782e-01 -3.32366884e-01 1.03966963e+00
3.64628941e-01 1.05258189e-01 -6.17700756e-01 -2.95655876e-01
5.74722290e-01 2.82831818e-01 -1.91508129e-01 -2.13456616e-01
-1.40582517e-01 1.64849818e-01 -6.48360372e-01 4.54660296e-01
8.30904663e-01 3.68225724e-01 1.46297455e-01 1.68082699e-01
-6.60722971e-01 -4.88488734e-01 -1.94321692e+00 9.32252705e-01
-4.11913991e-01 2.72430897e-01 8.69627416e-01 -1.00094008e+00
1.26250935e+00 5.45637190e-01 4.75965977e-01 -1.39102697e-01
7.02324927e-01 3.31540585e-01 -1.24533191e-01 -2.40274683e-01
2.92658210e-01 -2.42065936e-01 -2.46643320e-01 -8.70916694e-02
-4.24384087e-01 -9.00997519e-01 3.65456551e-01 -2.19011024e-01
4.61259633e-01 2.28116624e-02 8.86763036e-01 -8.00864041e-01
1.00065446e+00 -1.74992729e-03 9.17834461e-01 -3.39778326e-02
-2.29825452e-01 2.80978680e-01 5.02080858e-01 3.19621116e-02
-1.08113432e+00 -6.68399215e-01 -5.83912849e-01 7.04559684e-02
6.15827024e-01 -7.07125515e-02 -6.25461936e-01 -1.70983061e-01
4.89984512e-01 5.94293535e-01 -6.75467789e-01 -2.76914626e-01
-3.62583160e-01 -2.06244335e-01 6.47774711e-02 2.80140966e-01
-1.06532685e-01 -1.07503504e-01 -7.19353378e-01 3.13383579e-01
5.42915821e-01 -1.06261420e+00 -4.76672828e-01 -1.41372094e-02
-7.96792030e-01 -1.31979215e+00 -6.03826284e-01 -4.62937713e-01
9.96899903e-01 -2.09772512e-02 3.52712214e-01 -2.58077145e-01
-1.60123408e-01 3.95628333e-01 -1.09595455e-01 -9.89683449e-01
-2.55871803e-01 -5.76669753e-01 8.13768446e-01 1.18917115e-01
-4.39018190e-01 4.26772870e-02 -2.31009588e-01 7.31602192e-01
-7.83543169e-01 -4.33272749e-01 2.68769950e-01 6.96376741e-01
9.56934154e-01 3.38708580e-01 7.16727197e-01 -3.47790807e-01
6.51218295e-01 -3.89150858e-01 -1.61376941e+00 4.32076082e-02
-6.99693441e-01 3.20537724e-02 9.91603017e-01 -4.46723491e-01
-1.01318979e+00 4.23209995e-01 6.33767426e-01 -6.18041515e-01
1.56514838e-01 2.49852553e-01 -3.55784118e-01 -4.47977245e-01
4.70869064e-01 -7.55442381e-02 5.61046362e-01 -2.05289900e-01
3.38972330e-01 5.33854604e-01 2.88407058e-01 -5.02802312e-01
1.14497972e+00 4.68865246e-01 6.95413053e-01 -9.60703075e-01
-3.88771802e-01 -4.09125775e-01 -8.19800615e-01 -4.81571227e-01
5.36278129e-01 -5.05974472e-01 -9.25931275e-01 -9.50107872e-02
-9.24331307e-01 1.26625031e-01 -4.88695025e-01 6.40629947e-01
-9.83033597e-01 2.78453201e-01 -1.22994564e-01 -1.57051456e+00
-1.65442929e-01 -1.25918996e+00 7.24259973e-01 1.86749384e-01
-4.09401953e-02 -9.77469563e-01 1.25584275e-01 -2.82147735e-01
3.48280936e-01 9.27586555e-01 1.25035554e-01 -3.36222827e-01
-4.01040554e-01 -7.31467247e-01 2.17940509e-01 3.05318475e-01
-3.99136320e-02 2.28314877e-01 -3.31154495e-01 -7.00050175e-01
3.31604153e-01 -5.85790761e-02 1.53787911e-01 4.24979061e-01
2.85092920e-01 -4.63479519e-01 -6.04164243e-01 2.88292497e-01
1.84267235e+00 5.39015889e-01 1.80625170e-01 3.03107798e-01
3.72221977e-01 8.14671576e-01 1.13237858e+00 7.44662404e-01
3.24079692e-02 7.49273241e-01 7.63058662e-01 4.03613150e-01
5.68269789e-01 2.91588396e-01 2.45617986e-01 3.86932433e-01
1.47018265e-02 -2.10469410e-01 -6.73769951e-01 6.85325682e-01
-2.02610207e+00 -6.48503363e-01 -2.75761396e-01 2.58726406e+00
5.61326087e-01 -2.14029383e-02 4.96359877e-02 3.79359454e-01
1.13930392e+00 -3.32478970e-01 -5.58902264e-01 -8.71708930e-01
1.07537076e-01 -4.27183568e-01 9.57032382e-01 8.35354865e-01
-1.00345612e+00 3.10910016e-01 6.40800428e+00 5.22772789e-01
-8.54059041e-01 -3.03089559e-01 -2.61590987e-01 -1.19467661e-01
1.33377239e-01 -3.10605299e-02 -1.05092764e+00 5.16585827e-01
7.40545809e-01 -8.45976233e-01 3.04918855e-01 9.20269132e-01
6.22435808e-01 -3.15120220e-01 -7.49461472e-01 4.99143898e-01
4.58483920e-02 -7.49002278e-01 -3.24143380e-01 1.20738193e-01
9.33039546e-01 -5.85923612e-01 -4.28636819e-02 -5.48112355e-02
1.03091009e-01 -7.10233152e-01 8.44243109e-01 7.31568813e-01
7.16157794e-01 -9.83466089e-01 7.96652019e-01 4.60028440e-01
-1.28200555e+00 -3.49311173e-01 -4.61486161e-01 -1.64457843e-01
5.50588310e-01 6.45311773e-01 -8.29289317e-01 9.42060947e-01
2.16823474e-01 4.47221845e-01 8.03736895e-02 1.50587630e+00
-1.23488223e-02 5.29559366e-02 -7.43665874e-01 -3.62944871e-01
3.94697972e-02 -6.67105615e-01 1.14666283e+00 8.01028907e-01
5.82815349e-01 2.21131563e-01 4.36945081e-01 8.27677965e-01
5.32663524e-01 2.05920726e-01 -1.13834572e+00 2.90015250e-01
4.67607051e-01 1.06344557e+00 -5.79910278e-01 -1.96388930e-01
-1.93201840e-01 3.75094324e-01 -1.23009853e-01 1.94180503e-01
-7.56017625e-01 -8.53543758e-01 6.72994375e-01 1.76002514e-02
2.68075049e-01 -5.02279997e-01 -4.42030370e-01 -7.57394254e-01
7.32394457e-02 -3.47528577e-01 1.20776586e-01 -1.37553871e-01
-8.16441357e-01 4.73331153e-01 4.00506556e-01 -1.86659884e+00
-4.21978056e-01 -7.27156281e-01 -6.23332500e-01 8.46590579e-01
-1.24586439e+00 -5.70409119e-01 6.83464408e-02 3.11479390e-01
3.66367400e-01 5.09667024e-02 5.50455272e-01 -1.02863215e-01
-9.42546129e-01 6.45329878e-02 7.14554071e-01 -5.66562057e-01
5.77896059e-01 -1.37913907e+00 -5.17369032e-01 1.08730340e+00
-9.72751737e-01 6.61356688e-01 1.28210866e+00 -6.37258112e-01
-1.67781615e+00 -1.26868021e+00 3.29776913e-01 -2.70344883e-01
9.01479781e-01 -2.56808877e-01 -8.01158667e-01 5.01309395e-01
2.04484507e-01 2.54951537e-01 -1.44046634e-01 -5.84406674e-01
1.76780850e-01 -3.63657773e-02 -1.39461982e+00 5.76611221e-01
5.81164956e-01 1.90260619e-01 -3.61795127e-01 2.79013216e-01
6.86567128e-01 -6.55096233e-01 -1.25879109e+00 6.62269115e-01
4.00100678e-01 -1.58540577e-01 6.72360122e-01 -1.52675316e-01
-3.49386781e-01 -9.28423047e-01 -2.27477904e-02 -1.60972714e+00
-7.34974816e-02 -1.03828788e+00 -3.08586597e-01 9.43761051e-01
3.09117496e-01 -9.06966150e-01 2.89546937e-01 6.53041244e-01
-3.93373966e-01 -7.80584931e-01 -1.07172275e+00 -1.40180457e+00
1.27210140e-01 -1.19233064e-01 2.10335478e-01 6.31673276e-01
6.93126023e-01 -4.57822829e-02 -2.46308878e-01 6.24769330e-01
8.59889150e-01 -1.33136243e-01 4.65972662e-01 -1.47459984e+00
5.15225641e-02 -3.70385319e-01 -3.76635283e-01 -2.64594644e-01
8.21185112e-02 -3.19532543e-01 7.23270178e-01 -1.50178993e+00
-7.55245209e-01 -3.63909483e-01 6.58867834e-03 -6.49251193e-02
9.85531658e-02 5.47933988e-02 1.84739426e-01 -1.78382203e-01
-4.37701404e-01 7.73430169e-01 1.41180229e+00 7.65417591e-02
-3.32271814e-01 3.89976263e-01 -1.71900168e-01 8.53480935e-01
9.61833239e-01 -3.33367884e-02 -7.66402483e-01 6.04596920e-02
6.79763705e-02 3.35356236e-01 -1.61367655e-01 -1.10556734e+00
2.29096904e-01 -5.74586511e-01 -1.44509867e-01 -4.95906889e-01
5.22941351e-01 -1.35691082e+00 3.02454501e-01 6.54558122e-01
9.99169648e-02 -4.56276728e-04 3.60411644e-01 9.54633534e-01
-2.21507430e-01 -5.95902264e-01 1.19888616e+00 2.99168646e-01
-3.92148823e-01 1.66120157e-01 -5.40770769e-01 -4.35946882e-01
1.92164814e+00 -4.00430739e-01 -1.00925826e-01 -3.28177720e-01
-7.10041165e-01 4.64491874e-01 4.89021599e-01 1.76807567e-01
5.03380120e-01 -1.15892899e+00 -4.00821835e-01 1.99839305e-02
2.00133789e-02 3.71019453e-01 2.52432562e-02 7.68684030e-01
-2.52560765e-01 8.32466364e-01 -1.29312694e-01 -4.27724361e-01
-9.60253179e-01 8.89904797e-01 4.96281832e-01 2.48115659e-01
-4.10854191e-01 3.47192824e-01 -2.23319352e-01 -3.37823868e-01
1.75457925e-01 -6.48472965e-01 -3.21390897e-01 1.14920266e-01
3.53182793e-01 7.21476793e-01 -1.14357114e-01 -6.84676528e-01
-2.93150187e-01 8.40772212e-01 4.44753975e-01 -1.94730937e-01
8.96844387e-01 -3.97473574e-01 -1.10284790e-01 3.43321353e-01
7.50269055e-01 6.04995862e-02 -1.42457354e+00 -7.36544579e-02
4.72262502e-02 -3.01363468e-01 -7.21858963e-02 -2.66005307e-01
-6.63152754e-01 2.50967085e-01 2.88543999e-01 4.00000840e-01
8.61556649e-01 -5.16798019e-01 2.02690154e-01 4.99771535e-01
5.81939697e-01 -1.37591779e+00 -5.36689818e-01 7.61285424e-01
1.42409587e+00 -1.05038798e+00 2.82409310e-01 -8.67531896e-01
-4.43401396e-01 1.24025512e+00 6.83783889e-01 -5.31777620e-01
7.05153704e-01 6.68219328e-01 -3.69557828e-01 4.11035359e-01
-7.58635104e-01 -4.06676948e-01 2.47517675e-01 5.49493492e-01
-5.16756251e-02 -5.44545315e-02 -9.06946182e-01 3.92321438e-01
1.70430467e-01 -1.26635253e-01 1.02659440e+00 1.09953403e+00
-8.25551450e-01 -7.63091445e-01 -1.12099266e+00 2.29039475e-01
-1.28291070e-01 5.49565613e-01 -3.56916748e-02 1.05134571e+00
-6.47630915e-02 1.06008685e+00 -1.69026237e-02 -3.55209112e-02
9.06668007e-01 -2.43420109e-01 3.19923222e-01 -7.04289377e-01
-2.42397655e-02 3.72983105e-02 1.41510427e-01 -4.55788016e-01
-1.32768070e-02 -7.69015908e-01 -1.46440494e+00 6.95871189e-02
-9.17260170e-01 4.15255159e-01 7.73921013e-01 7.53968120e-01
1.58063605e-01 3.70481074e-01 7.22733378e-01 -9.01757300e-01
-1.08744347e+00 -9.15215671e-01 -7.52920210e-01 -2.71105528e-01
4.01652604e-01 -1.27313125e+00 -4.57656205e-01 -1.84873298e-01] | [5.269514083862305, 2.3273797035217285] |
60af21a1-1008-4dc7-9b67-c2da607fec2d | semantic-single-image-dehazing | 1804.05624 | null | http://arxiv.org/abs/1804.05624v2 | http://arxiv.org/pdf/1804.05624v2.pdf | Semantic Single-Image Dehazing | Single-image haze-removal is challenging due to limited information contained
in one single image. Previous solutions largely rely on handcrafted priors to
compensate for this deficiency. Recent convolutional neural network (CNN)
models have been used to learn haze-related priors but they ultimately work as
advanced image filters. In this paper we propose a novel semantic ap- proach
towards single image haze removal. Unlike existing methods, we infer color
priors based on extracted semantic features. We argue that semantic context can
be exploited to give informative cues for (a) learning color prior on clean
image and (b) estimating ambient illumination. This design allowed our model to
recover clean images from challenging cases with strong ambiguity, e.g.
saturated illumination color and sky regions in image. In experiments, we
validate our ap- proach upon synthetic and real hazy images, where our method
showed superior performance over state-of-the-art approaches, suggesting
semantic information facilitates the haze removal task. | ['ShaoDi You', 'Ziang Cheng', 'Viorela Ila', 'Hongdong Li'] | 2018-04-16 | null | null | null | null | ['single-image-haze-removal'] | ['computer-vision'] | [ 5.33032417e-01 -1.30779013e-01 4.97606695e-01 -3.68058175e-01
-6.36033535e-01 -3.46633613e-01 4.41038579e-01 -3.41602087e-01
-2.29753837e-01 7.54700065e-01 2.14288056e-01 -1.63268798e-03
1.33332446e-01 -7.25425363e-01 -8.01331282e-01 -1.14143705e+00
3.65545690e-01 -2.09433720e-01 3.41005147e-01 -3.99373651e-01
2.87214369e-01 8.13385844e-02 -1.77314925e+00 2.87686378e-01
1.26312435e+00 7.60003507e-01 6.67407632e-01 5.91246188e-01
-9.21261162e-02 9.12149966e-01 -7.18689263e-01 -9.69803482e-02
5.27029157e-01 -3.94560426e-01 -1.76853448e-01 3.13080877e-01
6.51463866e-01 -4.47046816e-01 -3.38641375e-01 1.47460449e+00
2.36687928e-01 1.98778972e-01 5.32837689e-01 -1.20866168e+00
-8.65279913e-01 -9.57040931e-04 -5.10627568e-01 5.83927035e-02
-5.05920202e-02 2.79064029e-01 5.06573319e-01 -1.15952909e+00
1.62471592e-01 9.17992055e-01 5.26540399e-01 4.74617392e-01
-8.10418010e-01 -7.16058016e-01 2.07353637e-01 3.08009893e-01
-1.39679325e+00 -5.47290683e-01 1.02375734e+00 -1.49316490e-01
4.10647750e-01 2.71860778e-01 5.64250648e-01 8.28038633e-01
1.04187585e-01 8.57339919e-01 1.50782394e+00 -6.12577915e-01
2.77924478e-01 1.96818978e-01 -1.55289412e-01 5.20003796e-01
5.43087900e-01 -2.39813533e-02 -5.90755999e-01 2.26472944e-01
5.98325789e-01 2.76825041e-01 -5.09464264e-01 -1.30225616e-02
-7.69451380e-01 5.76635480e-01 6.42996550e-01 -9.18637663e-02
-4.27702218e-01 2.01968879e-01 -5.36211848e-01 1.72003269e-01
6.36057079e-01 3.42866361e-01 -3.12656969e-01 4.54082936e-01
-8.72637630e-01 4.55358103e-02 4.85836327e-01 9.48077321e-01
1.25285995e+00 4.56071109e-01 -7.47029856e-02 8.98219109e-01
4.05895591e-01 9.84034419e-01 -1.33597523e-01 -1.21359146e+00
2.91825235e-01 8.26551765e-02 5.83104670e-01 -8.09624672e-01
7.88074583e-02 -3.27708900e-01 -8.81450891e-01 6.31912172e-01
3.76625061e-01 -6.07815087e-02 -1.63072383e+00 1.45484746e+00
2.20517337e-01 3.13797146e-01 1.58153519e-01 1.16110003e+00
5.88142812e-01 8.15315843e-01 -1.99262694e-01 -6.18974958e-03
1.28156412e+00 -1.18834853e+00 -9.93100524e-01 -6.15354002e-01
-1.20056823e-01 -9.74002004e-01 8.99077117e-01 7.85095811e-01
-9.78462577e-01 -4.83663142e-01 -1.06578290e+00 7.41122616e-03
-6.33521855e-01 -4.90116775e-02 4.85399246e-01 8.90069723e-01
-1.30270922e+00 3.30940962e-01 -6.04164720e-01 -2.24048361e-01
3.93413693e-01 1.04882628e-01 -9.96844023e-02 -7.45603919e-01
-1.12740874e+00 9.92139935e-01 2.29638100e-01 4.88142669e-01
-1.27403700e+00 -5.00861287e-01 -8.35428953e-01 -1.67152155e-02
5.37834525e-01 -9.61892426e-01 7.13092446e-01 -1.19350612e+00
-1.60555971e+00 5.57310402e-01 -2.80171514e-01 -3.11737567e-01
9.65262949e-02 -7.76575685e-01 -5.50742924e-01 4.70987588e-01
2.05313805e-02 6.56726778e-01 1.64899969e+00 -2.06229138e+00
-5.30884624e-01 -9.02948156e-02 2.19528243e-01 2.97945708e-01
-2.65887529e-01 -1.15494154e-01 -6.34844959e-01 -7.69683361e-01
9.46248025e-02 -7.53725052e-01 -2.37138763e-01 1.81187153e-01
-2.73625851e-01 3.79257649e-01 8.95132601e-01 -8.30251098e-01
7.89302886e-01 -2.02691793e+00 -1.06818348e-01 -1.72048598e-03
2.56477267e-01 5.50474346e-01 -1.04988560e-01 4.14534986e-01
2.12493584e-01 -1.50932506e-01 -5.26319385e-01 -5.42394936e-01
-6.98618293e-02 4.77769166e-01 -4.00892287e-01 7.30745256e-01
4.08487439e-01 6.36813998e-01 -9.30258989e-01 -2.55235493e-01
6.01073623e-01 8.46220553e-01 -4.58218515e-01 6.66018128e-01
-2.17299029e-01 5.75630903e-01 -2.08373711e-01 8.99324834e-01
1.20764649e+00 -2.97150295e-02 -3.30613524e-01 -2.36471429e-01
-1.75294116e-01 -9.56293792e-02 -9.87221718e-01 1.65221131e+00
-4.78525281e-01 6.54801786e-01 5.96244454e-01 -8.13348830e-01
7.64580786e-01 1.61015347e-01 2.07682904e-02 -6.96657121e-01
4.25786525e-02 3.34609225e-02 -4.72375572e-01 -7.73854315e-01
5.53914428e-01 -2.95635581e-01 4.34241295e-01 -9.15240776e-03
-9.06238109e-02 -6.15184426e-01 -3.14901441e-01 1.29829854e-01
6.39775574e-01 2.56870419e-01 2.82539390e-02 -3.15790266e-01
6.04698002e-01 -1.04145564e-01 5.83436251e-01 1.09659946e+00
-3.08299124e-01 1.33983684e+00 -2.96619862e-01 -3.97474676e-01
-9.37220752e-01 -1.16668046e+00 1.94398403e-01 9.74803686e-01
6.09306037e-01 -2.18579233e-01 -7.62580454e-01 -2.58367002e-01
-3.92432243e-01 8.41689706e-01 -6.86451197e-01 -1.02739960e-01
-4.71021682e-01 -9.45143342e-01 1.15696602e-01 1.37771919e-01
7.95375884e-01 -6.92665577e-01 -3.74667078e-01 -4.85146604e-03
-3.44197482e-01 -1.39727986e+00 -2.10361928e-01 1.71504259e-01
-6.22323751e-01 -9.58477139e-01 -7.57600725e-01 -5.74052334e-01
7.96869278e-01 1.10740924e+00 1.06469560e+00 2.68646151e-01
-3.04458320e-01 5.07002711e-01 -5.75783193e-01 -8.37143183e-01
-1.07243303e-02 -5.54081619e-01 -1.38599738e-01 4.38835055e-01
1.01774625e-01 -6.58758700e-01 -1.28483391e+00 2.04231858e-01
-1.36568451e+00 2.04928339e-01 8.42280686e-01 7.38329887e-01
2.87011147e-01 2.28177518e-01 1.39157236e-01 -8.21804166e-01
9.86063108e-02 -7.00326920e-01 -6.76785290e-01 2.04750687e-01
-7.04505086e-01 -9.15045738e-02 4.24453288e-01 -6.69900402e-02
-1.63505840e+00 4.03278247e-02 5.17539121e-02 -6.83752477e-01
-5.57314456e-01 1.59033671e-01 -2.52232552e-01 -3.97358418e-01
4.12722319e-01 5.69191098e-01 -1.95110664e-01 -5.52696288e-01
6.35427713e-01 5.67552328e-01 8.11384678e-01 -7.25963593e-01
1.26714718e+00 1.08222306e+00 2.47376151e-02 -1.11980677e+00
-1.13734055e+00 -6.53829992e-01 -5.89366376e-01 -1.86160758e-01
1.11803317e+00 -1.47597146e+00 -1.34259537e-01 6.35102093e-01
-1.03006530e+00 -4.02351737e-01 3.31141174e-01 2.60547400e-01
-1.03814237e-01 5.93570411e-01 -3.84290606e-01 -1.30306244e+00
-2.71206409e-01 -8.77119720e-01 1.31191659e+00 3.68039399e-01
5.63669384e-01 -1.04195046e+00 -2.19599664e-01 7.70686984e-01
8.27613771e-01 1.97123080e-01 6.12457275e-01 1.83017507e-01
-1.08572078e+00 1.18237413e-01 -4.78917718e-01 6.78446531e-01
3.59811455e-01 -2.03400806e-01 -1.51653874e+00 -2.31094167e-01
1.92818344e-01 -1.50190480e-02 1.19650149e+00 2.71055818e-01
1.16611290e+00 -2.30103999e-01 8.05039704e-02 1.06668162e+00
1.78125489e+00 -1.57381952e-01 1.01516843e+00 5.75338721e-01
1.00666595e+00 6.11137211e-01 7.47637272e-01 3.26630265e-01
3.75163168e-01 4.16901857e-01 1.02012467e+00 -4.76865619e-01
-6.01648211e-01 1.43877164e-01 3.44265133e-01 5.34546793e-01
-1.72868431e-01 -4.84908223e-01 -5.69879174e-01 8.02111685e-01
-1.81624174e+00 -5.05429685e-01 -1.54130623e-01 1.86454046e+00
6.97595477e-01 -1.50229529e-01 -4.70597833e-01 -2.09769860e-01
5.01633465e-01 3.97440910e-01 -8.84445533e-02 2.51411587e-01
-5.30864477e-01 2.78914183e-01 6.62641108e-01 6.40280724e-01
-1.05208111e+00 1.19648004e+00 6.15815830e+00 7.07839727e-01
-1.03152394e+00 1.59255177e-01 2.54646957e-01 7.12383464e-02
-4.96076226e-01 1.01487413e-01 -3.55504364e-01 5.58488309e-01
4.88795906e-01 5.77675581e-01 7.05627739e-01 4.06579912e-01
2.37171113e-01 -4.20367897e-01 -4.77634281e-01 1.15737200e+00
4.19691503e-01 -9.88033414e-01 6.95774779e-02 -2.08984554e-01
1.16166782e+00 1.22138932e-01 2.66079009e-01 -6.63222149e-02
4.39229697e-01 -1.03722620e+00 8.65930915e-01 7.45004296e-01
5.17796636e-01 -3.39310378e-01 7.05566108e-01 1.57198131e-01
-8.89286757e-01 3.21208350e-02 -5.08951724e-01 -2.26675883e-01
-1.08098900e-02 8.17332685e-01 -6.83466911e-01 8.04314852e-01
1.01136720e+00 6.62064612e-01 -5.62500536e-01 1.01915276e+00
-7.84424603e-01 7.71055162e-01 -3.13875347e-01 6.12307847e-01
2.76358783e-01 -5.37413299e-01 4.49964434e-01 1.22143281e+00
3.45963478e-01 4.35675263e-01 1.13605775e-01 1.01997542e+00
1.73446596e-01 -5.70177853e-01 -5.22227168e-01 2.61476129e-01
2.56625205e-01 1.14611888e+00 -6.81803346e-01 -2.99566925e-01
-7.25993693e-01 1.34705770e+00 -3.46773565e-02 9.66189444e-01
-7.39429235e-01 -3.27842444e-01 9.15690184e-01 -1.14612214e-01
6.42692924e-01 -4.09471095e-01 -4.15124327e-01 -1.49929857e+00
-1.95889995e-01 -5.83043754e-01 9.36846733e-02 -1.37777054e+00
-1.30186212e+00 2.57303327e-01 -1.69906225e-02 -1.06580913e+00
4.82904941e-01 -8.91971886e-01 -7.72117019e-01 9.65574324e-01
-2.44668198e+00 -1.43068635e+00 -9.28153157e-01 8.22447181e-01
6.81924045e-01 2.12807849e-01 5.49927533e-01 2.63179541e-01
-5.74299872e-01 5.31836338e-02 3.36635888e-01 -5.61069548e-02
9.55059409e-01 -1.45300293e+00 1.52645651e-02 1.63274395e+00
1.84929688e-02 6.93026185e-01 1.09663904e+00 -4.94544327e-01
-1.67097640e+00 -1.25246525e+00 1.18340708e-01 -5.42792916e-01
2.99262524e-01 -2.54195720e-01 -1.03904259e+00 4.38635200e-01
6.15071952e-01 3.13156039e-01 4.79931027e-01 -3.91113818e-01
-4.84235108e-01 -2.52571017e-01 -9.53290343e-01 5.90745628e-01
7.07596540e-01 -5.13797104e-01 -6.52640998e-01 1.32020563e-01
6.47681773e-01 -2.42019296e-01 -4.01684016e-01 3.67815256e-01
2.98792094e-01 -1.26086640e+00 1.27802742e+00 -1.19471853e-03
2.72559375e-01 -9.27832365e-01 -3.83031338e-01 -1.45757842e+00
-3.15584928e-01 -7.60122776e-01 -2.03432322e-01 1.05877316e+00
-7.93186249e-04 -5.09797573e-01 5.08010149e-01 4.05793190e-01
-5.13805449e-01 -1.24163434e-01 -2.62668699e-01 -5.68469405e-01
-1.03547402e-01 -3.41571838e-01 6.15439892e-01 9.57031250e-01
-8.09472919e-01 -9.34019983e-02 -1.04668224e+00 9.68146384e-01
1.21608913e+00 2.06572279e-01 8.65053713e-01 -9.15801466e-01
-5.97992726e-02 -8.74057338e-02 1.88286323e-02 -8.72961462e-01
-2.39840057e-02 -2.24328950e-01 5.70245743e-01 -1.63522708e+00
1.17019191e-01 -2.12702453e-01 -6.19666815e-01 4.81711149e-01
-8.00078452e-01 7.31376946e-01 1.65236130e-01 1.89131930e-01
-8.44916821e-01 8.33264828e-01 1.17178619e+00 -3.33304703e-01
1.35554224e-01 -3.43499660e-01 -8.59732211e-01 6.11920118e-01
9.38906789e-01 -5.12018800e-01 -4.53846276e-01 -9.33680534e-01
3.08825046e-01 -3.73631895e-01 6.09346986e-01 -9.08941090e-01
3.69917452e-01 -4.63172019e-01 6.45312130e-01 -4.40725595e-01
5.98349154e-01 -9.54424858e-01 5.48172295e-02 1.09380692e-01
2.68816322e-01 -6.25527978e-01 9.39456895e-02 9.43973124e-01
-4.88266170e-01 -6.79860264e-02 7.26868391e-01 -4.08490181e-01
-1.09448838e+00 1.91045716e-01 -3.68272066e-01 -1.76159725e-01
5.56123555e-01 -3.00728887e-01 -5.36288440e-01 -6.42707586e-01
-4.24950212e-01 1.23097777e-01 8.38168204e-01 2.25127980e-01
9.72451210e-01 -8.15390885e-01 -6.17021561e-01 2.91221529e-01
3.73026490e-01 2.14149654e-01 2.15394944e-01 7.40784585e-01
-7.88792610e-01 3.78669500e-02 -1.17371775e-01 -4.73052293e-01
-1.14617741e+00 6.86097205e-01 2.32114151e-01 4.46987867e-01
-7.54290581e-01 8.07945311e-01 7.46713102e-01 -2.13890392e-02
7.15757385e-02 -2.86717057e-01 8.38188156e-02 -5.08513927e-01
6.36553824e-01 2.26314262e-01 3.26197259e-02 -4.69425440e-01
-1.76922828e-01 7.44030297e-01 2.20754966e-01 2.89999545e-02
1.48996580e+00 -7.90425003e-01 -3.69612157e-01 1.27939448e-01
8.25866640e-01 1.59176096e-01 -1.68944550e+00 -4.91774440e-01
-3.44752789e-01 -1.08978474e+00 3.19312960e-01 -8.50591362e-01
-1.18300784e+00 1.05545974e+00 5.63480973e-01 -2.05806736e-02
1.50773072e+00 -2.35309809e-01 7.74335980e-01 3.53915870e-01
4.17048573e-01 -8.76381576e-01 2.60848165e-01 3.75766575e-01
5.38030863e-01 -1.72460079e+00 2.00931028e-01 -6.80536032e-01
-5.42125404e-01 1.19096839e+00 5.70127845e-01 -1.54969424e-01
6.01529777e-01 -5.45888729e-02 5.23707092e-01 -4.12136704e-01
-3.94098669e-01 -5.61337054e-01 4.26964462e-01 8.42597246e-01
1.59375276e-03 -1.26738280e-01 4.09841716e-01 2.51790941e-01
1.84450224e-01 -2.27457523e-01 7.27904022e-01 9.54588830e-01
-8.87681603e-01 -7.42491424e-01 -8.74714613e-01 -2.17960030e-01
-4.88038957e-01 -4.85865951e-01 -1.41054869e-01 4.96699274e-01
4.28699344e-01 1.44884729e+00 -3.48798543e-01 -1.30538121e-01
-2.08115250e-01 -1.36998743e-01 5.93293130e-01 -5.50232291e-01
-9.31168348e-03 2.48733670e-01 -1.50638655e-01 -5.82351387e-01
-6.92325473e-01 -1.24969251e-01 -7.42154360e-01 5.12970388e-02
-2.89213240e-01 1.36090629e-02 7.06614792e-01 9.37429905e-01
1.59876451e-01 5.25744021e-01 5.56588650e-01 -1.15457213e+00
9.67860371e-02 -8.79452348e-01 -8.49315703e-01 6.92457318e-01
9.86338913e-01 -6.52508974e-01 -6.57378078e-01 3.75466555e-01] | [10.906042098999023, -3.172971248626709] |
b1998cc9-3e91-4203-95c7-c8a11c339db0 | weisfeiler-leman-in-the-bamboo-novel-amr | 2108.11949 | null | https://arxiv.org/abs/2108.11949v1 | https://arxiv.org/pdf/2108.11949v1.pdf | Weisfeiler-Leman in the BAMBOO: Novel AMR Graph Metrics and a Benchmark for AMR Graph Similarity | Several metrics have been proposed for assessing the similarity of (abstract) meaning representations (AMRs), but little is known about how they relate to human similarity ratings. Moreover, the current metrics have complementary strengths and weaknesses: some emphasize speed, while others make the alignment of graph structures explicit, at the price of a costly alignment step. In this work we propose new Weisfeiler-Leman AMR similarity metrics that unify the strengths of previous metrics, while mitigating their weaknesses. Specifically, our new metrics are able to match contextualized substructures and induce n:m alignments between their nodes. Furthermore, we introduce a Benchmark for AMR Metrics based on Overt Objectives (BAMBOO), the first benchmark to support empirical assessment of graph-based MR similarity metrics. BAMBOO maximizes the interpretability of results by defining multiple overt objectives that range from sentence similarity objectives to stress tests that probe a metric's robustness against meaning-altering and meaning-preserving graph transformations. We show the benefits of BAMBOO by profiling previous metrics and our own metrics. Results indicate that our novel metrics may serve as a strong baseline for future work. | ['Anette Frank', 'Angel Daza', 'Juri Opitz'] | 2021-08-26 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [ 6.96309328e-01 5.31899095e-01 -2.36538917e-01 -6.57712579e-01
-4.21203703e-01 -7.56535113e-01 7.35374749e-01 9.56060052e-01
-3.33397925e-01 3.24166089e-01 5.96462309e-01 -3.68285090e-01
-4.18604732e-01 -6.50765955e-01 -9.49881151e-02 -3.62001657e-02
-5.35217822e-02 3.32867950e-01 1.36665091e-01 -6.71078682e-01
7.55793929e-01 5.62958717e-01 -1.53067148e+00 4.29268658e-01
9.68022823e-01 6.74581349e-01 -2.58148070e-02 6.53689802e-01
-1.52099669e-01 9.21751142e-01 -8.21467102e-01 -7.70879865e-01
-1.64636388e-01 -6.75055206e-01 -1.27899277e+00 -2.61399090e-01
6.90769136e-01 4.37730044e-01 -1.55285567e-01 9.76944089e-01
4.46309805e-01 2.25527108e-01 7.98533261e-01 -1.39655769e+00
-1.00427353e+00 8.90695214e-01 -2.72311985e-01 4.91290957e-01
1.01625073e+00 1.31717813e-03 1.56153154e+00 -6.19558096e-01
7.21941173e-01 1.34937692e+00 8.65052342e-01 4.32703674e-01
-1.29499233e+00 -1.40734091e-01 1.18043542e-01 2.86067933e-01
-1.25954211e+00 -3.83196712e-01 6.00090802e-01 -3.38428199e-01
1.31738424e+00 6.74611449e-01 2.52593517e-01 9.94904995e-01
1.06238224e-01 4.17156696e-01 9.70820069e-01 -5.74930429e-01
1.68063790e-01 -6.33539557e-02 6.01573408e-01 5.12867332e-01
5.24044275e-01 -4.20761555e-01 -3.58075649e-01 -2.92714357e-01
3.08661431e-01 -5.49318433e-01 -2.63797700e-01 -3.70027393e-01
-1.41604614e+00 8.42890263e-01 3.67562324e-01 8.62382293e-01
4.17958479e-03 -2.18681693e-02 5.07270575e-01 7.29283810e-01
3.35798144e-01 1.29226255e+00 -1.33803129e-01 -1.18346654e-01
-6.40757620e-01 2.30082870e-01 7.47846544e-01 1.00230932e+00
7.27215648e-01 -1.10934690e-01 -5.09011507e-01 9.15223360e-01
1.04183197e-01 8.22431073e-02 7.64582336e-01 -9.17635202e-01
5.03267229e-01 8.11942399e-01 -2.88041085e-01 -1.49949276e+00
-5.65695226e-01 -2.42030561e-01 -6.18259549e-01 -1.29588574e-01
1.96427137e-01 3.62681687e-01 -3.88710618e-01 1.82290077e+00
-1.26046866e-01 -2.31842399e-01 5.21844141e-02 6.84984207e-01
1.06529081e+00 2.09657714e-01 7.87310004e-02 -1.22567683e-01
1.24274909e+00 -7.50536084e-01 -6.36829853e-01 -3.72375935e-01
1.14340091e+00 -8.60695124e-01 1.68675601e+00 4.36261967e-02
-1.17570508e+00 -6.14657700e-01 -1.12581563e+00 -1.17380060e-01
-4.72314030e-01 -4.30176735e-01 6.35270059e-01 1.02997649e+00
-1.39397454e+00 1.00474489e+00 -2.79301822e-01 -7.07890689e-01
-5.75446784e-02 1.72959030e-01 -3.46511543e-01 1.64319098e-01
-1.27578831e+00 1.34528756e+00 4.83727306e-01 -3.48704934e-01
-9.02946368e-02 -4.86499518e-01 -8.88973892e-01 1.10547982e-01
2.69360334e-01 -6.90390348e-01 9.88709688e-01 -9.50446606e-01
-1.19986248e+00 1.11163163e+00 1.17973723e-01 -3.86624843e-01
2.35105157e-02 1.36146650e-01 -6.91917300e-01 2.16932327e-01
1.37740895e-02 5.98693371e-01 3.12033832e-01 -9.55920279e-01
-1.35047913e-01 -1.98916420e-01 3.70142549e-01 2.47364804e-01
-5.93305647e-01 1.54877588e-01 6.44658953e-02 -1.05155694e+00
-9.11989957e-02 -6.94405675e-01 -1.82596803e-01 -2.44814217e-01
-2.47950822e-01 -3.41733158e-01 2.92258173e-01 -4.47483361e-01
1.57854235e+00 -2.00843310e+00 3.68984371e-01 2.42147401e-01
5.07290781e-01 9.85506177e-02 -7.51398504e-01 8.66925716e-01
-3.16132993e-01 6.02054536e-01 -4.38414246e-01 -6.87402114e-02
1.57552212e-01 1.77444354e-01 -1.19597696e-01 1.31825149e-01
2.74863601e-01 1.09250629e+00 -1.08854926e+00 -5.48377156e-01
-1.19783811e-01 2.43594214e-01 -5.64502954e-01 -1.57378726e-02
9.36519578e-02 -3.94866709e-03 -7.07701668e-02 2.17664704e-01
2.42670044e-01 -3.08135003e-01 4.12435293e-01 -3.18492413e-01
3.99493277e-01 6.27306759e-01 -9.86151695e-01 1.68544388e+00
-2.78930783e-01 7.09013343e-01 -5.73855400e-01 -1.05426669e+00
1.19249606e+00 -7.64229298e-02 1.17864743e-01 -7.80291319e-01
-2.39150953e-02 1.44655239e-02 9.70486179e-02 -2.47550383e-01
1.16629422e+00 1.19956955e-02 -2.33208120e-01 5.99034488e-01
-9.21222940e-03 -4.93661284e-01 3.87670636e-01 4.04593110e-01
1.38278639e+00 -1.99427560e-01 7.43898451e-01 -4.71822649e-01
5.24624586e-01 -2.23627150e-01 2.10412666e-01 8.56310308e-01
-5.39917238e-02 6.96101367e-01 7.09274709e-01 -2.52344549e-01
-1.01182008e+00 -1.10929286e+00 1.78243101e-01 1.04687417e+00
1.37036994e-01 -9.26678777e-01 -6.98166788e-01 -9.00101364e-01
-1.29271612e-01 1.18402469e+00 -7.75993407e-01 -4.69665825e-01
-4.41445351e-01 -5.85860431e-01 1.03727508e+00 5.65954864e-01
-1.28935397e-01 -1.00387418e+00 -6.07540429e-01 -8.55904669e-02
-3.49049091e-01 -1.10204422e+00 -5.33230603e-01 -2.74533093e-01
-1.00335968e+00 -1.08409917e+00 -3.22905004e-01 -5.43983996e-01
5.64349294e-01 3.92763585e-01 1.74720490e+00 4.42494959e-01
-1.01452231e-01 6.71363652e-01 -9.15777564e-01 -4.58083861e-02
-7.84157693e-01 2.96322227e-01 -8.88475850e-02 -2.81560719e-01
2.48075485e-01 -6.13929331e-01 -3.56851250e-01 4.76721734e-01
-1.20707119e+00 -1.63860649e-01 5.86104274e-01 4.59913164e-01
2.61484176e-01 -4.22944188e-01 7.19016850e-01 -1.04775321e+00
1.27271366e+00 -4.05444145e-01 9.71306190e-02 4.85957146e-01
-1.08793962e+00 3.12248290e-01 5.13623118e-01 -2.77138352e-01
-5.40438712e-01 -6.04056120e-01 2.30995622e-02 5.96584938e-02
-5.32637397e-03 7.03409970e-01 1.76595915e-02 -7.08385557e-02
1.07546258e+00 -1.52778417e-01 4.92328359e-03 -3.87747824e-01
6.88436389e-01 5.15515864e-01 4.92228001e-01 -7.48464346e-01
5.42698920e-01 -4.90372069e-02 2.60850717e-03 -7.42775619e-01
-8.00204575e-01 -2.50533253e-01 -6.80327177e-01 -4.21690615e-03
5.68935990e-01 -3.17776829e-01 -4.23150808e-01 -1.42993957e-01
-1.09268975e+00 8.48818570e-02 -5.84317148e-01 1.97247460e-01
-6.18902624e-01 1.10578787e+00 -5.13134599e-01 -3.11205566e-01
-4.58198756e-01 -8.70715618e-01 8.38250339e-01 -2.42005333e-01
-1.11072206e+00 -1.17177093e+00 6.43084869e-02 4.91003573e-01
5.78357756e-01 4.58903551e-01 1.37231576e+00 -9.07931268e-01
1.59184650e-01 -1.20854564e-01 -1.81699798e-01 2.89993495e-01
1.98160544e-01 -4.65707630e-02 -7.15978324e-01 -3.96186143e-01
-1.28663689e-01 -2.70422697e-01 5.33271849e-01 -3.18917856e-02
1.01078045e+00 -4.26775903e-01 3.13958339e-02 1.57075226e-01
1.16115236e+00 -3.30720991e-02 7.94110298e-01 3.74262810e-01
7.09359348e-01 8.28284383e-01 5.53279638e-01 1.41764775e-01
3.78104508e-01 8.39425087e-01 2.79469728e-01 3.08553912e-02
-3.45718861e-01 -1.82303727e-01 4.22956288e-01 1.28533757e+00
-1.53038055e-01 -3.67981791e-01 -8.84394407e-01 3.93176407e-01
-1.65832889e+00 -8.47864509e-01 -3.00471634e-01 2.20240068e+00
7.26724088e-01 2.90467680e-01 2.48062000e-01 4.58654851e-01
8.09300125e-01 5.12573063e-01 -1.22223504e-01 -8.97778451e-01
-2.67920107e-01 4.65591550e-01 1.21569403e-01 5.39835632e-01
-5.24518311e-01 7.67856836e-01 7.24696302e+00 6.27983928e-01
-5.28630376e-01 -1.65992886e-01 2.90088266e-01 1.70469820e-01
-9.14494455e-01 2.34882250e-01 -2.04252720e-01 7.49749839e-02
1.00992513e+00 -3.82806122e-01 3.75560701e-01 6.71582758e-01
-1.14068069e-01 3.26983064e-01 -1.48737323e+00 7.57647693e-01
5.03874898e-01 -1.12198842e+00 4.64606583e-01 -2.52059728e-01
3.97748500e-01 -2.29948983e-01 -6.65622279e-02 1.17491484e-01
1.04302958e-01 -1.20087421e+00 5.76601446e-01 4.91124272e-01
6.33590877e-01 -6.94964051e-01 7.92914629e-01 -2.89662927e-01
-1.15271044e+00 2.88670272e-01 -3.82326454e-01 -4.09952372e-01
-1.18194431e-01 6.05787218e-01 -7.57337391e-01 8.20591331e-01
2.10653782e-01 7.60843933e-01 -1.33729100e+00 6.67232811e-01
-3.29858869e-01 3.76664162e-01 2.21387178e-01 -4.07179624e-01
6.78268597e-02 -2.33422220e-02 6.59274757e-01 1.53062820e+00
1.87068790e-01 -2.97040761e-01 2.34845094e-02 9.37897384e-01
5.95146157e-02 5.56039751e-01 -7.77877688e-01 -3.45031381e-01
6.99413002e-01 1.19315243e+00 -9.10582662e-01 -2.35724524e-01
-2.44949207e-01 9.24580097e-01 4.74952608e-01 -8.80111679e-02
-6.89558327e-01 -7.29885280e-01 6.71388566e-01 9.35630128e-02
-1.29962429e-01 -1.95630968e-01 -5.72306156e-01 -1.21185529e+00
3.30004375e-03 -1.25477862e+00 7.02729344e-01 -9.25848961e-01
-1.52250814e+00 8.02087367e-01 2.76428163e-01 -1.05835164e+00
-1.93344221e-01 -5.92259943e-01 -4.92177308e-01 4.61324543e-01
-1.20331800e+00 -8.89538407e-01 -3.83813709e-01 2.45178729e-01
2.53530979e-01 -4.04059663e-02 9.80243146e-01 1.11952879e-01
-3.48730028e-01 9.66977000e-01 -5.63814759e-01 -7.48950839e-02
6.38235629e-01 -1.31950796e+00 1.07874310e+00 8.35771978e-01
5.67049921e-01 7.86550164e-01 8.21595669e-01 -4.69937116e-01
-1.00530064e+00 -7.50883520e-01 1.14726710e+00 -6.71270132e-01
8.34286153e-01 -2.97669414e-02 -1.11114955e+00 5.52975118e-01
2.05371290e-01 -3.83505851e-01 8.07035923e-01 3.86389256e-01
-7.61734068e-01 2.02802077e-01 -9.95744884e-01 6.87487245e-01
1.49622774e+00 -6.76958144e-01 -1.14468658e+00 2.74029970e-01
8.89889598e-01 -2.24394593e-02 -1.19311476e+00 4.79525506e-01
3.45044255e-01 -1.10614419e+00 8.93638551e-01 -8.52698803e-01
5.77783287e-01 -1.94559380e-01 -3.16286027e-01 -1.38909733e+00
-6.72572851e-01 -6.43848956e-01 1.99871957e-01 1.34045660e+00
5.92732728e-01 -8.54347229e-01 4.20105100e-01 4.19549465e-01
-2.99265802e-01 -6.38669550e-01 -6.13214493e-01 -9.94774461e-01
4.95633518e-04 -3.33709747e-01 9.08291280e-01 1.52160525e+00
5.07988214e-01 7.41467476e-01 -3.12835760e-02 -2.35916406e-01
3.36749256e-01 -1.76304489e-01 5.48797488e-01 -1.31639373e+00
-3.58284324e-01 -9.27569091e-01 -9.75505710e-01 -3.80972385e-01
1.79739341e-01 -1.45781767e+00 -4.96629953e-01 -1.63130069e+00
2.19242468e-01 -1.14161767e-01 -4.61656481e-01 5.81097901e-01
-4.04282331e-01 1.64282292e-01 4.25183088e-01 1.57716632e-01
-5.78953564e-01 3.13414395e-01 9.56681013e-01 -1.31934389e-01
3.04714628e-02 -4.21704352e-01 -1.08066475e+00 5.59645295e-01
9.68710780e-01 -4.81503248e-01 -7.48013318e-01 -4.22435433e-01
6.89187467e-01 -3.68523300e-01 1.60485491e-01 -9.88782823e-01
1.68492161e-02 -2.33636037e-01 -1.43948138e-01 -5.43167964e-02
-3.95297855e-02 -2.95446664e-01 9.88930166e-02 3.55778158e-01
-8.28317225e-01 7.84730077e-01 -1.13932444e-02 3.13402802e-01
-2.64313519e-01 -5.87601602e-01 6.26578033e-01 -4.30110173e-04
-6.77085161e-01 -1.91049173e-01 -2.41765574e-01 5.75843751e-01
6.71043873e-01 -4.51433241e-01 -5.32219231e-01 -4.16883677e-01
-4.00931716e-01 -1.28853947e-01 7.61991322e-01 8.25055838e-01
7.43796170e-01 -1.34187746e+00 -8.16501200e-01 -2.77859330e-01
6.28341258e-01 -7.53710508e-01 -1.64959013e-01 6.39122844e-01
-4.26848680e-01 1.52119666e-01 -4.23112005e-01 -3.07552814e-01
-1.45690584e+00 7.44689822e-01 8.80963579e-02 -3.76972973e-01
-5.41572332e-01 4.68303233e-01 -3.58900987e-02 -6.04368925e-01
3.21236323e-03 -5.40090144e-01 -4.84312445e-01 1.52191952e-01
3.80710959e-01 5.95897555e-01 2.88955361e-01 -5.29586136e-01
-3.85403991e-01 5.69268882e-01 2.22178735e-02 -1.22467754e-02
1.13847649e+00 -5.90564385e-02 -4.27125424e-01 5.06184042e-01
1.08635068e+00 2.53971629e-02 -8.71058404e-02 -5.10630012e-02
6.80012763e-01 -4.19698536e-01 -2.59360909e-01 -7.69983411e-01
-6.39121473e-01 5.35689592e-01 2.38016412e-01 5.31297743e-01
1.10608745e+00 -4.26398451e-03 4.92233157e-01 6.06176972e-01
1.25601947e-01 -1.04501379e+00 4.13396657e-01 4.96094465e-01
1.11909628e+00 -6.43779218e-01 2.11690083e-01 -6.76148236e-01
-8.57548356e-01 1.07876754e+00 3.54204893e-01 4.50946949e-02
2.09573284e-03 -1.87158328e-03 -8.93540829e-02 -3.63144815e-01
-6.13860428e-01 -2.06449211e-01 6.64907813e-01 8.04025292e-01
8.20350349e-01 8.77073780e-02 -8.74729514e-01 3.52393001e-01
-6.00865543e-01 -3.33862185e-01 6.82334900e-01 8.36105049e-01
-3.76708329e-01 -1.35474777e+00 -8.52283910e-02 5.12720823e-01
-2.76762992e-01 -2.83806592e-01 -1.04478776e+00 8.65800798e-01
-4.23415601e-01 1.19111514e+00 -2.22866729e-01 -7.85369337e-01
5.73959887e-01 1.01963565e-01 6.54861033e-01 -8.06613982e-01
-8.73722613e-01 -7.82974243e-01 6.93661392e-01 -6.43006265e-01
-5.45842826e-01 -5.25550425e-01 -1.06792879e+00 -3.84008735e-01
-2.58092821e-01 2.67820477e-01 3.46392840e-01 9.12307858e-01
5.94119430e-01 4.23768133e-01 5.69565475e-01 -4.34530020e-01
-5.50597370e-01 -1.14034855e+00 -3.09343040e-01 1.00732505e+00
-2.41590351e-01 -4.61769074e-01 -4.29457635e-01 -1.41452819e-01] | [11.260128021240234, 9.2017240524292] |
26c407e4-2ce9-4a39-9d75-44592597fd7a | improving-open-domain-dialogue-evaluation | 2301.13372 | null | https://arxiv.org/abs/2301.13372v1 | https://arxiv.org/pdf/2301.13372v1.pdf | Improving Open-Domain Dialogue Evaluation with a Causal Inference Model | Effective evaluation methods remain a significant challenge for research on open-domain conversational dialogue systems. Explicit satisfaction ratings can be elicited from users, but users often do not provide ratings when asked, and those they give can be highly subjective. Post-hoc ratings by experts are an alternative, but these can be both expensive and complex to collect. Here, we explore the creation of automated methods for predicting both expert and user ratings of open-domain dialogues. We compare four different approaches. First, we train a baseline model using an end-to-end transformer to predict ratings directly from the raw dialogue text. The other three methods are variants of a two-stage approach in which we first extract interpretable features at the turn level that capture, among other aspects, user dialogue behaviors indicating contradiction, repetition, disinterest, compliments, or criticism. We project these features to the dialogue level and train a dialogue-level MLP regression model, a dialogue-level LSTM, and a novel causal inference model called counterfactual-LSTM (CF-LSTM) to predict ratings. The proposed CF-LSTM is a sequential model over turn-level features which predicts ratings using multiple regressors depending on hypotheses derived from the turn-level features. As a causal inference model, CF-LSTM aims to learn the underlying causes of a specific event, such as a low rating. We also bin the user ratings and perform classification experiments with all four models. In evaluation experiments on conversational data from the Alexa Prize SocialBot, we show that the CF-LSTM achieves the best performance for predicting dialogue ratings and classification. | ['Reza Ghanadan', 'Marilyn Walker', 'Yang Liu', 'Michael Johnston', 'Luke Dai', 'Cat P. Le'] | 2023-01-31 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [ 3.65999201e-03 6.10820293e-01 -3.42987835e-01 -9.48995352e-01
-8.73229444e-01 -4.64391828e-01 8.00215900e-01 1.97644442e-01
-2.19611496e-01 1.16022742e+00 8.19118321e-01 -3.48749936e-01
1.83661997e-01 -7.58465111e-01 -2.38216430e-01 -3.01925331e-01
-4.19545844e-02 5.42756915e-01 -1.51634514e-01 -6.71864688e-01
5.23935199e-01 -7.34217986e-02 -1.25301027e+00 9.61618006e-01
8.45904887e-01 1.08005345e+00 -2.71494538e-01 1.08594954e+00
-3.02000284e-01 1.33768928e+00 -9.40245748e-01 -5.83738327e-01
-1.93229437e-01 -7.74800599e-01 -1.39698970e+00 -4.03686985e-02
-1.92413002e-01 -5.99457741e-01 7.27711618e-02 5.96125603e-01
4.50573325e-01 3.03443849e-01 7.51941621e-01 -1.19560754e+00
-6.33195579e-01 9.66444850e-01 -1.18210003e-01 -7.08728656e-02
9.69820738e-01 1.22974848e-03 1.47237277e+00 -7.13085413e-01
2.44503528e-01 1.51595914e+00 4.44002777e-01 7.48285890e-01
-1.01525509e+00 -3.18513989e-01 1.25077292e-01 2.86914438e-01
-3.11609119e-01 -4.12499577e-01 7.35841095e-01 -4.54990625e-01
9.10353661e-01 3.41005594e-01 3.25174958e-01 1.52137816e+00
-1.17327981e-02 7.60741353e-01 1.39501071e+00 -2.94804484e-01
3.06544840e-01 5.36802113e-01 3.06318939e-01 6.01412773e-01
-7.74463892e-01 3.00472826e-02 -7.16975510e-01 -6.04183018e-01
4.68761384e-01 -9.59410593e-02 -3.76865238e-01 5.30015305e-02
-9.81592178e-01 1.34035027e+00 5.15766084e-01 1.33914113e-01
-4.47220653e-01 -3.61601114e-01 7.10457027e-01 8.71436059e-01
8.12746823e-01 6.41180336e-01 -9.55222249e-01 -4.76185828e-01
-4.77520019e-01 3.84000301e-01 1.52990019e+00 2.81747758e-01
5.35311222e-01 -4.04662609e-01 -4.90762264e-01 1.39075375e+00
3.31480265e-01 -3.94006968e-02 7.45615840e-01 -9.51112688e-01
4.74178076e-01 6.39218271e-01 3.36692870e-01 -8.84899855e-01
-3.87569636e-01 8.45608115e-02 -6.23610556e-01 -1.50070991e-02
4.23222065e-01 -7.87729740e-01 -3.98311615e-01 1.73493981e+00
2.04299137e-01 -2.85660803e-01 2.29988381e-01 1.05842078e+00
1.23147726e+00 8.04760098e-01 -8.97084251e-02 -6.65369153e-01
1.10415363e+00 -1.04551518e+00 -8.92275751e-01 1.87535957e-01
8.16115081e-01 -5.31122863e-01 1.14438021e+00 5.21550894e-01
-9.83564556e-01 -3.39190036e-01 -8.12763572e-01 -4.14386876e-02
-3.48761320e-01 -1.57076064e-02 6.02341473e-01 3.06518257e-01
-8.76985669e-01 8.81584167e-01 -4.81945574e-02 -3.24323654e-01
-2.50923902e-01 3.76664102e-01 -1.28862374e-02 5.46243250e-01
-1.89221120e+00 1.18418467e+00 -2.64904171e-01 1.28819600e-01
-7.16494381e-01 -4.62535143e-01 -8.02524090e-01 1.37278259e-01
1.60372913e-01 -3.04261178e-01 1.88430548e+00 -1.08227909e+00
-2.27531505e+00 7.55176902e-01 -1.99529245e-01 -3.06118667e-01
4.60797369e-01 -1.03278719e-01 -2.82199174e-01 -2.48017237e-01
2.83534378e-02 2.35185117e-01 6.53741181e-01 -1.19291520e+00
-7.69206107e-01 -3.29744518e-01 5.63221812e-01 4.77719188e-01
-3.48839104e-01 2.49613047e-01 4.83365208e-01 -3.73762846e-02
-4.81704891e-01 -7.60033369e-01 -3.38329852e-01 -5.43371618e-01
-4.52984363e-01 -9.31008637e-01 5.67891777e-01 -6.22383773e-01
1.17400467e+00 -1.65987027e+00 1.08229950e-01 -1.16767786e-01
1.44403681e-01 2.52981633e-01 2.36594900e-02 5.45917213e-01
-1.40374348e-01 1.91738188e-01 7.58894309e-02 -2.15885460e-01
-2.73699258e-02 3.27068307e-02 -4.13897276e-01 -5.20352311e-02
2.77260125e-01 6.52422786e-01 -1.02065337e+00 -3.21075708e-01
8.48240554e-02 -2.11113337e-02 -5.61557114e-01 9.99606788e-01
-4.30374086e-01 4.72868532e-01 -4.68311101e-01 1.25203282e-01
1.50081620e-01 -1.39810815e-01 2.23124847e-01 8.04712400e-02
-2.36234337e-01 1.08290744e+00 -5.24662614e-01 1.40445006e+00
-8.17812085e-01 4.80742931e-01 5.35070803e-03 -1.06404996e+00
1.16162777e+00 8.75005782e-01 1.97894365e-01 -3.98152471e-01
3.77865016e-01 5.40003181e-02 1.33504659e-01 -7.25280404e-01
4.65665907e-01 -3.15942705e-01 -4.54621345e-01 8.43285978e-01
1.92737237e-01 -1.13135152e-01 8.81962385e-03 2.88391680e-01
1.04344702e+00 -2.11030528e-01 5.07403910e-01 1.04071312e-01
7.89369166e-01 -7.56563544e-02 4.94512349e-01 5.76971591e-01
-2.93760210e-01 2.87572443e-01 1.15671396e+00 -5.60932398e-01
-7.64982939e-01 -5.78402460e-01 1.26996920e-01 1.62161970e+00
-2.50067770e-01 -2.47753322e-01 -6.49570227e-01 -1.05024314e+00
-2.30543077e-01 8.24281573e-01 -7.69741535e-01 -5.58098666e-02
-3.19523662e-01 -5.65963626e-01 2.56897897e-01 1.45310685e-01
3.64446402e-01 -1.29683566e+00 -1.94685429e-01 2.68917918e-01
-7.52693713e-01 -7.99761474e-01 -2.70142823e-01 2.42336005e-01
-6.28305495e-01 -9.75566506e-01 -4.25713122e-01 -5.11136651e-01
2.13218659e-01 -1.09041966e-01 1.18548024e+00 9.66253877e-02
3.02339673e-01 9.97970160e-03 -5.47338724e-01 -3.84778321e-01
-7.80753613e-01 -2.41422709e-02 7.38797113e-02 1.75686687e-01
3.23139727e-01 -5.71175218e-01 -4.30137336e-01 4.85132933e-01
-9.05567259e-02 -1.00558452e-01 2.48478502e-01 1.24466217e+00
-3.17867160e-01 -6.52911961e-01 1.10678720e+00 -1.22545898e+00
1.56920981e+00 -7.32911766e-01 9.78183001e-02 -1.95614416e-02
-5.79397857e-01 1.04166083e-01 9.47222531e-01 -3.70366514e-01
-1.43926311e+00 -3.10101658e-01 -4.05330420e-01 -2.48486106e-03
-2.71289617e-01 7.54198074e-01 9.71719027e-02 5.19479990e-01
8.76366317e-01 -2.77956009e-01 1.91050380e-01 -2.98024267e-01
5.14552057e-01 1.30675805e+00 1.93476826e-01 -4.84032214e-01
2.51487672e-01 -1.63605273e-01 -6.62719905e-01 -6.38423324e-01
-1.37915480e+00 -4.30995077e-01 -4.01471108e-01 -4.73451525e-01
7.25369692e-01 -6.46827281e-01 -1.09886122e+00 1.52167395e-01
-1.50118315e+00 -4.47619289e-01 -1.82075128e-02 2.46135488e-01
-6.61766827e-01 8.09440687e-02 -1.06484020e+00 -1.05046594e+00
-4.70328242e-01 -8.65315914e-01 5.46295047e-01 3.72102648e-01
-7.84634113e-01 -1.10664713e+00 2.93353140e-01 7.40554929e-01
3.21792394e-01 1.05825432e-01 9.51310754e-01 -1.26282036e+00
2.69959182e-01 -1.40998051e-01 -6.80626556e-02 4.36122775e-01
1.91736966e-01 5.45097366e-02 -1.08999610e+00 8.46065059e-02
2.85229385e-01 -1.13620818e+00 5.25251448e-01 1.13791451e-01
8.94722760e-01 -8.90019715e-01 5.35612665e-02 -2.49035686e-01
7.31563747e-01 2.43214875e-01 3.21361423e-01 -9.12646577e-02
3.72981042e-01 9.93023336e-01 8.93821239e-01 6.94910944e-01
6.25923276e-01 5.09295404e-01 3.37416679e-01 -2.49860242e-01
4.29448038e-01 -2.67527968e-01 6.51503146e-01 9.33615148e-01
-2.54651874e-01 -1.30504593e-01 -4.14912522e-01 4.28470165e-01
-2.14872432e+00 -1.06364942e+00 -4.53147948e-01 2.04635572e+00
1.27650511e+00 1.53500170e-01 3.69877428e-01 1.42965153e-01
7.66652405e-01 3.07189256e-01 -4.40846562e-01 -1.37683165e+00
4.46315795e-01 4.48374636e-03 -8.14079866e-02 8.02497447e-01
-6.96000516e-01 7.32940495e-01 6.21562672e+00 2.41151810e-01
-9.32225764e-01 3.01894397e-01 8.63950372e-01 -3.41980830e-02
-2.32332125e-01 2.91373692e-02 -4.23899323e-01 2.57158518e-01
1.22083223e+00 -2.17483357e-01 3.39833349e-01 9.16267395e-01
7.16471374e-01 -1.76425148e-02 -1.41016722e+00 5.13906777e-01
-8.76305345e-03 -9.18450296e-01 -2.07381636e-01 1.21688992e-02
6.24875665e-01 -2.39421368e-01 -3.23002934e-01 7.60681450e-01
8.99015486e-01 -9.98481333e-01 1.60279766e-01 3.58691841e-01
4.67199534e-01 -7.04842567e-01 8.96808982e-01 6.37835741e-01
-5.12491882e-01 -2.52144247e-01 -3.62167776e-01 -7.42217183e-01
1.81942686e-01 4.49999601e-01 -1.29617167e+00 2.25047171e-01
4.74598080e-01 6.29639804e-01 2.43587643e-02 3.01466554e-01
-6.02396488e-01 7.02220201e-01 1.92890108e-01 -6.50177777e-01
2.22587675e-01 -2.85036899e-02 3.18257540e-01 1.17418396e+00
-1.49219006e-01 3.98751110e-01 1.49843380e-01 8.77348006e-01
-3.13225210e-01 2.32086211e-01 -5.98413765e-01 1.19402252e-01
3.52569759e-01 1.49897480e+00 2.47522816e-02 -4.15474147e-01
-4.23278749e-01 8.44365120e-01 7.17093110e-01 1.33283615e-01
-7.64970481e-01 -2.75558233e-01 6.02865875e-01 -3.54999751e-01
-1.58032075e-01 3.82667601e-01 -2.40110725e-01 -1.14470279e+00
-3.69496644e-01 -1.03012836e+00 4.60625976e-01 -6.14619732e-01
-1.51601481e+00 6.87499821e-01 -3.63555908e-01 -9.07972217e-01
-8.76544535e-01 -4.02588576e-01 -1.05322921e+00 8.72906148e-01
-1.30283117e+00 -6.18265688e-01 -1.67127371e-01 6.07072294e-01
1.09670639e+00 -8.63929689e-02 1.14677477e+00 -1.15326010e-01
-3.67299169e-01 4.03688759e-01 -2.43514612e-01 3.11334252e-01
1.05191362e+00 -1.57196546e+00 6.57269126e-03 -9.50304698e-03
-3.66361976e-01 5.22095501e-01 8.69569659e-01 -3.18217158e-01
-7.96364307e-01 -6.90040112e-01 1.37806201e+00 -4.15496081e-01
7.81306922e-01 -3.71699661e-01 -8.36877465e-01 4.32677299e-01
6.40970826e-01 -5.64287961e-01 9.52668548e-01 8.87706697e-01
-1.48095608e-01 1.53171182e-01 -1.02464724e+00 4.64509457e-01
6.01456523e-01 -4.74336118e-01 -1.08064747e+00 4.75818545e-01
8.13352644e-01 -1.68176547e-01 -1.19197977e+00 2.30214432e-01
4.81722623e-01 -1.26160526e+00 4.38508391e-01 -9.24776316e-01
1.18349838e+00 3.62342775e-01 2.23653138e-01 -1.95561051e+00
-1.23975679e-01 -8.43633235e-01 7.80796912e-03 1.30472410e+00
8.82845819e-01 -4.34318751e-01 6.04909122e-01 1.00042272e+00
-6.03758954e-02 -9.99128222e-01 -5.69729209e-01 -8.54251757e-02
5.63813411e-02 -1.77876756e-01 4.02115554e-01 1.24916720e+00
1.03345287e+00 1.36553192e+00 -6.61367118e-01 -3.87077063e-01
1.78286985e-01 2.15882376e-01 8.41766059e-01 -1.34298635e+00
-2.75791645e-01 -4.57995087e-01 2.41195709e-01 -1.13734007e+00
3.28689903e-01 -4.83411163e-01 3.52422625e-01 -1.62463784e+00
1.07147917e-01 -2.88621932e-01 -1.97162889e-02 3.85160804e-01
-2.86715448e-01 -3.51568192e-01 -4.97542843e-02 5.94222993e-02
-4.33071315e-01 7.71772861e-01 1.19388878e+00 8.62412900e-02
-5.60601056e-01 4.46663260e-01 -8.26946318e-01 7.76348591e-01
8.96284461e-01 -2.38271564e-01 -2.92490095e-01 1.27497613e-01
2.76356161e-01 9.77546334e-01 -7.65614137e-02 -2.96311885e-01
1.63494125e-01 -3.96374166e-01 1.27005400e-02 -4.73674625e-01
3.76025498e-01 -2.07324296e-01 -7.04934001e-01 1.42450064e-01
-1.21472192e+00 -1.72755718e-01 -5.13422608e-01 4.73772883e-01
-4.53562051e-01 -3.07856560e-01 6.40084326e-01 -4.43172902e-01
-5.60625233e-02 -2.90601458e-02 -6.95519030e-01 -5.56798652e-02
5.66033065e-01 1.87166974e-01 -3.84227067e-01 -1.07087195e+00
-7.69892812e-01 5.35782278e-01 -2.11213395e-01 6.80082560e-01
6.31614864e-01 -1.11029887e+00 -9.06482220e-01 -2.98487872e-01
-7.42841586e-02 -2.63433576e-01 -5.25474362e-02 6.97894156e-01
2.32060939e-01 4.25438404e-01 7.35132769e-03 -3.34059864e-01
-1.30354035e+00 3.68187070e-01 3.01235318e-01 -7.44982779e-01
8.70696455e-02 9.00073528e-01 3.07063758e-02 -1.03412127e+00
2.66196281e-01 -1.31078839e-01 -9.34077740e-01 3.49335313e-01
3.83256137e-01 2.81060278e-01 4.95029837e-02 -3.81247222e-01
1.20733298e-01 -1.95872620e-01 -2.01053962e-01 -3.88197094e-01
1.29216552e+00 -1.89751908e-01 -3.07770908e-01 1.03323805e+00
1.24833369e+00 -2.79016703e-01 -8.27744305e-01 -3.53181154e-01
-3.45341638e-02 -2.55869001e-01 -1.23890094e-01 -1.13107097e+00
-6.52647316e-01 8.30059826e-01 1.48212332e-02 9.56369281e-01
6.39603913e-01 -1.29168168e-01 6.98384583e-01 3.81030470e-01
6.72109500e-02 -1.32137883e+00 4.22421008e-01 8.57207298e-01
1.19287014e+00 -1.36826718e+00 -2.58840650e-01 -1.70165777e-01
-1.09475362e+00 1.08715928e+00 9.37819362e-01 -5.17344885e-02
4.84255165e-01 -2.40354300e-01 3.88823509e-01 -2.08536357e-01
-1.52329957e+00 -5.74892107e-03 4.82544787e-02 1.66115850e-01
1.05842447e+00 2.32997745e-01 -7.34995008e-01 7.99851537e-01
-4.76699501e-01 -7.84185603e-02 9.23511028e-01 3.65618825e-01
-3.30245733e-01 -9.97157454e-01 -2.20049500e-01 7.57683694e-01
-4.17408049e-01 -1.57529879e-02 -9.78522897e-01 2.09508330e-01
-3.04730296e-01 1.67140424e+00 -1.97409064e-01 -8.04330707e-01
4.87045795e-01 3.23634714e-01 -9.15995762e-02 -7.72131205e-01
-1.11444199e+00 -3.78524005e-01 8.51513863e-01 -5.40599823e-01
-5.29426038e-01 -6.02325976e-01 -1.08201241e+00 -4.45273817e-01
-6.53522789e-01 6.79456532e-01 7.11697698e-01 1.24424613e+00
3.21148783e-02 3.69528025e-01 1.46965206e+00 -6.59961581e-01
-1.04470456e+00 -1.54452670e+00 -2.74157405e-01 5.10350704e-01
4.64053363e-01 -5.05656540e-01 -5.91887951e-01 -2.31091857e-01] | [12.84613037109375, 7.983914852142334] |
a7577f02-efea-4e3e-892b-173b24f4f569 | the-one-where-they-reconstructed-3d-humans | 2207.14279 | null | https://arxiv.org/abs/2207.14279v1 | https://arxiv.org/pdf/2207.14279v1.pdf | The One Where They Reconstructed 3D Humans and Environments in TV Shows | TV shows depict a wide variety of human behaviors and have been studied extensively for their potential to be a rich source of data for many applications. However, the majority of the existing work focuses on 2D recognition tasks. In this paper, we make the observation that there is a certain persistence in TV shows, i.e., repetition of the environments and the humans, which makes possible the 3D reconstruction of this content. Building on this insight, we propose an automatic approach that operates on an entire season of a TV show and aggregates information in 3D; we build a 3D model of the environment, compute camera information, static 3D scene structure and body scale information. Then, we demonstrate how this information acts as rich 3D context that can guide and improve the recovery of 3D human pose and position in these environments. Moreover, we show that reasoning about humans and their environment in 3D enables a broad range of downstream applications: re-identification, gaze estimation, cinematography and image editing. We apply our approach on environments from seven iconic TV shows and perform an extensive evaluation of the proposed system. | ['Angjoo Kanazawa', 'Matthew Tancik', 'Ethan Weber', 'Georgios Pavlakos'] | 2022-07-28 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [ 3.77285093e-01 -1.63413361e-01 1.13995269e-01 -4.33696359e-01
7.03814030e-02 -6.12231910e-01 7.75077999e-01 -3.08462214e-02
3.47171314e-02 2.10404381e-01 3.29844087e-01 2.22871363e-01
1.80191547e-01 -4.17742997e-01 -7.91212440e-01 -3.95128280e-01
-1.21785803e-02 4.83742058e-01 3.84855002e-01 -4.83754694e-01
3.41065735e-01 6.69008791e-01 -1.97476196e+00 1.12503536e-01
3.65663558e-01 7.67046154e-01 2.16895357e-01 8.17710817e-01
3.53650808e-01 6.51039183e-01 -6.94467068e-01 -4.50215995e-01
2.32289717e-01 -4.10984665e-01 -4.87294465e-01 6.21801853e-01
7.46686935e-01 -4.42027181e-01 -3.61608505e-01 8.92094731e-01
4.00471419e-01 3.53643537e-01 3.12656820e-01 -1.09591436e+00
-3.09062377e-02 7.92475492e-02 -5.89543939e-01 -3.38671021e-02
1.19288135e+00 -4.91304100e-02 4.93196070e-01 -5.30704975e-01
9.30824459e-01 1.18294060e+00 6.74792290e-01 3.02472144e-01
-1.02621031e+00 -3.90152574e-01 1.21785939e-01 3.75031531e-01
-1.27312601e+00 -4.65140671e-01 1.14551663e+00 -5.12154818e-01
4.63543117e-01 5.91412365e-01 1.24606144e+00 1.17990744e+00
6.56381920e-02 6.96746826e-01 1.14861608e+00 -6.65422082e-01
7.67793357e-02 4.68187302e-01 5.21805584e-02 5.95287859e-01
-2.05150336e-01 -1.31307825e-01 -7.72793949e-01 7.10949302e-02
8.73773694e-01 2.70176709e-01 -3.14362377e-01 -6.61027670e-01
-1.15290082e+00 5.02946116e-02 1.96985930e-01 2.70775586e-01
-3.39328170e-01 -3.92519385e-02 2.60933526e-02 1.41228676e-01
4.23582375e-01 6.84117526e-02 1.71305180e-01 -2.75705427e-01
-7.82249689e-01 5.23250341e-01 7.87081063e-01 1.01167488e+00
7.91200101e-01 -3.94775301e-01 2.48078033e-01 5.56688726e-01
5.94972968e-02 7.42622018e-01 2.34831423e-01 -9.00231421e-01
2.88246155e-01 6.06618226e-01 2.30931699e-01 -1.69705701e+00
-4.72938061e-01 5.88978529e-02 -4.70683485e-01 -8.90279934e-03
2.67077267e-01 2.31631443e-01 -5.48413634e-01 1.72228622e+00
8.41584921e-01 4.21170533e-01 -4.59295452e-01 9.92567182e-01
4.82412726e-01 2.57175267e-01 -4.32109624e-01 -5.87955564e-02
1.37880027e+00 -6.52764797e-01 -8.34324658e-01 -2.35576317e-01
2.08345115e-01 -7.09013999e-01 7.86732793e-01 4.05874074e-01
-1.05450189e+00 -7.73473501e-01 -8.85264933e-01 -9.24693868e-02
-3.26019287e-01 2.02893782e-02 4.29893047e-01 6.11167312e-01
-9.72876608e-01 5.85827112e-01 -7.73126066e-01 -1.02967405e+00
-1.55145392e-01 1.50208741e-01 -4.72097516e-01 -6.20830767e-02
-8.24806631e-01 1.02393699e+00 1.45513892e-01 2.68205404e-01
-6.44222200e-01 -2.73766637e-01 -7.81289995e-01 -2.54146993e-01
7.27515519e-01 -6.43366754e-01 1.04363132e+00 -7.61304677e-01
-1.49079812e+00 1.26973677e+00 -2.07320616e-01 -2.94370472e-01
7.35720992e-01 -3.98762435e-01 -4.64736193e-01 2.59973139e-01
-1.78916126e-01 8.43025669e-02 9.03590918e-01 -1.38437736e+00
-6.00151241e-01 -1.05722320e+00 3.24867487e-01 6.23626411e-01
-1.23868652e-01 -8.96355882e-02 -9.53395128e-01 -4.24828440e-01
3.66270363e-01 -1.25577188e+00 1.76618233e-01 -3.89951542e-02
-4.57260877e-01 2.63838172e-01 6.77725732e-01 -9.26457942e-01
1.27019405e+00 -2.10661244e+00 5.29030144e-01 5.02628565e-01
3.72823358e-01 -2.15599071e-02 4.20450985e-01 5.53208530e-01
1.06489658e-01 -2.96051234e-01 4.52880785e-02 -7.28781700e-01
3.98269258e-02 6.20638393e-03 -7.97444582e-02 6.68226004e-01
-4.06831056e-01 5.77497184e-01 -6.49626613e-01 -5.14174402e-01
6.10603750e-01 5.79369068e-01 -4.15130287e-01 2.86504120e-01
4.94429544e-02 9.14395154e-01 -4.14400816e-01 5.43299377e-01
6.29135787e-01 -6.61762729e-02 4.13275421e-01 -3.20413113e-01
-2.41435423e-01 -1.21243834e-01 -1.27172852e+00 1.94332004e+00
-1.07685573e-01 6.76802576e-01 4.95698256e-03 -7.65919089e-01
7.50503182e-01 1.24607861e-01 5.15940070e-01 -7.17445076e-01
3.92705888e-01 -2.76192486e-01 -4.36446995e-01 -9.31204915e-01
7.84485877e-01 2.53574729e-01 -9.06238630e-02 3.98510307e-01
-3.19685668e-01 -4.06168640e-01 -5.13103465e-03 1.41037926e-01
9.39612746e-01 3.46980691e-01 2.79876649e-01 1.86935470e-01
5.77981174e-01 -1.21445432e-01 8.67049620e-02 5.55715919e-01
-1.03664078e-01 7.18383789e-01 2.26111397e-01 -3.55893552e-01
-1.22025526e+00 -7.38966823e-01 1.14282675e-01 1.05692375e+00
6.07262135e-01 -5.65294564e-01 -1.05422091e+00 -1.77250728e-01
-1.79231316e-01 5.10098994e-01 -1.00418222e+00 1.79183297e-02
-6.35995746e-01 -2.52441734e-01 1.60805896e-01 1.53620794e-01
6.65027142e-01 -7.95879364e-01 -1.03647387e+00 -2.14878350e-01
-4.60506022e-01 -1.38220513e+00 -4.41609442e-01 -4.65871692e-01
-7.51667261e-01 -1.22733891e+00 -5.21186650e-01 -5.18768251e-01
6.16753340e-01 6.17663801e-01 1.01506436e+00 1.07771814e-01
-3.14151973e-01 1.00416160e+00 -4.37359959e-01 -1.80720836e-01
-4.00108606e-01 -3.97328019e-01 2.81791121e-01 4.33099389e-01
1.61202103e-01 -7.73160040e-01 -5.62940180e-01 5.71003675e-01
-5.15298128e-01 3.71803999e-01 1.20910138e-01 2.72323877e-01
6.00262463e-01 -4.41458561e-02 -3.23955983e-01 -1.05627739e+00
2.92791128e-01 -2.03651324e-01 -4.61140156e-01 3.91914546e-01
-4.34890836e-02 -3.23607445e-01 1.20362669e-01 -4.72417206e-01
-1.17844760e+00 2.46400431e-01 9.94311944e-02 -5.43645203e-01
-2.95469701e-01 1.05721792e-02 -1.87037557e-01 1.22848675e-02
5.74157596e-01 1.88042089e-01 8.33598897e-02 -5.83567500e-01
3.99985909e-01 5.18312871e-01 8.23924839e-01 -3.59410614e-01
8.68258476e-01 8.68700743e-01 -1.69255771e-02 -1.03331029e+00
-6.84821427e-01 -5.97394764e-01 -1.16461611e+00 -7.61448801e-01
9.46646810e-01 -6.67916059e-01 -9.59152699e-01 5.78512728e-01
-1.02795565e+00 -1.62642002e-01 -1.79835455e-03 4.02921706e-01
-7.27293551e-01 4.23078537e-01 -1.31942645e-01 -1.00835121e+00
1.53019249e-01 -1.03379178e+00 1.25531757e+00 1.38612702e-01
-3.51249874e-01 -9.49959219e-01 2.40840837e-01 5.27691364e-01
-8.41362253e-02 7.49548912e-01 5.10318577e-01 -2.19617724e-01
-5.32986522e-01 -2.67747790e-01 2.50377417e-01 -8.16660300e-02
3.86318602e-02 -1.84531733e-01 -1.15101600e+00 -4.42922749e-02
1.93743691e-01 1.69259027e-01 2.40529060e-01 2.34935731e-01
7.58776128e-01 -9.45637599e-02 -4.96747524e-01 6.15086019e-01
1.00484931e+00 1.89143598e-01 5.80873251e-01 2.40167320e-01
9.94100153e-01 9.89443243e-01 7.29320228e-01 6.34472668e-01
5.21771967e-01 1.16793358e+00 5.24620175e-01 4.05440480e-02
-7.41442367e-02 -4.89850879e-01 2.08105206e-01 6.93703353e-01
-5.70565701e-01 4.22580317e-02 -8.10601473e-01 1.42688826e-01
-1.87603688e+00 -1.18974590e+00 -1.40754670e-01 2.28442430e+00
2.79937923e-01 -9.63345617e-02 2.89248079e-01 1.79039478e-01
9.18018758e-01 1.08485736e-01 -6.01601899e-01 5.76205812e-02
-4.83206771e-02 -2.80107349e-01 2.94398695e-01 2.29713321e-01
-8.94278228e-01 8.52680147e-01 6.50496149e+00 2.34290987e-01
-8.76185656e-01 -1.28653899e-01 1.50044367e-01 -5.04937991e-02
4.76401672e-02 -2.07575262e-01 -4.64369625e-01 2.37415671e-01
4.17537332e-01 2.49861404e-02 5.56491494e-01 7.11244583e-01
2.59091318e-01 -5.08049250e-01 -1.42707264e+00 1.46869671e+00
6.95784807e-01 -9.38721180e-01 -3.40997815e-01 1.65449396e-01
5.79219460e-01 -3.79031003e-01 5.22403903e-02 -1.66139752e-01
-1.17645733e-01 -5.15296698e-01 1.18155670e+00 7.71520197e-01
6.02959633e-01 -5.40817678e-01 4.34518218e-01 6.14722788e-01
-1.06251299e+00 -1.01200931e-01 1.79059487e-02 -1.28058359e-01
3.93204421e-01 1.73762754e-01 -9.15247023e-01 4.90527987e-01
9.79312718e-01 8.19697857e-01 -8.19348037e-01 8.73706758e-01
-1.38152450e-01 -1.69597548e-02 -3.68080467e-01 4.88630161e-02
-4.54799175e-01 -3.85985464e-01 8.80716920e-01 8.59356046e-01
3.02992642e-01 3.53885591e-01 8.20864886e-02 5.82484066e-01
1.15175970e-01 -1.49327563e-02 -8.06179523e-01 3.23633105e-01
3.31069112e-01 9.42379713e-01 -7.47459114e-01 -3.09601098e-01
-1.80218801e-01 1.24886966e+00 1.82829887e-01 1.64954737e-01
-1.00318456e+00 7.74612278e-02 3.82978141e-01 4.64103967e-01
2.15279192e-01 -3.39828014e-01 1.18055917e-01 -1.38036764e+00
1.23355195e-01 -9.53484058e-01 8.22837204e-02 -1.30596578e+00
-7.20986247e-01 6.17384493e-01 3.73392045e-01 -1.33549309e+00
-5.08224845e-01 -2.75359929e-01 -2.99412578e-01 3.95116866e-01
-8.97822022e-01 -1.23337770e+00 -7.44610667e-01 8.21864665e-01
4.40040261e-01 1.37452722e-01 5.77768743e-01 2.44991094e-01
-5.49227715e-01 3.04411620e-01 8.11214224e-02 -2.11089224e-01
7.42097676e-01 -1.02655840e+00 2.88679153e-01 6.60153151e-01
3.48677039e-01 6.33578062e-01 1.03958893e+00 -6.82326138e-01
-1.78710210e+00 -5.11099100e-01 5.32956541e-01 -1.01760638e+00
3.65482539e-01 -6.91304862e-01 -5.70378900e-01 9.10385489e-01
-9.82274264e-02 -4.11844194e-01 3.57462466e-01 1.82158053e-01
-1.80283174e-01 -1.45292640e-01 -1.01963615e+00 7.20494628e-01
1.39328647e+00 -5.23592770e-01 -6.37546301e-01 3.00299197e-01
3.65968168e-01 -8.55830073e-01 -6.56591475e-01 3.53244841e-02
1.05131733e+00 -1.42110717e+00 1.02587664e+00 -9.45503339e-02
8.53756666e-02 -2.02540606e-01 -2.49322817e-01 -1.07654977e+00
7.57465139e-02 -6.35755241e-01 -2.07340017e-01 1.13252127e+00
-2.43010372e-01 -2.75196344e-01 7.00448990e-01 8.57801259e-01
8.28335807e-02 -2.59541065e-01 -6.35230303e-01 -2.35359401e-01
-7.21465588e-01 -6.25389576e-01 6.88748002e-01 7.28394151e-01
5.38409967e-03 1.23499900e-01 -8.75855982e-01 1.88461870e-01
6.41072273e-01 2.18875438e-01 1.54530954e+00 -1.30318809e+00
-2.67484367e-01 1.78801008e-02 -7.32548773e-01 -1.26388991e+00
3.54231190e-04 -3.27013552e-01 -6.57462031e-02 -1.01596153e+00
3.12518656e-01 -3.37240696e-01 2.71461278e-01 1.18649201e-02
1.88252300e-01 3.46872598e-01 4.38723564e-01 4.30626214e-01
-8.35071266e-01 3.13639671e-01 1.13737810e+00 2.09367797e-01
-2.65161484e-01 7.37039372e-02 -2.14692503e-01 1.01217914e+00
4.92828965e-01 -1.43106580e-01 -3.83687586e-01 -4.85731632e-01
2.08997175e-01 1.58559784e-01 6.20255113e-01 -1.22682631e+00
2.45323956e-01 -8.71598348e-02 4.06167984e-01 -9.82703865e-01
7.94876993e-01 -1.22739804e+00 6.16213024e-01 1.07752897e-01
-8.00374746e-02 9.75347906e-02 2.03242190e-02 7.75377333e-01
1.69031881e-02 -2.14925203e-02 1.89621136e-01 -2.78138071e-01
-8.11245799e-01 2.05581054e-01 -3.19349654e-02 -2.80851454e-01
1.09345722e+00 -6.77749455e-01 1.57539129e-01 -6.53464317e-01
-9.06134903e-01 4.67495099e-02 9.93203104e-01 5.03374338e-01
4.88463044e-01 -1.21739590e+00 -2.00614229e-01 4.08712655e-01
1.13018513e-01 -1.38124630e-01 7.82126665e-01 8.55325460e-01
-6.36335194e-01 8.95978808e-02 -3.57895434e-01 -9.22255218e-01
-1.75840485e+00 5.69905043e-01 2.39686757e-01 1.60345480e-01
-8.81434500e-01 3.20991844e-01 3.52949500e-01 -1.40444368e-01
3.54393840e-01 -1.98319525e-01 -4.26284820e-01 4.97771166e-02
7.52712309e-01 5.77358723e-01 -1.35176688e-01 -1.32879663e+00
-2.53868967e-01 1.14974582e+00 3.20334613e-01 -3.59178126e-01
1.31907320e+00 -8.90173852e-01 5.75290546e-02 8.46544266e-01
9.31178749e-01 2.37011671e-01 -1.28534603e+00 -3.62600178e-01
-2.84772873e-01 -8.47981632e-01 -2.86491901e-01 -3.66989911e-01
-7.39836395e-01 9.36009228e-01 7.50630498e-01 3.86089802e-01
1.25967264e+00 2.19481349e-01 4.84865963e-01 4.27015632e-01
7.68171668e-01 -1.11466062e+00 8.21141005e-02 2.85093069e-01
8.61474693e-01 -1.14155209e+00 2.46518910e-01 -5.10710716e-01
-7.18136191e-01 9.41452563e-01 3.55343997e-01 6.14017211e-02
3.90244246e-01 -7.98515454e-02 -2.12851703e-01 -3.72182488e-01
-2.21516922e-01 -2.27869749e-01 2.39737272e-01 8.25695336e-01
4.39985506e-02 -1.52990241e-02 3.02817494e-01 2.46438086e-01
-5.62234104e-01 -1.47276789e-01 4.99776989e-01 9.48199034e-01
-2.69375712e-01 -7.53453732e-01 -7.99444139e-01 -1.81545779e-01
-9.63654742e-02 3.57515186e-01 -6.03796303e-01 8.97589445e-01
2.04814702e-01 7.73454309e-01 -2.79052574e-02 -6.44088745e-01
7.13462114e-01 -8.56371745e-02 8.97569835e-01 -5.09253979e-01
-4.25287873e-01 1.27595097e-01 -3.52604948e-02 -7.54960060e-01
-9.07311261e-01 -8.94044757e-01 -8.02156627e-01 -5.33802390e-01
-2.41655976e-01 -2.37622306e-01 8.76720130e-01 8.46106827e-01
2.43089750e-01 3.12279820e-01 6.69146717e-01 -1.26528752e+00
-3.09124626e-02 -8.21282268e-01 -7.57655323e-01 9.22627985e-01
4.97594446e-01 -9.27054644e-01 -2.11360484e-01 4.83945668e-01] | [7.298220634460449, -0.8949249982833862] |
e33e026d-5ff8-4e9d-b8de-baa18424c51b | transition-based-dependency-parsing-with-1 | null | null | https://aclanthology.org/P13-1104 | https://aclanthology.org/P13-1104.pdf | Transition-based Dependency Parsing with Selectional Branching | null | ['Andrew McCallum', 'Jinho D. Choi'] | 2013-08-01 | null | null | null | acl-2013-8 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.270509719848633, 3.7760043144226074] |
c874aae2-638d-406c-ab37-8bd55c464ecd | sdst-successive-decoding-for-speech-to-text | 2009.09737 | null | https://arxiv.org/abs/2009.09737v4 | https://arxiv.org/pdf/2009.09737v4.pdf | Consecutive Decoding for Speech-to-text Translation | Speech-to-text translation (ST), which directly translates the source language speech to the target language text, has attracted intensive attention recently. However, the combination of speech recognition and machine translation in a single model poses a heavy burden on the direct cross-modal cross-lingual mapping. To reduce the learning difficulty, we propose COnSecutive Transcription and Translation (COSTT), an integral approach for speech-to-text translation. The key idea is to generate source transcript and target translation text with a single decoder. It benefits the model training so that additional large parallel text corpus can be fully exploited to enhance the speech translation training. Our method is verified on three mainstream datasets, including Augmented LibriSpeech English-French dataset, IWSLT2018 English-German dataset, and TED English-Chinese dataset. Experiments show that our proposed COSTT outperforms or on par with the previous state-of-the-art methods on the three datasets. We have released our code at \url{https://github.com/dqqcasia/st}. | ['Lei LI', 'Bo Xu', 'Hao Zhou', 'Qianqian Dong', 'Shuang Xu', 'Mingxuan Wang'] | 2020-09-21 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 3.17933768e-01 -5.89786321e-02 -2.74581730e-01 -3.44190270e-01
-1.78587496e+00 -6.55297816e-01 7.03217447e-01 -5.64281166e-01
-1.89100400e-01 7.59582579e-01 4.16422725e-01 -7.24702895e-01
5.44399202e-01 -1.89085573e-01 -7.15837181e-01 -5.29349267e-01
8.10479224e-01 6.16055608e-01 6.98516518e-02 -2.58855641e-01
-1.72414079e-01 -8.80730450e-02 -7.52641261e-01 6.26279831e-01
1.10768354e+00 6.54974759e-01 6.19612992e-01 2.71768868e-01
-3.33968848e-01 4.92048562e-01 -3.82840365e-01 -7.72203922e-01
1.72214836e-01 -8.33630860e-01 -7.61738658e-01 2.51411628e-02
9.23342779e-02 -1.17926896e-02 -3.53349298e-01 1.05513978e+00
8.92970562e-01 -1.81155995e-01 3.40485036e-01 -9.35077965e-01
-7.71114945e-01 9.77877975e-01 -4.41011965e-01 4.98445854e-02
3.52100521e-01 -7.24418685e-02 7.93375373e-01 -1.42791891e+00
5.32918274e-01 1.18151414e+00 3.07176918e-01 6.88728452e-01
-9.66901183e-01 -7.70863295e-01 -2.16867268e-01 1.18767619e-01
-1.60190129e+00 -1.18924630e+00 4.70324516e-01 -1.27022281e-01
1.06977391e+00 4.03095514e-01 1.88436598e-01 1.35749090e+00
2.72446480e-02 9.12631333e-01 1.20948982e+00 -6.57956839e-01
-1.64410487e-01 2.76998758e-01 -6.63775325e-01 4.42270577e-01
-4.89780605e-01 2.40303222e-02 -8.77883136e-01 1.05357498e-01
3.54834914e-01 -2.25580394e-01 -3.97160172e-01 2.84507662e-01
-1.69110203e+00 4.35010165e-01 -1.56821996e-01 4.10801530e-01
-1.95447668e-01 -2.45937407e-01 5.96435547e-01 5.96905589e-01
6.82354212e-01 -1.47584528e-01 -5.04601598e-01 -5.04668117e-01
-9.00351465e-01 -2.73672372e-01 6.52544677e-01 1.46467960e+00
6.52507186e-01 1.71559125e-01 -4.89376858e-02 1.18171430e+00
2.70690858e-01 1.12817001e+00 7.63181269e-01 -4.63249087e-01
1.30919588e+00 3.16254318e-01 -1.44533291e-01 -3.06979865e-01
2.24063545e-01 -4.62717265e-01 -8.10712695e-01 -6.91168964e-01
-6.93168491e-02 -4.04483289e-01 -7.23403275e-01 1.50614893e+00
2.48497531e-01 1.11049972e-01 5.25890052e-01 8.60325754e-01
6.99877322e-01 1.18841505e+00 -3.65363896e-01 -5.11653125e-01
1.11109543e+00 -1.45325375e+00 -8.80180895e-01 -4.32001412e-01
8.45259964e-01 -1.57465899e+00 1.24402630e+00 -1.04618296e-01
-1.13688004e+00 -6.04108095e-01 -6.63851678e-01 -6.13520555e-02
-8.58941078e-02 8.12329173e-01 2.14475263e-02 4.81113315e-01
-9.91647959e-01 2.70614270e-02 -8.56957972e-01 -4.58104789e-01
-6.04099557e-02 2.06754729e-01 -3.14061671e-01 -1.90600187e-01
-1.29733443e+00 9.27174091e-01 2.05381766e-01 1.97782755e-01
-7.85676122e-01 -3.09694022e-01 -7.08033025e-01 -1.80049092e-01
3.36646855e-01 -4.54494029e-01 1.60124195e+00 -9.39383745e-01
-2.09416366e+00 7.98044562e-01 -7.18615949e-01 -1.63986221e-01
5.62805116e-01 -7.96559975e-02 -6.05132639e-01 -1.55316845e-01
3.34677219e-01 4.30212498e-01 7.59570122e-01 -7.28515327e-01
-6.26535356e-01 -2.29645297e-01 -6.06402397e-01 4.82451826e-01
-3.27794492e-01 5.52967668e-01 -7.58398473e-01 -7.26001203e-01
1.63983136e-01 -1.23426950e+00 1.50212854e-01 -6.43510997e-01
-4.99971420e-01 -1.97682500e-01 6.77121878e-01 -1.01952493e+00
1.29812455e+00 -2.10098147e+00 3.30334395e-01 -3.13778728e-01
-5.32514095e-01 4.47922170e-01 -1.86490968e-01 1.03545165e+00
-3.89046874e-03 -1.55548938e-02 -3.25053781e-01 -6.72645688e-01
6.61597326e-02 1.26245171e-01 -5.15100539e-01 2.83413440e-01
1.16142789e-02 1.15615451e+00 -6.24741554e-01 -5.45983434e-01
1.10145554e-01 3.89004081e-01 -2.67111901e-02 3.33590448e-01
-8.09829216e-03 6.78899527e-01 -4.56699640e-01 6.55614614e-01
5.24878263e-01 -3.01261563e-02 2.44824708e-01 1.52356058e-01
-3.25681776e-01 1.06008744e+00 -5.92101634e-01 2.05471420e+00
-8.00608873e-01 5.17776847e-01 -5.26396446e-02 -7.95906961e-01
1.03501201e+00 9.39168870e-01 5.40368930e-02 -8.92039597e-01
1.03786640e-01 8.59828889e-01 -3.25694904e-02 -4.60145921e-01
2.70092785e-01 -2.76394844e-01 -2.06114888e-01 5.96508563e-01
8.60265270e-02 -2.00176924e-01 -8.02917592e-03 -6.83376193e-02
6.29489899e-01 3.19777876e-01 1.39207259e-01 -4.78670038e-02
5.92815936e-01 6.90367669e-02 6.29164815e-01 7.58439153e-02
7.19176531e-02 6.47131801e-01 1.56311393e-02 9.54860672e-02
-1.07537448e+00 -9.25835550e-01 2.19712675e-01 9.63757098e-01
-2.19720960e-01 -4.26296771e-01 -1.05053949e+00 -7.01981068e-01
-5.41987777e-01 7.67614067e-01 9.17429104e-03 3.30836996e-02
-8.16323757e-01 -5.89318752e-01 8.37677300e-01 1.94891214e-01
4.92907375e-01 -9.76002753e-01 3.33084792e-01 2.36495540e-01
-9.76160884e-01 -1.64533782e+00 -1.23691559e+00 -5.67333885e-02
-7.56592512e-01 -3.53940845e-01 -1.01425028e+00 -1.22013211e+00
5.09058058e-01 4.30549443e-01 7.98958421e-01 -3.35844219e-01
4.08972293e-01 -2.33865649e-01 -5.37602961e-01 -8.25022236e-02
-9.60810244e-01 5.12549698e-01 2.10193411e-01 2.78257877e-01
4.13483918e-01 -5.15071869e-01 -1.39568254e-01 6.45030618e-01
-4.08098668e-01 5.39687395e-01 9.53315079e-01 7.58782327e-01
6.22927666e-01 -4.67332006e-01 7.69356489e-01 -5.17583370e-01
5.41594088e-01 -3.04927766e-01 -4.81950790e-01 4.68705684e-01
-4.07507539e-01 -1.62328705e-01 7.75806427e-01 -4.04821396e-01
-1.24987602e+00 3.47819626e-02 -4.13590252e-01 -4.58604008e-01
1.09351523e-01 6.90277994e-01 -4.68370587e-01 2.06686154e-01
3.44481111e-01 9.64938462e-01 -2.39749089e-01 -6.84836268e-01
2.52897888e-01 1.40298212e+00 4.44727033e-01 -3.95499349e-01
8.17193091e-01 -5.73927201e-02 -4.98764038e-01 -5.73856294e-01
-5.65330267e-01 -3.53168815e-01 -7.97481239e-01 4.11293656e-02
6.11205280e-01 -1.20233881e+00 -1.39987871e-01 4.42832112e-01
-1.53095615e+00 -3.90300304e-01 1.52174890e-01 7.70489991e-01
-5.74669242e-01 2.27539167e-01 -7.06309378e-01 -6.22732759e-01
-6.20001018e-01 -1.39266145e+00 1.37156570e+00 -1.95668995e-01
-2.34448761e-02 -8.22266996e-01 1.30173907e-01 7.37107456e-01
3.64625335e-01 -4.34964448e-01 5.87316096e-01 -5.94998062e-01
-5.30019343e-01 -1.28307208e-01 -9.25839767e-02 5.68424463e-01
4.37701434e-01 -2.39888787e-01 -7.37259746e-01 -3.73865485e-01
3.90291400e-02 -3.03755671e-01 4.01219994e-01 -3.15237790e-02
4.37816948e-01 -4.11601096e-01 -1.82490692e-01 4.90020543e-01
1.05136657e+00 3.20050746e-01 5.30912876e-01 3.90622877e-02
6.76015556e-01 5.12052774e-01 6.72918022e-01 -7.81661943e-02
5.93236923e-01 1.03654885e+00 -2.72117198e-01 -1.00959584e-01
-3.80527526e-01 -6.34321928e-01 1.05193698e+00 2.04104877e+00
2.36707851e-01 -4.19937521e-01 -1.05387032e+00 5.99406779e-01
-1.87640250e+00 -6.67814255e-01 -1.73595399e-01 2.12201357e+00
1.16298354e+00 -1.12086520e-01 -9.30672064e-02 -2.70636499e-01
8.43865216e-01 7.00340001e-03 -1.32312700e-01 -3.13300908e-01
-1.53736919e-01 5.01164794e-02 3.49937230e-01 7.66558230e-01
-6.07317805e-01 1.49791420e+00 5.29149961e+00 1.21813118e+00
-1.45775843e+00 6.88877523e-01 5.63043773e-01 -6.22626618e-02
-2.89993167e-01 2.30431870e-01 -9.42436814e-01 6.23244941e-01
1.40051317e+00 -5.10194242e-01 5.80752730e-01 5.78691602e-01
6.48329079e-01 4.41469193e-01 -9.46312606e-01 1.07891119e+00
1.55328766e-01 -1.22760010e+00 1.21103421e-01 -2.78232377e-02
7.48583674e-01 4.97843534e-01 1.84863061e-02 3.95769626e-01
2.05410514e-02 -7.47640610e-01 7.90280998e-01 7.04729334e-02
1.26530480e+00 -5.99514186e-01 7.33756900e-01 6.21673286e-01
-1.28730881e+00 3.83515269e-01 -3.43063980e-01 2.40526065e-01
4.52453673e-01 3.84954870e-01 -9.85858381e-01 9.28294957e-01
3.54486495e-01 7.14253664e-01 -9.91079509e-02 3.94352436e-01
-4.21047479e-01 9.69522059e-01 -1.84230626e-01 9.06160474e-03
3.07231039e-01 -3.79872859e-01 5.53109944e-01 1.38576722e+00
8.39496791e-01 -2.60198236e-01 3.12157333e-01 5.58060408e-01
-2.52117306e-01 5.73766828e-01 -5.15057385e-01 -3.24724823e-01
6.91128373e-01 1.03710032e+00 -2.86706090e-01 -4.52718168e-01
-6.30706728e-01 1.41633379e+00 1.82975903e-01 4.21999484e-01
-8.17824841e-01 -2.93484718e-01 3.88014942e-01 -3.74423079e-02
2.02086851e-01 -3.43083709e-01 -5.18335216e-02 -1.42545164e+00
4.43068892e-01 -1.30533493e+00 -2.01883450e-01 -8.19117486e-01
-9.26395655e-01 1.07705247e+00 -3.97615194e-01 -1.67695570e+00
-4.69361573e-01 -8.39169845e-02 -3.78797710e-01 1.35585988e+00
-1.49478912e+00 -1.55523169e+00 1.83819294e-01 7.00931370e-01
1.11350703e+00 -4.80050355e-01 8.48141909e-01 7.47724771e-01
-6.53258324e-01 7.48718500e-01 2.76002169e-01 2.37194598e-01
9.69320357e-01 -6.74027562e-01 1.01935041e+00 1.11705613e+00
2.91501433e-01 5.52874863e-01 3.97610813e-01 -6.77569270e-01
-1.56861258e+00 -1.23849750e+00 1.66637874e+00 -4.26649064e-01
8.01409006e-01 -7.02908337e-01 -7.02059388e-01 8.36392760e-01
7.42010415e-01 -3.05357784e-01 5.65876067e-01 -4.02942568e-01
-9.06986073e-02 -2.11962074e-01 -5.97106934e-01 7.79743135e-01
1.05897760e+00 -8.75285447e-01 -4.23985243e-01 4.79029953e-01
1.13130438e+00 -5.90680659e-01 -7.65989006e-01 2.08573624e-01
3.06128174e-01 -5.24204493e-01 4.62694079e-01 -2.60046691e-01
3.33350688e-01 -4.64403749e-01 -5.29048562e-01 -1.64906931e+00
1.21199049e-01 -1.33414042e+00 2.05529332e-01 1.47791195e+00
9.87777770e-01 -8.07365179e-01 3.27670664e-01 -2.39343345e-02
-6.76725924e-01 -7.35487700e-01 -1.31727672e+00 -9.71256614e-01
1.43655270e-01 -3.57758045e-01 7.89560974e-01 8.92434895e-01
2.00886980e-01 7.77967632e-01 -6.74711406e-01 1.58463567e-01
3.20114702e-01 1.21218748e-01 8.82256627e-01 -4.57580328e-01
-2.98181504e-01 -1.79893315e-01 1.46450326e-01 -1.48402286e+00
2.30079144e-01 -1.26639795e+00 1.63350210e-01 -1.48040473e+00
1.30343378e-01 -3.00913185e-01 1.27451479e-01 5.28220594e-01
-8.64265636e-02 1.74924657e-01 7.78153539e-02 4.63846892e-01
-3.19274306e-01 9.45738614e-01 1.43454301e+00 7.32090324e-02
-1.27008826e-01 1.58611670e-01 -5.19189715e-01 1.67567730e-01
9.94953573e-01 -8.50513875e-01 -3.96199763e-01 -8.90675783e-01
-1.41367391e-01 5.32710612e-01 -1.85491934e-01 -7.17612624e-01
3.50432754e-01 -2.23378524e-01 -2.67296255e-01 -5.42012095e-01
3.69144171e-01 -5.48649848e-01 1.97938889e-01 2.09855303e-01
-3.46711457e-01 4.08594370e-01 1.74325749e-01 9.30376127e-02
-5.99317968e-01 5.18995784e-02 5.35354733e-01 1.40649742e-02
-1.65749967e-01 3.80049050e-01 -3.15958530e-01 6.44862056e-02
5.77740550e-01 1.54821053e-01 -3.85056317e-01 -4.87595618e-01
-4.40036893e-01 7.17160478e-02 2.85176426e-01 8.69589627e-01
5.46465158e-01 -1.65089798e+00 -1.24496722e+00 2.77134001e-01
1.63754746e-01 -3.98279935e-01 1.33991674e-01 1.31962264e+00
-1.57200560e-01 7.63021827e-01 3.36469412e-01 -6.15336955e-01
-1.29247737e+00 2.89238840e-01 3.99025202e-01 -2.81904377e-02
-4.49417800e-01 6.52939320e-01 4.06409353e-02 -9.17852581e-01
-1.52186034e-02 -7.94228911e-03 4.56992447e-01 -4.65858072e-01
4.72519100e-01 1.30130947e-01 3.10471386e-01 -1.04389393e+00
-3.17460805e-01 5.16900837e-01 -1.97698787e-01 -6.39039457e-01
1.01748753e+00 -6.80728316e-01 -2.72881925e-01 4.46044773e-01
1.37507308e+00 2.67894268e-01 -7.55525649e-01 -6.50624931e-01
9.66410339e-02 -3.65307748e-01 -6.75923675e-02 -8.68036628e-01
-8.64916265e-01 1.08064878e+00 2.12724656e-01 -1.74612299e-01
1.05019915e+00 5.06457947e-02 1.33685577e+00 3.93890619e-01
4.61212486e-01 -1.12462854e+00 -2.05198005e-01 6.69847310e-01
1.01624417e+00 -1.21930349e+00 -5.72186291e-01 -5.38030505e-01
-8.73332143e-01 9.37486649e-01 3.61507326e-01 4.61283863e-01
3.00705850e-01 3.70630264e-01 6.44688129e-01 4.18855011e-01
-9.56423700e-01 -3.88204716e-02 3.13173711e-01 3.40325922e-01
7.82153010e-01 2.11389899e-01 -2.07171291e-01 4.30415034e-01
-5.12988865e-01 4.24070507e-02 2.07824424e-01 6.92413986e-01
-1.79451689e-01 -1.63998675e+00 -2.20452160e-01 -6.07474744e-02
-5.31287611e-01 -7.00682640e-01 -5.38416386e-01 3.77081037e-01
-2.48000234e-01 1.20307207e+00 -3.51677626e-01 -4.90138620e-01
2.55690753e-01 3.73443663e-01 2.11441845e-01 -7.78654039e-01
-4.49824005e-01 5.95572531e-01 3.03333759e-01 -3.11777115e-01
-2.56415963e-01 -7.95000672e-01 -1.17169762e+00 -3.94583225e-01
-3.31399024e-01 4.51463252e-01 8.13352108e-01 1.00823486e+00
5.67067325e-01 3.01201046e-01 9.36835229e-01 -5.89411318e-01
-5.88772297e-01 -1.29540718e+00 -1.83923811e-01 -1.34804845e-01
2.18560487e-01 -3.01096635e-03 -2.34605879e-01 2.49770284e-01] | [14.491615295410156, 7.225666046142578] |
55e27da9-aae3-4fb5-9ef3-fe7576aec840 | mofi-learning-image-representations-from | 2306.07952 | null | https://arxiv.org/abs/2306.07952v2 | https://arxiv.org/pdf/2306.07952v2.pdf | MOFI: Learning Image Representations from Noisy Entity Annotated Images | We present MOFI, a new vision foundation model designed to learn image representations from noisy entity annotated images. MOFI differs from previous work in two key aspects: ($i$) pre-training data, and ($ii$) training recipe. Regarding data, we introduce a new approach to automatically assign entity labels to images from noisy image-text pairs. Our approach involves employing a named entity recognition model to extract entities from the alt-text, and then using a CLIP model to select the correct entities as labels of the paired image. The approach is simple, does not require costly human annotation, and can be readily scaled up to billions of image-text pairs mined from the web. Through this method, we have created Image-to-Entities (I2E), a new large-scale dataset with 1 billion images and 2 million distinct entities, covering rich visual concepts in the wild. Building upon the I2E dataset, we study different training recipes, including supervised pre-training, contrastive pre-training, and multi-task learning. For constrastive pre-training, we treat entity names as free-form text, and further enrich them with entity descriptions. Experiments show that supervised pre-training with large-scale fine-grained entity labels is highly effective for image retrieval tasks, and multi-task training further improves the performance. The final MOFI model achieves 86.66% mAP on the challenging GPR1200 dataset, surpassing the previous state-of-the-art performance of 72.19% from OpenAI's CLIP model. Further experiments on zero-shot and linear probe image classification also show that MOFI outperforms a CLIP model trained on the original image-text data, demonstrating the effectiveness of the I2E dataset in learning strong image representations. | ['Yinfei Yang', 'Zhe Gan', 'Xianzhi Du', 'Jon Shlens', 'Yantao Zheng', 'Shuangning Liu', 'Kun Duan', 'BoWen Zhang', 'Chen Chen', 'Aleksei Timofeev', 'Wentao Wu'] | 2023-06-13 | null | null | null | null | ['multi-task-learning'] | ['methodology'] | [ 2.09702730e-01 -2.57734694e-02 -4.66082804e-02 -5.08994818e-01
-1.28375065e+00 -6.56166375e-01 5.09906292e-01 -3.31645533e-02
-7.73198843e-01 4.91475850e-01 2.39936978e-01 6.97353557e-02
1.00567833e-01 -6.58883631e-01 -9.91096318e-01 -5.29653132e-01
1.29786566e-01 5.92854261e-01 2.63777971e-01 2.84768827e-02
-1.10593542e-01 1.66111544e-01 -1.66339183e+00 6.65645838e-01
5.22648156e-01 1.31711125e+00 2.86379874e-01 5.27886629e-01
-4.37246293e-01 8.58739197e-01 -5.54934025e-01 -6.84789717e-01
1.51831985e-01 -1.92067489e-01 -1.08593929e+00 6.10122047e-02
9.37062562e-01 -2.14123651e-01 -3.01122487e-01 8.67605329e-01
6.40770018e-01 -1.01389244e-01 7.41389334e-01 -1.27117181e+00
-1.03403127e+00 4.89573896e-01 -6.92823946e-01 1.29886672e-01
2.42904946e-01 -6.40287995e-02 1.28006911e+00 -1.36682713e+00
1.10009348e+00 1.02578235e+00 9.03406322e-01 6.72395945e-01
-1.30380583e+00 -9.05296266e-01 -1.43371951e-02 1.13524333e-01
-1.78690779e+00 -4.71556008e-01 9.36829448e-02 -5.19644678e-01
1.08829296e+00 1.34529501e-01 6.97053298e-02 1.18209767e+00
-2.56577551e-01 1.06818354e+00 1.05635059e+00 -3.34013760e-01
-6.78298697e-02 3.12715620e-01 2.48830155e-01 7.81106114e-01
1.07808493e-01 -8.37076157e-02 -3.91004473e-01 -2.02450529e-01
4.16901261e-01 -8.13343972e-02 -2.76665449e-01 -5.00546515e-01
-1.34789658e+00 6.76182508e-01 7.28745520e-01 3.71986896e-01
-1.18394576e-01 2.68577933e-01 5.26853919e-01 2.13792637e-01
2.42230296e-01 4.43379045e-01 -5.35301626e-01 2.17308730e-01
-9.80907261e-01 9.88325849e-02 7.05770254e-01 1.20154572e+00
1.00332844e+00 -3.86135846e-01 -3.05511832e-01 1.00366235e+00
-4.55333758e-03 7.39825428e-01 4.58626062e-01 -8.41510117e-01
4.61374253e-01 3.86778414e-01 1.42047316e-01 -9.76409733e-01
-2.54502296e-01 -2.17197567e-01 -8.17095101e-01 -1.96597591e-01
2.51786202e-01 -1.12393118e-01 -1.35541618e+00 1.79352999e+00
6.96637705e-02 1.93506405e-01 1.24381050e-01 6.63377464e-01
1.32994831e+00 6.81907773e-01 4.67260063e-01 3.08495373e-01
1.49969947e+00 -1.14287734e+00 -4.78384376e-01 -3.96515250e-01
7.38006949e-01 -6.60657108e-01 9.76777136e-01 4.99242134e-02
-7.04221129e-01 -7.05859661e-01 -6.10924482e-01 -1.78404853e-01
-8.71880412e-01 5.65463603e-01 5.22898197e-01 4.88081515e-01
-1.18049026e+00 2.81259000e-01 -3.56349200e-01 -5.09859383e-01
7.79732704e-01 3.83956492e-01 -8.62133682e-01 -4.23559785e-01
-1.00424266e+00 7.38550544e-01 6.19860053e-01 -2.04132780e-01
-9.57350075e-01 -6.96753860e-01 -1.01100349e+00 2.07998589e-01
4.63530451e-01 -6.48152530e-01 9.78115320e-01 -1.14570510e+00
-7.11303055e-01 1.51324546e+00 -1.68601543e-01 -4.31363493e-01
7.76791275e-02 -1.98441625e-01 -6.31764591e-01 4.17461187e-01
5.04522502e-01 1.30159557e+00 8.64890635e-01 -1.41913617e+00
-6.85418963e-01 -1.34051234e-01 -2.76027359e-02 -4.19692323e-02
-3.48888725e-01 2.26894408e-01 -1.05677116e+00 -5.40850818e-01
-1.77848056e-01 -9.82819855e-01 -1.97652474e-01 -1.08963717e-02
-3.75603169e-01 -2.55366117e-01 6.86386347e-01 -4.87391025e-01
1.05198109e+00 -2.42309165e+00 -1.81688234e-01 1.01690732e-01
4.03941691e-01 3.85704815e-01 -6.76499605e-01 2.73826867e-01
-2.83888608e-01 2.83905953e-01 -2.20465437e-01 -4.88319159e-01
7.68657960e-03 1.99806526e-01 -3.17892969e-01 9.23542529e-02
5.24612486e-01 1.40165174e+00 -8.84009123e-01 -6.15134180e-01
-3.94831039e-02 3.25615168e-01 -3.46716255e-01 2.66201973e-01
-1.29594162e-01 5.74301789e-03 -3.73903692e-01 8.10983360e-01
6.19988918e-01 -7.57778823e-01 -8.57830793e-02 -5.70245981e-01
2.70755351e-01 -3.17165792e-01 -1.28048575e+00 1.70600355e+00
-4.15236562e-01 6.00712061e-01 -2.06493642e-02 -1.01720059e+00
6.17609859e-01 2.48169363e-01 4.71133024e-01 -7.37250268e-01
4.79384437e-02 3.25326979e-01 -6.25261784e-01 -6.78074777e-01
4.68251795e-01 1.42274246e-01 -5.26235163e-01 1.05791964e-01
7.46127069e-01 6.49677664e-02 4.29187238e-01 5.80386639e-01
1.18593967e+00 5.16298488e-02 1.52447611e-01 -1.72138377e-03
3.99516463e-01 5.97623810e-02 4.86204475e-01 1.02499139e+00
-3.49337101e-01 8.67821097e-01 1.67303264e-01 -4.06875312e-01
-1.04817772e+00 -1.09202003e+00 -2.33752444e-01 1.23818195e+00
3.12113792e-01 -5.75397193e-01 -6.24996841e-01 -1.03510785e+00
-4.60200720e-02 2.15991840e-01 -8.21517467e-01 2.39588380e-01
-3.13197494e-01 -7.34840810e-01 7.31274843e-01 6.02368653e-01
6.79271936e-01 -1.21754324e+00 -2.12617338e-01 -8.08878709e-03
-4.93954450e-01 -1.53131247e+00 -6.27473772e-01 3.00063819e-01
-1.45702943e-01 -1.08117497e+00 -8.41671169e-01 -1.30647230e+00
6.93832874e-01 3.39984685e-01 1.61578202e+00 -9.52561200e-02
-7.13157415e-01 8.02517951e-01 -4.55421209e-01 -1.32346749e-01
-3.49859111e-02 -5.98183721e-02 -2.60921657e-01 1.12905493e-02
6.13558412e-01 -9.10421759e-02 -5.80828309e-01 4.04428214e-01
-1.10036969e+00 -8.23066831e-02 8.46088588e-01 1.23807788e+00
9.87636805e-01 -2.74522424e-01 5.46189010e-01 -1.08413029e+00
1.00168608e-01 -5.71154892e-01 -4.23096210e-01 5.85242629e-01
-3.47297162e-01 6.62791654e-02 2.56807804e-01 -4.69350219e-01
-1.07799578e+00 4.92290497e-01 -1.71789601e-01 -5.77570796e-01
-4.01329845e-01 3.77530515e-01 -1.57124385e-01 -1.67942449e-01
7.94863164e-01 2.72098202e-02 -5.46963394e-01 -3.29715967e-01
9.25825059e-01 8.39399457e-01 9.59006608e-01 -5.43754756e-01
7.40476191e-01 4.74675715e-01 -4.82228011e-01 -7.04689443e-01
-1.19050646e+00 -1.09928060e+00 -5.63062727e-01 1.06500834e-01
1.19694066e+00 -1.39437020e+00 -3.36340189e-01 4.40235287e-01
-1.02443445e+00 -2.78646857e-01 -3.20381522e-01 1.45633966e-01
-4.54763114e-01 3.02304000e-01 -7.55914450e-01 -2.88560480e-01
-4.45114315e-01 -1.02967191e+00 1.51569915e+00 2.11590111e-01
1.86106116e-01 -6.29798114e-01 8.34746510e-02 5.59476614e-01
3.32206428e-01 1.59407511e-01 5.82692862e-01 -7.82250285e-01
-6.62207127e-01 -1.70920610e-01 -7.97844470e-01 3.90856653e-01
-3.04533720e-01 -3.91075611e-01 -1.15978193e+00 -2.41876945e-01
-4.87289518e-01 -1.09732282e+00 1.27331793e+00 9.27700028e-02
1.24525881e+00 -4.27365974e-02 -7.00461924e-01 7.91529357e-01
1.81550574e+00 -3.02167773e-01 7.96412587e-01 3.26307803e-01
7.03277826e-01 3.65369856e-01 5.57842791e-01 1.21660993e-01
5.15492976e-01 5.54792225e-01 3.11803699e-01 -4.55735326e-01
-2.97600150e-01 -2.71588057e-01 1.69480905e-01 4.58475739e-01
3.79643977e-01 -3.68379205e-01 -9.20597553e-01 9.69432294e-01
-1.78658438e+00 -1.12769449e+00 1.09053575e-01 1.79309213e+00
8.46894324e-01 -2.71027476e-01 -1.86765835e-01 -5.99241495e-01
7.67407417e-01 8.19290355e-02 -4.67737138e-01 1.90725461e-01
-4.45109725e-01 4.82640475e-01 7.66662121e-01 4.56924215e-02
-1.72010601e+00 1.25252330e+00 6.44444180e+00 1.17783034e+00
-8.38627517e-01 2.99708754e-01 8.30595195e-01 6.10874780e-02
-7.98719898e-02 -2.10110739e-01 -1.15619481e+00 2.14864388e-01
1.02097869e+00 1.70542300e-01 1.85516611e-01 9.86843348e-01
-5.02730548e-01 -6.47582188e-02 -9.03621793e-01 1.41167533e+00
4.14068609e-01 -1.43774378e+00 2.40602717e-02 -9.24257636e-02
9.84471738e-01 6.88781679e-01 -7.24191070e-02 7.01881826e-01
5.77759802e-01 -1.06751156e+00 5.93699038e-01 4.07558233e-01
1.10132980e+00 -3.02427530e-01 9.63585079e-01 1.16068088e-01
-1.27556825e+00 -9.21132192e-02 -4.88410681e-01 6.79044604e-01
9.20860916e-02 5.30998528e-01 -5.09948432e-01 4.94961262e-01
1.22140813e+00 6.11585081e-01 -9.79829788e-01 1.10357583e+00
-6.64489865e-02 4.31122661e-01 -4.52015817e-01 2.64706820e-01
2.78426766e-01 1.01416342e-01 2.98112370e-02 1.64375412e+00
2.78930336e-01 4.69659939e-02 4.52138543e-01 7.60482669e-01
-6.81845725e-01 1.59842730e-01 -6.60391092e-01 -1.48780987e-01
3.53408247e-01 1.63045573e+00 -7.22117186e-01 -6.60482824e-01
-6.71460748e-01 1.51877236e+00 6.52177036e-01 5.12451172e-01
-8.30260634e-01 -6.12607718e-01 4.81668860e-01 -2.71199226e-01
9.33576465e-01 7.72971287e-02 2.39105240e-01 -1.40833044e+00
-5.55779897e-02 -7.84506202e-01 5.10214806e-01 -9.92187083e-01
-1.57437074e+00 8.89181674e-01 -1.64273500e-01 -9.58250642e-01
-3.57412770e-02 -8.59400213e-01 -1.80658683e-01 5.75703919e-01
-1.67458272e+00 -1.53992164e+00 -4.41480815e-01 8.11788917e-01
3.25010628e-01 -1.87134687e-02 1.16149318e+00 6.12028837e-01
-5.09803832e-01 7.84576356e-01 4.38406259e-01 6.46994889e-01
1.09882581e+00 -1.24429238e+00 3.91669780e-01 7.64715254e-01
7.74155378e-01 4.57653850e-01 1.32172376e-01 -4.77057725e-01
-1.03539681e+00 -1.34143114e+00 8.69823277e-01 -7.44889081e-01
7.14158773e-01 -5.96317053e-01 -7.30544865e-01 7.73165226e-01
2.71878153e-01 6.69662952e-01 8.46219838e-01 7.91866109e-02
-8.41269374e-01 4.11761142e-02 -1.08049417e+00 2.40784615e-01
1.14280403e+00 -8.80559325e-01 -6.01536751e-01 4.98642415e-01
8.24562788e-01 -2.90225178e-01 -1.01488030e+00 4.19936687e-01
3.86438549e-01 -6.40434027e-01 1.25624621e+00 -6.40274227e-01
2.94896573e-01 -3.12131166e-01 -4.12629098e-01 -1.01012492e+00
-3.60160440e-01 -1.36460036e-01 8.47339705e-02 1.48609626e+00
5.89088261e-01 -1.39716625e-01 6.41134262e-01 4.76191074e-01
1.49593636e-01 -5.14835238e-01 -7.18852043e-01 -8.03124845e-01
-6.61020130e-02 -4.08625007e-01 2.25144222e-01 9.88041878e-01
-2.91000783e-01 5.27791440e-01 -3.91988039e-01 2.08391726e-01
4.67536837e-01 1.61683738e-01 8.23940456e-01 -9.79096353e-01
-2.96110749e-01 -7.36786947e-02 -4.50872213e-01 -1.06223428e+00
3.33342046e-01 -1.04701447e+00 3.16926718e-01 -1.59466326e+00
5.51180422e-01 -6.11408532e-01 -4.92418975e-01 8.50774646e-01
-3.85550618e-01 8.61083806e-01 2.88480222e-01 4.03118521e-01
-1.24740362e+00 3.82840276e-01 6.87204063e-01 -4.35635835e-01
2.00733647e-01 -4.76950794e-01 -6.52173400e-01 5.23861110e-01
4.35285479e-01 -5.82451820e-01 -1.34183913e-01 -5.16615391e-01
1.92251936e-01 -3.34223479e-01 3.80291402e-01 -1.02917552e+00
1.98906213e-01 2.64140606e-01 4.03836221e-01 -5.24306715e-01
2.58091360e-01 -8.61692488e-01 3.14899795e-02 -1.31779581e-01
-3.68668050e-01 -1.70866787e-01 2.36756936e-01 7.61848927e-01
-4.81181502e-01 -4.02234495e-01 6.35906398e-01 -3.68043274e-01
-1.45937538e+00 5.39905667e-01 -1.15783073e-01 4.35624957e-01
9.33722019e-01 1.46008819e-01 -4.87902373e-01 -9.55563709e-02
-9.89448845e-01 3.77429664e-01 4.39525813e-01 4.45422083e-01
4.74370807e-01 -1.28555238e+00 -5.44543147e-01 8.09937343e-03
7.57913411e-01 -2.22055241e-01 3.52616131e-01 5.11932909e-01
-2.31178388e-01 3.68848979e-01 -6.72752084e-03 -7.45836794e-01
-1.32098866e+00 8.33686173e-01 1.85519382e-01 -4.95162338e-01
-5.09243906e-01 1.12800384e+00 5.11724055e-01 -6.44192219e-01
2.64547616e-01 -1.41236007e-01 -1.61788240e-01 1.77451223e-01
7.23108053e-01 -1.78362593e-01 1.93865579e-02 -8.79251361e-01
-2.64610201e-01 7.46112585e-01 -2.34555691e-01 -8.75302404e-03
1.35408545e+00 -3.18069816e-01 -6.26921467e-03 1.30058020e-01
1.58181655e+00 -4.05943813e-03 -9.38374519e-01 -6.16550982e-01
1.72376752e-01 -3.16763073e-01 2.15732250e-02 -1.00691974e+00
-1.00230634e+00 6.91601753e-01 8.70916247e-01 -2.32235834e-01
1.19941819e+00 4.59038883e-01 8.72344255e-01 7.68245518e-01
4.73830462e-01 -1.04753220e+00 3.03028673e-01 4.87119436e-01
5.52130997e-01 -1.71431077e+00 -2.99178779e-01 -4.61898178e-01
-8.25580060e-01 7.02587783e-01 6.74660623e-01 4.78061847e-02
6.24928832e-01 1.59899041e-01 1.80116296e-01 -4.70061719e-01
-5.08733749e-01 -7.86097109e-01 3.50803614e-01 7.34717906e-01
1.99287936e-01 -9.85207409e-02 1.50797248e-01 6.67202532e-01
3.22631836e-01 1.45514965e-01 2.07789689e-01 7.39046752e-01
-3.51173967e-01 -1.04010141e+00 -1.64430588e-01 5.25821090e-01
-5.56869030e-01 -4.51286137e-01 -2.14861989e-01 6.80452526e-01
4.96623546e-01 7.63428152e-01 8.73903483e-02 -3.22546989e-01
3.61653656e-01 2.70456165e-01 2.29316562e-01 -6.63322747e-01
-5.64537883e-01 -1.32663399e-01 1.60884976e-01 -7.27964103e-01
-6.90359890e-01 -3.71315986e-01 -1.08645940e+00 1.54788956e-01
-6.29271507e-01 1.05291702e-01 5.27755618e-01 7.78374732e-01
7.60633945e-01 1.83106095e-01 4.23836350e-01 -7.55031526e-01
-1.34082764e-01 -7.37401307e-01 -4.09065604e-01 9.59326446e-01
1.24788083e-01 -6.08461201e-01 -2.76919454e-01 3.32153112e-01] | [10.111721992492676, 1.6540344953536987] |
e0a12eab-2859-43f7-9c96-50a56bedfb10 | learning-robust-deep-equilibrium-models | 2304.12707 | null | https://arxiv.org/abs/2304.12707v2 | https://arxiv.org/pdf/2304.12707v2.pdf | Learning Robust Deep Equilibrium Models | Deep equilibrium (DEQ) models have emerged as a promising class of implicit layer models in deep learning, which abandon traditional depth by solving for the fixed points of a single nonlinear layer. Despite their success, the stability of the fixed points for these models remains poorly understood. Recently, Lyapunov theory has been applied to Neural ODEs, another type of implicit layer model, to confer adversarial robustness. By considering DEQ models as nonlinear dynamic systems, we propose a robust DEQ model named LyaDEQ with guaranteed provable stability via Lyapunov theory. The crux of our method is ensuring the fixed points of the DEQ models are Lyapunov stable, which enables the LyaDEQ models to resist minor initial perturbations. To avoid poor adversarial defense due to Lyapunov-stable fixed points being located near each other, we add an orthogonal fully connected layer after the Lyapunov stability module to separate different fixed points. We evaluate LyaDEQ models on several widely used datasets under well-known adversarial attacks, and experimental results demonstrate significant improvement in robustness. Furthermore, we show that the LyaDEQ model can be combined with other defense methods, such as adversarial training, to achieve even better adversarial robustness. | ['Yao Zhao', 'Ting Liu', 'Shikui Wei', 'Haoyu Chu'] | 2023-04-25 | null | null | null | null | ['adversarial-defense'] | ['adversarial'] | [-5.94862819e-01 2.33222663e-01 2.36062482e-02 2.78472900e-01
-3.43402714e-01 -1.10137415e+00 4.08019900e-01 -3.31992567e-01
-2.31271192e-01 6.87266350e-01 -2.07448050e-01 -5.42036951e-01
-2.94275314e-01 -4.94287163e-01 -9.44617569e-01 -9.96431589e-01
-3.19767267e-01 -2.94597685e-01 2.54084855e-01 -7.57259905e-01
6.57623634e-02 8.22454453e-01 -5.65816283e-01 -3.42332900e-01
9.05694962e-01 8.02379429e-01 -6.55072093e-01 5.56088924e-01
4.35375690e-01 6.72650754e-01 -7.23510981e-01 -5.18697023e-01
6.71510100e-01 -6.97423369e-02 -5.45871198e-01 -5.67299485e-01
3.47308397e-01 -4.67607170e-01 -9.14708793e-01 1.55476487e+00
6.07815444e-01 -7.86407143e-02 6.65879548e-01 -1.72430658e+00
-7.00589657e-01 4.68781024e-01 -2.19756365e-01 3.50053944e-02
-2.37606883e-01 2.49832764e-01 8.12717319e-01 -7.51295745e-01
1.96664751e-01 1.27364075e+00 1.03832841e+00 1.06387198e+00
-1.25146532e+00 -1.13198352e+00 5.38382947e-01 -1.52725410e-02
-1.51858437e+00 -4.45650101e-01 9.29224372e-01 -4.16686803e-01
7.50831723e-01 3.58103782e-01 3.70536625e-01 1.13762546e+00
8.67860675e-01 4.87923712e-01 8.48049939e-01 1.99724913e-01
1.75906047e-01 4.14544307e-02 1.84870675e-01 8.97343874e-01
1.74888596e-01 5.41893780e-01 1.19356260e-01 -3.07547420e-01
9.98689115e-01 -9.78727825e-03 -7.12691665e-01 -3.66230577e-01
-7.30050802e-01 9.35026705e-01 9.02526855e-01 -1.55268908e-01
1.60940036e-01 3.54547322e-01 4.96771872e-01 5.79871595e-01
2.08361268e-01 6.07134163e-01 -3.95300776e-01 4.58902776e-01
-3.38928968e-01 3.69066238e-01 7.36734450e-01 8.10144961e-01
4.77033108e-01 5.24463832e-01 2.64702767e-01 3.32280189e-01
4.22860652e-01 4.88369614e-01 3.38029534e-01 -9.97357786e-01
3.91496927e-01 5.95034242e-01 -6.83502331e-02 -1.37981284e+00
-2.83249289e-01 -6.57223165e-01 -1.32659686e+00 7.73289979e-01
2.87154913e-01 -3.22186619e-01 -7.79959381e-01 2.14322805e+00
1.00121081e-01 5.23300111e-01 3.68543029e-01 8.15025091e-01
4.09592181e-01 7.74840176e-01 -4.56430405e-01 -1.22599967e-01
5.64126611e-01 -4.58037257e-01 -5.28220236e-01 2.87279963e-01
4.65974271e-01 -1.88401952e-01 8.07789505e-01 2.45253906e-01
-1.13667417e+00 -2.11567715e-01 -1.33342946e+00 2.03036577e-01
-4.69493240e-01 -3.67641032e-01 -6.03324734e-03 3.98314804e-01
-1.17587364e+00 7.72798657e-01 -8.78405631e-01 4.59603786e-01
3.26190293e-01 8.80023420e-01 -3.82377416e-01 6.60175681e-01
-1.89002550e+00 7.71355569e-01 3.94664675e-01 4.84669894e-01
-1.43408048e+00 -8.91851664e-01 -7.74358749e-01 2.21351311e-02
3.82095993e-01 -3.57506663e-01 8.49580228e-01 -7.27956414e-01
-1.70230341e+00 4.54034060e-01 3.75992328e-01 -6.05607688e-01
8.60278010e-01 -3.75991940e-01 -1.40994668e-01 5.12317307e-02
-3.36529523e-01 3.22514057e-01 1.11649561e+00 -1.27573073e+00
1.21960253e-01 -7.30545297e-02 5.78861773e-01 -1.66189641e-01
-5.60702384e-01 -1.46600902e-01 -3.11430227e-02 -8.29344273e-01
2.80848164e-02 -1.15927148e+00 -4.27251995e-01 4.46812809e-01
-5.65612614e-01 -1.10253535e-01 1.09466195e+00 -4.73547786e-01
1.22788608e+00 -2.17627954e+00 4.75499153e-01 3.31540912e-01
6.76383734e-01 8.61856937e-01 -1.60185233e-01 2.92247325e-01
-3.13236415e-01 4.38529938e-01 -3.32053274e-01 -1.72640700e-02
1.43964931e-01 2.41988122e-01 -1.03365791e+00 8.17620754e-01
3.76254499e-01 1.02624750e+00 -7.63337791e-01 6.73535466e-02
-1.72796100e-02 5.01213193e-01 -6.32440269e-01 -3.43000144e-02
2.48058084e-02 2.87303180e-01 -4.46540743e-01 1.68621719e-01
8.25733244e-01 3.56450766e-01 -2.46655852e-01 7.32646659e-02
3.92004773e-02 5.62584214e-03 -1.17887354e+00 8.73056710e-01
-1.58452585e-01 5.69375217e-01 5.36957800e-01 -1.12033916e+00
7.38165617e-01 5.25962114e-01 2.97909766e-01 -2.55866587e-01
5.09649992e-01 2.46189356e-01 2.95768548e-02 -4.76845019e-02
-4.93953191e-02 -9.44669247e-02 -2.61656672e-01 1.00353710e-01
-1.90076515e-01 -8.86137262e-02 -4.43592370e-01 4.70887840e-01
7.73180485e-01 -2.82843381e-01 -8.98522660e-02 -5.94872773e-01
9.75775421e-01 -3.20120603e-01 8.46889317e-01 5.46555758e-01
-5.62705159e-01 4.91702110e-01 8.50942075e-01 -2.65098721e-01
-9.47749853e-01 -1.27721465e+00 -1.63059160e-01 3.50200295e-01
4.02286321e-01 -2.45335579e-01 -1.01094139e+00 -8.27158749e-01
2.01697163e-02 5.48960119e-02 -8.80838752e-01 -8.92391264e-01
-8.24366510e-01 -4.31931436e-01 1.13070440e+00 6.02007091e-01
6.53105557e-01 -6.88731551e-01 -1.84786052e-01 1.35930270e-01
7.86258951e-02 -8.26612413e-01 -6.88647687e-01 2.31181189e-01
-6.70789778e-01 -1.15726840e+00 -7.73163199e-01 -7.48303354e-01
7.41821766e-01 -3.02556809e-02 5.66111326e-01 3.33685637e-01
-8.16698186e-03 1.93952620e-01 1.78220198e-01 -3.94270450e-01
-7.13245451e-01 -1.31474704e-01 7.84651220e-01 -3.88738252e-02
-5.45229197e-01 -5.78160167e-01 -4.15935397e-01 5.94620049e-01
-1.01335800e+00 -3.19964737e-01 -6.65815026e-02 8.92353237e-01
4.66584474e-01 3.30739498e-01 9.05583084e-01 -2.70838648e-01
7.49239206e-01 -5.19861162e-01 -1.01246548e+00 1.41833708e-01
-4.12967712e-01 1.00566991e-01 1.35405838e+00 -8.70165050e-01
-2.77433962e-01 -2.26906374e-01 -3.22833031e-01 -1.10561204e+00
5.06692767e-01 4.44150120e-01 -5.39844692e-01 -6.29466057e-01
6.30469620e-01 1.23444252e-01 3.88812900e-01 -1.10492669e-01
3.16559881e-01 2.77107030e-01 7.55295098e-01 -4.65472490e-01
1.60027337e+00 3.03096086e-01 1.90254048e-01 -5.68731427e-01
-6.02551937e-01 8.40819813e-03 -5.43515027e-01 -2.63503730e-01
5.84547937e-01 -7.56636977e-01 -1.03994262e+00 1.01822889e+00
-1.02420461e+00 -5.44467628e-01 -1.64881945e-01 5.32696443e-03
-3.76598805e-01 4.40019667e-01 -7.99004495e-01 -7.92843759e-01
-4.94763255e-01 -1.25158155e+00 4.96776521e-01 2.07367823e-01
1.59577534e-01 -1.41305304e+00 2.98955664e-02 -4.09202933e-01
1.90637320e-01 6.70109153e-01 8.58005762e-01 -4.10904348e-01
-3.80962104e-01 -5.49421847e-01 1.46372974e-01 6.63846076e-01
-9.07612592e-02 2.26049185e-01 -7.61454403e-01 -8.21198761e-01
3.70317370e-01 -3.65576029e-01 6.89425766e-01 1.48668274e-01
9.43851173e-01 -9.16035235e-01 -2.41434842e-01 1.15527105e+00
1.33862638e+00 1.57772139e-01 5.79953372e-01 2.79170096e-01
7.66022146e-01 3.39856297e-01 2.70950515e-02 -6.07433468e-02
-5.49817365e-03 3.93309921e-01 9.35110569e-01 -1.86212525e-01
2.56679893e-01 -6.58123046e-02 9.74214137e-01 8.92300963e-01
1.47726670e-01 3.65198590e-02 -7.21063912e-01 3.81971389e-01
-1.83036983e+00 -9.78893459e-01 8.14864263e-02 2.08907008e+00
8.07639360e-01 4.71285909e-01 8.06332976e-02 5.27307212e-01
5.81292152e-01 5.34597300e-02 -1.04322553e+00 -5.63179255e-01
-2.66424567e-01 -2.76541412e-02 6.27127171e-01 7.74878263e-01
-1.22393417e+00 9.05476451e-01 6.66100121e+00 5.66553712e-01
-1.52091312e+00 -2.48494878e-01 2.64303178e-01 5.61334193e-02
-6.16727285e-02 -2.30839357e-01 -7.74584293e-01 3.60081613e-01
6.50766134e-01 -5.75231671e-01 4.51769531e-01 7.06248760e-01
2.17502862e-01 7.98741937e-01 -8.00114930e-01 4.57652301e-01
-2.04617411e-01 -1.33257222e+00 8.82513914e-03 3.28832537e-01
9.68805611e-01 -1.50386408e-01 6.27450764e-01 4.03041542e-01
5.27381301e-01 -1.13296282e+00 8.23594868e-01 3.89049560e-01
7.52393484e-01 -1.04417729e+00 5.21217644e-01 5.66038370e-01
-1.16529095e+00 -3.26152831e-01 -6.05684876e-01 -4.80969734e-02
-1.46173507e-01 -1.56775117e-02 -2.15294883e-01 5.95405400e-01
6.34982049e-01 8.93841684e-01 -4.38402534e-01 6.82663143e-01
-4.14396703e-01 5.53737223e-01 -2.96498179e-01 2.08920464e-01
4.91930306e-01 -5.12823313e-02 1.05072618e+00 7.02911139e-01
-8.19659084e-02 1.39200419e-01 4.72818268e-03 9.91490006e-01
-1.54574364e-01 -2.75917143e-01 -7.22372353e-01 7.66311139e-02
2.73955315e-01 9.22062397e-01 -3.01814437e-01 2.80523673e-02
9.76035371e-02 7.60541737e-01 4.51692581e-01 6.42750800e-01
-1.31743240e+00 -7.40720153e-01 1.37993395e+00 -1.47995561e-01
1.52459601e-02 -6.04410410e-01 -1.72819316e-01 -1.40295398e+00
-1.57481968e-01 -1.02434182e+00 7.62015358e-02 -4.34654951e-01
-1.24351549e+00 5.80715060e-01 -8.49722922e-02 -1.44756460e+00
-5.23245335e-02 -8.08027983e-01 -1.00999022e+00 9.85640168e-01
-1.28405201e+00 -6.55139148e-01 7.93780982e-02 1.08154798e+00
7.26283193e-02 -2.19726279e-01 6.41396880e-01 1.00319050e-01
-1.03946066e+00 1.23401380e+00 4.09089029e-01 5.07991195e-01
4.66201156e-01 -1.20996225e+00 6.40552819e-01 1.27019310e+00
-2.61418551e-01 1.12718618e+00 6.06082499e-01 -4.93369907e-01
-1.44928515e+00 -1.32895589e+00 8.85781497e-02 -5.44750273e-01
9.74011064e-01 -5.40294826e-01 -1.41455650e+00 7.20144391e-01
-1.64182141e-01 6.41700476e-02 9.36308131e-02 -5.61833799e-01
-4.97825980e-01 -1.59871578e-01 -1.10115278e+00 1.06249797e+00
8.89882684e-01 -5.43512344e-01 -5.18991709e-01 -7.58694410e-02
9.81574595e-01 -5.20883322e-01 -6.79748654e-01 5.35454869e-01
4.10667062e-01 -4.99724090e-01 1.26759696e+00 -8.22095335e-01
2.58512169e-01 -4.79023635e-01 1.52736470e-01 -1.48188698e+00
-3.22813898e-01 -1.18472266e+00 -6.29732013e-01 1.04621518e+00
1.65656731e-01 -1.08097363e+00 5.29735088e-01 4.67042834e-01
-2.93615401e-01 -1.02175200e+00 -1.05652976e+00 -1.32409871e+00
1.04200876e+00 -1.10764943e-01 3.27825248e-01 9.03332472e-01
-7.91593120e-02 -5.89104071e-02 -4.96052533e-01 7.77318656e-01
5.78351915e-01 -5.49537361e-01 6.76458538e-01 -1.15333605e+00
1.15931593e-01 -7.18733966e-01 -5.67457020e-01 -8.50059211e-01
6.18392110e-01 -1.07658541e+00 1.14690259e-01 -1.12034643e+00
-4.69648629e-01 -4.68518227e-01 -7.03112364e-01 6.03329837e-01
-2.75420457e-01 3.95039469e-01 2.70801514e-01 2.43613183e-01
2.85918601e-02 7.55580008e-01 1.20011008e+00 -3.52162063e-01
-3.53623003e-01 3.20711061e-02 -6.65873528e-01 9.21375811e-01
8.89443934e-01 -2.95927614e-01 -6.25842452e-01 -4.10960853e-01
1.23967059e-01 -1.37307152e-01 6.42971158e-01 -1.13251734e+00
2.95682967e-01 -1.58737153e-01 1.41533241e-01 -1.32523939e-01
2.99639255e-01 -9.79746222e-01 2.01442633e-02 1.12397218e+00
-3.64088207e-01 1.41361520e-01 4.87567067e-01 5.46050191e-01
-3.54420058e-02 -5.99446222e-02 1.16671729e+00 4.59633350e-01
-2.36940831e-01 8.28268111e-01 -4.37488616e-01 2.53986567e-01
1.20922577e+00 -1.18423961e-01 -4.15709883e-01 -3.48432600e-01
-7.51495183e-01 4.73254621e-01 3.18657935e-01 4.05719370e-01
8.13426018e-01 -1.30280304e+00 -6.41919732e-01 4.46597159e-01
-4.20919836e-01 -1.41344462e-02 9.43358317e-02 7.22603440e-01
-5.47336698e-01 2.12129906e-01 -1.99161440e-01 -3.88128877e-01
-1.23404050e+00 8.61389637e-01 1.12347746e+00 9.06075351e-03
-7.03887701e-01 8.66868675e-01 5.59964120e-01 -3.75640661e-01
5.96634209e-01 -2.45487794e-01 -6.01836108e-02 -2.41450533e-01
3.95970315e-01 2.83003181e-01 -4.09895778e-01 -7.51224875e-01
-5.84010959e-01 7.21483767e-01 -9.95473936e-02 -1.07807204e-01
1.07084703e+00 4.26788963e-02 -2.48916950e-02 6.69207275e-02
1.39067817e+00 1.56947315e-01 -1.68421757e+00 -5.50490199e-03
-4.18342710e-01 1.08303856e-02 -1.86640099e-01 -4.47155148e-01
-1.31830978e+00 1.16258252e+00 3.16251338e-01 4.55093592e-01
1.14458263e+00 -5.15726745e-01 8.69065225e-01 4.00378257e-01
1.82822630e-01 -5.96966267e-01 1.57227293e-01 1.09089553e+00
1.19728518e+00 -8.17239583e-01 -1.75876349e-01 -1.71500906e-01
-2.04060510e-01 1.10052288e+00 8.51317406e-01 -8.39182317e-01
9.67778981e-01 3.01548719e-01 2.99133301e-01 2.89549142e-01
-6.52789116e-01 4.81247187e-01 3.30179185e-01 4.62767452e-01
-2.44414181e-01 -3.47399265e-01 -9.05468985e-02 6.65491641e-01
-2.61775494e-01 -5.10671794e-01 6.40420854e-01 6.81945801e-01
-1.88406885e-01 -8.22009563e-01 -3.04790318e-01 -2.78994054e-01
-7.34097660e-01 1.61701635e-01 -5.75160384e-01 8.76380563e-01
-1.93966970e-01 6.51579201e-01 -3.44472915e-01 -7.56117046e-01
3.36181074e-01 -1.93740562e-01 1.45356834e-01 -1.53599948e-01
-7.76126564e-01 -2.04816028e-01 -6.65192068e-01 -5.30285418e-01
3.37889165e-01 -1.78656518e-01 -1.51418293e+00 -5.87807834e-01
-4.43869114e-01 2.49684989e-01 2.21991539e-01 8.22750986e-01
2.34280065e-01 6.25496268e-01 1.10492575e+00 -7.28749335e-01
-1.08088028e+00 -3.65245581e-01 -3.68383884e-01 1.13605961e-01
1.01001060e+00 -5.98372281e-01 -8.80817354e-01 -2.72606432e-01] | [5.642575740814209, 7.856022834777832] |
7b54ffe8-3d36-4524-aeff-938de66f0b3b | netflix-and-forget-efficient-and-exact | 2302.06676 | null | https://arxiv.org/abs/2302.06676v1 | https://arxiv.org/pdf/2302.06676v1.pdf | Netflix and Forget: Efficient and Exact Machine Unlearning from Bi-linear Recommendations | People break up, miscarry, and lose loved ones. Their online streaming and shopping recommendations, however, do not necessarily update, and may serve as unhappy reminders of their loss. When users want to renege on their past actions, they expect the recommender platforms to erase selective data at the model level. Ideally, given any specified user history, the recommender can unwind or "forget", as if the record was not part of training. To that end, this paper focuses on simple but widely deployed bi-linear models for recommendations based on matrix completion. Without incurring the cost of re-training, and without degrading the model unnecessarily, we develop Unlearn-ALS by making a few key modifications to the fine-tuning procedure under Alternating Least Squares optimisation, thus applicable to any bi-linear models regardless of the training procedure. We show that Unlearn-ALS is consistent with retraining without \emph{any} model degradation and exhibits rapid convergence, making it suitable for a large class of existing recommenders. | ['Chong Wang', 'Kevin Yao', 'Xin Yang', 'Jiankai Sun', 'Mimee Xu'] | 2023-02-13 | null | null | null | null | ['matrix-completion'] | ['methodology'] | [ 3.16037238e-01 1.35007754e-01 -3.11120391e-01 -4.84421104e-01
-3.69625211e-01 -4.91596252e-01 1.36724621e-01 -3.55760939e-02
-4.83306378e-01 7.31152117e-01 3.06197494e-01 -3.49539220e-01
-4.30881619e-01 -4.39676553e-01 -9.01258171e-01 -6.62993133e-01
-1.62523240e-01 6.19545579e-01 -2.64692128e-01 -3.85766655e-01
2.30967581e-01 5.85672259e-02 -1.34738004e+00 4.01964486e-01
7.35608101e-01 6.02075100e-01 9.47922394e-02 7.00512588e-01
2.17264891e-01 9.21532691e-01 -1.35378420e-01 -8.30352128e-01
3.88896883e-01 -8.64423215e-02 -6.30055249e-01 1.99622795e-01
5.11771917e-01 -6.36483908e-01 -5.70677578e-01 7.44363129e-01
2.01808244e-01 6.23943686e-01 3.41881037e-01 -1.11983252e+00
-6.57201827e-01 9.05833006e-01 -2.88805127e-01 1.92165315e-01
2.90057033e-01 -2.25458801e-01 1.22410166e+00 -9.35594976e-01
4.02662575e-01 9.06063855e-01 1.08660591e+00 5.13953865e-01
-1.52430129e+00 -5.06713629e-01 8.37989986e-01 1.08659737e-01
-1.13143420e+00 -8.99710774e-01 4.79216963e-01 -3.33781004e-01
8.62781286e-01 7.04495430e-01 4.58184481e-01 1.25213933e+00
-6.12012893e-02 8.14863384e-01 5.58824897e-01 -2.36749440e-03
7.55980164e-02 5.90216994e-01 3.53236586e-01 2.88930833e-01
4.11417812e-01 2.91547514e-02 -7.71475792e-01 -4.85003501e-01
2.76829183e-01 5.48519492e-01 -2.92605668e-01 -5.03732681e-01
-8.18168700e-01 6.74547732e-01 -1.22418433e-01 9.56503674e-02
-4.40347314e-01 -2.71539949e-02 4.87373471e-02 8.34871233e-01
4.93936926e-01 3.85466039e-01 -7.90836751e-01 -4.53746527e-01
-1.19432724e+00 3.59200627e-01 1.13807249e+00 9.27832246e-01
6.54668331e-01 1.88747328e-02 2.81188399e-01 9.77481008e-01
1.97076753e-01 3.52290571e-01 5.57269156e-01 -9.86269534e-01
2.52372861e-01 2.78421640e-01 4.51383680e-01 -1.10181785e+00
-4.96386528e-01 -7.28379965e-01 -8.46441448e-01 -3.51580322e-01
3.35880816e-01 -2.49199271e-01 -2.83369869e-01 1.75334382e+00
2.09285408e-01 4.53607738e-01 -1.46257669e-01 6.64485931e-01
1.75712988e-01 5.39671123e-01 -3.04427117e-01 -5.73781073e-01
6.68266892e-01 -1.02457047e+00 -5.50317168e-01 -5.70292294e-01
6.93722785e-01 -5.83443582e-01 1.01133227e+00 1.06651485e+00
-1.17566836e+00 -3.08816105e-01 -8.91791940e-01 9.10601616e-02
2.11744150e-03 -6.03047907e-02 8.39415014e-01 6.75541818e-01
-9.02930737e-01 1.11433756e+00 -9.01542127e-01 -4.19064015e-01
6.61443397e-02 5.15450120e-01 -2.57739693e-01 -1.92482755e-01
-8.79968464e-01 7.39130795e-01 2.53069215e-02 3.77798557e-01
-4.69702661e-01 -8.70119989e-01 -4.36685294e-01 1.13663226e-01
5.74498832e-01 -5.86146653e-01 1.39435923e+00 -1.04992473e+00
-1.44151115e+00 3.92848641e-01 -2.35698730e-01 -4.85793740e-01
5.53888500e-01 -6.46000385e-01 -7.40575254e-01 -5.90226650e-01
-3.26522261e-01 8.85052234e-02 1.21777368e+00 -1.06469309e+00
-5.98865926e-01 -4.17600065e-01 1.48884520e-01 4.02639776e-01
-6.74432158e-01 -2.53441572e-01 -4.02998537e-01 -6.61268294e-01
1.41795069e-01 -1.16592395e+00 -2.86462396e-01 -1.45938814e-01
-1.92003414e-01 3.54030669e-01 4.08445299e-01 -7.75928140e-01
1.69077146e+00 -2.20027637e+00 1.36566490e-01 3.25313240e-01
4.06672060e-02 2.03631952e-01 -2.54501373e-01 6.27023220e-01
5.42667089e-03 -5.94169460e-02 -9.86626446e-02 -7.63149261e-01
6.92088678e-02 5.26111364e-01 -4.90134597e-01 6.91627443e-01
-4.10475075e-01 4.59597677e-01 -8.45895946e-01 2.05609828e-01
-1.46233693e-01 3.03901076e-01 -9.75457668e-01 6.81571141e-02
-7.52771273e-02 3.38112622e-01 -2.48099416e-01 2.95976132e-01
5.42130053e-01 -3.84070754e-01 4.29623693e-01 3.59466039e-02
1.08976692e-01 5.93845010e-01 -1.65291965e+00 1.62997258e+00
-3.83660525e-01 1.60562679e-01 4.14228052e-01 -1.25130510e+00
4.79121268e-01 1.61801174e-01 6.89833701e-01 -4.86174196e-01
-2.64431953e-01 3.56681310e-02 -1.47145823e-01 -4.66377258e-01
8.02712560e-01 -9.21408683e-02 2.23987952e-01 8.58206451e-01
-5.05319163e-02 7.09021807e-01 1.50100574e-01 3.95849705e-01
1.13191819e+00 1.65146351e-01 1.64423455e-02 3.50178629e-02
2.35306278e-01 -3.87239069e-01 7.46565223e-01 1.05348659e+00
3.38174284e-01 3.89836818e-01 1.25531480e-01 -3.85855198e-01
-8.34466040e-01 -8.07025135e-01 -1.20851755e-01 1.75651836e+00
-8.70841220e-02 -6.53518200e-01 -1.32858112e-01 -5.64792931e-01
2.69242644e-01 9.44052517e-01 -3.78566682e-01 -3.88068557e-01
-4.26432937e-01 -1.09633231e+00 9.86383706e-02 1.23084888e-01
-2.98202753e-01 -5.48051178e-01 -8.55017547e-03 5.83688021e-01
-2.02933446e-01 -6.63450718e-01 -9.19470012e-01 5.30247837e-02
-1.22733653e+00 -7.65211105e-01 -3.30110282e-01 -2.71337211e-01
7.24203646e-01 6.15351439e-01 1.21748483e+00 2.73884237e-01
2.67138898e-01 5.83823264e-01 -2.99602747e-01 4.84405346e-02
-2.71120131e-01 2.35503167e-01 5.25988460e-01 3.52176994e-01
2.02817440e-01 -1.09846687e+00 -5.85091770e-01 4.54642981e-01
-6.51428521e-01 -2.00711444e-01 4.34451342e-01 7.91701376e-01
3.37158650e-01 7.61783123e-02 5.55438697e-01 -1.65198684e+00
5.17231822e-01 -7.68346429e-01 -2.81595528e-01 1.54468521e-01
-1.35648072e+00 3.01316567e-02 6.61402404e-01 -8.12310874e-01
-9.23633039e-01 -2.73311317e-01 1.09414350e-05 -2.67542750e-01
3.48048747e-01 4.45566624e-01 -1.61948183e-03 4.38226946e-02
6.56554461e-01 1.38811126e-01 -1.24835089e-01 -1.02217162e+00
6.19015038e-01 7.29732513e-01 4.32985425e-01 -2.75373966e-01
7.47472823e-01 3.43095243e-01 -5.79497099e-01 -6.82316065e-01
-1.00059044e+00 -7.41963625e-01 -4.05644655e-01 -2.88759694e-02
-1.08450539e-01 -9.26309049e-01 -9.76164222e-01 1.81437299e-01
-5.70236266e-01 -4.43419784e-01 -3.22537601e-01 4.69348669e-01
-3.33240271e-01 6.28152251e-01 -6.13095343e-01 -8.70815217e-01
-2.39543587e-01 -5.45594871e-01 2.93646365e-01 5.22697940e-02
-3.91437441e-01 -1.06773603e+00 1.67562768e-01 4.42052275e-01
4.87422466e-01 -5.32178104e-01 5.72774649e-01 -9.69968677e-01
-2.32499257e-01 -4.01601076e-01 2.49917910e-01 5.14674127e-01
9.39652920e-02 -3.52585465e-01 -7.22963870e-01 -9.41409469e-01
9.91166756e-02 -7.80441388e-02 6.62193656e-01 7.84623176e-02
9.84995782e-01 -1.13453925e+00 -2.82386154e-01 7.84264326e-01
9.80241895e-01 -1.13143444e-01 3.55958998e-01 2.20379397e-01
5.40703833e-01 3.82312268e-01 4.16719735e-01 8.28443229e-01
5.36548853e-01 6.21204138e-01 2.81838298e-01 2.04126984e-01
4.07082200e-01 -4.60871041e-01 7.92312264e-01 1.03656948e+00
-7.06372922e-03 -2.14964241e-01 -1.75276309e-01 3.30395550e-01
-2.14838839e+00 -9.96202469e-01 -1.70344245e-02 2.73654699e+00
8.27632368e-01 1.93626925e-01 3.09386402e-01 -1.63539097e-01
3.89859885e-01 1.23727009e-01 -1.08157825e+00 -5.10126352e-01
-4.66355011e-02 -5.42087620e-03 8.34041476e-01 7.55062103e-01
-8.81566048e-01 6.33932233e-01 6.50749207e+00 3.49273264e-01
-8.69523287e-01 2.83647001e-01 2.99758196e-01 -7.40957677e-01
-5.38375556e-01 2.85094559e-01 -9.29971993e-01 5.32402337e-01
1.26606727e+00 -3.25547112e-03 1.29026985e+00 8.76028478e-01
2.76916146e-01 6.79560602e-02 -1.41352439e+00 9.56951499e-01
2.52751023e-01 -9.11541402e-01 -3.26729059e-01 1.67139262e-01
6.65605962e-01 1.82770476e-01 1.34310409e-01 5.81990302e-01
4.86384153e-01 -7.48577476e-01 5.01597762e-01 9.15485322e-01
2.65557259e-01 -4.32012975e-01 2.08324641e-01 7.57336617e-01
-5.00728667e-01 -4.61746484e-01 -6.62279427e-01 -2.58717537e-01
3.76351923e-01 5.79401791e-01 -6.24753237e-01 3.51774156e-01
5.87873638e-01 1.03901768e+00 -3.17857593e-01 8.56081367e-01
9.64839980e-02 1.03250539e+00 -5.62683880e-01 3.75556618e-01
-1.66760087e-01 -5.78408003e-01 6.18583441e-01 1.00058222e+00
4.90968615e-01 5.52828312e-02 1.92639008e-01 2.61269212e-01
-1.35107175e-01 1.82043016e-01 -2.44551316e-01 -9.80216786e-02
4.16897893e-01 1.12703753e+00 -1.34641938e-02 -1.78278342e-01
-6.23845398e-01 1.08895504e+00 3.16281259e-01 5.02320826e-01
-6.27968311e-01 1.94826394e-01 7.99593031e-01 4.82972056e-01
3.42252463e-01 -1.58862948e-01 3.31389951e-03 -1.59304321e+00
1.26320403e-02 -1.23615551e+00 4.65866357e-01 -4.90182906e-01
-1.40929091e+00 1.60488889e-01 -3.63737673e-01 -1.04778433e+00
-4.26468551e-01 -1.28097102e-01 -2.27216497e-01 5.24365366e-01
-1.25731242e+00 -7.27934122e-01 2.54178315e-01 5.47231197e-01
3.75798076e-01 1.02239372e-02 7.39343345e-01 7.34478056e-01
-6.26298726e-01 8.66146445e-01 7.32230246e-01 -5.27530134e-01
1.08592939e+00 -9.35039878e-01 2.66314805e-01 8.21858585e-01
4.13276911e-01 9.96287346e-01 1.04845178e+00 -5.59283912e-01
-1.65438211e+00 -9.49433148e-01 9.55091476e-01 -5.75541615e-01
6.95965827e-01 -4.99823213e-01 -1.14661670e+00 1.23243999e+00
-3.90949130e-01 -4.19132173e-01 9.20559704e-01 9.39475656e-01
-2.60247707e-01 -5.28705597e-01 -8.98668706e-01 5.37324011e-01
1.25039101e+00 -3.60599041e-01 -4.06657904e-01 5.99596143e-01
4.02934611e-01 -3.02469522e-01 -9.30757880e-01 1.14452899e-01
7.99462080e-01 -7.43637025e-01 1.12194157e+00 -1.22880995e+00
-1.86148420e-01 6.53304160e-02 -2.13497475e-01 -1.03521621e+00
-6.25301003e-01 -1.21110582e+00 -6.33030057e-01 1.24314356e+00
7.04086959e-01 -5.29946327e-01 9.65523124e-01 1.09527516e+00
5.12158358e-03 -4.20536190e-01 -7.43196666e-01 -5.76466084e-01
-1.65351495e-01 -6.29540563e-01 4.52283353e-01 9.73467708e-01
1.96931630e-01 4.45300102e-01 -1.42070627e+00 3.34497720e-01
4.93790925e-01 7.59993866e-02 9.32447135e-01 -1.26098263e+00
-1.04267144e+00 -1.31078899e-01 3.01907301e-01 -1.56871688e+00
-2.80349225e-01 -8.94894242e-01 -1.71108529e-01 -9.90334272e-01
9.31280628e-02 -7.12621868e-01 -6.32736921e-01 5.88537753e-01
-1.21725336e-01 3.26423794e-02 1.08905047e-01 6.05877519e-01
-8.74642015e-01 3.15715253e-01 7.97737777e-01 1.02891866e-02
-5.64143300e-01 8.52541387e-01 -1.31207263e+00 5.33374488e-01
6.02037728e-01 -6.85004115e-01 -6.97560549e-01 -3.15778524e-01
9.07383978e-01 9.99653563e-02 -2.25087646e-02 -6.24453664e-01
2.91076690e-01 -9.94029716e-02 -1.22955646e-02 -1.96253657e-01
4.72502679e-01 -8.70539904e-01 7.18721867e-01 1.68674409e-01
-5.95413566e-01 1.16453618e-01 -2.35913500e-01 9.38391805e-01
4.61604029e-01 -4.22013104e-01 4.35305119e-01 2.72181258e-02
-2.80709535e-01 5.01287997e-01 -4.88573849e-01 -8.44576284e-02
4.72279042e-01 -8.06161463e-02 -6.43082187e-02 -7.00140059e-01
-1.25985706e+00 4.62520063e-01 5.70645988e-01 4.75347817e-01
2.97425300e-01 -9.97005641e-01 -4.84435678e-01 2.44090632e-01
-1.61748797e-01 -3.80999714e-01 5.64334571e-01 1.00190449e+00
3.03149611e-01 2.18322381e-01 3.02100271e-01 -1.85254321e-01
-1.07488453e+00 8.75271976e-01 1.91318884e-01 -3.26413631e-01
-6.61609709e-01 7.94484794e-01 -1.55515462e-01 -6.03865445e-01
5.10860682e-01 -2.97829121e-01 -2.36579821e-01 1.86157405e-01
6.65386736e-01 4.68388349e-01 2.85148829e-01 -3.83508027e-01
-7.02740774e-02 4.22520936e-02 -6.51370406e-01 1.43158197e-01
1.64394939e+00 -6.14865005e-01 2.11515557e-02 7.96853781e-01
8.62498760e-01 2.07151428e-01 -1.48045003e+00 -6.11620903e-01
-1.11063205e-01 -5.27941227e-01 -5.13438992e-02 -7.89907932e-01
-8.81666481e-01 4.80111748e-01 2.81862289e-01 2.48711035e-01
9.52597439e-01 -4.62881774e-01 8.92640710e-01 8.51198256e-01
3.60616446e-01 -1.11669576e+00 -3.33073914e-01 3.38547349e-01
6.00406766e-01 -9.31509316e-01 4.06577677e-01 7.11020082e-02
-6.59127772e-01 7.20225275e-01 1.89710170e-01 -3.27572674e-02
9.30813253e-01 -1.30771622e-02 -2.95700878e-01 2.55461544e-01
-1.22475255e+00 3.07584196e-01 4.63994592e-02 4.91840988e-01
2.59712219e-01 -5.05688675e-02 -2.79726028e-01 9.68649983e-01
-1.93083838e-01 2.02560052e-01 6.43462002e-01 7.28643656e-01
-4.83243257e-01 -1.37390518e+00 -4.19639945e-02 8.74152899e-01
-4.93671685e-01 -2.93926090e-01 1.19884327e-01 3.60556126e-01
-6.88289776e-02 1.04341853e+00 -1.31807417e-01 -4.59464580e-01
4.91549224e-01 1.65048763e-01 2.31835708e-01 -5.06139219e-01
-5.91849327e-01 2.08212305e-02 2.91401654e-01 -5.78649402e-01
-1.56166628e-01 -1.12030995e+00 -6.22138619e-01 -8.64738345e-01
-4.25937533e-01 2.43859991e-01 3.20555031e-01 9.78088319e-01
5.71919441e-01 9.70963016e-02 7.77953327e-01 -5.42648375e-01
-1.11144388e+00 -9.41132367e-01 -7.02035606e-01 3.98747921e-01
3.46918076e-01 -5.40710330e-01 -3.96753967e-01 1.39499620e-01] | [10.009511947631836, 5.603546142578125] |
caf82247-68b3-4e69-a1a8-1d77df8ac2e9 | constructing-datasets-for-multi-hop-reading | 1710.06481 | null | http://arxiv.org/abs/1710.06481v2 | http://arxiv.org/pdf/1710.06481v2.pdf | Constructing Datasets for Multi-hop Reading Comprehension Across Documents | Most Reading Comprehension methods limit themselves to queries which can be
answered using a single sentence, paragraph, or document. Enabling models to
combine disjoint pieces of textual evidence would extend the scope of machine
comprehension methods, but currently there exist no resources to train and test
this capability. We propose a novel task to encourage the development of models
for text understanding across multiple documents and to investigate the limits
of existing methods. In our task, a model learns to seek and combine evidence -
effectively performing multi-hop (alias multi-step) inference. We devise a
methodology to produce datasets for this task, given a collection of
query-answer pairs and thematically linked documents. Two datasets from
different domains are induced, and we identify potential pitfalls and devise
circumvention strategies. We evaluate two previously proposed competitive
models and find that one can integrate information across documents. However,
both models struggle to select relevant information, as providing documents
guaranteed to be relevant greatly improves their performance. While the models
outperform several strong baselines, their best accuracy reaches 42.9% compared
to human performance at 74.0% - leaving ample room for improvement. | ['Sebastian Riedel', 'Johannes Welbl', 'Pontus Stenetorp'] | 2017-10-17 | constructing-datasets-for-multi-hop-reading-1 | https://aclanthology.org/Q18-1021 | https://aclanthology.org/Q18-1021.pdf | tacl-2018-1 | ['multi-hop-reading-comprehension'] | ['natural-language-processing'] | [ 5.78525424e-01 4.14437234e-01 -2.63819277e-01 -6.18832290e-01
-1.64701080e+00 -8.52142394e-01 8.65672767e-01 7.81316698e-01
-6.27750933e-01 8.32192302e-01 5.55805266e-01 -6.65658534e-01
-2.79195398e-01 -6.14113510e-01 -7.23746955e-01 -5.52633293e-02
2.52467364e-01 7.88232207e-01 6.33518040e-01 -3.92270476e-01
7.46475160e-01 -2.50837356e-01 -1.44802582e+00 7.69996941e-01
1.13442266e+00 3.37239325e-01 3.67340177e-01 9.17439938e-01
-4.59279507e-01 7.33260930e-01 -7.81142056e-01 -5.63331127e-01
-2.99701631e-01 -4.61417317e-01 -1.41885149e+00 -4.27873731e-01
9.65599835e-01 -3.33795011e-01 2.59021342e-01 6.30854487e-01
3.49699378e-01 -7.15138391e-02 7.98098505e-01 -9.64877605e-01
-5.33164561e-01 8.39297950e-01 -3.78299057e-01 4.66612756e-01
1.00694239e+00 -3.04872215e-01 1.27077317e+00 -7.13930190e-01
6.87130809e-01 1.19887030e+00 4.96315747e-01 6.08157516e-01
-1.29529572e+00 -2.19287977e-01 2.56145328e-01 2.43562981e-01
-9.18771327e-01 -6.17615938e-01 3.71209830e-01 -2.40044281e-01
1.29446507e+00 5.81252158e-01 2.07456291e-01 1.28263974e+00
-4.44319248e-02 1.06838679e+00 1.04998291e+00 -8.84644747e-01
-2.80148629e-02 2.38508910e-01 5.24830163e-01 6.33574784e-01
2.49841437e-01 -4.27523434e-01 -9.82481003e-01 -2.66949505e-01
6.35598823e-02 -6.13907039e-01 -3.68916482e-01 -5.45523874e-02
-1.23079503e+00 7.65816331e-01 9.79109406e-02 3.96162212e-01
-1.29266698e-02 -2.29750410e-01 2.83879697e-01 6.30066276e-01
3.75799835e-01 9.16096032e-01 -6.34981811e-01 -2.02572629e-01
-1.07677603e+00 7.56160796e-01 1.27772141e+00 8.66394460e-01
5.04914820e-01 -8.00521255e-01 -3.44841421e-01 9.50541854e-01
3.08637977e-01 1.77253157e-01 2.44875506e-01 -1.03343582e+00
1.07462573e+00 4.96405810e-01 1.96256816e-01 -1.08533180e+00
-4.32465911e-01 -2.88708240e-01 -2.02539220e-01 -3.36073190e-01
7.06862330e-01 8.61296244e-03 -5.12387097e-01 1.79014885e+00
1.56325758e-01 -3.96366209e-01 1.53832525e-01 6.10515952e-01
7.80341446e-01 5.88807881e-01 2.39731312e-01 -2.04992324e-01
1.34887075e+00 -1.06016636e+00 -6.53105080e-01 -5.70530176e-01
1.02540672e+00 -9.10292745e-01 1.20445919e+00 4.74897474e-01
-1.50591826e+00 -3.36990774e-01 -1.06821215e+00 -4.35985476e-01
-4.10200268e-01 -2.61084735e-01 3.34767133e-01 4.89645272e-01
-1.03585172e+00 4.18071777e-01 -5.34163654e-01 -4.55310434e-01
1.04922213e-01 1.06647104e-01 -2.03613713e-01 -3.89777690e-01
-1.25124252e+00 1.26144969e+00 4.87010121e-01 -2.27998063e-01
-6.00444496e-01 -5.77951550e-01 -6.74708366e-01 7.28153512e-02
5.20144403e-01 -9.76765990e-01 1.61861908e+00 -6.62415087e-01
-8.76774788e-01 1.03008425e+00 -5.61404943e-01 -3.65702301e-01
4.77139384e-01 -5.72630763e-01 -1.35302022e-01 4.26422179e-01
4.54495102e-01 9.49289620e-01 5.24553895e-01 -1.25330997e+00
-8.02151024e-01 -3.51584852e-01 2.40778595e-01 3.98577183e-01
-4.15006995e-01 2.67085165e-01 -4.64862168e-01 -4.28694546e-01
1.63022116e-01 -3.85321259e-01 7.19346851e-02 -3.14011693e-01
-4.92841840e-01 -5.56868792e-01 4.84112948e-01 -8.79589081e-01
1.43155563e+00 -1.52560973e+00 2.14484900e-01 -5.49188666e-02
2.03959137e-01 -8.52651335e-03 -3.21720958e-01 7.37024009e-01
4.17012900e-01 4.70456243e-01 -4.25567299e-01 -4.16231304e-01
2.20607370e-01 1.87930137e-01 -4.05042559e-01 -1.99667692e-01
3.56850088e-01 9.88441050e-01 -9.26912010e-01 -8.24015081e-01
-3.71264219e-01 5.86376898e-02 -5.00406146e-01 2.61137843e-01
-8.22126389e-01 -4.40926924e-02 -6.07939720e-01 6.69478834e-01
3.89818460e-01 -3.57649624e-01 1.04209989e-01 2.44717732e-01
1.40574962e-01 9.87470031e-01 -7.18743205e-01 1.84745967e+00
-2.01110959e-01 7.12000966e-01 7.68742785e-02 -9.80488181e-01
7.69566894e-01 1.97752282e-01 2.95391809e-02 -9.75921690e-01
-3.47382396e-01 4.45048898e-01 -3.10822129e-02 -9.39455152e-01
6.91473126e-01 8.07550102e-02 -1.30111873e-01 7.25305915e-01
-9.54633877e-02 -3.29109102e-01 4.48521286e-01 6.19457304e-01
1.29269636e+00 1.37947172e-01 -1.84876155e-02 -3.46465528e-01
4.35842335e-01 4.53814059e-01 -1.24847412e-01 1.39150822e+00
3.37060727e-02 4.82143104e-01 6.15280151e-01 -5.74233662e-03
-9.51167524e-01 -8.96398842e-01 -5.59045412e-02 1.59070981e+00
9.75864232e-02 -6.65913403e-01 -7.90820837e-01 -8.77479136e-01
-6.36237785e-02 1.04783511e+00 -4.14489895e-01 5.59467748e-02
-7.62963116e-01 -6.85187638e-01 6.47865593e-01 5.56853414e-01
2.52737731e-01 -8.91233206e-01 -8.00822437e-01 3.50967616e-01
-6.92989707e-01 -9.65631366e-01 -1.37036487e-01 3.11660171e-01
-7.97450662e-01 -1.08345890e+00 -4.42011625e-01 -8.53409708e-01
2.13828936e-01 2.66053468e-01 1.79851854e+00 5.24451196e-01
6.30234834e-03 6.65070117e-01 -5.66976428e-01 -6.25730932e-01
-5.67865968e-01 6.12094820e-01 -5.00917792e-01 -7.99829006e-01
7.75691390e-01 -1.84655607e-01 -4.91805553e-01 2.05302924e-01
-9.21475053e-01 2.15328634e-01 5.83434165e-01 8.24705780e-01
1.22462623e-01 -4.30157065e-01 1.04466689e+00 -8.69552433e-01
1.24588954e+00 -7.07086802e-01 -1.45052761e-01 7.26615191e-01
-7.36275434e-01 1.67654812e-01 2.38631397e-01 -7.81097934e-02
-1.17274845e+00 -4.35553014e-01 -2.36631736e-01 5.56187749e-01
-2.84529239e-01 9.65521038e-01 1.71583608e-01 4.48289782e-01
9.15891469e-01 -2.63350853e-03 -1.55950323e-01 -5.87346137e-01
4.44511116e-01 6.01482511e-01 5.40214777e-01 -1.08753395e+00
4.26451951e-01 2.21669041e-02 -4.52881604e-01 -6.57808483e-01
-1.41791582e+00 -6.30189359e-01 -6.15953505e-01 -2.02017166e-02
6.58650875e-01 -6.61887527e-01 -2.90366411e-01 -9.76515636e-02
-1.36537993e+00 -1.92824304e-01 8.33186209e-02 2.59283185e-01
-5.22249460e-01 3.60444665e-01 -5.10108054e-01 -7.02818811e-01
-2.95636684e-01 -7.39567757e-01 1.26494646e+00 1.14342459e-01
-7.64515698e-01 -1.12517953e+00 1.56681031e-01 8.74605238e-01
4.53229010e-01 -1.46916732e-01 1.32733607e+00 -1.13207984e+00
-6.06457829e-01 -8.12218115e-02 -1.75841853e-01 -4.82868999e-02
-2.88736701e-01 -9.47690085e-02 -8.89441609e-01 -2.46002942e-01
-2.25395598e-02 -9.51696634e-01 1.05910230e+00 -1.13508468e-02
1.10556221e+00 -2.96218544e-01 -5.63410938e-01 -1.14318311e-01
1.10485148e+00 -9.12727863e-02 4.06125486e-01 6.62333906e-01
2.73396432e-01 1.05850780e+00 4.44896787e-01 -7.71892369e-02
8.38322401e-01 5.68691850e-01 1.76242128e-01 1.46474853e-01
1.34024918e-02 -4.20459300e-01 1.90465018e-01 9.66214776e-01
2.61827946e-01 -5.79483390e-01 -1.07091677e+00 7.60692596e-01
-1.84811950e+00 -9.35595274e-01 -2.40647063e-01 1.99419820e+00
1.21502602e+00 4.72547829e-01 -4.46708985e-02 9.60217975e-03
1.80577010e-01 1.04182698e-01 -3.63993227e-01 -4.89403367e-01
-2.39678502e-01 3.61151069e-01 -7.35396072e-02 8.84000182e-01
-8.61910760e-01 7.49485254e-01 7.35275507e+00 4.99405056e-01
-3.98229361e-01 -1.82493418e-01 5.27342796e-01 -2.27384567e-02
-8.26240063e-01 1.06525943e-01 -9.09819961e-01 1.03203408e-01
9.81143892e-01 -2.81086098e-02 5.00547094e-03 3.44709516e-01
2.86970800e-03 -5.75341463e-01 -1.44889688e+00 2.48570114e-01
4.73636240e-01 -1.14876366e+00 1.50177270e-01 -3.57345521e-01
4.91226256e-01 -1.14648320e-01 -2.41905347e-01 4.68245834e-01
1.46427602e-01 -1.33843136e+00 4.66043353e-01 6.67606413e-01
1.44631937e-01 -3.53739470e-01 5.97065926e-01 8.65802407e-01
-5.04921198e-01 4.42223251e-02 -1.60915911e-01 -1.92724347e-01
1.85395181e-01 2.97326110e-02 -8.41063201e-01 5.28705537e-01
6.81824744e-01 3.99034739e-01 -1.10395050e+00 9.02371109e-01
-3.65627527e-01 7.63993025e-01 -3.76079679e-01 -4.60745037e-01
3.22564214e-01 1.78951889e-01 4.84526783e-01 1.40345013e+00
9.90686789e-02 -1.53152123e-02 2.07771271e-01 7.96489418e-01
-1.42402485e-01 2.51878440e-01 -5.80787897e-01 -8.96263588e-03
5.67218661e-01 7.35772073e-01 -3.51964146e-01 -5.28939903e-01
-6.45062983e-01 6.83012307e-01 7.18504190e-01 3.07684600e-01
-3.12851429e-01 -4.52954233e-01 -1.01905450e-01 -2.18215380e-02
-4.06990312e-02 -2.16069162e-01 -4.61773157e-01 -1.17892766e+00
5.21920025e-01 -1.24791491e+00 7.64092267e-01 -8.90386522e-01
-1.41611147e+00 2.40270913e-01 4.48689312e-01 -4.64542478e-01
-6.81813121e-01 -4.89783227e-01 -5.18033206e-01 1.01703608e+00
-1.67149293e+00 -1.01444316e+00 -2.64203578e-01 2.64738619e-01
8.50117445e-01 8.64144713e-02 9.13080275e-01 2.59476639e-02
-1.58523187e-01 5.65991938e-01 -5.19650280e-02 -1.45640314e-01
9.54854429e-01 -1.44793594e+00 3.13117355e-01 6.65762544e-01
3.93407881e-01 8.72473896e-01 8.77357125e-01 -6.26309693e-01
-1.09932661e+00 -2.78430015e-01 1.57527089e+00 -1.07538974e+00
6.21062517e-01 -3.16313624e-01 -1.52414787e+00 6.74541473e-01
7.69927263e-01 -9.72568214e-01 7.68127263e-01 5.90556204e-01
-5.21071911e-01 3.22500259e-01 -6.83852792e-01 5.83133340e-01
8.64468753e-01 -6.20871902e-01 -1.40461707e+00 4.96183813e-01
7.96694458e-01 -2.84511000e-01 -6.45866156e-01 4.93304700e-01
5.57714880e-01 -1.09786189e+00 8.14467430e-01 -1.04922354e+00
7.86125958e-01 -3.56678218e-02 -2.01171890e-01 -1.12309968e+00
2.00815916e-01 -4.45245206e-01 -2.50419766e-01 1.12509966e+00
9.91365314e-01 -3.81680340e-01 6.30348921e-01 8.96959901e-01
-1.87381908e-01 -7.50822902e-01 -6.86724901e-01 -4.69288290e-01
6.52040482e-01 -4.51295435e-01 3.32830757e-01 8.56780708e-01
5.24918318e-01 8.00351024e-01 1.17277354e-01 -9.88730043e-02
4.40364212e-01 1.75844610e-01 6.39963388e-01 -1.15156126e+00
-2.21257553e-01 -7.13464141e-01 4.14823771e-01 -1.55053473e+00
1.57886758e-01 -8.87656033e-01 9.32410881e-02 -1.96830904e+00
4.90913361e-01 -3.74660939e-01 1.43327201e-02 3.66196275e-01
-6.52830541e-01 -2.74385571e-01 -2.06906274e-02 1.11767054e-01
-9.53501284e-01 2.16914371e-01 1.05432379e+00 -2.57701665e-01
4.25510295e-02 -1.61134258e-01 -1.01553738e+00 5.69142103e-01
9.02094841e-01 -3.58635098e-01 -6.35299325e-01 -8.50707471e-01
5.26953459e-01 1.17572926e-01 4.77385342e-01 -6.39414132e-01
5.84076464e-01 -6.79063499e-02 4.14554149e-01 -7.97856987e-01
2.36046478e-01 -3.77324551e-01 -5.18025815e-01 6.82094246e-02
-1.10044515e+00 3.34535718e-01 2.54776359e-01 4.70487833e-01
-3.95401478e-01 -7.91855991e-01 1.75302923e-01 -3.60219747e-01
-4.26027328e-01 -4.26899582e-01 -3.65735739e-01 4.97568071e-01
5.08683145e-01 -6.98075145e-02 -8.16569567e-01 -3.71905774e-01
-4.91162270e-01 6.53755248e-01 2.47888550e-01 6.81357205e-01
6.36484146e-01 -7.57291138e-01 -9.94266987e-01 -1.40099809e-01
3.94525141e-01 8.53256956e-02 -1.04769610e-01 5.49577594e-01
-2.12532341e-01 7.69985139e-01 1.12790413e-01 -5.59242010e-01
-1.26049376e+00 2.81505257e-01 1.92471698e-01 -3.78470272e-01
-3.86782378e-01 8.24087679e-01 -3.30739319e-01 -4.04827714e-01
4.87243146e-01 -1.61477402e-01 -5.11719227e-01 2.88280636e-01
5.62285185e-01 1.81471720e-01 2.54540294e-01 -2.47233454e-02
-9.31078047e-02 4.38310206e-01 -4.76494938e-01 -4.44038451e-01
1.12306440e+00 -3.69478792e-01 -2.21847177e-01 4.81556088e-01
9.90367711e-01 -2.87180934e-02 -7.28511870e-01 -3.10161233e-01
6.40404224e-01 -4.26169753e-01 -3.18797439e-01 -1.29730868e+00
4.94613647e-02 8.85766923e-01 1.98404072e-03 6.63387120e-01
9.03963983e-01 3.24556828e-01 7.21732378e-01 8.70056510e-01
6.69426993e-02 -1.12338901e+00 3.08116436e-01 6.29948795e-01
8.20123792e-01 -1.59776163e+00 1.11940779e-01 -2.70307481e-01
-3.93648028e-01 1.11785114e+00 9.01111364e-01 5.01994908e-01
1.30029067e-01 9.01363492e-02 1.32997558e-01 -5.72912157e-01
-1.32388282e+00 -1.35800123e-01 5.58482170e-01 4.31891561e-01
7.94101536e-01 -2.82965541e-01 -4.17878717e-01 2.94124544e-01
-3.19246113e-01 -2.89182961e-01 3.50740403e-01 1.20038557e+00
-7.96197712e-01 -1.20773506e+00 -3.17139179e-01 7.11517334e-01
-5.17318368e-01 -1.89257652e-01 -7.73812711e-01 8.53890419e-01
-3.95152181e-01 1.39349818e+00 -1.09963201e-01 1.14185512e-01
3.46313655e-01 4.13377613e-01 5.34991443e-01 -5.64792335e-01
-6.27889633e-01 -8.32145587e-02 6.50341392e-01 -2.56301850e-01
-6.02284551e-01 -6.87106729e-01 -8.62740219e-01 -7.71825314e-02
-3.82998556e-01 4.09528226e-01 5.35857558e-01 1.20993459e+00
2.34834746e-01 2.50290543e-01 1.26255572e-01 -7.83835277e-02
-7.66049981e-01 -1.04619312e+00 1.33091837e-01 3.76719326e-01
3.55955094e-01 -3.22493196e-01 -1.69302270e-01 2.10991487e-01] | [11.234124183654785, 8.066829681396484] |
d40df6d9-d7e3-4561-a5fb-a1bef49fff21 | sequence-to-sequence-models-can-directly | 1703.08581 | null | http://arxiv.org/abs/1703.08581v2 | http://arxiv.org/pdf/1703.08581v2.pdf | Sequence-to-Sequence Models Can Directly Translate Foreign Speech | We present a recurrent encoder-decoder deep neural network architecture that
directly translates speech in one language into text in another. The model does
not explicitly transcribe the speech into text in the source language, nor does
it require supervision from the ground truth source language transcription
during training. We apply a slightly modified sequence-to-sequence with
attention architecture that has previously been used for speech recognition and
show that it can be repurposed for this more complex task, illustrating the
power of attention-based models. A single model trained end-to-end obtains
state-of-the-art performance on the Fisher Callhome Spanish-English speech
translation task, outperforming a cascade of independently trained
sequence-to-sequence speech recognition and machine translation models by 1.8
BLEU points on the Fisher test set. In addition, we find that making use of the
training data in both languages by multi-task training sequence-to-sequence
speech translation and recognition models with a shared encoder network can
improve performance by a further 1.4 BLEU points. | ['Ron J. Weiss', 'Navdeep Jaitly', 'Jan Chorowski', 'Yonghui Wu', 'Zhifeng Chen'] | 2017-03-24 | null | null | null | null | ['sequence-to-sequence-speech-recognition'] | ['speech'] | [ 6.34741068e-01 3.67334217e-01 -1.19657062e-01 -5.90283155e-01
-1.72332275e+00 -6.84013426e-01 7.01829433e-01 -4.58487868e-01
-4.47827697e-01 7.53281653e-01 4.13710684e-01 -9.65068102e-01
7.33844995e-01 -2.41312712e-01 -1.10905147e+00 -5.71163714e-01
4.15431082e-01 9.05516267e-01 -1.75512731e-01 -3.01015228e-01
-3.09411854e-01 1.05456792e-01 -8.57994616e-01 6.09812438e-01
6.12541258e-01 6.28433824e-01 3.13032240e-01 1.17056346e+00
-1.95489433e-02 7.71169543e-01 -6.25117540e-01 -4.49539065e-01
1.76508397e-01 -1.02171767e+00 -8.95177364e-01 9.18384269e-02
3.43118578e-01 -4.43372607e-01 -5.29070795e-01 6.99430943e-01
6.07971251e-01 -2.87691392e-02 6.49388850e-01 -5.77198327e-01
-1.12687957e+00 6.54082596e-01 -1.36677265e-01 3.17946196e-01
3.39164525e-01 2.52134800e-01 1.04664397e+00 -1.20401919e+00
4.60053593e-01 1.16392612e+00 4.69368786e-01 5.06467700e-01
-1.17189538e+00 -4.37599242e-01 -1.95746437e-01 -1.95879325e-01
-9.65585053e-01 -1.10411596e+00 7.45558739e-02 -2.36508161e-01
1.84487939e+00 1.38784021e-01 7.26789311e-02 1.47264206e+00
2.40887716e-01 8.01492810e-01 7.85135746e-01 -6.88299417e-01
-1.57114699e-01 -4.84271459e-02 -3.38331759e-01 6.62249744e-01
-2.03756005e-01 2.18447164e-01 -4.83007967e-01 2.34477192e-01
4.82102036e-01 -3.43263149e-01 -2.68198490e-01 3.00539732e-01
-1.35495067e+00 7.50576198e-01 3.08476120e-01 3.60312521e-01
-1.84403300e-01 1.74887136e-01 5.43959141e-01 7.48412788e-01
6.50914192e-01 3.81891102e-01 -6.54899180e-01 -3.42612565e-01
-1.14900148e+00 -3.25523347e-01 8.81295919e-01 1.04953051e+00
5.93338609e-01 5.62952816e-01 -1.14992879e-01 8.89381766e-01
1.45831227e-01 9.39361989e-01 8.83536220e-01 -6.41279697e-01
7.85979986e-01 -8.99209306e-02 -1.68718934e-01 -9.04007927e-02
-9.54477210e-03 -6.22149169e-01 -6.06433630e-01 -8.26832280e-02
2.06691310e-01 -4.22446400e-01 -1.20833504e+00 1.52498055e+00
-2.82591999e-01 -3.67822163e-02 4.55381036e-01 7.92322934e-01
5.40814996e-01 1.16100073e+00 -3.11733246e-01 -1.04927279e-01
9.89022970e-01 -1.48484182e+00 -5.87549508e-01 -6.21738195e-01
8.29065382e-01 -8.77152205e-01 1.04917800e+00 -9.61451009e-02
-1.39361596e+00 -6.23179436e-01 -1.09456778e+00 -2.21565336e-01
-1.02131724e-01 3.00993979e-01 -6.95271417e-02 5.82635880e-01
-1.46304274e+00 3.79934102e-01 -9.86339331e-01 -5.13184011e-01
9.79394838e-03 3.31067264e-01 -4.55715567e-01 6.14051968e-02
-1.20340550e+00 1.23922682e+00 2.13859066e-01 -8.71146172e-02
-1.35142624e+00 -3.38780344e-01 -9.29217339e-01 3.39061558e-01
3.33897434e-02 -6.13595128e-01 1.93430698e+00 -1.54712892e+00
-1.96321380e+00 8.16343129e-01 -5.96360624e-01 -6.64857924e-01
4.04807389e-01 -9.38069597e-02 -5.73311865e-01 1.13991000e-01
1.06169745e-01 5.07853270e-01 8.50720286e-01 -6.82821155e-01
-4.03642565e-01 2.86279581e-02 -3.26754421e-01 2.15127274e-01
-2.60369852e-02 4.99803901e-01 -1.81238621e-01 -6.37196302e-01
-3.16042215e-01 -1.03447843e+00 6.85363412e-02 -4.58609581e-01
-2.67638892e-01 -5.25335036e-03 5.92238307e-01 -1.20089746e+00
8.48630786e-01 -1.94520640e+00 2.69700915e-01 -2.74238229e-01
-3.13151121e-01 4.17952627e-01 -5.64172506e-01 6.78203464e-01
-3.20769250e-01 1.92484394e-01 -4.36029553e-01 -7.47825742e-01
-1.33819744e-01 2.03331068e-01 -5.14638484e-01 3.88690054e-01
6.43239796e-01 1.25724542e+00 -7.38186002e-01 1.40140980e-01
-1.59199923e-01 5.27724147e-01 -4.02744621e-01 4.39319909e-01
-1.32082701e-01 3.97585034e-01 1.93963230e-01 2.65367478e-01
6.19646013e-02 -3.83940518e-01 2.58431464e-01 4.59373772e-01
5.29889278e-02 1.17353070e+00 -6.00393191e-02 1.86146414e+00
-9.32332337e-01 1.10256338e+00 4.18291502e-02 -1.00567150e+00
8.07429552e-01 9.19418335e-01 -1.39830396e-01 -7.70052373e-01
3.90221528e-03 7.04361677e-01 4.13178504e-01 -2.43616164e-01
2.04432383e-01 -4.00201201e-01 -3.48771252e-02 7.96500802e-01
4.78021741e-01 -6.78685158e-02 -2.05196947e-01 5.43854106e-03
1.16057265e+00 8.26085955e-02 1.86405092e-01 -2.46727373e-02
3.80749345e-01 -1.72633216e-01 1.63811788e-01 6.20866179e-01
8.32967386e-02 8.22300196e-01 2.63872236e-01 -2.74718344e-01
-1.61788797e+00 -9.26062286e-01 2.34843314e-01 1.45354009e+00
-7.65078723e-01 -1.55216724e-01 -8.67326021e-01 -6.74612999e-01
-4.14549977e-01 9.51096892e-01 -3.80905151e-01 -1.25537962e-01
-7.20049441e-01 -4.66073036e-01 1.05848742e+00 6.45321786e-01
6.39106035e-02 -1.13443470e+00 -8.15438759e-03 4.07068193e-01
-4.05746430e-01 -1.26675200e+00 -8.63885880e-01 4.65638816e-01
-5.29120982e-01 -3.23726684e-01 -1.14937997e+00 -1.05930281e+00
4.92792696e-01 3.63553129e-02 1.24527109e+00 -8.65515023e-02
3.56259674e-01 -1.22018613e-01 -3.67284805e-01 -8.05761889e-02
-1.34408045e+00 6.00113630e-01 1.43118337e-01 1.64188549e-03
3.37059528e-01 -5.02734661e-01 -1.74641218e-02 2.85898566e-01
-6.02379382e-01 1.46055207e-01 9.01955009e-01 1.16789317e+00
1.39518306e-01 -7.89951801e-01 8.15527618e-01 -4.32148367e-01
5.71837664e-01 -4.13245708e-01 -4.29989815e-01 2.29801148e-01
-3.60717297e-01 3.95666867e-01 7.43574798e-01 -1.60758391e-01
-8.04422915e-01 8.87433812e-02 -4.85733539e-01 -5.12574553e-01
-1.44323900e-01 5.33314347e-01 -5.47551028e-02 2.75508434e-01
6.77908123e-01 6.78021073e-01 2.69836903e-01 -4.42671180e-01
3.83520693e-01 1.17444396e+00 5.93905509e-01 -9.79233235e-02
5.00187457e-01 -1.63809836e-01 -5.07464111e-01 -6.87095523e-01
-7.36507475e-01 -1.78099409e-01 -6.29493296e-01 3.93368810e-01
1.03760850e+00 -1.23159790e+00 -3.64376873e-01 1.81934282e-01
-1.62427664e+00 -6.66614234e-01 9.03124735e-02 4.36650872e-01
-8.10418844e-01 1.46391660e-01 -8.78213227e-01 -6.56238675e-01
-5.02200305e-01 -1.33139122e+00 1.26916301e+00 -5.31240463e-01
-2.41967574e-01 -8.83011281e-01 1.46020740e-01 2.89213777e-01
6.41840339e-01 -5.22073150e-01 8.03762555e-01 -1.19191790e+00
-5.31651616e-01 -1.68191895e-01 -7.59541318e-02 7.48405039e-01
2.35538512e-01 -2.54427999e-01 -1.00702882e+00 -4.76686925e-01
5.79746440e-02 -4.88524288e-01 1.07556486e+00 1.31557435e-01
5.03782988e-01 -5.56563497e-01 -8.33254214e-03 5.14834285e-01
9.78098571e-01 3.04260731e-01 5.54260612e-01 -1.07870214e-01
5.56589842e-01 3.87150019e-01 -2.20383942e-01 -4.44561034e-01
1.59014240e-01 7.47068405e-01 -1.16583541e-01 -1.89167172e-01
-3.60494405e-01 -4.60242331e-01 1.07870591e+00 1.35883677e+00
2.67137170e-01 -6.32811129e-01 -1.10524976e+00 6.90100551e-01
-1.65303230e+00 -1.04050779e+00 2.24538535e-01 2.04486656e+00
8.50561082e-01 2.42830470e-01 8.95322710e-02 -1.67443633e-01
6.59100354e-01 7.74738565e-02 -3.97622496e-01 -8.90937388e-01
-5.99080510e-02 3.61621261e-01 5.33552051e-01 1.11329746e+00
-7.85055399e-01 1.21908569e+00 7.20903921e+00 6.55390084e-01
-1.41627431e+00 4.67492580e-01 7.71913946e-01 -1.28244922e-01
-4.08796221e-01 -1.49035558e-01 -7.22247124e-01 3.72430533e-01
2.00405931e+00 -9.09082443e-02 7.82749414e-01 6.34332180e-01
2.09561959e-01 4.28200781e-01 -1.36855471e+00 8.12135518e-01
3.33124697e-01 -1.36099184e+00 1.43174902e-01 4.72892933e-02
8.41011345e-01 7.58904696e-01 -4.65765260e-02 4.47645545e-01
4.58907366e-01 -1.48288047e+00 7.31964350e-01 1.07146069e-01
1.40240014e+00 -5.70108175e-01 7.21342564e-01 5.27558684e-01
-7.27042913e-01 3.18832994e-01 -1.33221939e-01 -1.49272531e-01
4.45302993e-01 2.30393589e-01 -1.54086542e+00 5.05157769e-01
7.57070482e-02 5.20918250e-01 -2.10816517e-01 3.94755363e-01
-3.91094506e-01 1.10844827e+00 -2.27267087e-01 -1.39824659e-01
5.81435144e-01 9.81016979e-02 4.67881948e-01 1.63312745e+00
5.55490971e-01 -3.76287013e-01 2.30372744e-03 6.82993293e-01
-4.98013258e-01 -3.94494049e-02 -8.44465256e-01 -5.62250972e-01
1.55198887e-01 6.93029523e-01 -2.65004575e-01 -6.33060992e-01
-6.37893975e-01 1.55173838e+00 4.68318462e-01 6.48826599e-01
-5.88133335e-01 -5.73296010e-01 5.99959314e-01 2.22758483e-02
6.40349150e-01 -4.57687706e-01 -2.01658085e-01 -1.32825911e+00
-1.78965982e-02 -1.17349744e+00 -4.04239923e-01 -9.94221985e-01
-1.11126268e+00 1.25540221e+00 -6.27529323e-01 -9.58399177e-01
-1.11294401e+00 -5.97255290e-01 -5.13657928e-01 1.57840669e+00
-1.25823867e+00 -1.03187370e+00 4.27234173e-01 1.82801351e-01
1.12593567e+00 -6.25603497e-01 1.11192036e+00 2.78595984e-01
-3.33350539e-01 7.74812520e-01 3.90312254e-01 5.79612851e-01
7.08049357e-01 -9.40608859e-01 1.53155196e+00 1.09270430e+00
5.76643944e-01 4.81929153e-01 4.95973349e-01 -4.69037622e-01
-1.15309429e+00 -1.04869831e+00 1.52154005e+00 -7.02463686e-01
7.06796169e-01 -7.08288908e-01 -8.56646717e-01 1.24128366e+00
8.52272570e-01 -2.15470448e-01 6.08121216e-01 -3.03610526e-02
-4.80901152e-01 1.48584440e-01 -5.60286224e-01 5.07578909e-01
9.87744868e-01 -1.16738069e+00 -7.27013707e-01 3.87530595e-01
1.17567241e+00 -4.66749400e-01 -3.76155317e-01 -7.87178148e-03
4.64029878e-01 -5.31792581e-01 6.11684918e-01 -1.04623055e+00
7.37526953e-01 -2.12935992e-02 -4.54652011e-01 -1.73455453e+00
-2.49533594e-01 -9.27297592e-01 1.89180687e-01 8.17189336e-01
1.16541302e+00 -6.38075709e-01 3.43000442e-01 6.42714053e-02
-6.39226437e-01 -6.49027228e-01 -1.17964756e+00 -9.36527193e-01
3.65906268e-01 -2.88426429e-01 4.49303150e-01 6.31638229e-01
5.06523177e-02 9.83223259e-01 -4.56102908e-01 1.04905330e-02
3.56468931e-03 -1.95677832e-01 5.85080206e-01 -5.79120278e-01
-6.06485903e-01 -3.77076566e-01 -9.51927751e-02 -1.53645742e+00
5.19284129e-01 -1.41511810e+00 4.10539538e-01 -1.51824272e+00
-9.33577791e-02 1.19110078e-01 -4.81902622e-02 5.97847462e-01
-2.92432904e-02 1.71244845e-01 2.47561023e-01 2.29110524e-01
-2.91008472e-01 6.64448619e-01 9.19189572e-01 -1.31276265e-01
2.31501758e-01 -1.41966015e-01 -7.33357549e-01 1.89763576e-01
6.11328661e-01 -6.45270944e-01 -2.03287140e-01 -9.10073698e-01
2.22910102e-02 3.69929135e-01 4.56743874e-02 -7.52856433e-01
2.66343728e-02 2.43318737e-01 2.93190002e-01 -1.62572727e-01
3.99489194e-01 -4.43144649e-01 -1.93388328e-01 4.36697721e-01
-6.71315491e-01 4.01665062e-01 2.25873724e-01 3.37748051e-01
-2.66718149e-01 -1.80330127e-01 6.52122855e-01 -1.06844857e-01
-1.96512327e-01 -2.27523167e-02 -7.90480375e-01 7.42311031e-02
3.97411555e-01 8.49719569e-02 -2.39444882e-01 -8.85040939e-01
-6.87321126e-01 -2.07949921e-01 4.04796213e-01 6.01624489e-01
3.02528858e-01 -1.33062363e+00 -1.20465672e+00 5.46869278e-01
-9.37852040e-02 -4.86576915e-01 -3.63995403e-01 7.59980261e-01
-4.27044183e-01 8.97838891e-01 5.04103079e-02 -6.75158978e-01
-9.11257505e-01 4.13653195e-01 6.27595067e-01 -1.14880547e-01
-2.78682768e-01 8.32876861e-01 1.13996185e-01 -7.48079836e-01
-9.75707099e-02 -4.46383297e-01 7.00883985e-01 -4.49088871e-01
5.79332054e-01 -6.59745410e-02 4.77584332e-01 -7.75818586e-01
-3.28592449e-01 2.08011806e-01 -1.39831200e-01 -7.62463987e-01
1.16364050e+00 -1.36461511e-01 9.98757780e-02 6.52344763e-01
1.64489400e+00 7.29126856e-02 -1.23119628e+00 -1.72361016e-01
-7.39659369e-02 2.67020725e-02 -1.01195864e-01 -1.13147497e+00
-6.44431651e-01 1.31071651e+00 7.68875480e-02 9.56422556e-03
7.66746163e-01 1.74089838e-02 9.72242832e-01 5.61810791e-01
8.26260969e-02 -7.01562583e-01 -6.26517832e-02 1.10450077e+00
1.12916052e+00 -1.29695117e+00 -6.98484600e-01 8.64695162e-02
-7.53964961e-01 1.10781443e+00 4.64538671e-02 -7.49389082e-02
2.89356619e-01 4.83193815e-01 3.16266656e-01 4.10210669e-01
-1.09279120e+00 -9.52885970e-02 3.59126270e-01 4.34976935e-01
7.75416195e-01 8.88163671e-02 2.19713524e-01 2.51505524e-01
-3.26819211e-01 -9.55832675e-02 4.49114323e-01 4.98172879e-01
-4.65022206e-01 -1.20265055e+00 -1.22839123e-01 1.37252271e-01
-5.87074518e-01 -6.64000154e-01 -5.72200418e-01 2.04081163e-01
-5.40011108e-01 1.14449143e+00 3.28135669e-01 -3.78771454e-01
9.70720202e-02 7.16828346e-01 2.79842257e-01 -8.71546984e-01
-7.12010324e-01 2.67627120e-01 4.88748521e-01 -2.51039565e-01
-3.73261049e-02 -5.92099488e-01 -1.04097164e+00 -2.75602490e-01
-1.74228907e-01 1.29371837e-01 8.33102405e-01 1.12495911e+00
5.96079588e-01 6.82111919e-01 6.63006365e-01 -8.66582990e-01
-7.63290703e-01 -1.40189612e+00 7.31779858e-02 -3.30912205e-03
1.00342107e+00 1.22276768e-01 -3.51378560e-01 2.17692748e-01] | [14.495596885681152, 7.187502861022949] |
eceb11af-94f0-488a-aebf-652376f096a0 | beyond-ood-state-actions-supported-cross | 2306.12755 | null | https://arxiv.org/abs/2306.12755v1 | https://arxiv.org/pdf/2306.12755v1.pdf | Beyond OOD State Actions: Supported Cross-Domain Offline Reinforcement Learning | Offline reinforcement learning (RL) aims to learn a policy using only pre-collected and fixed data. Although avoiding the time-consuming online interactions in RL, it poses challenges for out-of-distribution (OOD) state actions and often suffers from data inefficiency for training. Despite many efforts being devoted to addressing OOD state actions, the latter (data inefficiency) receives little attention in offline RL. To address this, this paper proposes the cross-domain offline RL, which assumes offline data incorporate additional source-domain data from varying transition dynamics (environments), and expects it to contribute to the offline data efficiency. To do so, we identify a new challenge of OOD transition dynamics, beyond the common OOD state actions issue, when utilizing cross-domain offline data. Then, we propose our method BOSA, which employs two support-constrained objectives to address the above OOD issues. Through extensive experiments in the cross-domain offline RL setting, we demonstrate BOSA can greatly improve offline data efficiency: using only 10\% of the target data, BOSA could achieve {74.4\%} of the SOTA offline RL performance that uses 100\% of the target data. Additionally, we also show BOSA can be effortlessly plugged into model-based offline RL and noising data augmentation techniques (used for generating source-domain data), which naturally avoids the potential dynamics mismatch between target-domain data and newly generated source-domain data. | ['Donglin Wang', 'Sibo Gai', 'Yachen Kang', 'Zifeng Zhuang', 'Zhenyu Wei', 'Ziqi Zhang', 'Jinxin Liu'] | 2023-06-22 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [-2.57147640e-01 6.65754303e-02 -6.61531508e-01 -5.25493063e-02
-8.49744618e-01 -7.32129991e-01 6.45351350e-01 -4.44477610e-02
-6.46663904e-01 1.00966358e+00 1.83841959e-01 -7.23382056e-01
-1.53994024e-01 -5.25262177e-01 -8.22685897e-01 -5.19085348e-01
-4.49552506e-01 7.39768863e-01 -6.08621305e-03 -3.40409249e-01
-1.78527355e-01 3.85424942e-01 -1.43718266e+00 -2.95411974e-01
9.37709272e-01 8.02168608e-01 3.15692186e-01 6.80589139e-01
-1.73281804e-01 9.84889150e-01 -8.69205117e-01 3.25009733e-01
7.50018418e-01 -4.42429453e-01 -5.20824432e-01 2.81214006e-02
1.17321037e-01 -8.48156989e-01 -7.68229306e-01 6.15606189e-01
6.94851577e-01 4.95612264e-01 2.31850967e-01 -1.45821953e+00
-2.33820468e-01 4.84470904e-01 -3.52462858e-01 1.85231492e-01
2.22203657e-01 7.54442692e-01 6.83346868e-01 -5.11997283e-01
3.69310260e-01 1.32556188e+00 3.01446587e-01 8.20435166e-01
-1.33881509e+00 -8.24583948e-01 4.96058583e-01 -2.43487850e-01
-7.74081826e-01 -5.27766109e-01 7.40497887e-01 -1.70438215e-01
9.86356437e-01 -1.75123572e-01 7.54346073e-01 1.46215713e+00
-1.29618675e-01 1.10562074e+00 1.31337261e+00 -2.33877897e-01
5.45202315e-01 -8.95318314e-02 -4.51715998e-02 5.74601293e-01
-3.57862711e-02 6.83732748e-01 -5.88142991e-01 -1.92266971e-01
9.05598462e-01 -1.24045245e-01 1.16796881e-01 -2.96960294e-01
-8.91380966e-01 7.34322131e-01 1.07186526e-01 -1.52406439e-01
-1.53347060e-01 2.87982643e-01 4.99288708e-01 6.94121420e-01
2.40508929e-01 4.03709412e-01 -8.14182878e-01 -9.42091942e-01
-4.56726283e-01 5.24956286e-01 8.16382110e-01 1.24092627e+00
7.38991141e-01 5.88353395e-01 -2.69476473e-01 6.65024281e-01
9.55631509e-02 7.17696249e-01 7.63473213e-01 -1.02448142e+00
1.01228631e+00 4.70702827e-01 6.04561865e-01 -4.65589315e-01
-4.63635504e-01 -4.00342733e-01 -4.20618176e-01 1.62787020e-01
6.50360465e-01 -6.05178177e-01 -9.20693219e-01 2.15525842e+00
5.55479228e-01 1.50998339e-01 3.83199155e-01 6.62076652e-01
1.68673664e-01 6.28777921e-01 -7.55519718e-02 -5.11879385e-01
8.35506141e-01 -1.06860530e+00 -7.69942880e-01 -3.99723828e-01
6.55270576e-01 -2.86679029e-01 1.53395522e+00 4.09549594e-01
-1.06375277e+00 -5.59856832e-01 -7.94212162e-01 3.11535597e-01
-1.30188763e-01 -4.84570749e-02 7.13480532e-01 4.59323198e-01
-9.09940183e-01 5.38211882e-01 -1.17696190e+00 -8.87650028e-02
2.08255813e-01 6.30394638e-01 -2.33757384e-02 6.03086576e-02
-1.29148388e+00 8.82461488e-01 4.21891361e-01 -1.75093964e-01
-1.46835089e+00 -8.33121002e-01 -8.57112825e-01 -2.63084531e-01
1.03053391e+00 -1.82808697e-01 1.70536053e+00 -7.70812392e-01
-2.10305929e+00 -9.76272821e-02 -5.71169443e-02 -6.24615312e-01
6.05838776e-01 -5.00521362e-01 -4.75180089e-01 -1.42524362e-01
-2.13507876e-01 3.61714065e-01 9.05126929e-01 -1.12373531e+00
-7.40779817e-01 -1.90285891e-01 1.81685120e-01 4.46784735e-01
-3.40453535e-01 -4.87068206e-01 -6.02254689e-01 -6.65147126e-01
-4.37399119e-01 -1.02983582e+00 -2.62108743e-01 -3.72885495e-01
-1.82839274e-01 -2.92363405e-01 1.06360304e+00 -4.61284041e-01
1.27126491e+00 -2.12990856e+00 -2.14402661e-01 1.29606813e-01
4.84035872e-02 3.73031020e-01 -4.05506521e-01 6.64482236e-01
2.51354426e-01 -1.53128207e-01 -6.27443241e-03 -4.76193249e-01
2.41376072e-01 5.59515893e-01 -7.35111415e-01 2.78225303e-01
9.55839083e-02 7.43834734e-01 -1.17416084e+00 -1.44889250e-01
3.10695112e-01 -1.69936359e-01 -7.74113357e-01 5.94222844e-01
-6.94214284e-01 7.38973737e-01 -5.25870442e-01 5.66240489e-01
4.69416618e-01 1.27685135e-02 4.84811932e-01 1.54597357e-01
-1.52230501e-01 3.99623305e-01 -1.12683630e+00 1.65082312e+00
-6.04841888e-01 2.75128841e-01 5.94405755e-02 -9.58337903e-01
8.88592720e-01 8.34378898e-02 8.31788540e-01 -1.10012484e+00
3.38034071e-02 1.70228690e-01 1.58563510e-01 -4.04667050e-01
5.82856059e-01 -6.99044019e-02 -3.54332894e-01 5.85585773e-01
-1.62719823e-02 3.99366915e-02 4.60900478e-02 1.05933100e-01
1.24785912e+00 4.77496982e-01 2.98721969e-01 -2.75888667e-02
-1.22532852e-01 1.51839569e-01 7.78022408e-01 9.54130173e-01
-5.58733642e-01 -1.50610387e-01 5.48904479e-01 -2.23016933e-01
-9.75973427e-01 -1.06313920e+00 3.00660789e-01 1.10898280e+00
1.60705954e-01 -3.09495866e-01 -4.50776100e-01 -9.96729434e-01
4.63824898e-01 7.30672419e-01 -4.76080209e-01 -3.74179840e-01
-6.10346138e-01 -3.72552902e-01 7.24478006e-01 5.42677999e-01
5.91421247e-01 -9.48825955e-01 -5.73312700e-01 5.80449462e-01
-3.35885584e-02 -1.00668061e+00 -6.96941912e-01 5.09793401e-01
-9.05306101e-01 -9.95041728e-01 -3.91476870e-01 -2.55158037e-01
4.70422477e-01 2.18131557e-01 8.50602984e-01 -4.28672880e-01
6.17053770e-02 6.23565495e-01 -3.37981105e-01 -3.92508119e-01
-6.41324699e-01 8.16911738e-03 6.26089990e-01 -3.54974747e-01
1.76169038e-01 -5.39600253e-01 -4.93853509e-01 3.28991950e-01
-7.83279359e-01 -1.39578462e-01 4.74681765e-01 1.03565133e+00
3.20597768e-01 -1.11799492e-02 9.31583464e-01 -7.87711084e-01
8.52378845e-01 -6.30797923e-01 -8.37036788e-01 3.39922644e-02
-7.57693768e-01 4.46087390e-01 1.27685416e+00 -9.45843577e-01
-8.55926394e-01 -5.93132712e-02 7.69708604e-02 -9.14528906e-01
-1.50126368e-02 3.89373362e-01 -5.86077161e-02 2.88615286e-01
5.39153278e-01 4.76240605e-01 5.02301097e-01 -5.42219400e-01
3.67166460e-01 6.18939042e-01 2.92509973e-01 -1.15170527e+00
8.49178374e-01 2.16225564e-01 -2.63308108e-01 -5.50106645e-01
-8.39707911e-01 -2.93645501e-01 -1.59415510e-02 -1.19365387e-01
3.20206046e-01 -1.01535249e+00 -1.01312149e+00 4.72482264e-01
-4.47084457e-01 -1.26066387e+00 -7.34111011e-01 4.53911066e-01
-1.00482202e+00 1.58163026e-01 -3.57547402e-01 -1.25311363e+00
-1.71308462e-02 -1.14621687e+00 7.83744812e-01 1.25057101e-01
6.88277036e-02 -9.88225877e-01 3.46733272e-01 9.12438706e-02
3.26765060e-01 1.37269452e-01 7.67516971e-01 -6.79752469e-01
-2.74596393e-01 2.18213841e-01 1.94989502e-01 3.65141749e-01
4.46186900e-01 -3.59759927e-01 -7.50153780e-01 -9.10329819e-01
-9.23944414e-02 -9.01539564e-01 3.42848450e-01 8.13971832e-02
1.39339936e+00 -5.77752471e-01 1.04667127e-01 2.98689246e-01
1.22139442e+00 5.70026755e-01 1.78845704e-01 2.62138337e-01
5.03278196e-01 2.11137921e-01 1.27591777e+00 1.16550732e+00
6.26889229e-01 7.31941283e-01 4.79124099e-01 -8.75437409e-02
6.51455745e-02 -9.08039987e-01 9.71148431e-01 8.01428020e-01
2.95408070e-01 -2.68551290e-01 -6.78196847e-01 4.92588878e-01
-1.95214069e+00 -9.02421415e-01 5.23615360e-01 2.45933080e+00
1.26266897e+00 1.23230591e-01 6.85155034e-01 -9.60051790e-02
2.35224217e-01 1.54533744e-01 -1.41737914e+00 -3.58400613e-01
2.09148884e-01 3.78324300e-01 7.57425368e-01 4.62104112e-01
-8.16773593e-01 1.14319229e+00 6.49416733e+00 9.05551016e-01
-1.15358555e+00 -1.13105578e-02 2.29774430e-01 -2.10412771e-01
-9.89801660e-02 7.21339136e-02 -1.01465988e+00 7.22160339e-01
1.15209520e+00 -7.25791082e-02 1.07473385e+00 9.52192724e-01
6.48391068e-01 -1.78097337e-01 -1.24159706e+00 9.10098672e-01
-3.60746086e-01 -8.81592214e-01 -2.40195185e-01 2.92480558e-01
6.58105135e-01 7.15728626e-02 1.03518046e-01 1.25807869e+00
1.03273761e+00 -7.57254958e-01 7.75977075e-01 3.01107526e-01
1.01448596e+00 -9.34181809e-01 9.93258804e-02 9.31779206e-01
-1.17550647e+00 -6.14069939e-01 -2.03025103e-01 -1.86889291e-01
-1.46824270e-01 2.11372554e-01 -1.01242971e+00 4.95862573e-01
3.16820621e-01 7.12445080e-01 -1.96919188e-01 4.65175956e-01
7.02336207e-02 9.04145598e-01 -5.15469491e-01 2.51679998e-02
4.56448436e-01 -6.75339699e-02 4.10717845e-01 6.00592852e-01
2.02143595e-01 -2.91678160e-02 8.12000990e-01 5.70527315e-01
2.83999518e-02 -4.02947068e-01 -7.64580429e-01 -3.40255141e-01
8.57303023e-01 6.62342072e-01 4.99748439e-02 -3.14374030e-01
-2.06146389e-01 7.47269928e-01 5.85796952e-01 5.35369992e-01
-1.04662097e+00 -1.52944177e-01 9.91131246e-01 -1.64047495e-01
6.82838559e-02 -6.45655692e-01 3.42989653e-01 -1.24425375e+00
-1.96759239e-01 -1.40246284e+00 3.42953414e-01 -2.79387146e-01
-1.29071474e+00 2.80008316e-02 1.97924361e-01 -1.50130928e+00
-5.73735237e-01 -3.45601618e-01 -2.48994321e-01 5.80283940e-01
-1.53990448e+00 -6.56876683e-01 -5.15752956e-02 8.90818059e-01
8.26162636e-01 -2.13915214e-01 5.70514560e-01 3.00269246e-01
-7.32190430e-01 8.92322838e-01 4.53109652e-01 -1.50324374e-01
8.11825573e-01 -1.35135663e+00 3.27351302e-01 5.07389486e-01
-1.47723973e-01 5.94701588e-01 6.72247529e-01 -7.12777793e-01
-2.02940464e+00 -1.25587261e+00 5.78953885e-02 -3.81069422e-01
7.64742553e-01 -4.91636246e-01 -5.26449323e-01 7.00778067e-01
-2.20798358e-01 7.95123726e-02 4.05313015e-01 -3.30463843e-03
-2.91613229e-02 -1.42027289e-01 -1.17981946e+00 8.31743300e-01
1.17910874e+00 -4.13567275e-01 -4.16389048e-01 1.35622248e-01
9.65231061e-01 -7.49599040e-01 -8.54218066e-01 2.22376034e-01
2.55127907e-01 -5.10760427e-01 7.68599212e-01 -9.80854928e-01
-1.32186964e-01 -2.11944297e-01 -6.50819615e-02 -1.47803807e+00
2.09679365e-01 -1.06545591e+00 -8.44041765e-01 1.11212873e+00
4.26503159e-02 -9.07049537e-01 7.29510128e-01 6.17573977e-01
-1.77414298e-01 -8.16605747e-01 -8.91111732e-01 -1.31368065e+00
1.21765748e-01 -5.51066518e-01 6.70148373e-01 7.90626407e-01
8.97993147e-03 9.99787543e-03 -7.19512224e-01 7.46707395e-02
4.66436237e-01 1.50121346e-01 1.41494846e+00 -5.49828410e-01
-7.79365897e-01 1.84577168e-03 3.21429402e-01 -1.72984743e+00
2.30077341e-01 -5.61804950e-01 2.53339767e-01 -1.15548337e+00
-2.43187174e-01 -9.74721491e-01 -2.40459308e-01 7.74637103e-01
-3.72354314e-02 -4.87284303e-01 4.11002129e-01 2.85994172e-01
-6.42353237e-01 1.08463275e+00 1.54762661e+00 1.70817256e-01
-7.03029573e-01 1.00039780e-01 -5.84047139e-01 3.69495779e-01
9.02013421e-01 -3.93682271e-01 -9.68411982e-01 -2.65775800e-01
-1.68459401e-01 5.81121862e-01 1.18131280e-01 -9.43085551e-01
-1.07114471e-01 -7.42273808e-01 -1.13218583e-01 -5.83636999e-01
3.79909486e-01 -9.19744253e-01 -1.54086560e-01 6.19914234e-01
-3.58146816e-01 9.78710428e-02 4.98420596e-01 9.90836740e-01
7.86762461e-02 1.86706439e-01 5.38921058e-01 3.98912281e-03
-6.35677218e-01 5.74120820e-01 -5.37641883e-01 6.02131128e-01
1.15318573e+00 -5.21000698e-02 -4.25813198e-01 -5.69318533e-01
-6.35890722e-01 7.21228480e-01 3.34807247e-01 4.31073517e-01
2.62696922e-01 -1.24930406e+00 -3.34043354e-02 3.57801557e-01
-1.33456916e-01 2.01104537e-01 1.86501577e-01 7.10183859e-01
1.73772424e-02 2.65763849e-01 -7.01709138e-03 -3.41523319e-01
-6.22341275e-01 5.22123814e-01 2.72866726e-01 -6.98786318e-01
-6.60190940e-01 3.22075307e-01 -1.52449980e-01 -7.59425402e-01
4.91953999e-01 -6.24944925e-01 1.81765452e-01 -1.56074062e-01
1.09162800e-01 4.66925472e-01 -1.98284701e-01 1.49929613e-01
-1.88757293e-02 -2.91901454e-02 -1.97582208e-02 -3.13939393e-01
1.18386364e+00 -1.37196947e-02 7.67368078e-01 7.09006906e-01
8.15148532e-01 7.88847059e-02 -2.10936093e+00 -2.57402450e-01
-7.55707845e-02 -4.60138470e-01 -2.23779976e-01 -1.10529315e+00
-7.77968168e-01 7.00642169e-01 5.54497421e-01 1.19033031e-01
9.40238297e-01 -3.23043346e-01 1.02121329e+00 6.11783266e-01
8.34256947e-01 -1.54044104e+00 4.71546054e-01 8.97667587e-01
5.72484493e-01 -1.30850732e+00 -1.99821666e-01 2.88264543e-01
-8.80131304e-01 7.29359090e-01 1.01363826e+00 -2.32361376e-01
3.98491502e-01 3.90345365e-01 -1.01977065e-01 1.95519373e-01
-1.19680023e+00 -3.28155637e-01 -3.25399518e-01 6.43290758e-01
-2.23313212e-01 1.28511906e-01 -5.23310751e-02 6.62773848e-01
-8.84752125e-02 1.04449823e-01 3.82219434e-01 1.32290995e+00
-3.30744028e-01 -1.41796148e+00 -1.87806740e-01 5.04682600e-01
-4.85801250e-02 1.72346100e-01 -3.85609195e-02 1.12936163e+00
-1.04531057e-01 8.20813000e-01 1.92903101e-01 -4.79341030e-01
3.81946594e-01 1.12609779e-02 3.58697385e-01 -4.81481701e-01
-6.20738804e-01 7.09048808e-02 1.71957910e-01 -8.43518555e-01
-2.92061809e-02 -4.76323217e-01 -1.60165441e+00 -3.46678853e-01
-1.90718696e-01 1.12950727e-01 5.38311660e-01 8.71860862e-01
7.37223685e-01 5.20299792e-01 1.10055017e+00 -7.22373664e-01
-1.45125675e+00 -8.95914972e-01 -5.01529634e-01 2.39857286e-01
5.42938411e-01 -1.08182812e+00 -4.83942419e-01 -4.36707109e-01] | [4.069471836090088, 2.157606601715088] |
12a1561b-595a-4433-95f5-0b85a59e712e | overcoming-barriers-to-data-sharing-with | 2012.03769 | null | https://arxiv.org/abs/2012.03769v3 | https://arxiv.org/pdf/2012.03769v3.pdf | Overcoming Barriers to Data Sharing with Medical Image Generation: A Comprehensive Evaluation | Privacy concerns around sharing personally identifiable information are a major practical barrier to data sharing in medical research. However, in many cases, researchers have no interest in a particular individual's information but rather aim to derive insights at the level of cohorts. Here, we utilize Generative Adversarial Networks (GANs) to create derived medical imaging datasets consisting entirely of synthetic patient data. The synthetic images ideally have, in aggregate, similar statistical properties to those of a source dataset but do not contain sensitive personal information. We assess the quality of synthetic data generated by two GAN models for chest radiographs with 14 different radiology findings and brain computed tomography (CT) scans with six types of intracranial hemorrhages. We measure the synthetic image quality by the performance difference of predictive models trained on either the synthetic or the real dataset. We find that synthetic data performance disproportionately benefits from a reduced number of unique label combinations. Our open-source benchmark also indicates that at low number of samples per class, label overfitting effects start to dominate GAN training. We additionally conducted a reader study in which trained radiologists do not perform better than random on discriminating between synthetic and real medical images for intermediate levels of resolutions. In accordance with our benchmark results, the classification accuracy of radiologists increases at higher spatial resolution levels. Our study offers valuable guidelines and outlines practical conditions under which insights derived from synthetic medical images are similar to those that would have been derived from real imaging data. Our results indicate that synthetic data sharing may be an attractive and privacy-preserving alternative to sharing real patient-level data in the right settings. | ['Patrick Schwab', 'Stefan Bauer', 'Benedikt Dietz', 'Tobias Hepp', 'Sergios Gatidis', 'Jürgen Hetzel', 'August DuMont Schütte'] | 2020-11-29 | null | null | null | null | ['medical-image-generation'] | ['medical'] | [ 6.34225249e-01 6.74325883e-01 -7.33887553e-02 -4.96126711e-01
-1.33207786e+00 -6.27221167e-01 3.79448533e-01 2.16896623e-01
-6.71637893e-01 1.01515317e+00 3.85083139e-01 -3.24795723e-01
1.57184586e-01 -7.75761783e-01 -6.99516118e-01 -7.90032804e-01
1.68071225e-01 4.35997814e-01 -3.30844730e-01 2.81595826e-01
-3.37320894e-01 3.33578497e-01 -1.00790501e+00 4.67729121e-01
6.33004665e-01 7.72777736e-01 -2.98330009e-01 5.44882298e-01
2.99164474e-01 7.62853384e-01 -8.94078553e-01 -7.63980091e-01
6.66395009e-01 -7.46228278e-01 -4.99937356e-01 1.85147002e-01
4.09859836e-01 -4.76935923e-01 -2.80790150e-01 9.71659482e-01
8.49652469e-01 -3.78333092e-01 5.88490129e-01 -1.18577576e+00
-7.70435154e-01 5.56387901e-01 -4.66695070e-01 -9.77955610e-02
1.37042165e-01 5.96168041e-01 4.81882811e-01 -2.07288206e-01
8.33507180e-01 6.46580338e-01 8.26112568e-01 7.43503869e-01
-1.47586513e+00 -7.59844363e-01 -4.94026929e-01 -7.01433659e-01
-1.13678157e+00 -3.44854385e-01 3.17696154e-01 -5.11356413e-01
9.50504839e-02 6.52786613e-01 6.87818825e-01 1.42577600e+00
6.17539048e-01 1.86369315e-01 1.57613289e+00 -1.64427608e-01
1.51046410e-01 5.63173831e-01 -2.72920161e-01 5.47333241e-01
6.26179934e-01 1.56363264e-01 -2.87931204e-01 -6.38485551e-01
8.88857424e-01 4.60513011e-02 -5.49933851e-01 -3.82280886e-01
-1.51591051e+00 8.72643411e-01 3.07423800e-01 1.21167511e-01
-4.18791801e-01 1.35373488e-01 3.59620363e-01 2.31710702e-01
3.25024307e-01 6.13923192e-01 -3.48839127e-02 1.58358350e-01
-7.88720667e-01 1.66669726e-01 5.52111030e-01 8.02639425e-01
4.22451831e-02 -4.57481220e-02 -4.53722388e-01 6.45632625e-01
-1.48704812e-01 4.88988042e-01 8.07760656e-01 -1.01064122e+00
2.43804783e-01 1.30692676e-01 1.25581145e-01 -8.57957065e-01
-1.47044986e-01 -5.99025369e-01 -9.62842286e-01 3.21969867e-01
6.72258496e-01 -3.75681400e-01 -9.53072846e-01 1.94936776e+00
4.44578230e-02 -4.74206984e-01 2.29605019e-01 6.74395025e-01
8.63108099e-01 4.74906061e-03 3.21106941e-01 -7.59987012e-02
1.56971335e+00 -4.43705440e-01 -6.17528975e-01 8.21105838e-02
5.77019751e-01 -5.99373996e-01 1.01442170e+00 1.08421020e-01
-1.31495619e+00 -2.29570389e-01 -9.07274842e-01 2.54406720e-01
-4.94425222e-02 -2.09593609e-01 4.54307675e-01 1.24115992e+00
-1.05640280e+00 3.69423598e-01 -6.96777225e-01 -2.14597896e-01
9.69498396e-01 1.87717840e-01 -6.08898759e-01 -4.16099504e-02
-1.09113073e+00 7.05571949e-01 9.71414521e-02 -4.96453524e-01
-7.89791107e-01 -1.15978348e+00 -5.94550550e-01 -2.39921406e-01
-6.58165738e-02 -1.10258138e+00 1.01928377e+00 -1.18090785e+00
-8.46674442e-01 1.34701240e+00 2.92256743e-01 -6.97874725e-01
1.22074080e+00 3.55722159e-01 -4.40274805e-01 1.67729989e-01
2.23343238e-01 9.44796741e-01 7.42893517e-01 -1.49017417e+00
-2.76271135e-01 -3.30656409e-01 -2.22465828e-01 3.32483910e-02
5.49498294e-03 -1.90754861e-01 8.69959965e-02 -1.09197307e+00
-3.00374627e-01 -1.06131315e+00 -3.55467945e-01 3.05361807e-01
-4.80341405e-01 7.04287291e-01 4.12265569e-01 -7.12212384e-01
6.28361583e-01 -2.15252423e+00 -5.19084096e-01 2.80401260e-01
5.44549704e-01 -2.30877614e-03 1.51942357e-01 -2.52994653e-02
-3.42998892e-01 5.95170677e-01 -5.54371476e-01 -1.21128544e-01
-3.01385552e-01 3.77052580e-03 -5.36549352e-02 5.72663426e-01
5.23031391e-02 1.12113786e+00 -7.39551604e-01 -4.68867302e-01
-1.37956426e-01 4.97483432e-01 -4.57810968e-01 1.68604210e-01
2.25671425e-01 1.00921381e+00 -2.93121845e-01 4.14123684e-01
7.20265627e-01 -4.48607236e-01 4.89922501e-02 -1.16037410e-02
5.16663969e-01 -2.33359516e-01 -5.13167739e-01 1.46291435e+00
-2.30607793e-01 4.61607605e-01 -7.05804899e-02 -3.74256253e-01
6.53664649e-01 5.43221295e-01 6.92111075e-01 -6.38043463e-01
6.66729510e-02 -5.56594506e-03 3.02233189e-01 -2.26568401e-01
2.49968171e-01 -6.41672730e-01 -1.21440269e-01 8.06954086e-01
-3.84341300e-01 -2.92640328e-01 -5.54194152e-01 7.02242255e-02
1.38719285e+00 -4.11654830e-01 2.06387192e-01 -1.44963756e-01
-1.04248315e-01 8.11472014e-02 4.66770291e-01 9.48201120e-01
-3.48077506e-01 1.14605546e+00 5.88483334e-01 -2.70132959e-01
-1.39947820e+00 -1.37410986e+00 -5.73861301e-01 1.53413445e-01
-4.10997301e-01 6.63825795e-02 -8.83228660e-01 -7.75693119e-01
-7.64464587e-02 8.45524728e-01 -9.44893837e-01 -2.83488691e-01
-1.84549019e-01 -9.81098413e-01 9.88304079e-01 3.25863957e-01
4.42180753e-01 -8.99566233e-01 -1.11493778e+00 1.72997545e-02
-8.22735578e-02 -9.37469363e-01 -6.23766184e-01 -5.93762398e-02
-7.67525494e-01 -1.10085857e+00 -1.19477868e+00 -3.17403764e-01
1.02841425e+00 -1.57701299e-01 1.21714771e+00 -1.25214066e-02
-6.90421700e-01 6.74778879e-01 -3.57275933e-01 -7.65266895e-01
-9.63599265e-01 -2.15810344e-01 -3.63273412e-01 -7.19224066e-02
5.16622514e-02 -4.42751110e-01 -9.09055889e-01 3.62808108e-01
-1.17057776e+00 2.63526235e-02 6.61048830e-01 1.06596649e+00
7.18554437e-01 -2.27223244e-02 6.58696294e-01 -1.61503506e+00
7.68932402e-01 -7.22617030e-01 -1.04371689e-01 2.12124124e-01
-7.31346548e-01 -3.09578460e-02 5.41708469e-01 -2.12494731e-01
-1.15113461e+00 -1.23082027e-01 1.73517287e-01 -3.52475226e-01
-2.23810896e-01 1.64254948e-01 1.42931014e-01 -5.59825152e-02
9.92773473e-01 -4.98335846e-02 4.71762568e-01 -6.85655624e-02
3.43831629e-02 6.21691704e-01 5.09467542e-01 -5.00376821e-01
4.41617817e-01 7.37413883e-01 -6.96316957e-02 -4.46269602e-01
-4.78074819e-01 1.97711512e-01 -1.71628281e-01 7.17189014e-02
9.94390607e-01 -9.08636630e-01 -3.61851662e-01 3.86292636e-01
-5.99187315e-01 -2.82146841e-01 -1.04782343e+00 4.84640360e-01
-6.10636592e-01 3.70017476e-02 -5.08849621e-01 -4.82188761e-01
-4.33987796e-01 -1.33689213e+00 9.81092572e-01 -3.64399105e-02
-5.01184404e-01 -9.55799043e-01 -1.16863139e-01 7.01249421e-01
5.03281832e-01 1.07246625e+00 1.03371966e+00 -7.39270806e-01
-4.76132125e-01 -3.34785938e-01 -1.30179869e-02 3.85002077e-01
4.86696631e-01 -3.29930305e-01 -1.07623374e+00 -5.36225319e-01
2.64574349e-01 -3.96934539e-01 5.03998756e-01 4.53765869e-01
1.50029492e+00 -3.98720920e-01 -1.27904877e-01 6.28514230e-01
1.32722580e+00 4.53648716e-01 8.60868335e-01 -1.19574759e-02
4.98674482e-01 7.89152980e-01 2.34949917e-01 4.35196549e-01
1.04233131e-01 2.76733249e-01 2.00945333e-01 -5.44440150e-01
-1.86914101e-01 -3.62193942e-01 -3.10377534e-02 2.92980552e-01
2.07676604e-01 -3.55157465e-01 -8.61216784e-01 5.98198056e-01
-1.12812805e+00 -7.76806891e-01 1.82372089e-02 2.57161522e+00
9.70502079e-01 -4.35339883e-02 1.38771638e-01 -2.51343220e-01
7.97897220e-01 -5.90395778e-02 -8.07385862e-01 -3.02918881e-01
-2.66254812e-01 3.46689016e-01 1.08158076e+00 -5.92039861e-02
-6.85348153e-01 3.10290940e-02 7.15711594e+00 4.84033555e-01
-1.05387378e+00 4.24808204e-01 1.44396734e+00 -4.26233053e-01
-7.39668489e-01 -3.17929685e-01 4.20131674e-03 6.61775172e-01
1.19656050e+00 -3.62988532e-01 -2.88698934e-02 5.36842823e-01
-1.96360108e-02 -1.18812293e-01 -1.24457705e+00 8.53751957e-01
-1.68503389e-01 -1.49755788e+00 1.06519781e-01 5.14496982e-01
8.85093451e-01 -1.27151564e-01 7.02925026e-01 -4.10272628e-01
6.62142217e-01 -1.61886799e+00 6.23610914e-01 7.30822682e-01
1.50887585e+00 -7.40665436e-01 9.66244042e-01 5.49667403e-02
-3.62330437e-01 3.52697343e-01 -4.89197522e-02 6.93289399e-01
-6.98846653e-02 5.02592981e-01 -1.13527107e+00 4.45681721e-01
7.55976856e-01 1.51396811e-01 -6.94994509e-01 6.54004276e-01
2.15070099e-01 5.69327772e-01 -6.55840784e-02 4.30013210e-01
-5.52716851e-02 6.31429479e-02 3.77515763e-01 8.23923528e-01
3.43332410e-01 1.84445903e-01 -4.44441348e-01 8.96067441e-01
-2.77037412e-01 8.32981467e-02 -9.48567867e-01 -8.99093822e-02
4.67910528e-01 9.99660313e-01 -7.13184834e-01 -8.68153423e-02
-3.84848654e-01 7.40174294e-01 -1.37317613e-01 2.67809723e-02
-7.09948778e-01 3.98456305e-02 4.18641984e-01 5.17809570e-01
-7.99671113e-02 4.80402827e-01 -5.27885675e-01 -9.29760993e-01
-9.20116082e-02 -1.21315503e+00 5.33748865e-01 -9.81605232e-01
-1.43242824e+00 6.11995816e-01 -4.76636961e-02 -1.31926870e+00
-3.30896407e-01 -1.28922924e-01 -2.55971193e-01 1.09252322e+00
-8.50051761e-01 -1.04529321e+00 -3.04892600e-01 5.21656752e-01
-4.21474762e-02 -2.59659410e-01 1.18147957e+00 -6.59912974e-02
-9.83810797e-02 1.16357207e+00 3.16737056e-01 2.96036690e-01
8.29912186e-01 -1.06157410e+00 2.87420988e-01 5.05423903e-01
-1.29128352e-01 6.41507447e-01 5.49028695e-01 -7.06827462e-01
-8.73925567e-01 -1.24885607e+00 8.51434693e-02 -7.28395998e-01
1.55796334e-01 -4.17475356e-03 -7.49414384e-01 9.53139246e-01
2.90603548e-01 1.12532131e-01 1.25884557e+00 -6.27658129e-01
-1.23482592e-01 1.21952176e-01 -2.07781291e+00 5.51741838e-01
8.05916488e-01 -3.99850011e-01 -2.53310263e-01 2.93356180e-01
7.15533495e-01 -4.94348496e-01 -1.45745504e+00 2.08888441e-01
5.37616789e-01 -1.05925536e+00 8.31562638e-01 -6.24746323e-01
6.14200592e-01 7.96991885e-02 -1.83002725e-01 -1.43105912e+00
-5.30155525e-02 -4.25965041e-01 7.65041649e-01 1.11367512e+00
5.45440495e-01 -1.00150955e+00 1.12158704e+00 1.32592881e+00
2.75668532e-01 -7.15123594e-01 -9.30974126e-01 -5.51162004e-01
3.56200635e-01 -1.01334900e-01 8.18510413e-01 1.11484182e+00
-4.49836552e-01 -3.18082035e-01 -1.80856064e-01 -1.69745669e-01
9.75203276e-01 -2.04154462e-01 6.57160580e-01 -8.75279844e-01
-3.74405891e-01 -1.10010430e-01 -5.10863781e-01 5.68527021e-02
-6.27901629e-02 -1.02431977e+00 -4.24407274e-01 -1.07092965e+00
4.30725783e-01 -9.41204309e-01 -2.36689806e-01 4.01099235e-01
-5.43635562e-02 4.62182045e-01 2.49904990e-01 2.92346895e-01
1.89916104e-01 7.55297318e-02 1.63501298e+00 -8.40129703e-02
2.60970533e-01 -1.09184511e-01 -1.20004690e+00 4.06367391e-01
7.57294774e-01 -6.99784517e-01 -6.13195181e-01 -2.12257460e-01
-9.57365781e-02 4.04422522e-01 6.58030868e-01 -9.27995503e-01
-5.51855192e-02 4.89780866e-02 6.86307192e-01 -8.73596147e-02
3.62527147e-02 -1.01535225e+00 1.10053217e+00 6.88438416e-01
-6.39319479e-01 9.47869122e-02 1.30193964e-01 5.76000094e-01
-4.25194092e-02 -1.69026494e-01 1.02376997e+00 -5.85826933e-01
3.42057794e-02 2.61815012e-01 -1.84020087e-01 4.10814583e-01
1.29584002e+00 -4.33215052e-01 -3.81091356e-01 -6.35161459e-01
-8.12937379e-01 -1.75285921e-01 1.00332105e+00 1.10238669e-02
4.72160846e-01 -1.25851965e+00 -9.61991668e-01 1.53106660e-01
1.78329304e-01 1.56589836e-01 4.20277447e-01 7.09844887e-01
-6.20613456e-01 1.94375619e-01 -4.65215504e-01 -4.83164579e-01
-1.02756155e+00 4.80217308e-01 5.39577663e-01 -3.84579003e-01
-5.59093237e-01 7.60719478e-01 4.96437132e-01 -2.73296177e-01
-1.68659061e-01 -8.05095136e-02 4.18166906e-01 -4.51759957e-02
3.38251740e-01 1.84621140e-01 2.31615543e-01 -6.47031963e-01
-1.15872413e-01 -1.25079341e-02 -2.91978240e-01 -2.13193148e-01
1.28248274e+00 8.00362006e-02 1.93146393e-01 2.99110204e-01
1.29863393e+00 2.70167202e-01 -1.25190651e+00 2.90220212e-02
-6.17633879e-01 -6.91265166e-01 -1.69462815e-01 -1.06645441e+00
-1.47458875e+00 3.56878489e-01 9.02446806e-01 1.66315511e-01
1.10832691e+00 4.74924315e-03 6.07530415e-01 -3.78325075e-01
6.00113451e-01 -5.99693239e-01 -2.09181905e-01 -6.29341483e-01
8.37054849e-01 -1.23446035e+00 1.79715186e-01 -2.40911171e-01
-9.93126214e-01 5.44776082e-01 2.57306188e-01 1.00589529e-01
4.09025818e-01 4.49175596e-01 3.38231295e-01 -1.89760789e-01
-4.91869539e-01 5.31073749e-01 -1.03574932e-01 9.49355066e-01
3.95025045e-01 3.40501726e-01 -2.69036531e-01 5.56631207e-01
-6.64050400e-01 -9.96123347e-03 8.15530181e-01 1.04050612e+00
3.77495140e-01 -1.21746612e+00 -5.81366956e-01 1.08852148e+00
-9.30229604e-01 -8.46373588e-02 -2.97500551e-01 8.74385178e-01
1.24532878e-01 6.36400640e-01 1.49144202e-01 -1.07795902e-01
2.79300421e-01 2.10462753e-02 3.37948442e-01 -5.61506212e-01
-9.38310623e-01 -3.86846781e-01 8.82228911e-02 -4.69030410e-01
-4.15991157e-01 -9.90314543e-01 -9.43261087e-01 -3.32515419e-01
7.28912950e-02 6.37464644e-03 4.67124760e-01 2.91672409e-01
5.20415306e-01 7.51935124e-01 5.05398512e-01 -3.65838930e-02
-4.81933892e-01 -4.56703126e-01 -6.65223479e-01 7.91819930e-01
2.95607805e-01 -2.08106190e-01 -2.93963730e-01 1.87099174e-01] | [14.15540885925293, -1.8616517782211304] |
15e96c93-1ba3-46b6-8d93-18355611ea73 | sub-optimal-multi-phase-path-planning-a | 1601.05744 | null | http://arxiv.org/abs/1601.05744v1 | http://arxiv.org/pdf/1601.05744v1.pdf | Sub-Optimal Multi-Phase Path Planning: A Method for Solving Rubik's Revenge | Rubik's Revenge, a 4x4x4 variant of the Rubik's puzzles, remains to date as
an unsolved puzzle. That is to say, we do not have a method or successful
categorization to optimally solve every one of its approximately $7.401 \times
10^{45}$ possible configurations. Rubik's Cube, Rubik's Revenge's predecessor
(3x3x3), with its approximately $4.33 \times 10^{19}$ possible configurations,
has only recently been completely solved by Rokicki et. al, further finding
that any configuration requires no more than 20 moves. With the sheer dimension
of Rubik's Revenge and its total configuration space, a brute-force method of
finding all optimal solutions would be in vain. Similar to the methods used by
Rokicki et. al on Rubik's Cube, in this paper we develop a method for solving
arbitrary configurations of Rubik's Revenge in phases, using a combination of a
powerful algorithm known as IDA* and a useful definition of distance in the
cube space. While time-series results were not successfully gathered, it will
be shown that this method far outweighs current human-solving methods and can
be used to determine loose upper bounds for the cube space. Discussion will
suggest that this method can also be applied to other puzzles with the proper
transformations. | ['Jared Weed'] | 2016-01-20 | null | null | null | null | ['rubik-s-cube'] | ['graphs'] | [ 4.12999094e-02 9.93548036e-02 1.07503332e-01 3.42576474e-01
-7.21579909e-01 -1.17270029e+00 -2.28117526e-01 -3.79851967e-01
-2.09502310e-01 1.27097273e+00 -4.12448168e-01 -8.41724217e-01
-1.19285154e+00 -7.76222050e-01 -5.38209915e-01 -9.70954716e-01
-4.73033935e-01 8.71173561e-01 7.16107339e-02 -4.99394357e-01
7.63329387e-01 5.58242798e-01 -1.21258163e+00 1.78396124e-02
6.31598651e-01 7.47557461e-01 3.60418171e-01 8.25676441e-01
2.22913921e-02 4.40276384e-01 -9.25622702e-01 -4.40705568e-02
5.71687818e-01 -4.51781273e-01 -1.17075145e+00 -7.06477761e-02
-2.70093858e-01 2.65439954e-02 -2.45160088e-01 5.89083195e-01
1.47266269e-01 1.17799796e-01 5.37934184e-01 -1.38318563e+00
-4.52142179e-01 2.91712254e-01 -9.56619978e-01 2.92457819e-01
7.82837570e-01 1.64037079e-01 9.84267890e-01 -3.91051531e-01
7.35942006e-01 6.84842646e-01 6.13126159e-01 3.71968567e-01
-1.02343583e+00 -7.21250594e-01 -1.23422801e-01 6.26731634e-01
-1.66042328e+00 -1.50733721e-02 6.73385739e-01 9.34043527e-02
1.24029398e+00 8.36014628e-01 8.41215134e-01 4.83762532e-01
2.22245783e-01 3.64663243e-01 1.21234596e+00 -6.14374399e-01
3.55739951e-01 -4.24430907e-01 -1.34112522e-01 3.19138527e-01
6.82450891e-01 1.52156213e-02 -1.06871270e-01 -3.25648844e-01
8.04933846e-01 -5.66727102e-01 -3.72639745e-01 -3.72949570e-01
-9.27343488e-01 9.51871932e-01 7.31249899e-02 4.13927913e-01
-1.32940814e-01 2.81696349e-01 -4.77967411e-02 3.42067599e-01
-1.52808368e-01 9.21093583e-01 -6.35653257e-01 -4.98312145e-01
-6.93869054e-01 7.16954350e-01 1.03157651e+00 9.97447252e-01
3.41521472e-01 -2.60662645e-01 5.58222651e-01 2.90701747e-01
-1.62170082e-01 2.66197860e-01 -1.98250443e-01 -1.42381060e+00
8.42640638e-01 5.33983469e-01 5.39636850e-01 -9.42386866e-01
-7.55684972e-01 -7.73569494e-02 -5.55356860e-01 5.75950742e-01
9.49032187e-01 -2.52509356e-01 -4.05621022e-01 1.46362185e+00
1.85486555e-01 -6.42428100e-02 2.40719523e-02 9.03270185e-01
1.70360595e-01 7.72907376e-01 -7.89873183e-01 -5.62322199e-01
1.32384753e+00 -6.01675093e-01 -3.40335757e-01 -5.88447638e-02
5.67233264e-01 -9.41979408e-01 8.05624366e-01 7.99884617e-01
-1.51330388e+00 1.52496919e-01 -1.25839114e+00 3.79553318e-01
-4.63725477e-02 -4.48432565e-01 1.05394387e+00 8.48607421e-01
-1.07853627e+00 4.91511762e-01 -4.96716112e-01 -1.63003772e-01
9.96084437e-02 8.65737081e-01 -2.88554549e-01 -3.24828804e-01
-7.07435548e-01 9.81123149e-01 3.19729120e-01 1.14952214e-01
-2.39864066e-01 -4.24079984e-01 -4.33813095e-01 1.78628713e-01
9.26624835e-01 -5.49217045e-01 9.31385159e-01 -2.76436776e-01
-1.19829309e+00 5.32310665e-01 -1.15816683e-01 2.36257929e-02
1.57940120e-01 7.49194682e-01 -3.17188017e-02 1.79501310e-01
8.14071074e-02 1.29413500e-01 8.81987810e-02 -1.36659873e+00
-5.71027458e-01 -4.35274541e-01 3.80593538e-01 1.62770376e-01
3.05101454e-01 -2.42411774e-02 -1.84527695e-01 -3.04681957e-01
7.59718060e-01 -1.14476919e+00 -4.12261546e-01 -5.12159348e-01
-2.92564332e-01 -3.23778659e-01 2.62126058e-01 -4.58696991e-01
1.21894145e+00 -1.65004468e+00 4.44160223e-01 6.78163409e-01
1.82831541e-01 -5.39509021e-02 -2.65183806e-01 8.08019578e-01
-2.07469448e-01 1.94661707e-01 -2.05351681e-01 3.46047789e-01
7.08363578e-02 3.74323547e-01 1.08947478e-01 5.26149988e-01
-2.90080458e-01 8.00635517e-01 -7.13089168e-01 -1.06773198e-01
-2.44011387e-01 -6.74481913e-02 -6.63645267e-01 -2.96289772e-01
4.97663319e-02 5.42953331e-03 -2.05049828e-01 7.78784573e-01
9.16630328e-01 -1.43730596e-01 4.93821919e-01 3.37790877e-01
-2.97823608e-01 -2.58962531e-02 -1.43472111e+00 1.43506598e+00
1.04619861e-02 5.56029081e-01 4.69456345e-01 -1.10058022e+00
8.45055938e-01 2.56904513e-01 7.96802878e-01 -8.48816931e-01
1.94806814e-01 3.98287922e-01 3.78093243e-01 -5.79786003e-01
5.99819958e-01 -2.12543815e-01 -3.70436251e-01 9.08424139e-01
-6.52037740e-01 -3.62170130e-01 6.50034249e-01 2.05665797e-01
1.70504236e+00 -1.44479156e-01 4.40555140e-02 -4.66029376e-01
1.63958892e-01 4.46656495e-01 6.30519867e-01 6.67289913e-01
5.05136214e-02 7.24667251e-01 8.14476252e-01 -3.94084334e-01
-1.15552831e+00 -1.05805576e+00 4.42341790e-02 3.95563632e-01
4.22513098e-01 -4.72321361e-01 -6.33727610e-01 -6.82345703e-02
1.85448129e-03 7.73785055e-01 -7.00577676e-01 2.34091341e-01
-7.95842469e-01 -8.59482527e-01 2.77310610e-01 2.19237670e-01
5.31404242e-02 -1.05618012e+00 -8.52913439e-01 4.05976683e-01
-3.01988959e-01 -6.18807673e-01 -2.68166453e-01 3.87443632e-01
-7.39248216e-01 -1.70200419e+00 -3.90109122e-01 -6.19894207e-01
6.39243484e-01 7.27839291e-01 9.61155951e-01 1.90519065e-01
-7.54523039e-01 4.32259560e-01 -5.53870916e-01 -1.55256659e-01
6.77229241e-02 -1.16646625e-01 4.24737521e-02 -9.96779501e-01
1.65216580e-01 -9.35164511e-01 -6.97064340e-01 6.72434807e-01
-7.01470375e-01 -2.09237322e-01 3.32051516e-01 7.04191923e-01
5.06957829e-01 7.57928073e-01 5.19959331e-01 -4.23673511e-01
8.06336761e-01 -6.23674512e-01 -7.84185708e-01 2.26529509e-01
-6.39553547e-01 -1.68296620e-02 5.51788330e-01 -4.43726748e-01
-6.37757480e-01 -8.15787390e-02 2.57533640e-01 -1.72293499e-01
3.48159701e-01 4.34582889e-01 -3.79940905e-02 -4.26812530e-01
5.05841196e-01 3.23256128e-03 -2.51377791e-01 -4.47508752e-01
1.23993918e-01 3.85481417e-01 7.28934467e-01 -7.62881458e-01
8.42539489e-01 2.62360841e-01 2.92247087e-01 -3.73374373e-01
3.77985038e-04 -2.35140339e-01 -2.39361539e-01 4.98697907e-02
5.07777750e-01 -3.28419134e-02 -1.86660993e+00 1.04312062e-01
-1.21969557e+00 -2.51219541e-01 -3.25874150e-01 1.49573013e-01
-8.98370087e-01 5.39347887e-01 -4.64799047e-01 -1.07162774e+00
4.74171825e-02 -8.79872620e-01 4.92100805e-01 2.77815670e-01
-3.91510844e-01 -5.61733484e-01 5.36222830e-02 8.71427000e-01
3.00978303e-01 3.43190789e-01 1.08669555e+00 -1.29116014e-01
-9.09548104e-01 -1.90040886e-01 -2.98448265e-01 -3.72115135e-01
1.55916559e-02 -1.28622711e-01 -1.55200986e-02 -5.77074945e-01
3.03147763e-01 5.57766967e-02 1.64450392e-01 3.36148798e-01
9.86090064e-01 -4.25295711e-01 -3.35515410e-01 5.43889940e-01
1.50004935e+00 1.12718499e+00 1.03638041e+00 8.29918146e-01
6.54940158e-02 5.40392220e-01 7.02833772e-01 6.71462536e-01
5.30766368e-01 7.49325275e-01 6.98563159e-01 3.93169910e-01
3.20576370e-01 2.51961499e-01 -6.60295039e-02 6.02431238e-01
-5.54140091e-01 -4.33781236e-01 -8.00725162e-01 4.75430310e-01
-1.66046703e+00 -1.24706578e+00 -3.46115917e-01 1.92838895e+00
5.80445051e-01 1.62978083e-01 4.91803624e-02 7.46527910e-01
4.80920553e-01 -2.06386775e-01 -5.97142994e-01 -8.97017121e-01
-1.38770938e-01 5.49388647e-01 7.69392312e-01 5.58372915e-01
-3.17070633e-01 4.60586131e-01 6.79425812e+00 8.78216386e-01
-3.18340987e-01 -2.47658178e-01 3.53435397e-01 -5.28377533e-01
-3.60968232e-01 4.04269636e-01 -3.16674888e-01 5.25129378e-01
4.52118427e-01 -5.04848659e-01 1.14429402e+00 4.77928102e-01
2.05734178e-01 -8.73663902e-01 -9.16457415e-01 1.15402043e+00
1.11596517e-01 -1.38333297e+00 -7.18024135e-01 4.60092753e-01
5.52910805e-01 -8.70077193e-01 -6.99804304e-03 3.90604176e-02
2.09190413e-01 -1.36842573e+00 3.72733653e-01 3.74189354e-02
6.20383561e-01 -1.00283444e+00 4.80166316e-01 5.23596048e-01
-1.10139751e+00 -3.75720263e-01 -4.67865795e-01 -5.37828088e-01
2.32416689e-01 7.67347217e-02 -8.90486062e-01 6.57623887e-01
6.70311391e-01 -3.87705058e-01 8.43617618e-02 1.10782576e+00
2.16789603e-01 1.74844742e-01 -6.54714823e-01 -1.85452789e-01
3.87415707e-01 -5.63095689e-01 5.98922253e-01 4.79154885e-01
7.56387174e-01 1.10681379e+00 -4.13927972e-01 7.60998368e-01
5.47355652e-01 -3.72325107e-02 -3.38946670e-01 1.25556588e-01
6.21584475e-01 8.15673411e-01 -1.11462605e+00 2.65616000e-01
4.76638861e-02 8.12766492e-01 -1.11084603e-01 3.06877702e-01
-8.22924554e-01 -7.47918427e-01 6.21140063e-01 2.68511176e-01
3.60088855e-01 -5.83272576e-01 -6.84828162e-01 -7.19324350e-01
2.81376541e-01 -9.29988205e-01 4.55314875e-01 -1.13054752e+00
-9.48930562e-01 5.42660654e-01 2.30787799e-01 -9.60337460e-01
-5.29914320e-01 -7.54716575e-01 -7.39680588e-01 8.75179172e-01
-8.38235021e-01 -6.19504571e-01 2.16054227e-02 7.03872085e-01
4.38945144e-01 -1.90419048e-01 8.30258608e-01 -3.70240323e-02
-3.49562526e-01 6.02476716e-01 3.09165865e-01 -5.96737921e-01
-8.83614123e-02 -9.42853093e-01 2.31614575e-01 5.04965425e-01
-7.64513239e-02 3.55421513e-01 1.07348490e+00 -3.94277722e-01
-2.15739632e+00 -2.18904037e-02 6.86449051e-01 -5.40969670e-01
7.37413108e-01 -4.97870743e-02 -6.63363338e-01 6.11333251e-01
9.69286785e-02 -5.27662575e-01 3.72601599e-01 6.92805499e-02
-4.50685434e-02 -1.26662161e-02 -1.39286911e+00 6.38858914e-01
1.39105248e+00 5.45408987e-02 -4.09887731e-01 4.37515497e-01
2.70865649e-01 -6.19547784e-01 -6.87376082e-01 2.07369879e-01
5.84640384e-01 -1.20177877e+00 1.01476574e+00 -3.61406326e-01
-1.95200443e-02 -2.40636051e-01 -2.54123002e-01 -1.07086682e+00
-6.20443642e-01 -8.93202722e-01 4.04567838e-01 6.78796291e-01
3.10372770e-01 -7.63794422e-01 1.29404330e+00 7.49650240e-01
-5.56452572e-03 -1.07784498e+00 -1.53811061e+00 -1.05194879e+00
3.04960489e-01 -5.07590175e-01 6.55330241e-01 8.16227198e-01
6.43032670e-01 7.30704470e-03 -3.34845603e-01 -2.95900889e-02
7.15778232e-01 3.72959971e-01 6.34629309e-01 -9.82536972e-01
-5.65510929e-01 -4.69037384e-01 -3.18512589e-01 -7.25666106e-01
-1.89992845e-01 -6.46869004e-01 -2.82141715e-01 -1.62909806e+00
2.23820359e-01 -1.12891984e+00 1.67118117e-01 4.53619301e-01
4.54488099e-01 3.31957996e-01 4.77075905e-01 6.16747513e-02
-4.18855786e-01 -1.89668030e-01 1.44082105e+00 -3.44779864e-02
-3.23065400e-01 -1.43971918e-02 -9.87002552e-01 2.94773072e-01
1.01375294e+00 -4.93156880e-01 -5.85721791e-01 -2.64504433e-01
7.66577661e-01 8.59320223e-01 -1.07748091e-01 -8.12391281e-01
4.73310471e-01 -5.84361613e-01 7.48739094e-02 -6.67835891e-01
6.52063549e-01 -8.03819239e-01 9.59741473e-01 4.88639176e-01
4.39549237e-01 4.53204036e-01 3.56171459e-01 2.03276455e-01
3.58243436e-01 -6.59396052e-01 2.65703112e-01 -1.74534798e-01
-3.14225823e-01 -1.90958828e-01 -5.10259092e-01 -1.58668593e-01
1.50168955e+00 -8.24621260e-01 -6.20688081e-01 -5.15748084e-01
-8.25655758e-01 3.05083126e-01 5.06614745e-01 -2.56101578e-03
7.29560912e-01 -1.09390104e+00 -6.46539032e-01 4.89474237e-02
-5.05598843e-01 2.46375473e-03 5.02861559e-01 7.74191320e-01
-9.14467633e-01 5.61391890e-01 -4.23527211e-01 -1.49614215e-01
-1.28353822e+00 9.07537222e-01 -4.28236537e-02 -4.31936502e-01
-5.62492609e-01 7.07032442e-01 -2.10762039e-01 8.30696598e-02
-1.33684427e-01 -1.28343683e-02 2.75711209e-01 -1.21062994e-01
4.04382616e-01 7.87649930e-01 -1.20321997e-01 -3.15727323e-01
-5.48705220e-01 8.15204263e-01 2.14304686e-01 -2.17948258e-01
1.63713837e+00 -1.10589601e-01 -5.03981531e-01 -2.30232641e-01
9.88946140e-01 1.77931279e-01 -7.49575973e-01 5.49429297e-01
-5.69990836e-03 -7.06276596e-01 -6.36035323e-01 -9.45802271e-01
-1.07477367e+00 3.43001813e-01 8.59912559e-02 4.71406549e-01
1.37372315e+00 8.46560448e-02 8.09209645e-01 5.59422374e-01
9.71309960e-01 -9.04443800e-01 -4.70163189e-02 4.33278561e-01
1.09236622e+00 -6.58355594e-01 2.44454056e-01 -5.84591150e-01
-2.92335212e-01 1.23916197e+00 3.74622434e-01 -2.58929729e-01
1.15742840e-01 7.30480671e-01 -2.96692997e-01 -2.58865237e-01
-9.01243210e-01 -1.20901717e-02 -3.82704139e-01 5.97773910e-01
-2.27249905e-01 1.69677794e-01 -8.49325836e-01 9.17376399e-01
-6.19874418e-01 -3.37068848e-02 1.03112972e+00 1.12999153e+00
-5.66947162e-01 -1.30090356e+00 -1.13448429e+00 3.68665636e-01
-4.14246172e-02 2.04345614e-01 -1.77639827e-01 1.10333085e+00
1.72074303e-01 1.29535460e+00 -4.61288169e-03 -3.98843348e-01
4.66859519e-01 -2.12884367e-01 1.10447586e+00 -6.12414926e-02
-3.20984781e-01 -8.37716684e-02 1.45985603e-01 -5.69821596e-01
1.29561216e-01 -6.32657409e-01 -1.36852694e+00 -1.16906428e+00
-2.30722487e-01 5.70334196e-01 5.67135870e-01 7.09016442e-01
5.55891218e-03 6.35211170e-02 6.60043716e-01 -9.18850541e-01
-3.33724290e-01 -3.17496806e-01 -9.98996556e-01 -7.02899769e-02
-3.83544147e-01 -7.14482188e-01 -5.45189023e-01 -5.78229010e-01] | [4.840744495391846, 1.8322033882141113] |
0baf844b-65e3-4556-bcec-fcfa1a67b34d | mcmlsd-a-probabilistic-algorithm-and | 2001.01788 | null | https://arxiv.org/abs/2001.01788v1 | https://arxiv.org/pdf/2001.01788v1.pdf | MCMLSD: A Probabilistic Algorithm and Evaluation Framework for Line Segment Detection | Traditional approaches to line segment detection typically involve perceptual grouping in the image domain and/or global accumulation in the Hough domain. Here we propose a probabilistic algorithm that merges the advantages of both approaches. In a first stage lines are detected using a global probabilistic Hough approach. In the second stage each detected line is analyzed in the image domain to localize the line segments that generated the peak in the Hough map. By limiting search to a line, the distribution of segments over the sequence of points on the line can be modeled as a Markov chain, and a probabilistically optimal labelling can be computed exactly using a standard dynamic programming algorithm, in linear time. The Markov assumption also leads to an intuitive ranking method that uses the local marginal posterior probabilities to estimate the expected number of correctly labelled points on a segment. To assess the resulting Markov Chain Marginal Line Segment Detector (MCMLSD) we develop and apply a novel quantitative evaluation methodology that controls for under- and over-segmentation. Evaluation on the YorkUrbanDB and Wireframe datasets shows that the proposed MCMLSD method outperforms prior traditional approaches, as well as more recent deep learning methods. | ['Emilio J. Almazàn', 'Yiming Qian', 'Ron Tal', 'James H. Elder'] | 2020-01-06 | null | null | null | null | ['line-segment-detection'] | ['computer-vision'] | [ 1.24175310e-01 2.26891667e-01 -1.85946032e-01 -2.68491089e-01
-9.52269554e-01 -5.66267252e-01 5.58698714e-01 6.20116115e-01
-5.98369062e-01 4.30291653e-01 -3.82512003e-01 -1.43110454e-01
-4.58078738e-03 -7.29628325e-01 -8.06749463e-01 -6.94902837e-01
-8.47201273e-02 9.25099492e-01 1.08204246e+00 3.79750669e-01
4.48753744e-01 7.57161617e-01 -1.27018321e+00 2.79240403e-02
6.62155628e-01 9.93429601e-01 2.96913236e-02 8.30496550e-01
-3.25227559e-01 6.26824498e-01 -5.93574941e-01 -1.79045215e-01
2.90589720e-01 -4.68819588e-01 -7.50227869e-01 3.20745081e-01
4.42253172e-01 -2.93681234e-01 7.37711862e-02 1.00126350e+00
4.46202010e-01 1.35781229e-01 9.16952729e-01 -1.19862056e+00
3.61050606e-01 6.07566178e-01 -1.19260478e+00 1.49556577e-01
4.74903017e-01 -1.78827763e-01 1.12775242e+00 -8.25209200e-01
6.34951234e-01 1.15429425e+00 7.67088652e-01 -1.93161279e-01
-1.39675486e+00 -2.65003562e-01 -1.38347760e-01 5.14251813e-02
-1.60176206e+00 4.69216704e-03 6.62786067e-01 -6.75483227e-01
5.26497960e-01 -5.67580126e-02 7.91969895e-01 1.85212463e-01
1.48313478e-01 1.10768437e+00 1.07819772e+00 -9.29481566e-01
4.69233930e-01 -2.59760395e-02 2.05326378e-01 9.21544611e-01
-2.72147171e-02 1.62692025e-01 -4.94796216e-01 -2.16375478e-02
8.74829292e-01 -5.11203110e-01 1.12213686e-01 -8.83607447e-01
-1.10074282e+00 7.63808429e-01 3.33120823e-01 1.00902572e-01
-4.88343418e-01 2.31103256e-01 2.45352730e-01 -3.36186975e-01
2.85957634e-01 4.97046486e-02 -6.15180917e-02 1.54568097e-02
-1.69865048e+00 4.72663790e-01 9.55616772e-01 7.70948231e-01
9.51851010e-01 -4.97674346e-01 -3.71795118e-01 5.96832812e-01
6.59199297e-01 1.93878502e-01 1.10450108e-02 -1.05061781e+00
1.27840966e-01 4.84289259e-01 2.33813286e-01 -9.23214734e-01
-3.81202728e-01 -3.39571953e-01 -2.42522359e-01 7.27710843e-01
9.22095478e-01 -1.11520454e-01 -1.02140391e+00 1.18773925e+00
5.63314140e-01 7.25353137e-02 -4.06988651e-01 6.17781818e-01
1.52945057e-01 5.86008012e-01 1.34317242e-02 -1.17766649e-01
1.39203882e+00 -8.74977350e-01 -4.35096771e-01 1.23983305e-02
4.17352140e-01 -9.91287291e-01 6.38443112e-01 5.62633812e-01
-1.07925749e+00 -4.68900144e-01 -1.21018565e+00 5.19677550e-02
-8.51814896e-02 1.93124801e-01 2.81320572e-01 7.89367557e-01
-9.85552311e-01 5.91775298e-01 -1.05518436e+00 -2.58660942e-01
4.41140711e-01 9.69462693e-02 1.25700757e-01 7.21551329e-02
-6.36663258e-01 6.89749181e-01 8.83854210e-01 2.61728391e-02
-7.39103675e-01 -3.02953124e-01 -8.38875592e-01 -6.37069624e-03
2.61775881e-01 -4.84807253e-01 1.23192215e+00 -8.10970664e-01
-1.34549332e+00 9.70464170e-01 -1.68729395e-01 -8.58525276e-01
1.00675142e+00 -4.19928402e-01 2.27199674e-01 4.97660607e-01
2.40541920e-01 1.06829417e+00 7.62585640e-01 -1.51004803e+00
-1.19131827e+00 -9.38287843e-03 -2.12181538e-01 2.70051390e-01
4.83545750e-01 6.24389574e-02 -9.50998485e-01 -2.86175579e-01
4.21553910e-01 -1.04463422e+00 -2.88312048e-01 4.83185425e-02
-7.76837468e-01 -4.33862358e-01 7.66349494e-01 -5.31053185e-01
1.11929500e+00 -1.93917942e+00 -3.21363688e-01 7.15718508e-01
1.50657341e-01 -1.03240877e-01 4.34360325e-01 3.05328965e-01
3.64498720e-02 -1.38228923e-01 -3.95956635e-01 -4.14502859e-01
-8.45051035e-02 -5.53389788e-02 -1.07457168e-01 8.05103958e-01
-8.07993785e-02 4.44518685e-01 -7.70663321e-01 -1.08607936e+00
5.61642826e-01 2.77339995e-01 -2.91319132e-01 4.21273112e-02
-2.67568737e-01 7.10429996e-02 -2.55853117e-01 3.97189349e-01
8.21328521e-01 4.71863151e-02 -1.71759829e-01 -2.33307019e-01
-5.13662934e-01 -1.74183771e-01 -1.59614778e+00 1.44942284e+00
5.03983423e-02 9.01860535e-01 -6.01389110e-01 -7.48141170e-01
1.00521970e+00 6.23020567e-02 6.42390072e-01 -2.10546106e-01
9.17550623e-02 1.07259355e-01 -2.08986059e-01 -1.36112109e-01
6.18939817e-01 6.79672658e-02 -8.74852538e-02 3.78654391e-01
4.09666412e-02 -2.92300224e-01 4.59501714e-01 2.39811167e-01
8.83065462e-01 6.70848191e-01 4.98633385e-01 -1.41232505e-01
3.02936941e-01 4.00448084e-01 3.58264834e-01 1.22026038e+00
-4.20635730e-01 8.12800407e-01 7.72470891e-01 -1.90951645e-01
-1.12804008e+00 -1.18747091e+00 -1.73527479e-01 8.58448923e-01
4.76789773e-01 -7.06041083e-02 -1.14801276e+00 -5.75361371e-01
-2.57781893e-01 9.79262412e-01 -4.90199298e-01 2.79117107e-01
-5.22538602e-01 -6.05670989e-01 4.92624342e-01 3.72073889e-01
5.27834713e-01 -9.28896964e-01 -1.12272584e+00 3.52949530e-01
-6.25284091e-02 -1.07037485e+00 -2.82349676e-01 3.51494044e-01
-7.50834227e-01 -9.17540967e-01 -8.65933120e-01 -7.40262210e-01
5.80532312e-01 2.25490462e-02 9.49514210e-01 -2.51092374e-01
-3.33567679e-01 1.33188799e-01 -3.69303286e-01 -2.22526953e-01
-5.14384270e-01 -7.46460631e-02 -5.74258864e-01 -8.51441920e-02
3.45944256e-01 7.83114508e-02 -6.70842886e-01 3.56983602e-01
-4.15894598e-01 1.11352906e-01 6.73365474e-01 5.17126143e-01
9.29627538e-01 3.17225814e-01 -1.70152590e-01 -7.03108847e-01
3.97945791e-01 -4.33323868e-02 -9.49644148e-01 2.04292104e-01
-4.33304399e-01 1.55127510e-01 -9.58943442e-02 -1.72260821e-01
-1.02584934e+00 6.09200239e-01 -9.50816274e-02 -3.32584083e-01
-5.71210980e-01 4.38268572e-01 4.70092073e-02 1.12044215e-02
7.11263239e-01 5.43172732e-02 -3.23962063e-01 -2.14217186e-01
6.25127017e-01 2.71391869e-01 7.43748605e-01 -3.78501534e-01
6.43372297e-01 7.00892031e-01 1.16914138e-01 -8.19560051e-01
-6.79108143e-01 -8.60924602e-01 -1.03644204e+00 -7.02611506e-01
1.12768126e+00 -5.07437766e-01 -4.47236896e-01 5.02768517e-01
-1.23462057e+00 -2.83228785e-01 -5.25166214e-01 5.06348252e-01
-8.74045432e-01 6.90123737e-01 -5.34396529e-01 -1.04496217e+00
6.60887063e-02 -1.14457846e+00 1.13167024e+00 4.05529499e-01
-3.64123970e-01 -8.80626142e-01 2.66482115e-01 3.79969850e-02
-3.51770818e-01 5.85488856e-01 7.19314516e-01 -6.33917570e-01
-6.72448575e-01 -6.46457314e-01 -3.95285517e-01 2.44261578e-01
-4.19232041e-01 4.75354820e-01 -8.09439063e-01 1.67330787e-01
-3.36247742e-01 6.64652810e-02 7.47558057e-01 8.82892191e-01
7.47260392e-01 1.59747690e-01 -6.62939906e-01 2.27176383e-01
1.59018385e+00 3.45964313e-01 6.68668985e-01 5.19419551e-01
6.11659408e-01 6.90460086e-01 9.52140212e-01 4.62751508e-01
1.96767882e-01 7.45215297e-01 2.30819270e-01 -3.70362222e-01
-2.29280159e-01 -4.60810125e-01 1.30016848e-01 7.70350248e-02
5.01903594e-02 -4.80804443e-01 -1.17243946e+00 5.91578603e-01
-1.84690344e+00 -9.46175814e-01 -5.43338060e-01 2.35377884e+00
5.33478677e-01 7.64504790e-01 4.79946494e-01 4.81514454e-01
1.01860440e+00 -1.67143106e-01 -3.09441060e-01 -2.64874369e-01
3.35103720e-02 -7.38837048e-02 1.02454841e+00 6.37138128e-01
-1.41620219e+00 9.90577817e-01 6.68170643e+00 1.22554147e+00
-6.47266746e-01 -1.13201775e-01 7.13504434e-01 5.35583854e-01
1.49215519e-01 4.14189607e-01 -9.34334695e-01 3.04547966e-01
5.15633225e-01 9.91718844e-02 -1.79752633e-01 9.04369354e-01
4.51003194e-01 -9.07538295e-01 -1.05400014e+00 7.74995983e-01
6.42390642e-03 -1.09327173e+00 -2.54644543e-01 1.07190281e-01
6.54580951e-01 3.32547165e-02 -1.83459550e-01 -1.61669299e-01
5.51711440e-01 -5.55673480e-01 1.18232024e+00 7.32077956e-01
2.64236480e-01 -8.84782970e-01 6.24751449e-01 5.09756923e-01
-1.10989392e+00 2.67295897e-01 -6.69747368e-02 3.82735014e-01
5.77749074e-01 7.62398064e-01 -1.49391329e+00 3.84003162e-01
6.09912694e-01 2.25435197e-01 -5.92995942e-01 1.73796296e+00
-4.31437075e-01 8.11419249e-01 -7.29692459e-01 1.25349775e-01
3.52235615e-01 -2.83029437e-01 6.13856137e-01 1.51039648e+00
1.21417597e-01 -4.19516414e-01 3.87794465e-01 8.57704937e-01
4.36424762e-01 1.09295323e-01 -1.25306696e-01 3.72299314e-01
4.70917851e-01 1.16331089e+00 -1.71246147e+00 -3.87875497e-01
-1.94535777e-01 1.04450560e+00 1.26930013e-01 2.72311360e-01
-7.76421845e-01 -4.38193172e-01 -3.63331944e-01 2.34975323e-01
4.10366535e-01 -4.33342427e-01 -4.16873634e-01 -3.61327916e-01
-2.59120733e-01 -2.77512759e-01 3.50115061e-01 -7.06682444e-01
-7.42475748e-01 4.38892692e-01 3.94432247e-01 -1.26216543e+00
-4.93078560e-01 -2.66341269e-01 -6.18789613e-01 6.89948320e-01
-1.28581154e+00 -1.04774451e+00 -2.14295030e-01 1.94931269e-01
6.48944795e-01 3.08739871e-01 2.51144737e-01 4.68438268e-02
-4.31395024e-01 4.08715248e-01 6.74860477e-02 2.36981153e-01
4.04159218e-01 -1.71283650e+00 3.76554161e-01 1.21131027e+00
2.79464543e-01 1.67129576e-01 1.09344447e+00 -6.93349540e-01
-1.76033914e-01 -6.66243613e-01 7.24643230e-01 -1.94449112e-01
4.05092657e-01 -2.15836570e-01 -7.83198476e-01 4.22322184e-01
1.21188290e-01 -2.57917017e-01 3.87367010e-01 -2.68343806e-01
1.30923212e-01 3.02437305e-01 -1.12547231e+00 4.62505460e-01
3.87555420e-01 -1.18723691e-01 -5.28568089e-01 2.52090216e-01
3.16561967e-01 -4.28913295e-01 -6.30334198e-01 2.46836811e-01
6.72200680e-01 -1.40276039e+00 8.07808578e-01 2.79460251e-01
1.62582636e-01 -7.09437609e-01 2.57217854e-01 -9.06342745e-01
-1.18390031e-01 -5.62000096e-01 1.30899131e-01 1.27085853e+00
3.49707872e-01 -1.54573694e-01 9.93264437e-01 3.18112105e-01
-9.30906273e-03 -6.09163702e-01 -1.06340683e+00 -5.29164851e-01
-3.10772836e-01 -6.31702900e-01 1.56988874e-01 3.59939247e-01
-3.44456315e-01 -1.59745216e-02 -7.44717196e-02 2.92909294e-01
9.45642650e-01 4.59725149e-02 7.76950061e-01 -1.06509268e+00
-3.54891747e-01 -6.24614596e-01 -7.06365585e-01 -1.39883149e+00
-2.05285445e-01 -4.84849066e-01 6.02183223e-01 -1.73022509e+00
1.60308629e-02 -5.95822275e-01 4.35981154e-03 1.31592274e-01
-4.57510911e-02 8.29630047e-02 6.19841479e-02 3.90404642e-01
-6.67768300e-01 -1.29974689e-02 7.72094548e-01 1.87778458e-01
-2.54196554e-01 2.95748204e-01 1.27165377e-01 1.17988288e+00
5.03262401e-01 -6.29128695e-01 -1.40426099e-01 1.06699295e-01
9.22105610e-02 5.84744625e-02 4.16417152e-01 -1.29100120e+00
5.79587162e-01 2.13217959e-01 5.09095430e-01 -1.32133842e+00
1.88167125e-01 -6.91880405e-01 -2.11320639e-01 4.19084668e-01
-4.42119360e-01 -1.38353601e-01 -4.19107974e-02 7.57206678e-01
-1.13058798e-01 -7.59454191e-01 1.05517602e+00 1.18542304e-02
-8.62067223e-01 -4.73343283e-02 -6.63709760e-01 -2.78808117e-01
1.45334733e+00 -7.40209579e-01 3.29840779e-01 -3.24996203e-01
-8.26364458e-01 1.91088170e-01 5.01755536e-01 -3.94429751e-02
5.66236198e-01 -9.21831906e-01 -5.30458927e-01 -2.10427642e-02
-2.87420093e-03 1.81990758e-01 8.69088620e-02 8.18296969e-01
-1.17645538e+00 4.59579863e-02 1.21029817e-01 -1.05884135e+00
-1.18362224e+00 4.25829262e-01 5.04430711e-01 -2.69955754e-01
-5.95054448e-01 8.85353446e-01 8.84622708e-03 1.18618228e-01
2.78450549e-01 -3.12286496e-01 -3.19333196e-01 2.57031679e-01
4.98201512e-02 6.47586882e-01 -1.28537491e-01 -9.06261086e-01
-1.81509569e-01 7.04702556e-01 -1.38222203e-01 -7.04230845e-01
7.38298416e-01 -3.35851640e-01 1.94053590e-01 6.50605381e-01
9.08023834e-01 4.67797257e-02 -1.58000851e+00 -2.23846138e-02
4.73110944e-01 -4.41488355e-01 1.97780192e-01 -5.75883031e-01
-5.94368398e-01 7.22609103e-01 9.18260336e-01 4.75099415e-01
7.95675337e-01 2.10823923e-01 5.03237069e-01 -1.81193516e-01
1.86178684e-01 -1.19142056e+00 -9.62403342e-02 1.07508846e-01
3.50222647e-01 -9.04531002e-01 1.98219329e-01 -7.71095097e-01
-6.33330345e-01 1.37232828e+00 2.11519912e-01 -4.11291867e-01
6.12583995e-01 4.03085232e-01 1.53812738e-02 -2.65465230e-01
-7.51520991e-02 -2.91771173e-01 2.68294483e-01 4.11500037e-01
2.83697456e-01 8.09569936e-03 -3.24444205e-01 -1.85666114e-01
-2.00990975e-01 7.74269924e-02 5.60444057e-01 9.62234974e-01
-8.36605310e-01 -8.98082495e-01 -7.18485892e-01 2.20164835e-01
-3.14713717e-01 1.69890508e-01 -1.17052667e-01 8.32719684e-01
3.71703863e-01 8.38465929e-01 4.31341767e-01 2.66249832e-02
2.38398775e-01 -1.20544367e-01 4.21851754e-01 -4.08591598e-01
-3.73587292e-03 7.07853079e-01 -7.44649693e-02 -4.33622926e-01
-5.12065053e-01 -1.04151475e+00 -1.51560259e+00 3.51018995e-01
-7.97993243e-01 3.64378802e-02 7.68785954e-01 9.60369945e-01
-2.46778861e-01 2.81902462e-01 3.82025331e-01 -1.11622083e+00
-7.76652396e-02 -6.67183280e-01 -6.10030055e-01 1.13062561e-01
1.05116323e-01 -6.10178828e-01 -3.31720352e-01 3.23986590e-01] | [8.30703353881836, -1.40459144115448] |
b9ce67b1-da3b-4f5c-b269-bca934532467 | can-cognate-prediction-be-modelled-as-a-low | null | null | https://aclanthology.org/2021.findings-acl.75 | https://aclanthology.org/2021.findings-acl.75.pdf | Can Cognate Prediction Be Modelled as a Low-Resource Machine Translation Task? | null | ['Benoît Sagot', 'Rachel Bawden', 'Clémentine Fourrier'] | null | null | null | null | findings-acl-2021-8 | ['cognate-prediction'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.368314266204834, 3.712765693664551] |
0fe4ae8d-2308-4596-b412-6eb21d7e3a1d | cardiac-mri-segmentation-with-strong | 1907.02865 | null | https://arxiv.org/abs/1907.02865v2 | https://arxiv.org/pdf/1907.02865v2.pdf | Cardiac MRI Segmentation with Strong Anatomical Guarantees | Recent publications have shown that the segmentation accuracy of modern-day convolutional neural networks (CNN) applied on cardiac MRI can reach the inter-expert variability, a great achievement in this area of research. However, despite these successes, CNNs still produce anatomically inaccurate segmentations as they provide no guarantee on the anatomical plausibility of their outcome, even when using a shape prior. In this paper, we propose a cardiac MRI segmentation method which always produces anatomically plausible results. At the core of the method is an adversarial variational autoencoder (aVAE) whose latent space encodes a smooth manifold on which lies a large spectrum of valid cardiac shapes. This aVAE is used to automatically warp anatomically inaccurate cardiac shapes towards a close but correct shape. Our method can accommodate any cardiac segmentation method and convert its anatomically implausible results to plausible ones without affecting its overall geometric and clinical metrics. With our method, CNNs can now produce results that are both within the inter-expert variability and always anatomically plausible. | ['Pierre-Marc Jodoin', 'Thierry Judge', 'Olivier Bernard', 'Youssef Skandarani', 'Nathan Painchaud', 'Alain Lalande'] | 2019-07-05 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 8.75092596e-02 6.51411176e-01 2.59298444e-01 -3.64884734e-01
-7.67727315e-01 -8.05787146e-01 4.12664533e-01 -1.50200352e-01
-3.32734376e-01 6.39070094e-01 -5.68825006e-02 -4.92085665e-01
-4.41671424e-02 -7.34347999e-01 -7.10291624e-01 -7.85320342e-01
1.15383536e-01 6.36986375e-01 1.99587375e-01 -9.66651961e-02
-9.81844664e-02 7.63953626e-01 -7.17549920e-01 2.10378272e-03
8.66890848e-01 9.06672835e-01 -4.04842526e-01 7.21998096e-01
1.85501650e-01 2.86729842e-01 -4.84690338e-01 -6.15331531e-01
3.84792179e-01 -3.95332307e-01 -9.97930586e-01 -1.81639567e-01
4.75889415e-01 -3.46601665e-01 -9.59599018e-02 1.18081534e+00
3.59275371e-01 -1.38500750e-01 9.64809418e-01 -9.81200814e-01
-7.67258346e-01 5.75861454e-01 -1.58321038e-01 1.35073289e-01
-1.86370745e-01 3.03309739e-01 6.37339652e-01 -6.26703382e-01
6.72093153e-01 8.19514930e-01 1.11607695e+00 7.14753270e-01
-1.45967746e+00 -4.20445144e-01 -4.90956485e-01 -3.65457654e-01
-1.43115497e+00 -2.73046106e-01 8.76013398e-01 -7.70967782e-01
3.37010741e-01 4.92512405e-01 7.91072845e-01 1.09697843e+00
7.47056663e-01 2.44459331e-01 8.67830992e-01 -4.02526669e-02
3.49665701e-01 -8.45589582e-03 -2.51346916e-01 5.22915483e-01
2.26026237e-01 2.35178173e-01 3.05266827e-01 -2.04187870e-01
1.28379440e+00 -5.99796847e-02 -4.67190593e-01 -5.37568390e-01
-1.33289456e+00 1.01505005e+00 7.31179774e-01 5.93310237e-01
-4.71596599e-01 2.13829130e-01 3.14272046e-01 -1.05670258e-01
2.58130878e-01 6.55183077e-01 -1.39293104e-01 1.11379199e-01
-1.29779553e+00 1.20182551e-01 5.85990787e-01 4.92372215e-01
1.22351259e-01 2.78778285e-01 -1.66707143e-01 5.82288504e-01
3.04830909e-01 3.09694022e-01 4.93319184e-01 -1.15511096e+00
-1.40684634e-01 3.71860325e-01 -1.24694079e-01 -1.20631087e+00
-4.60852742e-01 -5.75825810e-01 -1.16018498e+00 4.70484287e-01
6.63167536e-01 -1.95163324e-01 -9.43646669e-01 1.53339255e+00
3.49414825e-01 -7.75132701e-02 -3.68093625e-02 1.07849252e+00
7.70177245e-01 2.69880325e-01 3.83984059e-01 1.19248927e-02
1.12122202e+00 -6.24719024e-01 -5.60953319e-01 -1.35188764e-02
3.86687785e-01 -8.17376792e-01 9.85512197e-01 3.00956368e-01
-1.27682555e+00 -3.54946882e-01 -1.10545278e+00 1.31531041e-02
5.43186031e-02 -2.97065318e-01 5.66424251e-01 9.72183466e-01
-1.45742977e+00 1.07943463e+00 -1.07862997e+00 8.76062270e-03
7.46662438e-01 3.68994206e-01 -3.67888987e-01 4.42811459e-01
-1.01823246e+00 1.08961213e+00 5.48177838e-01 3.00281942e-01
-7.80958593e-01 -1.02471495e+00 -7.38790929e-01 -9.45920199e-02
1.36144817e-01 -8.71008635e-01 1.00246239e+00 -1.21742415e+00
-1.41527641e+00 1.04575717e+00 4.81997520e-01 -5.84742308e-01
9.40465987e-01 2.92845726e-01 -3.74947548e-01 3.70120853e-01
-1.34337991e-01 1.05067801e+00 9.22046959e-01 -1.29161012e+00
1.26374170e-01 -1.87156171e-01 -3.25031132e-01 -2.82384574e-01
4.61469069e-02 -1.87915012e-01 -2.34762065e-05 -1.02596581e+00
2.90027410e-01 -1.19190907e+00 -5.68166375e-01 4.63143498e-01
-4.46466446e-01 1.48707494e-01 4.24157292e-01 -8.05975378e-01
7.17765093e-01 -2.00140333e+00 2.45702520e-01 2.10288778e-01
5.99395335e-01 2.15371907e-01 2.61491150e-01 -1.87449917e-01
-2.89639354e-01 5.05111933e-01 -8.36668730e-01 -1.04063347e-01
-2.53757745e-01 3.33290368e-01 -1.26947641e-01 8.00038755e-01
2.89608747e-01 1.26347423e+00 -8.00401688e-01 -6.48991227e-01
4.01595622e-01 7.69476831e-01 -7.98002660e-01 5.92981391e-02
5.16032800e-02 9.30477381e-01 -3.23318541e-01 3.74947876e-01
6.87301219e-01 -1.00554638e-01 2.29919087e-02 -2.68070608e-01
1.15874290e-01 -6.67231560e-01 -6.24362230e-01 1.69094932e+00
-2.77545631e-01 4.90171134e-01 1.58906683e-01 -9.27115679e-01
7.72156060e-01 3.82458240e-01 8.22391212e-01 -2.12232798e-01
5.08608937e-01 4.87938523e-01 2.11558953e-01 -2.71148711e-01
1.49883419e-01 -7.80155063e-01 1.87366411e-01 2.55818218e-01
1.10373698e-01 -3.78446043e-01 -4.10803646e-01 -7.24357888e-02
7.18186796e-01 5.10048307e-02 1.57275926e-02 -6.85410261e-01
4.32075381e-01 -1.26924075e-03 4.19838309e-01 4.81273949e-01
-6.11593425e-01 1.27493405e+00 5.44677436e-01 -7.27751791e-01
-1.54699111e+00 -1.47892785e+00 -6.84674978e-01 1.51742101e-01
-1.64350912e-01 4.20794338e-01 -1.22818398e+00 -6.95209265e-01
-2.96059430e-01 6.85350060e-01 -8.76866400e-01 -4.28856045e-01
-5.44543266e-01 -6.06535375e-01 9.78216052e-01 6.70700371e-01
1.44792855e-01 -1.20123243e+00 -8.92028451e-01 2.68981099e-01
-1.73357233e-01 -1.04017818e+00 -4.55429733e-01 -2.25587368e-01
-1.05967665e+00 -1.10934103e+00 -1.25877523e+00 -5.23280859e-01
7.95007348e-01 -7.86035955e-01 1.26152277e+00 2.87066162e-01
-5.07813454e-01 1.57008946e-01 -1.54558524e-01 -3.58105391e-01
-1.06841326e+00 -1.58824235e-01 -1.01436414e-01 8.33423659e-02
-1.98013559e-01 -6.85312390e-01 -9.58314240e-01 4.04442921e-02
-1.19759929e+00 -1.20519638e-01 3.29341561e-01 6.50141478e-01
7.94313729e-01 -3.54808003e-01 4.41835999e-01 -8.46862674e-01
4.79270935e-01 -4.39027280e-01 -3.50406140e-01 1.01460777e-01
-7.38382459e-01 6.32533133e-02 7.15862632e-01 -2.84930259e-01
-6.93343818e-01 -3.37690897e-02 -4.58290696e-01 -8.13545942e-01
-2.87053257e-01 3.09490830e-01 3.95769268e-01 -1.76058397e-01
1.06140637e+00 1.64067820e-01 3.01813126e-01 -1.66913688e-01
3.61024499e-01 1.83420449e-01 8.74790847e-01 -4.48652059e-01
5.72028697e-01 7.52477884e-01 2.84730077e-01 -4.53962326e-01
-5.58017313e-01 3.49572524e-02 -1.00177634e+00 -4.23243731e-01
1.33154547e+00 -3.00217092e-01 -4.49048996e-01 3.33015800e-01
-1.04613006e+00 -4.06931132e-01 -5.54550886e-01 3.35251451e-01
-6.62183523e-01 5.94162107e-01 -6.10768378e-01 -3.64762098e-01
-6.55693293e-01 -1.52421749e+00 9.27558482e-01 -3.03789340e-02
-5.13988853e-01 -1.44701147e+00 -1.43161137e-02 2.48428494e-01
5.86050987e-01 1.05898523e+00 1.01466250e+00 -6.75124109e-01
-1.58030346e-01 -2.53312916e-01 -4.19195965e-02 5.48780203e-01
7.14797452e-02 3.53289068e-01 -9.01110351e-01 -3.23263913e-01
1.99241132e-01 8.92118141e-02 5.24530053e-01 9.36610162e-01
1.31046569e+00 -2.47215927e-01 -1.10597931e-01 8.30079973e-01
1.41572571e+00 1.81014091e-01 7.70900726e-01 -5.24928644e-02
7.20734358e-01 5.03021717e-01 -1.08880997e-01 -1.62875071e-01
-6.12232313e-02 4.83280838e-01 6.70679450e-01 -5.00151515e-01
-3.06473643e-01 1.27064884e-01 -7.98103437e-02 7.89272308e-01
-2.39879131e-01 1.17043957e-01 -1.32847393e+00 5.98702073e-01
-1.32537341e+00 -7.48630643e-01 -3.63781810e-01 2.14801311e+00
8.92131388e-01 1.30578622e-01 -6.53918386e-02 1.11717515e-01
6.23939931e-01 -8.68375748e-02 -5.23913383e-01 -7.46247172e-01
1.78470388e-01 3.84998947e-01 4.16713834e-01 4.39307004e-01
-1.11561918e+00 6.49601996e-01 6.76590014e+00 6.42466903e-01
-1.47923386e+00 3.64288390e-01 9.58542287e-01 2.43407026e-01
-4.39450115e-01 -3.18259984e-01 9.53191593e-02 3.86791438e-01
9.54295397e-01 -1.68581400e-02 2.38610655e-01 8.30814242e-01
6.90414235e-02 1.77547827e-01 -1.04887211e+00 7.91979134e-01
1.61022518e-03 -1.70946002e+00 8.45731571e-02 3.68681476e-02
8.37531328e-01 -1.39505014e-01 3.10212284e-01 5.54169938e-02
-2.67811213e-02 -1.59165311e+00 8.28838885e-01 7.02865005e-01
1.26242244e+00 -9.27288949e-01 8.25180054e-01 2.09600523e-01
-6.26468658e-01 3.73593688e-01 -2.36033738e-01 5.28818369e-01
2.34098881e-01 3.75762135e-01 -8.96292686e-01 3.61464947e-01
6.41283631e-01 1.79061860e-01 -4.81290340e-01 8.92090261e-01
1.30338103e-01 4.54930574e-01 -2.28002995e-01 3.02748919e-01
3.36596966e-01 -3.12141001e-01 7.00362742e-01 8.72342825e-01
1.20087214e-01 9.10753161e-02 -2.00196467e-02 1.43804467e+00
-2.10733595e-03 2.25781038e-01 -5.67968726e-01 1.92307800e-01
-2.29509212e-02 1.14651442e+00 -1.22975504e+00 -2.67458975e-01
9.33693349e-02 9.11029875e-01 -6.88863732e-03 1.90446720e-01
-9.85234797e-01 1.37670413e-01 2.95259416e-01 1.16588719e-01
-7.08619580e-02 5.93886971e-02 -7.65994668e-01 -9.91043806e-01
-5.60831651e-02 -6.78322196e-01 2.16414303e-01 -7.43346751e-01
-1.27583969e+00 8.81399751e-01 -1.18996024e-01 -1.01411510e+00
-1.41865030e-01 -5.81590235e-01 -6.07793689e-01 8.25630724e-01
-9.54676270e-01 -1.09830964e+00 -5.09013943e-02 3.68557185e-01
3.32858980e-01 3.20927911e-02 1.03320456e+00 2.36051261e-01
-1.10115223e-01 5.57447135e-01 -4.52191420e-02 4.46823716e-01
4.18155342e-01 -1.30552757e+00 2.14252248e-01 7.61290908e-01
1.47465214e-01 6.84648156e-01 8.05345535e-01 -6.37771070e-01
-7.18433619e-01 -1.14367998e+00 3.21885318e-01 -7.35072255e-01
5.05537093e-01 2.41770595e-01 -1.02916861e+00 7.67878890e-01
8.13685283e-02 4.11871850e-01 5.76726079e-01 -2.30630070e-01
-8.64872262e-02 3.60082686e-01 -1.46985948e+00 7.79962182e-01
5.61762810e-01 -3.25010091e-01 -7.70645082e-01 2.83885926e-01
5.06203771e-01 -7.61800170e-01 -1.38457239e+00 7.04452932e-01
5.70979714e-01 -1.01989317e+00 1.15850282e+00 -6.82897806e-01
7.91139841e-01 -1.68200023e-02 2.49073908e-01 -1.26896453e+00
-1.00919753e-01 -3.53140026e-01 2.25224152e-01 5.94688773e-01
3.75346124e-01 -4.42581654e-01 6.99351192e-01 9.24717367e-01
-3.37419361e-01 -9.79673386e-01 -1.35984242e+00 -5.73647559e-01
1.00736475e+00 -5.12179255e-01 4.12516445e-01 1.25300133e+00
-5.64743161e-01 -3.54123145e-01 -7.02203531e-03 1.54838532e-01
9.07709181e-01 -1.46474630e-01 2.61848778e-01 -1.28896415e+00
-8.00524801e-02 -8.75637412e-01 -4.67640549e-01 -1.72784895e-01
6.07273839e-02 -1.33144403e+00 -5.42392321e-02 -1.17526829e+00
4.72444743e-02 -5.50945640e-01 -2.42162928e-01 3.37975949e-01
1.29779264e-01 5.80629826e-01 1.28087446e-01 3.09321076e-01
9.56523940e-02 2.50533998e-01 1.68340683e+00 -1.43515512e-01
-1.18768454e-01 -1.38073385e-01 -4.94465560e-01 1.06343257e+00
7.69147813e-01 -7.21621156e-01 -1.84042200e-01 -3.40901881e-01
-8.73204023e-02 2.79524028e-01 9.72283661e-01 -1.01364851e+00
-6.51124911e-03 1.72239751e-01 7.07086682e-01 -1.32018447e-01
-8.18466023e-02 -9.33890045e-01 5.94905615e-01 6.99385762e-01
-3.33639592e-01 -1.98317528e-01 2.76254773e-01 3.66009831e-01
4.67971154e-03 -2.52368659e-01 1.17205930e+00 -1.49096817e-01
-4.87993807e-02 5.59594214e-01 -1.83635145e-01 3.23595077e-01
1.14114487e+00 -3.82896423e-01 3.54062736e-01 -1.97385758e-01
-1.31623209e+00 -2.91451991e-01 6.09398186e-01 3.26402068e-01
5.40043116e-01 -1.28282046e+00 -8.79071772e-01 2.66262084e-01
-2.03682885e-01 2.68909663e-01 5.10816395e-01 1.00308478e+00
-1.16789675e+00 1.63134828e-01 -4.38119441e-01 -1.06234694e+00
-7.72768140e-01 6.32069707e-01 1.00784326e+00 -2.07538679e-02
-9.31727469e-01 8.40166330e-01 8.51716101e-02 -4.18299466e-01
-2.17781693e-01 -4.57554460e-01 -1.59693137e-01 -2.72712827e-01
3.22941095e-02 -1.02381557e-01 8.53885263e-02 -8.37572336e-01
-2.78495759e-01 5.99330723e-01 3.89847577e-01 -1.01446472e-02
1.28709531e+00 2.71686614e-01 -8.22888687e-02 2.08577156e-01
1.08582616e+00 8.40259492e-02 -1.42543352e+00 4.35583621e-01
-3.00504059e-01 -2.27870703e-01 3.61300915e-01 -8.16749930e-01
-1.63882244e+00 9.72446680e-01 8.32740903e-01 2.74926186e-01
7.51962543e-01 4.49495763e-02 8.36686075e-01 -2.72356242e-01
2.18756348e-01 -7.79562056e-01 -2.69337445e-01 -3.66313793e-02
1.24727333e+00 -1.11708474e+00 -1.63718522e-01 -3.47316384e-01
-6.75096393e-01 1.21219349e+00 1.85857669e-01 -4.34063107e-01
8.22000921e-01 2.01232612e-01 3.99321109e-01 -3.51446331e-01
5.44764549e-02 4.46601272e-01 5.17726660e-01 7.28833199e-01
3.34601372e-01 2.73485035e-01 -2.41841719e-01 6.80277288e-01
-3.56162965e-01 -1.25434369e-01 4.66250628e-01 3.40205789e-01
4.88400236e-02 -8.64295959e-01 -4.77530062e-01 1.60643473e-01
-8.84637713e-01 -7.00961053e-02 -4.25134692e-03 9.20706153e-01
1.69276163e-01 4.29832757e-01 3.98482345e-02 4.46506813e-02
2.18861282e-01 4.38862778e-02 5.23767054e-01 -3.73804897e-01
-7.36751914e-01 -3.42623191e-03 -4.39572871e-01 -5.69500804e-01
-9.15275142e-02 -5.64031661e-01 -1.45129597e+00 -3.86171639e-01
-3.90325151e-02 -1.13010019e-01 6.17912591e-01 8.73674691e-01
8.81600901e-02 7.18711913e-01 2.74459958e-01 -8.03613245e-01
-6.46878898e-01 -6.10559344e-01 -3.54124516e-01 7.38031089e-01
2.85537750e-01 -5.37974775e-01 -2.18519136e-01 2.65714586e-01] | [14.184260368347168, -2.2490007877349854] |
12009b12-a12a-4254-8950-2e73caaa6631 | optimizing-bi-encoder-for-named-entity | 2208.14565 | null | https://arxiv.org/abs/2208.14565v2 | https://arxiv.org/pdf/2208.14565v2.pdf | Optimizing Bi-Encoder for Named Entity Recognition via Contrastive Learning | We present a bi-encoder framework for named entity recognition (NER), which applies contrastive learning to map candidate text spans and entity types into the same vector representation space. Prior work predominantly approaches NER as sequence labeling or span classification. We instead frame NER as a representation learning problem that maximizes the similarity between the vector representations of an entity mention and its type. This makes it easy to handle nested and flat NER alike, and can better leverage noisy self-supervision signals. A major challenge to this bi-encoder formulation for NER lies in separating non-entity spans from entity mentions. Instead of explicitly labeling all non-entity spans as the same class $\texttt{Outside}$ ($\texttt{O}$) as in most prior methods, we introduce a novel dynamic thresholding loss. Experiments show that our method performs well in both supervised and distantly supervised settings, for nested and flat NER alike, establishing new state of the art across standard datasets in the general domain (e.g., ACE2004, ACE2005) and high-value verticals such as biomedicine (e.g., GENIA, NCBI, BC5CDR, JNLPBA). We release the code at github.com/microsoft/binder. | ['Hoifung Poon', 'Jianfeng Gao', 'Hao Cheng', 'Sheng Zhang'] | 2022-08-30 | null | null | null | null | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [ 1.42766073e-01 1.79067820e-01 -2.67123699e-01 -3.77473295e-01
-1.08423400e+00 -9.00944352e-01 1.15329236e-01 6.55650377e-01
-8.80969584e-01 9.49961245e-01 5.09098113e-01 -2.72746205e-01
6.17264658e-02 -6.86230004e-01 -7.20388353e-01 -5.05405188e-01
8.79477337e-02 4.25169110e-01 -2.22588897e-01 5.95292496e-03
-2.95028463e-02 2.74695873e-01 -1.21272564e+00 3.42179865e-01
6.49648368e-01 6.62160695e-01 7.39647150e-02 5.97496450e-01
-3.85163456e-01 6.44352853e-01 -6.68106794e-01 -8.73464942e-01
-1.49170579e-02 -3.01327616e-01 -1.25959444e+00 -3.18461657e-01
3.76087666e-01 2.21225873e-01 -5.37213147e-01 9.64858770e-01
8.44442368e-01 3.57449166e-02 7.53989935e-01 -8.45265627e-01
-7.82381892e-01 8.69659066e-01 -4.38134134e-01 3.48215312e-01
2.36402184e-01 -2.08874136e-01 1.35972238e+00 -9.40451026e-01
9.89324033e-01 7.50228763e-01 1.10991371e+00 6.50857985e-01
-1.17658734e+00 -4.43745553e-01 -1.80212989e-01 1.52841747e-01
-1.48475718e+00 -4.56580341e-01 1.65095076e-01 -5.40542662e-01
1.00322843e+00 3.70981455e-01 3.71439680e-02 1.13921249e+00
1.16854245e-02 9.09644485e-01 7.16244996e-01 -3.44598979e-01
2.59094536e-01 -1.56847417e-01 6.20809555e-01 6.02074087e-01
2.74622738e-01 -2.85709977e-01 -4.46205348e-01 -3.43418211e-01
4.29866612e-01 -1.93650797e-02 -2.95973748e-01 1.07712507e-01
-1.34445548e+00 7.77609110e-01 2.83082873e-01 4.45329100e-01
-4.72863346e-01 -3.69297899e-02 8.65171015e-01 1.06107416e-02
3.55909914e-01 4.58516508e-01 -9.51040745e-01 -2.14962691e-01
-1.06980908e+00 -2.45556176e-01 9.33079720e-01 1.10996068e+00
6.74220622e-01 -2.16863111e-01 -5.42596757e-01 9.98072445e-01
-5.49484231e-02 1.15759000e-01 5.62691212e-01 -5.98629355e-01
4.52994943e-01 4.35961097e-01 4.75312620e-02 -4.60868299e-01
-6.43845677e-01 -2.38202184e-01 -1.03944814e+00 -4.83078241e-01
2.84934044e-01 -5.90870798e-01 -1.06352818e+00 1.76820707e+00
3.85050744e-01 3.26459438e-01 2.71678269e-01 5.83390951e-01
1.26110065e+00 5.65491438e-01 4.11391735e-01 -1.23494454e-01
1.84631121e+00 -6.38730586e-01 -8.33124459e-01 -7.32797533e-02
8.46894145e-01 -6.25494957e-01 6.83714509e-01 -6.91995844e-02
-8.49763870e-01 -3.02027702e-01 -6.84882820e-01 -5.37070155e-01
-6.92671180e-01 4.58088517e-01 4.92262900e-01 4.78250116e-01
-8.64305377e-01 7.13146985e-01 -9.06959414e-01 -4.14998978e-01
3.37169170e-01 1.70980304e-01 -5.94065249e-01 -8.38654041e-02
-1.53659797e+00 9.41659510e-01 7.28194356e-01 1.42086926e-03
-4.21829939e-01 -8.68639529e-01 -1.22708666e+00 1.62564293e-01
1.24098033e-01 -6.05604529e-01 1.01279747e+00 -3.47440451e-01
-9.16974366e-01 1.15435505e+00 -3.05422097e-01 -5.82388043e-01
1.07110180e-01 -2.64297545e-01 -6.46722198e-01 -1.43332100e-02
3.98596138e-01 6.13160551e-01 -3.39047313e-02 -8.16249132e-01
-4.45168525e-01 -2.98874319e-01 -1.71614215e-01 7.00675184e-03
-4.68981802e-01 1.62856042e-01 -3.94181311e-01 -7.98919857e-01
-1.44129708e-01 -6.11556411e-01 -4.29247230e-01 -1.85892597e-01
-8.57231498e-01 -6.00033820e-01 2.47777551e-01 -8.61670256e-01
1.46527374e+00 -2.02973342e+00 4.39276770e-02 -6.17562383e-02
3.03794026e-01 1.96997598e-01 -6.57428727e-02 7.39827156e-01
-4.36769038e-01 4.50119942e-01 -4.58927959e-01 -2.33131871e-01
6.02015071e-02 2.44769886e-01 -1.21384673e-01 6.31690383e-01
2.50772268e-01 8.47820699e-01 -1.02857268e+00 -6.50107622e-01
-1.74570501e-01 7.22245216e-01 -2.46027410e-01 2.75767058e-01
6.16772808e-02 -1.32303620e-02 -2.43124649e-01 5.99889159e-01
4.24689382e-01 -4.87024695e-01 3.37784410e-01 -3.82377088e-01
-1.71302095e-01 5.26310623e-01 -1.29235852e+00 1.67764568e+00
-2.53736854e-01 4.16161984e-01 -5.94427325e-02 -1.02175558e+00
8.36216986e-01 7.18887448e-01 6.16769433e-01 -4.33037616e-02
2.26369221e-02 1.24164432e-01 -4.00504410e-01 -4.11815703e-01
5.33828855e-01 -1.45616651e-01 -3.52416158e-01 9.02088135e-02
5.18338263e-01 4.82784510e-01 4.73022938e-01 9.05359685e-02
1.39465570e+00 3.12445499e-02 7.59077311e-01 -1.70241714e-01
2.16539025e-01 1.94785409e-02 1.08869362e+00 6.83958054e-01
-1.42320186e-01 6.41385555e-01 4.82276618e-01 -2.84029454e-01
-1.03602624e+00 -1.02822518e+00 -5.64828277e-01 1.25491250e+00
-1.81168944e-01 -6.27235472e-01 -5.85543871e-01 -1.00502229e+00
-5.56773394e-02 6.95900142e-01 -7.09847033e-01 7.58051500e-03
-6.46100521e-01 -9.09104943e-01 1.30303741e+00 7.74595201e-01
1.00077830e-01 -1.11373734e+00 -3.46292347e-01 4.00477022e-01
-4.58123177e-01 -9.81209517e-01 -8.67205381e-01 9.24948394e-01
-5.29438853e-01 -9.39733744e-01 -1.04627657e+00 -1.06749225e+00
4.33754802e-01 -3.70462209e-01 1.27744675e+00 -3.19586694e-01
-5.38956821e-01 2.09188312e-01 -3.94618750e-01 -2.93578625e-01
-1.35781035e-01 2.74482310e-01 -4.54137521e-03 -2.49378592e-01
6.38656557e-01 -8.05186108e-02 -5.06164432e-01 1.11451037e-01
-9.44358766e-01 -3.71736616e-01 5.17597795e-01 1.21196365e+00
9.97078896e-01 -3.13492477e-01 6.72210097e-01 -1.16856360e+00
3.45549494e-01 -9.36927080e-01 -2.32569039e-01 5.07902861e-01
-4.57407475e-01 2.74504095e-01 7.45198667e-01 -2.85100728e-01
-8.63927662e-01 2.18642607e-01 -5.78980088e-01 -3.07145774e-01
-4.39159632e-01 8.19797516e-01 -2.89197475e-01 8.32558095e-01
7.02420771e-01 3.20498884e-01 -6.13772213e-01 -7.62379646e-01
4.98338312e-01 1.02604890e+00 8.04697037e-01 -4.08071488e-01
3.12573880e-01 9.89652947e-02 -3.84835333e-01 -8.16792667e-01
-9.70616996e-01 -9.75500584e-01 -7.21685946e-01 4.93896842e-01
1.17452264e+00 -1.08150911e+00 -5.36743999e-01 -4.49857488e-02
-1.20704627e+00 1.22785931e-02 -4.14950341e-01 4.51698661e-01
-2.70062387e-01 5.20802081e-01 -1.08666992e+00 -5.79538822e-01
-6.99564159e-01 -7.76757956e-01 1.27675331e+00 2.15312675e-01
-4.25597817e-01 -9.63726163e-01 3.45077574e-01 2.31115624e-01
-4.09864262e-02 3.21516335e-01 8.53528976e-01 -1.18517828e+00
1.82521671e-01 -2.31339931e-01 -3.54172885e-02 1.20709784e-01
9.60398465e-02 -2.63316721e-01 -8.63210142e-01 -1.46409556e-01
-4.32451487e-01 -4.28391039e-01 1.21604538e+00 1.51282579e-01
1.11630762e+00 -4.21849906e-01 -5.72788060e-01 7.34331131e-01
1.40856445e+00 -1.53793227e-02 6.28040373e-01 2.73072392e-01
6.53062046e-01 4.44865555e-01 4.21574146e-01 4.46599364e-01
3.72954369e-01 4.66248482e-01 -8.50842521e-02 -2.68077463e-01
3.36471722e-02 -3.06601584e-01 2.38565594e-01 7.51133144e-01
2.40336850e-01 -5.56739926e-01 -1.06106055e+00 9.38777387e-01
-1.60716724e+00 -1.08163714e+00 -2.75470186e-02 1.98497081e+00
1.42769253e+00 -2.99644768e-01 -5.14443181e-02 -1.17112197e-01
1.00364363e+00 -6.02896251e-02 -6.17909789e-01 -3.05717975e-01
-3.08735669e-01 5.39269745e-01 7.00567186e-01 5.22232130e-02
-1.48391461e+00 8.84819806e-01 5.68397522e+00 7.83188462e-01
-7.98867702e-01 1.91467583e-01 7.10273743e-01 1.57296970e-01
-2.68615097e-01 -2.71675110e-01 -1.21028733e+00 6.23241723e-01
1.20999372e+00 -2.26537704e-01 4.07739356e-03 8.03746581e-01
-2.22740635e-01 5.16568244e-01 -1.30995595e+00 1.07050133e+00
-6.39989460e-03 -1.67056262e+00 -3.61047834e-01 -4.07126024e-02
4.82916147e-01 4.87129480e-01 -2.78733879e-01 5.09733438e-01
6.89835846e-01 -1.21594787e+00 4.78175938e-01 3.55472416e-01
1.32207561e+00 -6.04558289e-01 1.05272055e+00 1.46198064e-01
-1.27095926e+00 2.68242478e-01 -4.98185188e-01 6.80700302e-01
1.86810926e-01 8.79776001e-01 -7.87475944e-01 8.01060140e-01
8.37777793e-01 7.12996244e-01 -2.12005377e-01 9.81370330e-01
-2.04009712e-01 7.64416397e-01 -3.63775939e-01 1.08677074e-01
-8.51246789e-02 2.20907703e-01 3.26127321e-01 1.93194711e+00
1.33446380e-01 1.43164888e-01 2.13636532e-01 5.66287637e-01
-6.82383716e-01 4.03114498e-01 -3.49320710e-01 -1.80771291e-01
6.86316371e-01 1.22234297e+00 -6.91740870e-01 -4.15522635e-01
-3.79239947e-01 1.15806222e+00 7.69444048e-01 2.90090948e-01
-8.06608021e-01 -9.86472487e-01 8.27178240e-01 -3.73133987e-01
7.48547673e-01 1.18661046e-01 -2.85630673e-01 -1.47822571e+00
-1.63071260e-01 -6.99184716e-01 1.08791244e+00 -3.46973419e-01
-1.76062691e+00 7.27414668e-01 -3.52567345e-01 -1.09419227e+00
-9.95923951e-02 -6.09146595e-01 -3.81501138e-01 7.24157929e-01
-1.27246344e+00 -9.01271045e-01 3.34294647e-01 3.51847887e-01
3.24419349e-01 1.02562055e-01 1.14732599e+00 4.46666569e-01
-7.97640920e-01 1.08056426e+00 5.32172680e-01 8.83716047e-01
9.37649071e-01 -1.51957607e+00 4.63180125e-01 6.39560461e-01
3.58596325e-01 8.78658473e-01 3.73068184e-01 -6.70042753e-01
-1.19500268e+00 -1.39534986e+00 1.33673751e+00 -5.70737958e-01
4.69211042e-01 -5.06666720e-01 -9.65232253e-01 8.41877222e-01
2.91150659e-01 1.37190342e-01 1.25775659e+00 4.07631814e-01
-6.39938891e-01 3.54094177e-01 -1.21704006e+00 4.27219212e-01
1.06309497e+00 -5.86899102e-01 -8.38225603e-01 3.61981153e-01
7.62573004e-01 -4.02285606e-01 -1.48531377e+00 1.24423645e-01
2.02384859e-01 -3.46577853e-01 1.04066241e+00 -9.90517259e-01
4.14390743e-01 -3.43990207e-01 -2.51032561e-01 -1.30367267e+00
-3.84026170e-01 -4.49503928e-01 -2.79687971e-01 1.52684808e+00
6.45120084e-01 -2.94337928e-01 6.69165969e-01 3.53683770e-01
-3.88023704e-01 -7.81335473e-01 -1.14803588e+00 -7.42962599e-01
2.25625142e-01 -6.60604089e-02 4.87699687e-01 1.37834096e+00
3.86370033e-01 5.80880463e-01 -2.54314125e-01 2.93383390e-01
3.93785506e-01 2.64925268e-02 1.08520128e-01 -1.17722678e+00
-1.89559102e-01 -2.00259045e-01 -4.51901883e-01 -9.41854596e-01
2.50446498e-01 -1.34804988e+00 3.30943674e-01 -1.77137339e+00
2.85595894e-01 -3.27964962e-01 -6.78610921e-01 1.03425777e+00
-5.16803145e-01 1.33665681e-01 -1.47907332e-01 1.24234438e-01
-7.27392793e-01 3.04232061e-01 3.91296685e-01 -4.01741415e-02
5.38643934e-02 -3.98043871e-01 -8.45577657e-01 4.85094041e-01
6.94980919e-01 -7.33146727e-01 2.29795516e-01 -5.02126157e-01
1.56216234e-01 1.17072344e-01 -7.28795305e-02 -6.78891122e-01
3.52298737e-01 -4.91914898e-02 5.62683821e-01 -6.05544567e-01
-3.48041356e-02 -4.15333122e-01 4.60786968e-02 2.76907653e-01
-7.69177973e-01 1.12044374e-02 1.03509992e-01 7.33171284e-01
-1.93637326e-01 -3.66399795e-01 6.98008955e-01 -1.59660682e-01
-7.22257137e-01 4.67948318e-01 -3.29846263e-01 6.40351415e-01
7.70970106e-01 2.18983088e-02 -3.67927074e-01 1.00006275e-01
-8.88754129e-01 3.40610027e-01 2.28087515e-01 2.33497590e-01
3.23646665e-01 -1.13097131e+00 -9.60443914e-01 -2.39857420e-01
3.83266658e-01 -3.59612815e-02 3.10768783e-01 6.91176653e-01
-1.52178451e-01 3.84387553e-01 1.34272471e-01 -2.88660735e-01
-1.18698013e+00 5.81049860e-01 2.12659568e-01 -5.02783358e-01
-5.77010512e-01 1.06823349e+00 5.20667881e-02 -8.08605492e-01
3.39220941e-01 -1.70425996e-01 -3.93725306e-01 2.91637421e-01
5.83417833e-01 3.74962807e-01 2.57784039e-01 -6.70625210e-01
-5.73707700e-01 8.14665109e-02 -1.24322601e-01 2.42252305e-01
1.53093100e+00 4.93520536e-02 -1.13473162e-01 3.96504879e-01
1.55105686e+00 2.16304846e-02 -6.91341221e-01 -1.87397003e-01
5.59112251e-01 2.26922631e-01 -1.61376297e-01 -8.05326223e-01
-7.25749314e-01 7.36112297e-01 5.04908979e-01 -1.43896952e-01
7.63336837e-01 3.54595006e-01 9.10062253e-01 5.16616046e-01
1.06904395e-01 -1.02542436e+00 -4.43871647e-01 7.58762300e-01
3.61394346e-01 -1.14519882e+00 -2.09893614e-01 -1.89932629e-01
-7.24130273e-01 1.00174236e+00 3.61128777e-01 1.50973544e-01
4.71076041e-01 4.71651107e-01 6.16012067e-02 -2.07180694e-01
-6.62404001e-01 -3.81343931e-01 2.27456629e-01 4.72505718e-01
1.09733784e+00 8.08593258e-02 -5.57649553e-01 1.10197687e+00
-1.45688042e-01 2.22042631e-02 5.31553626e-01 1.03383732e+00
-2.05694050e-01 -1.05624115e+00 -1.16356336e-01 8.22140336e-01
-1.09666371e+00 -5.40949345e-01 -1.16547778e-01 3.94617945e-01
2.99357116e-01 8.03217173e-01 1.02317944e-01 -1.46826342e-01
3.78993064e-01 4.87357646e-01 -1.02432430e-01 -9.13110435e-01
-8.62951815e-01 -8.30908306e-03 4.57518399e-01 -1.82931170e-01
-9.07779112e-02 -7.30244219e-01 -1.62662876e+00 7.25640953e-02
-4.78725046e-01 6.28768682e-01 3.86845112e-01 6.36059225e-01
6.34477794e-01 3.81477118e-01 3.45094681e-01 -8.25833604e-02
-6.33174062e-01 -8.70743155e-01 -5.84646702e-01 6.10210955e-01
1.99263319e-01 -4.19506252e-01 -2.00435117e-01 3.37736011e-01] | [8.811552047729492, 8.895841598510742] |
c6c44e5e-ec19-4686-b228-2909fac1520c | 190409406 | 1904.09406 | null | https://arxiv.org/abs/1904.09406v3 | https://arxiv.org/pdf/1904.09406v3.pdf | DeepMoD: Deep learning for Model Discovery in noisy data | We introduce DeepMoD, a Deep learning based Model Discovery algorithm. DeepMoD discovers the partial differential equation underlying a spatio-temporal data set using sparse regression on a library of possible functions and their derivatives. A neural network approximates the data and constructs the function library, but it also performs the sparse regression. This construction makes it extremely robust to noise, applicable to small data sets, and, contrary to other deep learning methods, does not require a training set. We benchmark our approach on several physical problems such as the Burgers', Korteweg-de Vries and Keller-Segel equations, and find that it requires as few as $\mathcal{O}(10^2)$ samples and works at noise levels up to $75\%$. Motivated by these results, we apply DeepMoD directly on noisy experimental time-series data from a gel electrophoresis experiment and find that it discovers the advection-diffusion equation describing this system. | ['Pierre Sens', 'Gert-Jan Both', 'Subham Choudhury', 'Remy Kusters'] | 2019-04-20 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [-2.05935374e-01 -2.70493656e-01 2.23665565e-01 1.03888707e-03
-4.61093873e-01 -3.17645222e-01 4.10814404e-01 2.74314731e-01
-5.69064260e-01 1.00303805e+00 -3.98714781e-01 -3.06383789e-01
-4.14780766e-01 -6.91969633e-01 -7.08490968e-01 -1.02618730e+00
-7.71135211e-01 6.82174087e-01 1.34583026e-01 -2.19970316e-01
2.42806897e-01 7.61837006e-01 -1.30784047e+00 4.20831004e-03
5.58787823e-01 1.24584496e+00 3.93194631e-02 9.77630317e-01
4.70005022e-03 9.76349294e-01 -5.78970790e-01 6.72009736e-02
2.02475190e-01 -4.51373756e-01 -7.59662390e-01 -1.36637866e-01
-5.33238612e-02 -2.42192566e-01 -6.40204668e-01 7.47540057e-01
4.97264057e-01 4.44849372e-01 8.86096895e-01 -9.29093659e-01
-6.03538156e-01 6.95846304e-02 -3.85272592e-01 6.78709745e-01
-8.45816880e-02 2.74211079e-01 7.01894879e-01 -9.39295709e-01
8.28034222e-01 9.15679634e-01 1.02372169e+00 3.04248333e-01
-1.68572962e+00 -1.55469820e-01 -2.09170997e-01 -2.66729355e-01
-1.35383976e+00 -3.19997400e-01 5.42967141e-01 -4.99974966e-01
1.27342343e+00 8.03120658e-02 6.32935941e-01 1.00172722e+00
7.40876675e-01 6.85889423e-01 1.13006282e+00 -1.65490910e-01
6.38848722e-01 -2.31454998e-01 1.74826309e-02 8.20480108e-01
6.71327189e-02 5.79017699e-02 -1.89488485e-01 -4.29640025e-01
1.09455192e+00 5.23954220e-02 -3.54710817e-01 -2.24258795e-01
-9.35531318e-01 8.60284984e-01 2.12849990e-01 5.45428514e-01
-5.54435670e-01 3.65374774e-01 3.08054388e-01 6.23794138e-01
4.95592475e-01 8.28129411e-01 -8.11961651e-01 -2.63429731e-01
-9.28079009e-01 6.45388067e-01 1.04995644e+00 6.33214295e-01
7.84713924e-01 2.51814902e-01 4.92292315e-01 6.79437578e-01
-3.71438488e-02 3.65261555e-01 3.96811664e-01 -1.28409255e+00
-1.22013472e-01 3.44042212e-01 3.87730420e-01 -1.04212499e+00
-8.30948949e-01 -5.27003050e-01 -1.21445870e+00 2.25696713e-01
5.41986883e-01 -6.22107208e-01 -8.46646845e-01 1.30195117e+00
1.11663997e-01 3.04143846e-01 -3.81440041e-03 8.84399354e-01
6.20344579e-01 9.32202578e-01 -3.56400311e-01 -6.76097453e-01
6.55053139e-01 -4.05908972e-01 -6.89543128e-01 8.22672546e-02
8.10824811e-01 -5.66754103e-01 6.61240458e-01 6.01540387e-01
-1.31133020e+00 -2.97977865e-01 -8.48186910e-01 -3.97646092e-02
-3.36717129e-01 -3.21730405e-01 9.41662848e-01 -1.93100169e-01
-1.16247857e+00 1.25787139e+00 -1.10750496e+00 -1.58563539e-01
3.87114614e-01 6.28891706e-01 -2.96042114e-01 1.06430128e-01
-7.66091168e-01 6.60682201e-01 -1.29298806e-01 7.04789087e-02
-9.68112886e-01 -8.64782989e-01 -7.47398674e-01 -8.14778358e-03
4.58342992e-02 -5.20792782e-01 1.16434109e+00 -5.48629820e-01
-1.53650868e+00 6.51647270e-01 -3.49141568e-01 -6.82134569e-01
2.99040109e-01 1.22430354e-01 -5.06685436e-01 1.44409299e-01
-1.08237073e-01 2.03211144e-01 7.57265568e-01 -1.16117013e+00
-2.15964496e-01 -1.46501362e-01 -1.40373170e-01 -5.59670269e-01
8.53310898e-02 -8.11496973e-02 1.56309605e-02 -6.50268734e-01
4.44310278e-01 -8.55758965e-01 -6.04996204e-01 -3.94652132e-03
-1.54556766e-01 -8.62686038e-02 8.27343762e-01 -4.93697613e-01
8.95938396e-01 -1.97268558e+00 3.39594811e-01 3.52003753e-01
4.66378897e-01 1.97764322e-01 -7.22785816e-02 7.81268537e-01
-8.26844424e-02 8.74819830e-02 -6.55613482e-01 -3.04196179e-01
-2.16613322e-01 3.21481228e-01 -1.58725426e-01 9.14376140e-01
3.04389924e-01 6.77133858e-01 -7.93489814e-01 -3.24268043e-02
-7.13950545e-02 5.73762834e-01 -3.81802857e-01 1.16182320e-01
-4.25649226e-01 6.27011478e-01 -2.14895189e-01 5.40063977e-01
7.40133464e-01 -6.97920024e-01 5.99059723e-02 1.16326302e-01
-2.00386479e-01 6.83857501e-02 -1.12867737e+00 1.54677379e+00
-3.38551193e-01 8.97893190e-01 3.61318558e-01 -1.62333429e+00
1.16469514e+00 2.64566898e-01 1.04181588e+00 -7.23537207e-01
3.39678705e-01 4.95600939e-01 9.34509188e-02 -6.34454012e-01
6.16571084e-02 -3.60424787e-01 1.79777533e-01 5.24473608e-01
2.52426922e-01 -2.02209085e-01 2.12128758e-01 5.10214791e-02
1.54425192e+00 -1.05653837e-01 -2.59157289e-02 -8.24043989e-01
3.68389726e-01 9.77212042e-02 6.46120489e-01 7.60558009e-01
-1.54011071e-01 3.60845536e-01 8.05019140e-01 -1.12331378e+00
-1.18785512e+00 -8.03207099e-01 -3.49853277e-01 7.26772726e-01
8.44410658e-02 -2.39721686e-01 -5.40027738e-01 1.53836058e-02
2.16140389e-01 2.41595984e-01 -6.86916113e-01 -4.74602804e-02
-7.86130548e-01 -9.98941004e-01 4.78522509e-01 3.73314053e-01
3.94237131e-01 -1.11148763e+00 -3.74710381e-01 5.60194612e-01
3.73573273e-01 -8.87052238e-01 1.64538369e-01 5.69462538e-01
-9.40714598e-01 -1.11750829e+00 -6.78579390e-01 -9.26619470e-01
5.67949772e-01 -3.93668979e-01 1.08462441e+00 1.16652764e-01
-6.88144922e-01 1.67231828e-01 5.65476045e-02 -1.06193438e-01
-4.71410483e-01 -1.35096699e-01 3.36035758e-01 -1.22642286e-01
3.92428219e-01 -9.68078852e-01 -5.46395004e-01 8.96872878e-02
-7.10019767e-01 -5.11766255e-01 1.03139117e-01 1.01241684e+00
7.71447420e-01 5.31291246e-01 5.89563370e-01 -4.16927338e-01
6.92494690e-01 -6.57867968e-01 -1.01185393e+00 -3.70688856e-01
-4.16492879e-01 2.00138986e-01 9.92577374e-01 -5.72031796e-01
-6.11585855e-01 -1.59359369e-02 -5.05834073e-02 -4.63240951e-01
-1.69295996e-01 5.93113184e-01 5.60290754e-01 -1.94431156e-01
7.49155223e-01 4.66214895e-01 2.00254366e-01 -7.00809717e-01
-1.80323973e-01 2.29286656e-01 5.55759251e-01 -6.15037739e-01
5.76773763e-01 6.77255213e-01 4.15084064e-01 -1.25724065e+00
-3.15096945e-01 -1.28428847e-01 -5.23173988e-01 1.95264056e-01
5.04982054e-01 -5.61034322e-01 -1.32214010e+00 4.56488073e-01
-1.19003701e+00 -7.25769997e-01 -3.10232073e-01 5.58398604e-01
-7.66381562e-01 5.34878559e-02 -1.20751858e+00 -9.98615801e-01
-9.47827995e-02 -9.05912876e-01 8.14583659e-01 -4.43381667e-02
-2.60186493e-01 -1.26880705e+00 2.15512604e-01 -3.52307469e-01
6.35054946e-01 6.39065564e-01 9.19710040e-01 -3.41410786e-01
-5.98810494e-01 -1.88679740e-01 -6.84664920e-02 2.65926659e-01
1.48486653e-02 1.58151612e-01 -7.22049177e-01 -4.12952721e-01
3.95437211e-01 -1.42104134e-01 1.06323242e+00 7.58942068e-01
1.25828111e+00 -3.35356802e-01 -2.71112263e-01 8.54658008e-01
1.41434324e+00 3.19371670e-01 4.30646092e-01 -1.39984731e-02
3.80130976e-01 2.82302111e-01 1.20161185e-02 5.58362365e-01
-5.72152510e-02 2.16292635e-01 2.36589700e-01 -2.38820970e-01
4.66391653e-01 1.49326846e-01 1.63904339e-01 8.07932734e-01
-1.99114904e-01 -1.96298361e-01 -1.00627720e+00 5.06112337e-01
-1.84461379e+00 -9.40435469e-01 -2.02920556e-01 1.88737988e+00
6.76307201e-01 7.25441985e-03 1.24984831e-01 2.61708409e-01
2.73985624e-01 6.79202676e-02 -8.38743567e-01 -5.73525786e-01
-3.18070769e-01 6.93862379e-01 4.35328573e-01 6.27224743e-01
-1.08244586e+00 8.07819605e-01 7.33408642e+00 3.18274021e-01
-1.46707046e+00 -1.04874998e-01 6.57686710e-01 -1.62055269e-01
-1.99097231e-01 -2.62290925e-01 -3.12661588e-01 4.17115152e-01
1.18005931e+00 -2.05830202e-01 7.73253560e-01 5.34573436e-01
2.97810882e-01 -2.15723544e-01 -1.09598339e+00 1.13693583e+00
-4.78378445e-01 -1.78637218e+00 -6.05083466e-01 2.26105869e-01
7.18105853e-01 3.25493723e-01 -3.13559398e-02 1.19086206e-01
4.91194814e-01 -1.39542902e+00 1.24457434e-01 6.29777014e-01
2.96001583e-01 -7.23524153e-01 7.05914259e-01 5.98938465e-01
-8.98232222e-01 -2.36485735e-01 -5.03398120e-01 -4.23779249e-01
8.48563090e-02 9.49550211e-01 -5.80476522e-01 -1.59743205e-02
7.36037791e-01 8.00498724e-01 1.09034538e-01 8.27800512e-01
3.74728441e-01 5.47072768e-01 -5.81636548e-01 -8.63309577e-02
2.90921748e-01 -3.08431536e-01 3.86550188e-01 1.15131330e+00
4.83026803e-01 4.91982818e-01 9.62041467e-02 1.13878262e+00
-2.15702444e-01 -9.30061340e-02 -7.63483703e-01 -4.65461463e-01
1.59233987e-01 8.41496468e-01 -8.63962293e-01 -2.48325154e-01
-2.90520359e-02 9.53213573e-01 2.56195575e-01 8.94665182e-01
-4.43584859e-01 -4.90504593e-01 9.56430495e-01 2.21390665e-01
6.19457304e-01 -7.77509749e-01 -9.32744443e-02 -1.09562683e+00
-5.00831008e-03 -7.13252187e-01 1.06844865e-01 -5.84304869e-01
-1.32842052e+00 3.65151167e-01 -3.40228111e-01 -7.18690217e-01
-2.57478088e-01 -8.41780841e-01 -3.82152468e-01 9.03329432e-01
-1.14490104e+00 -2.90653110e-01 1.28429368e-01 5.29221296e-01
1.42638698e-01 -2.07817778e-01 9.83238041e-01 1.95579335e-01
-5.39709449e-01 1.93374977e-02 6.33727431e-01 1.06981158e-01
6.93139359e-02 -1.16456413e+00 5.06721199e-01 4.29106414e-01
-1.75916880e-01 6.58111870e-01 9.34736907e-01 -4.26518291e-01
-1.85694361e+00 -7.46634662e-01 6.85567439e-01 -3.44240308e-01
8.56080055e-01 -4.72108305e-01 -1.23001289e+00 5.20271182e-01
-2.87728552e-02 6.59899712e-01 4.87302750e-01 -1.27733186e-01
3.30157094e-02 -9.20003876e-02 -1.28262448e+00 2.60932684e-01
1.02098691e+00 -5.60685039e-01 -1.14407137e-01 5.59421360e-01
5.51426411e-01 -4.70419079e-01 -1.24053562e+00 2.20309123e-01
3.70850265e-01 -1.00800037e+00 8.56980324e-01 -1.02361345e+00
4.93814886e-01 -1.50426300e-02 -1.14282511e-01 -1.36073017e+00
-4.19695765e-01 -1.11832511e+00 -5.29486537e-01 4.39261883e-01
3.72047216e-01 -7.35492766e-01 6.29081726e-01 4.75324035e-01
-1.04246199e-01 -1.08080769e+00 -1.05805242e+00 -9.57641423e-01
4.45529044e-01 -3.39840353e-01 5.34047782e-01 1.14803731e+00
-4.29769680e-02 -4.21059132e-02 -2.90348053e-01 2.58517683e-01
5.05026281e-01 1.37326077e-01 6.02917552e-01 -1.36381209e+00
-4.99539584e-01 -3.56022358e-01 -3.56958359e-01 -1.08037686e+00
3.37599546e-01 -5.01708984e-01 1.74280237e-02 -1.26662993e+00
-4.72975761e-01 -5.61588287e-01 -3.53819013e-01 1.77844614e-01
4.94060516e-01 -1.21442676e-02 -3.66644114e-01 2.03114316e-01
-3.71651798e-01 4.91167456e-01 1.17841363e+00 -1.55531704e-01
-3.85861486e-01 -1.19210772e-01 -3.85442853e-01 7.80193388e-01
7.21965075e-01 -4.74476516e-01 -1.40504688e-01 -3.06726545e-01
2.60638356e-01 2.95618206e-01 3.81794602e-01 -9.08309758e-01
1.94983304e-01 -3.44513029e-01 4.94326532e-01 -2.63373852e-01
4.21463549e-01 -5.69359183e-01 -3.78041598e-03 4.92796540e-01
-2.43995115e-01 1.26227498e-01 3.78944546e-01 4.05140787e-01
-7.89280012e-02 8.84690210e-02 9.77995038e-01 -2.64408559e-01
-4.28998291e-01 4.42349285e-01 -7.67116606e-01 8.77301842e-02
7.11529315e-01 1.08820565e-01 -2.11309001e-01 -4.13387150e-01
-9.16726947e-01 -1.53006837e-02 4.59798157e-01 -9.95630324e-02
4.89968956e-01 -1.09467101e+00 -6.53421462e-01 5.38904667e-01
-4.77915734e-01 2.91306674e-01 1.50666803e-01 1.02466083e+00
-8.28692853e-01 2.50714064e-01 2.21989956e-02 -7.31360614e-01
-5.05757570e-01 5.52065670e-01 7.06364214e-01 -7.54116923e-02
-8.72568488e-01 1.05241895e+00 -1.73695132e-01 -2.67698318e-01
4.31549065e-02 -6.72110260e-01 3.21987987e-01 -3.06043297e-01
4.99360442e-01 4.91870254e-01 -2.75179837e-02 -4.23079133e-01
-3.80306095e-01 3.78680915e-01 1.00309178e-01 8.39420259e-02
1.78748775e+00 3.10469329e-01 -4.87665772e-01 6.02140307e-01
1.74565125e+00 -4.17624027e-01 -1.42069435e+00 -2.97684371e-01
7.87568688e-02 -1.47934660e-01 1.53674498e-01 -3.17168027e-01
-9.29113865e-01 9.74162221e-01 4.27330673e-01 7.52666712e-01
8.48869503e-01 -9.30114388e-02 1.01110697e+00 8.83943200e-01
1.41435266e-01 -1.17380309e+00 4.55113873e-02 9.54761982e-01
7.78598249e-01 -1.23653162e+00 -2.22507492e-02 5.29152416e-02
3.79254371e-02 1.27918708e+00 2.66123205e-01 -5.44641554e-01
1.02954018e+00 6.92312360e-01 -9.13005546e-02 -4.36571628e-01
-8.82377207e-01 1.14149660e-01 -1.13915406e-01 5.26997447e-01
5.10495543e-01 -2.28588432e-01 3.13856080e-02 2.65393257e-01
-1.50472634e-02 2.86255032e-01 5.24215937e-01 1.07074511e+00
-4.26451474e-01 -7.48113155e-01 -1.32530034e-01 5.19708097e-01
-3.18287045e-01 1.18322626e-01 -1.49154559e-01 7.77157784e-01
1.19905481e-02 5.79281271e-01 3.50997180e-01 -1.28928989e-01
8.26307014e-02 5.19692563e-02 2.86448151e-01 -2.56291538e-01
-1.58498913e-01 1.46654233e-01 -1.08938977e-01 -5.99285841e-01
-3.28297794e-01 -5.72847962e-01 -1.64683950e+00 -7.71374524e-01
1.76964208e-01 3.21501493e-01 5.30709922e-01 1.18217456e+00
5.88009596e-01 3.88284326e-01 5.79328716e-01 -1.05992794e+00
-3.40459555e-01 -7.43422508e-01 -9.52277660e-01 2.14761734e-01
1.01710963e+00 -6.40932679e-01 -6.54946208e-01 -1.65398866e-02] | [6.605240345001221, 3.53838849067688] |
9d400093-ed25-48e6-a3c4-56bad21f21df | multi-modal-facial-action-unit-detection-with | 2303.10590 | null | https://arxiv.org/abs/2303.10590v3 | https://arxiv.org/pdf/2303.10590v3.pdf | Multi-modal Facial Action Unit Detection with Large Pre-trained Models for the 5th Competition on Affective Behavior Analysis in-the-wild | Facial action unit detection has emerged as an important task within facial expression analysis, aimed at detecting specific pre-defined, objective facial expressions, such as lip tightening and cheek raising. This paper presents our submission to the Affective Behavior Analysis in-the-wild (ABAW) 2023 Competition for AU detection. We propose a multi-modal method for facial action unit detection with visual, acoustic, and lexical features extracted from the large pre-trained models. To provide high-quality details for visual feature extraction, we apply super-resolution and face alignment to the training data and show potential performance gain. Our approach achieves the F1 score of 52.3% on the official validation set of the 5th ABAW Challenge. | ['Mohammad Soleymani', 'Xinrui Wang', 'Di Chang', 'Minh Tran', 'Yufeng Yin'] | 2023-03-19 | null | null | null | null | ['face-alignment', 'action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.95204073e-01 6.20080009e-02 -1.98549509e-01 -6.82189822e-01
-1.23357046e+00 -4.60339099e-01 4.87316281e-01 -2.24964395e-01
-5.27236700e-01 2.28592277e-01 6.27240062e-01 8.08692932e-01
4.98990178e-01 4.52565849e-02 -2.93993056e-01 -7.15344727e-01
-2.89532691e-01 -4.96077165e-02 -2.16685936e-01 -3.35251898e-01
-2.30026558e-01 6.23637140e-01 -1.78225148e+00 8.13133419e-01
-1.44014463e-01 1.49466312e+00 -7.58112788e-01 7.16507137e-01
4.75388974e-01 5.33142209e-01 -4.33737725e-01 -5.71653187e-01
-5.46183847e-02 -5.88098347e-01 -8.46533418e-01 1.52517468e-01
7.76526093e-01 -4.89141405e-01 2.83908136e-02 8.77609074e-01
7.19515800e-01 -1.68265142e-02 7.76217282e-01 -1.53925133e+00
-4.38125730e-02 1.19409589e-02 -9.43282902e-01 1.38262704e-01
6.35124028e-01 9.13916007e-02 1.06631970e+00 -1.20586967e+00
8.42543006e-01 1.29688323e+00 5.96911907e-01 1.24956310e+00
-1.20551646e+00 -8.25408340e-01 -1.83375716e-01 2.44148403e-01
-1.55080914e+00 -1.22670126e+00 5.70442021e-01 -3.80643904e-01
1.24436927e+00 1.79171279e-01 5.56124747e-01 1.28557253e+00
-3.20730627e-01 9.53103781e-01 9.24107909e-01 -3.80525798e-01
-7.86176473e-02 -9.92815346e-02 -2.45114282e-01 7.05363333e-01
-5.90630889e-01 8.67510121e-03 -1.01337099e+00 -4.70957667e-01
1.99118540e-01 -5.90851843e-01 -6.76374659e-02 2.23021299e-01
-4.14334685e-01 8.03406715e-01 -2.91881692e-02 3.28624666e-01
-5.81993759e-01 2.10767180e-01 7.82460213e-01 2.06621177e-02
7.62119412e-01 4.70272824e-03 -3.26126963e-01 -7.35695183e-01
-8.08197200e-01 2.26562738e-01 2.02378586e-01 5.03855348e-01
4.64174360e-01 -1.10404894e-01 -3.85920852e-01 1.16031301e+00
4.19866055e-01 5.75485289e-01 3.37278068e-01 -1.08187699e+00
-9.90112796e-02 5.08332610e-01 -3.76739427e-02 -7.70764112e-01
-6.45659268e-01 5.83023787e-01 -4.17741746e-01 4.06527400e-01
2.39993095e-01 -3.29093307e-01 -6.60304368e-01 2.01125121e+00
6.07471526e-01 -1.92503515e-03 -2.56356876e-02 7.94798851e-01
1.07883334e+00 4.52409595e-01 4.22119617e-01 -4.94614571e-01
1.60429263e+00 -5.70623159e-01 -8.10083747e-01 -1.13351807e-01
8.24023068e-01 -7.83553600e-01 5.90481818e-01 3.82161230e-01
-1.10950303e+00 -3.21078509e-01 -6.45455003e-01 4.94325608e-02
-7.83804208e-02 3.80677551e-01 4.08591509e-01 6.83866441e-01
-1.03766513e+00 1.34820014e-01 -6.59843385e-01 -4.68908876e-01
8.10888112e-01 5.50940871e-01 -1.08323276e+00 3.67524296e-01
-9.44556236e-01 8.62865448e-01 -1.47037223e-01 1.21277452e-01
-5.83852410e-01 -3.96518648e-01 -1.03115618e+00 -2.89490849e-01
8.49065036e-02 2.12239206e-01 1.50604367e+00 -1.52790999e+00
-1.70202219e+00 1.62290156e+00 -5.83063185e-01 -1.02377340e-01
1.37650177e-01 -3.36721927e-01 -5.35906613e-01 4.06102717e-01
-2.59056956e-01 8.68720710e-01 1.12186372e+00 -7.05915928e-01
-5.30058682e-01 -6.98988557e-01 -3.87635589e-01 1.19985931e-01
-4.54595864e-01 1.03210521e+00 -3.47289979e-01 -2.84451574e-01
-3.30374748e-01 -8.43098283e-01 3.83789212e-01 1.53015032e-01
1.96394354e-01 -8.06183875e-01 7.56155908e-01 -8.56493711e-01
1.17669702e+00 -2.57595396e+00 1.45637840e-01 2.68075168e-01
2.65255272e-01 3.83475542e-01 -4.11772549e-01 -2.21481938e-02
-2.25059271e-01 -2.35679582e-01 2.87107408e-01 -8.59164596e-01
9.81857777e-02 -1.26180977e-01 1.32914305e-01 7.18346894e-01
4.72188592e-01 1.04099584e+00 -6.39022648e-01 -7.07945764e-01
-5.63269258e-02 5.09696007e-01 -5.37104607e-01 3.16535830e-01
1.41441450e-01 -8.06873962e-02 -5.69815114e-02 1.11280930e+00
4.60898489e-01 4.21140254e-01 4.80885021e-02 -4.10605669e-01
2.95873553e-01 -1.83214813e-01 -7.96649694e-01 1.44127047e+00
-2.01723680e-01 8.86300981e-01 5.60872436e-01 -6.42302513e-01
7.93512642e-01 7.12055683e-01 7.91829884e-01 -8.41943920e-01
6.94783568e-01 1.66973583e-02 -1.31856099e-01 -5.76634347e-01
1.11036479e-01 -3.10653836e-01 -1.25532001e-01 2.28333011e-01
5.37693560e-01 1.44167617e-01 3.65816727e-02 -3.83442827e-02
9.62292850e-01 2.09036797e-01 6.33897960e-01 -2.98440456e-02
6.13295496e-01 -5.59295893e-01 5.17288566e-01 -7.14665949e-02
-7.54646063e-01 5.21709502e-01 6.80909038e-01 -2.97937036e-01
-5.38539171e-01 -7.32221484e-01 -5.85942380e-02 1.53495586e+00
-6.33615971e-01 -8.52724433e-01 -1.02112150e+00 -7.58776486e-01
-1.15358122e-02 1.08870573e-01 -1.08281553e+00 -3.32947075e-01
-3.45282227e-01 -5.37675619e-01 1.00499976e+00 5.83016992e-01
1.97599933e-01 -1.41562879e+00 -4.90372628e-01 -1.85922667e-01
-3.55748743e-01 -1.43918407e+00 -5.93118429e-01 -3.31776470e-01
-6.09172545e-02 -1.07426822e+00 -5.33597291e-01 -7.17022657e-01
4.25777346e-01 -4.31443453e-01 7.69349813e-01 -2.10403800e-01
-7.72205830e-01 6.57599449e-01 -3.71666819e-01 -5.73659301e-01
-4.64072138e-01 -5.42021573e-01 3.83627236e-01 7.85865247e-01
9.34840858e-01 -1.65581837e-01 -3.48520428e-01 2.89853185e-01
-4.28028762e-01 -3.43271405e-01 2.14823022e-01 6.31846428e-01
5.10968566e-01 -8.42247605e-01 4.96402800e-01 -2.04262942e-01
4.85518754e-01 1.32182648e-03 -3.95882547e-01 1.56170428e-01
-1.41629577e-01 -4.44009304e-01 -1.43286008e-02 -5.30369043e-01
-8.29452693e-01 5.09864092e-01 -5.37933290e-01 -6.24252737e-01
-3.45540524e-01 -4.36798185e-02 -3.60336632e-01 -3.07657301e-01
6.67917848e-01 -3.26373987e-02 3.34769279e-01 -3.57051939e-02
2.24319696e-01 9.56073880e-01 5.08211493e-01 -2.49260306e-01
3.08216512e-01 5.19914210e-01 1.17878735e-01 -1.05382347e+00
-7.21919060e-01 -7.63835311e-01 -7.32716441e-01 -7.42185891e-01
1.11863828e+00 -9.26319122e-01 -1.19179118e+00 7.61821330e-01
-1.20245504e+00 -2.52506405e-01 4.45126779e-02 1.81477502e-01
-7.15112090e-01 2.15164423e-01 -5.08615315e-01 -1.16349030e+00
-6.94186985e-01 -8.91266704e-01 1.66967320e+00 4.50868979e-02
-1.01842499e+00 -5.21708488e-01 4.88501430e-01 4.25520092e-01
8.84298906e-02 5.54548681e-01 2.60203540e-01 -4.83423591e-01
5.24790108e-01 -4.89856929e-01 -7.84997568e-02 7.30332255e-01
2.46771574e-01 3.41686964e-01 -1.47834289e+00 -1.53346397e-02
-4.63317961e-01 -1.08235061e+00 7.28524566e-01 1.47640452e-01
9.52860653e-01 -1.84982389e-01 -1.50090918e-01 4.54150319e-01
7.58293569e-01 -2.51614042e-02 6.88528597e-01 -3.38294208e-01
3.01508963e-01 7.94787049e-01 8.44443679e-01 7.76612461e-01
-2.19115257e-01 1.16031015e+00 3.74766380e-01 -6.97768480e-02
-2.79302187e-02 1.51446864e-01 7.25907326e-01 -8.93657655e-02
-5.37164032e-01 3.06821525e-01 -6.89653993e-01 4.50910509e-01
-1.72465730e+00 -1.29371870e+00 2.42988542e-01 1.86671019e+00
9.25521135e-01 -2.79604048e-01 7.33647346e-01 -1.11999564e-01
3.93341511e-01 2.14666739e-01 -4.45658341e-02 -8.47096562e-01
-1.94761395e-01 5.56896925e-01 -6.72160387e-02 5.73489130e-01
-1.43362188e+00 1.12994838e+00 6.24590158e+00 9.21816528e-01
-1.22163391e+00 1.90433264e-01 6.24218702e-01 -6.16208076e-01
4.14151669e-01 -8.43182921e-01 -1.07282734e+00 1.08259819e-01
1.05768418e+00 -5.01058474e-02 1.78171590e-01 9.38709378e-01
2.93436587e-01 -2.10780129e-01 -1.12455642e+00 1.60009980e+00
6.48081720e-01 -1.00815296e+00 -2.90260315e-01 1.60465807e-01
4.67646211e-01 7.50403255e-02 5.58703393e-02 2.01428801e-01
-2.48833895e-01 -1.28641701e+00 4.10549819e-01 2.74533957e-01
1.16713786e+00 -9.49553430e-01 6.12113297e-01 -2.06983149e-01
-1.19936085e+00 8.41597766e-02 8.48302245e-02 2.22438321e-01
-1.64056439e-02 -2.64750183e-01 -7.61958182e-01 -1.65234894e-01
7.91970789e-01 6.94018722e-01 -3.48011315e-01 4.79197204e-01
-7.20181987e-02 7.20318019e-01 -6.01810932e-01 -1.58138350e-01
8.40348378e-03 2.21265584e-01 4.35690701e-01 1.43453097e+00
-7.31806606e-02 3.09692085e-01 -2.28043482e-01 4.32536155e-01
-2.24668220e-01 4.55075741e-01 -6.39062405e-01 -2.42768884e-01
-2.06177115e-01 1.79590607e+00 -1.01153798e-01 4.85153645e-02
-5.73940456e-01 1.26213121e+00 3.08760107e-01 3.73514928e-02
-8.67450118e-01 -1.63422134e-02 1.32459736e+00 -1.16293132e-01
-1.60350855e-02 2.53189266e-01 3.70326430e-01 -8.40258777e-01
7.45843397e-03 -9.98463929e-01 6.12512887e-01 -5.72369218e-01
-1.01529765e+00 6.91129804e-01 -6.39946759e-02 -9.59765673e-01
-6.58682227e-01 -9.96556640e-01 -5.47595143e-01 7.63113678e-01
-1.08968115e+00 -1.49215055e+00 -3.20512682e-01 8.14019382e-01
4.14582372e-01 -1.58259690e-01 1.32083857e+00 2.62650013e-01
-4.90183234e-01 1.20812201e+00 -3.49392802e-01 5.67261338e-01
9.07143235e-01 -6.88575566e-01 -2.96855103e-02 5.60257137e-01
3.14180166e-01 -3.39603238e-02 6.15153670e-01 -1.31424755e-01
-1.09685779e+00 -9.02898133e-01 9.95189965e-01 -6.70755029e-01
6.58270895e-01 -7.63627648e-01 -6.08853936e-01 5.73094845e-01
-3.37276831e-02 2.64911085e-01 1.02422416e+00 2.86942452e-01
-5.96364141e-01 -1.32483929e-01 -1.15102398e+00 4.01823789e-01
8.34322751e-01 -8.22606385e-01 -2.57723212e-01 1.68671787e-01
-1.89739078e-01 -2.29179382e-01 -8.66326392e-01 6.08614087e-01
1.07346404e+00 -8.44406903e-01 6.88942432e-01 -1.02965462e+00
4.01071250e-01 3.11351180e-01 -2.80503124e-01 -9.83764231e-01
-1.14237733e-01 -1.03686023e+00 -2.21672952e-02 1.41828299e+00
3.05580467e-01 -1.14842571e-01 8.79354239e-01 6.18533015e-01
3.89772326e-01 -8.23245049e-01 -1.32392108e+00 -4.14036453e-01
-2.42207006e-01 -4.55733210e-01 -3.44147496e-02 5.53791165e-01
6.45696759e-01 3.86640549e-01 -4.92221832e-01 -1.67950243e-01
3.87625039e-01 -6.35773465e-02 8.16871047e-01 -9.70439196e-01
-8.34676698e-02 -6.32373512e-01 -9.46583509e-01 -3.29345405e-01
6.59493923e-01 -5.47445416e-01 1.07306905e-01 -8.22753072e-01
3.16607773e-01 6.50553405e-01 -2.99561948e-01 9.85382676e-01
3.49823944e-02 9.00827885e-01 2.84079351e-02 -4.63427991e-01
-7.47175276e-01 6.29220009e-01 6.62230194e-01 -7.74337053e-02
3.79060470e-02 4.38046604e-02 -3.15330118e-01 9.18453634e-01
7.28054225e-01 -1.29515246e-01 8.22369531e-02 2.22941965e-01
5.22644958e-03 -1.67458519e-01 3.06520522e-01 -5.63291907e-01
-3.32923532e-01 -1.19733498e-01 3.57340276e-01 -2.39822850e-01
9.87783074e-01 -5.01178682e-01 -6.34921014e-01 2.47225657e-01
-2.45757729e-01 -1.60217389e-01 6.92832351e-01 -4.45482545e-02
-3.10077190e-01 1.93126917e-01 1.30961180e+00 3.08098853e-01
-9.80805457e-01 3.40191752e-01 -5.22027791e-01 6.47446364e-02
1.27078760e+00 -1.20602220e-01 5.06397374e-02 -6.18450940e-01
-1.05415916e+00 1.18248582e-01 5.20566991e-03 6.83612943e-01
7.88801789e-01 -1.34220064e+00 -1.07894433e+00 3.66892964e-01
6.60329580e-01 -8.29604685e-01 3.30405354e-01 1.06784892e+00
6.19522370e-02 4.16444018e-02 -4.81799453e-01 -5.12984276e-01
-2.57413483e+00 1.67295888e-01 6.73644066e-01 1.91530153e-01
-3.75178829e-02 1.35421145e+00 9.50726569e-02 7.18539208e-02
2.14002430e-01 2.47385308e-01 -3.95345390e-01 5.75671375e-01
9.89503384e-01 2.03246862e-01 7.19783977e-02 -1.35191023e+00
-7.41196215e-01 6.47646248e-01 1.74607038e-01 -1.85026482e-01
1.10779834e+00 8.17646384e-02 -2.57404417e-01 3.36390913e-01
1.67222464e+00 3.77545319e-02 -8.72658193e-01 -1.42089888e-01
-8.97500739e-02 -2.55897224e-01 -1.90290483e-03 -6.23668790e-01
-8.57947648e-01 9.09722805e-01 8.64442408e-01 -2.59982407e-01
1.29555249e+00 5.00461400e-01 4.45399255e-01 1.59764051e-01
7.69520849e-02 -1.40833855e+00 1.80590734e-01 2.82235324e-01
1.22779393e+00 -1.44651222e+00 -2.03908965e-01 -3.17121774e-01
-7.97219098e-01 1.13242602e+00 8.66965115e-01 3.79737839e-02
7.54298329e-01 5.15557349e-01 4.00586396e-01 -2.65654266e-01
-8.95334005e-01 -5.99663317e-01 6.95337892e-01 6.00610077e-01
5.35300910e-01 -9.30677876e-02 1.05817467e-02 6.37436628e-01
7.77071640e-02 -5.43337837e-02 -1.38531744e-01 5.49133182e-01
-2.98772871e-01 -8.21672916e-01 -1.76449493e-02 2.30667353e-01
-7.91582704e-01 1.36158288e-01 -1.01671219e+00 5.79386890e-01
2.31886040e-02 7.66121149e-01 7.11126029e-02 -3.67930889e-01
4.51105535e-01 6.45018101e-01 8.44340146e-01 -4.05993909e-01
-5.85103214e-01 3.64380538e-01 4.54164982e-01 -1.25417864e+00
-6.51551843e-01 -9.91873145e-01 -1.24773550e+00 5.17065711e-02
7.43170008e-02 -1.67397857e-01 4.37695205e-01 9.41795230e-01
3.96303236e-01 -1.72575280e-01 5.74045777e-01 -9.91651595e-01
-3.09368223e-01 -1.26921690e+00 -6.43990755e-01 7.74648726e-01
5.23520112e-01 -6.71824813e-01 -3.11622798e-01 1.21515818e-01] | [13.55714225769043, 1.9141325950622559] |
33381ce4-1fd7-46a2-931f-983ec7bdeb33 | reconstruct-from-top-view-a-3d-lane-detection-1 | null | null | https://openaccess.thecvf.com/content/CVPR2022W/WAD/papers/Li_Reconstruct_From_Top_View_A_3D_Lane_Detection_Approach_Based_CVPRW_2022_paper.pdf | https://openaccess.thecvf.com/content/CVPR2022W/WAD/papers/Li_Reconstruct_From_Top_View_A_3D_Lane_Detection_Approach_Based_CVPRW_2022_paper.pdf | Reconstruct from Top View: A 3D Lane Detection Approach based on Geometry Structure Prior | In this paper, we propose an advanced approach in targeting the problem of monocular 3D lane detection by leveraging geometry structure underneath the process of 2D to 3D lane reconstruction. Inspired by previous methods, we first analyze the geometry heuristic between the 3D lane and its 2D representation on the ground and propose to impose explicit supervision based on the structure prior, which makes it achievable to build inter-lane and intra-lane relationships to facilitate the reconstruction of 3D lanes from local to global. Second, to reduce the structure loss in 2D lane representation, we directly extract top view lane information from front view images, which tremendously eases the confusion of distant lane features in previous methods. Furthermore, we propose a novel task-specific data augmentation method by synthesizing new training data for both segmentation and reconstruction tasks in our pipeline, to counter the imbalanced data distribution of camera pose and ground slope to improve generalization on unseen data. Our work marks the first attempt to employ the geometry
prior information into DNN-based 3D lane detection and makes it achievable for detecting lanes in an extra-long distance, doubling the original detection range. The proposed method can be smoothly adopted by other frameworks without extra costs. Experimental results show that our work outperforms state-of-the-art approaches by 3.8% F-Score on Apollo 3D synthetic dataset at real-time speed of 82 FPS without introducing extra parameters. | ['Guangliang Cheng', 'Ya Wang', 'Jia Shi', 'Chenguang Li'] | 2022-06-20 | null | null | null | cvpr-2022-6 | ['3d-lane-detection', 'lane-detection'] | ['computer-vision', 'computer-vision'] | [ 9.21326503e-02 1.70944512e-01 -2.79227830e-03 -4.01555628e-01
-5.48587143e-01 -7.50233769e-01 4.12652522e-01 -2.13164508e-01
-3.63577873e-01 1.80942848e-01 -2.51968652e-01 -6.49295986e-01
4.14382458e-01 -6.69835031e-01 -8.83430362e-01 -4.30583060e-01
2.70385534e-01 3.76344353e-01 9.45818782e-01 -3.00092906e-01
4.48358208e-01 9.61912513e-01 -1.74138641e+00 -6.41328469e-02
9.82943833e-01 7.97776341e-01 4.03988093e-01 4.53657687e-01
-1.91203132e-01 4.82623070e-01 -2.40913123e-01 -2.77092814e-01
7.25756168e-01 -6.31173030e-02 -1.70345023e-01 6.95443332e-01
1.05787385e+00 -6.76665306e-01 -3.39851707e-01 8.23199987e-01
3.41798633e-01 -8.01972970e-02 5.31633794e-01 -1.00807548e+00
1.74302563e-01 -3.39150161e-01 -8.98588181e-01 -5.48259020e-02
3.26943368e-01 4.74930346e-01 6.50547743e-01 -9.63341296e-01
6.15033686e-01 1.18878281e+00 8.37909043e-01 4.97117303e-02
-1.04627633e+00 -3.89134765e-01 4.38756824e-01 -7.08626509e-02
-1.23605561e+00 -3.68877083e-01 1.14318132e+00 -4.33301389e-01
6.27571583e-01 1.98331266e-03 4.61231560e-01 8.95612955e-01
-1.65218875e-01 9.66675460e-01 1.15290356e+00 -3.97121966e-01
-1.72842130e-01 -1.02974172e-03 4.31961454e-02 9.67692852e-01
3.59486669e-01 1.08273312e-01 -1.92827180e-01 5.20143270e-01
9.74711061e-01 -1.67013511e-01 -1.67989850e-01 -8.70947957e-01
-1.08774626e+00 4.94193345e-01 4.61874038e-01 -2.19043002e-01
1.09929763e-01 -2.37939909e-01 2.59022653e-01 -3.66626978e-02
3.41903865e-01 1.95790067e-01 -2.60055065e-01 1.62813708e-01
-7.88498998e-01 3.49404961e-01 3.73821318e-01 1.28576469e+00
1.09428036e+00 -2.26881579e-02 4.41717744e-01 5.57777286e-01
3.28784794e-01 7.13589191e-01 -7.16242939e-02 -1.06362557e+00
9.10345197e-01 7.94691801e-01 1.89902529e-01 -1.10539269e+00
-6.77006662e-01 -4.55126107e-01 -3.96700889e-01 2.80840933e-01
8.81856501e-01 3.96866947e-02 -1.07573569e+00 1.04053903e+00
6.18383408e-01 -1.10813854e-02 -1.48577988e-01 1.15144360e+00
4.65260953e-01 4.88483846e-01 -5.09989619e-01 2.19645664e-01
1.11794543e+00 -1.19861221e+00 -2.87732065e-01 -8.02873194e-01
7.98549652e-01 -9.69701171e-01 1.02080262e+00 2.77353585e-01
-7.93437719e-01 -7.20363259e-01 -1.04012489e+00 -3.90712857e-01
-2.51850933e-01 6.04259014e-01 5.35932541e-01 5.11460900e-01
-8.44244421e-01 -5.62479580e-03 -4.75489795e-01 -3.55723828e-01
1.04051873e-01 6.21909685e-02 -4.88790005e-01 -3.06422234e-01
-8.20653081e-01 7.48791397e-01 4.53157961e-01 3.86281431e-01
-5.52667379e-01 -6.21016920e-01 -1.13317609e+00 -2.71984816e-01
9.07155991e-01 -5.28024912e-01 8.19930911e-01 -5.04688621e-01
-1.40022957e+00 1.10415900e+00 -9.56797898e-02 -3.05617809e-01
8.94887626e-01 -2.14257538e-01 -1.77845538e-01 1.97751611e-01
7.36013949e-02 7.58162916e-01 7.04313219e-01 -1.53475571e+00
-1.01581252e+00 -3.54859531e-01 -4.02640551e-03 3.76121193e-01
2.93653101e-01 -6.71938241e-01 -8.47614944e-01 -3.51546615e-01
5.59008598e-01 -9.89383876e-01 -3.35915744e-01 2.18596622e-01
-5.77509522e-01 7.89386779e-02 1.06153071e+00 -5.09989798e-01
7.17379630e-01 -2.19974017e+00 -2.06314012e-01 1.70764685e-01
4.32412624e-02 3.82650912e-01 -3.14629562e-02 3.49807739e-01
3.22601527e-01 -2.26233885e-01 -2.98619121e-01 -6.22463405e-01
-5.43232486e-02 3.17204684e-01 -5.95345259e-01 6.05327785e-01
3.44000936e-01 7.03425109e-01 -6.60378575e-01 -3.92754912e-01
6.90464079e-01 1.80343911e-01 -4.13348794e-01 3.99172872e-01
-4.16855998e-02 4.36361939e-01 -3.69082123e-01 7.21846461e-01
1.32866049e+00 2.32690945e-01 3.60762402e-02 -3.26437920e-01
-6.04493737e-01 3.56964946e-01 -1.20041108e+00 1.59966552e+00
-4.45842683e-01 6.25741541e-01 7.18525499e-02 -8.17825973e-01
1.30602491e+00 -5.15480042e-01 2.75761951e-02 -6.85976923e-01
9.86984596e-02 3.89956385e-01 -2.78494447e-01 -3.88225496e-01
5.75205088e-01 3.09711456e-01 -2.72770733e-01 -1.04680069e-01
-3.65819752e-01 -5.60853839e-01 8.44254270e-02 -4.91669662e-02
6.75048828e-01 4.57767963e-01 -5.62182404e-02 6.92553148e-02
7.39924490e-01 1.59418866e-01 6.46932721e-01 6.65204704e-01
-2.69358993e-01 8.21370065e-01 5.40691435e-01 -4.59693074e-01
-1.32016182e+00 -1.13390112e+00 4.88754697e-02 4.62721735e-01
6.15667760e-01 1.23123294e-02 -6.54170275e-01 -9.23535883e-01
2.36117527e-01 5.10475338e-01 -3.09766024e-01 2.37364992e-01
-7.34896779e-01 -3.29359442e-01 4.28966790e-01 5.73378980e-01
8.00140679e-01 -2.81755060e-01 -7.75377691e-01 9.52164549e-03
-1.97948366e-02 -1.71957898e+00 -4.50554520e-01 2.58867979e-01
-8.88480842e-01 -1.07837188e+00 -4.84931469e-01 -6.99799538e-01
5.76884627e-01 8.72149885e-01 7.35658765e-01 -1.86144993e-01
-1.70114100e-01 3.26848738e-02 -3.20530176e-01 -7.08319098e-02
-2.42711574e-01 1.64261937e-01 -1.50559664e-01 1.08273037e-01
1.01519346e-01 -3.88294995e-01 -7.21352339e-01 8.03004742e-01
-4.86406088e-01 4.70187992e-01 9.11402881e-01 6.26963675e-01
5.85693777e-01 -1.87477618e-01 1.76727459e-01 -7.66096115e-01
-2.73548186e-01 -1.13226250e-01 -1.11249137e+00 -2.25994945e-01
-4.86291021e-01 -1.06648348e-01 6.77071810e-01 -3.04448400e-02
-1.24199152e+00 6.50038064e-01 -8.70192796e-02 -6.74441695e-01
-4.60680187e-01 -5.70243299e-02 -4.14494157e-01 -2.40256473e-01
3.22656929e-01 3.39613944e-01 2.42544562e-01 -4.86957818e-01
4.79582489e-01 6.58828795e-01 7.02675045e-01 -2.06463665e-01
1.05568790e+00 9.90482807e-01 3.66562665e-01 -8.68552327e-01
-9.70837414e-01 -7.36474335e-01 -1.11572599e+00 -4.53190476e-01
9.09007668e-01 -1.03725624e+00 -4.55355227e-01 5.86849868e-01
-1.08055687e+00 -2.87984192e-01 1.33397371e-01 2.38240227e-01
-6.07299685e-01 6.64834142e-01 -3.22552264e-01 -6.86021149e-01
1.13366634e-01 -1.26907039e+00 1.54635262e+00 1.57776829e-02
3.52339923e-01 -8.35664749e-01 -4.20303255e-01 8.05956602e-01
3.83474976e-02 2.74509668e-01 5.35740376e-01 -2.30813116e-01
-1.23921430e+00 -4.46499556e-01 -4.94243979e-01 4.08950269e-01
-2.14601755e-01 -1.26539066e-01 -1.13460338e+00 1.87589694e-03
-2.83188403e-01 -1.13708995e-01 9.15309131e-01 1.58246867e-02
8.48196149e-01 3.13970357e-01 -3.85543346e-01 7.53491104e-01
1.22696340e+00 -1.42684549e-01 6.64848328e-01 5.35432398e-01
1.15039742e+00 9.51229453e-01 1.21821868e+00 8.37348253e-02
6.90163374e-01 8.19702983e-01 7.03596473e-01 -4.67351735e-01
-3.64298314e-01 -6.10858917e-01 4.17795509e-01 4.65358585e-01
1.27061367e-01 -2.04722553e-01 -8.96844625e-01 4.90296781e-01
-1.85269713e+00 -6.21569097e-01 -7.71687150e-01 2.25939369e+00
1.30616948e-01 5.71741164e-01 1.84860110e-01 -3.07537336e-02
4.93829191e-01 1.93571880e-01 -4.49000746e-01 -2.73078293e-01
-2.30121017e-01 -6.49060547e-01 1.09412730e+00 6.42671466e-01
-1.28351557e+00 1.30809045e+00 5.20180511e+00 8.05445313e-01
-1.22883546e+00 -3.34173530e-01 5.43176532e-01 3.83512378e-01
-1.50705293e-01 1.38222784e-01 -1.29790366e+00 3.31640482e-01
5.26760817e-01 5.29155254e-01 3.11183315e-02 8.54248703e-01
5.17937839e-01 -5.62896132e-01 -9.50499237e-01 8.84075761e-01
1.73926607e-01 -1.11568499e+00 -6.73728511e-02 3.06010514e-01
6.32178664e-01 1.75231293e-01 3.37899216e-02 1.59250684e-02
-1.80213720e-01 -4.83676136e-01 9.93111610e-01 2.82539040e-01
7.23669827e-01 -8.50174487e-01 7.93351889e-01 7.85018086e-01
-1.27114391e+00 -7.52441064e-02 -3.33267957e-01 -1.10406652e-02
3.83953929e-01 6.21289551e-01 -1.47206700e+00 7.87328720e-01
3.05951953e-01 8.45246434e-01 -7.93624341e-01 9.11299407e-01
-3.36431593e-01 2.93869585e-01 -4.57068652e-01 3.13261986e-01
5.68765283e-01 -3.04046035e-01 7.03005552e-01 1.23335266e+00
1.88392058e-01 -2.54963547e-01 4.95819777e-01 6.66011453e-01
2.60205090e-01 -1.43413916e-01 -9.70577478e-01 6.87317133e-01
4.23634350e-01 1.25613129e+00 -8.79124463e-01 -2.85932980e-02
-3.78198296e-01 8.80405247e-01 3.02354753e-01 1.92172155e-01
-8.75591159e-01 -4.20856923e-01 4.70715880e-01 6.50453866e-01
5.46784341e-01 -5.96656561e-01 -2.10321337e-01 -1.06230247e+00
3.41181755e-01 -3.27785164e-01 -8.20280388e-02 -7.87618339e-01
-9.50575709e-01 3.26234698e-01 -4.95855324e-02 -1.68893993e+00
-1.64900273e-01 -8.34709466e-01 -2.20534384e-01 8.48233521e-01
-2.03755593e+00 -1.44946826e+00 -4.60677266e-01 2.63189614e-01
6.42075777e-01 1.94562286e-01 1.04116909e-01 3.28051776e-01
-3.90063465e-01 4.48491335e-01 -1.06206186e-01 -2.23436952e-02
9.04942036e-01 -1.07402790e+00 5.85165143e-01 1.20092392e+00
-3.57310385e-01 1.84624016e-01 5.56472123e-01 -6.56000614e-01
-1.45182896e+00 -1.31324172e+00 7.44886518e-01 -6.17750466e-01
5.11271000e-01 -7.90499210e-01 -8.34883630e-01 4.86391991e-01
-2.84196347e-01 8.80390555e-02 1.89707540e-02 -3.28582019e-01
-3.38612139e-01 -4.28371012e-01 -1.05482411e+00 6.02762699e-01
1.44043756e+00 -3.44537824e-01 -3.86383891e-01 9.88875031e-02
6.12161696e-01 -6.31366491e-01 -4.39180344e-01 5.47649205e-01
3.81805867e-01 -1.33195853e+00 1.03038836e+00 1.03955388e-01
2.19563574e-01 -9.46341813e-01 -1.35317802e-01 -9.18082237e-01
1.18939906e-01 -4.63064939e-01 1.72248527e-01 1.08990693e+00
2.04435289e-01 -5.77619672e-01 9.13902938e-01 2.65351921e-01
-7.41608799e-01 -6.87058985e-01 -8.25949550e-01 -6.85331702e-01
-2.22081333e-01 -5.76574981e-01 2.35685736e-01 6.85774326e-01
-6.32604897e-01 2.27727160e-01 -3.66196871e-01 6.91244841e-01
8.39372933e-01 3.64920318e-01 1.38883841e+00 -1.14131498e+00
-2.45758127e-02 -3.29485148e-01 -6.92796648e-01 -1.95482600e+00
1.19297184e-01 -8.25014651e-01 1.69574082e-01 -1.17141581e+00
-3.73149186e-01 -5.90712488e-01 3.52082253e-01 1.99528843e-01
5.54441027e-02 2.08198532e-01 2.84304410e-01 1.66780531e-01
-5.33372879e-01 3.88830423e-01 1.38227546e+00 2.67742783e-01
-1.61395580e-01 -1.72696989e-02 -2.82824904e-01 8.68033528e-01
3.94040257e-01 8.56065676e-02 -4.04273808e-01 -5.66726446e-01
-1.77439004e-01 3.16367298e-02 4.48514342e-01 -9.94229794e-01
2.71897763e-01 7.70851001e-02 3.69982064e-01 -1.25346398e+00
4.28566933e-01 -9.74791408e-01 -3.43826354e-01 1.23536140e-01
1.19270027e-01 -1.25933483e-01 2.59473413e-01 7.53235757e-01
-5.05008101e-02 -8.80730152e-02 7.28191853e-01 7.34389573e-02
-1.18674469e+00 8.08238797e-03 -3.41680974e-01 -1.30790874e-01
1.23328018e+00 -7.07807779e-01 -3.17211926e-01 -1.41804844e-01
-3.43831748e-01 6.06955826e-01 9.61908817e-01 3.21152687e-01
7.92080402e-01 -8.60344350e-01 -3.19016188e-01 5.99584043e-01
3.15639824e-01 5.85635424e-01 3.51924568e-01 1.05744553e+00
-9.97752786e-01 5.24074137e-01 -3.50851491e-02 -1.00892949e+00
-1.09229267e+00 4.73750919e-01 4.00696903e-01 -1.10913236e-02
-9.06114638e-01 4.25920576e-01 3.17425996e-01 -5.59909225e-01
1.24215834e-01 -4.47243780e-01 1.17573977e-01 3.87352183e-02
3.26253295e-01 3.23991448e-01 2.34315425e-01 -7.82353818e-01
-2.80871212e-01 1.08043206e+00 -1.50381565e-01 2.00819165e-01
1.14411342e+00 -6.01401269e-01 2.79900968e-01 3.79059672e-01
1.11872315e+00 3.00318062e-01 -1.92849064e+00 1.96158767e-01
-9.51617807e-02 -8.02115977e-01 7.56252855e-02 -4.75609034e-01
-8.49237204e-01 1.08690810e+00 5.88657200e-01 -1.09055489e-01
8.76580119e-01 -9.53419954e-02 7.75030911e-01 3.04144472e-01
5.34549654e-01 -8.83459210e-01 -1.57058194e-01 5.75893641e-01
3.55722725e-01 -1.43701124e+00 -1.25678957e-01 -1.02424753e+00
-6.08098149e-01 1.50381529e+00 7.31322706e-01 -2.42596671e-01
1.32640570e-01 8.04886371e-02 2.11181194e-01 -6.89302396e-04
-2.44448408e-01 -3.70737642e-01 3.89959157e-01 5.57221472e-01
-1.39705062e-01 -1.91480532e-01 5.65424226e-02 -5.33948988e-02
1.43581524e-01 -3.51150155e-01 7.11919129e-01 8.52393925e-01
-6.51149929e-01 -1.02476287e+00 -4.67061490e-01 -3.84840891e-02
2.76625007e-01 3.20061952e-01 -3.42394829e-01 1.17752051e+00
2.93775946e-01 6.07566535e-01 2.61773258e-01 -4.05334890e-01
6.87635839e-01 -1.05082445e-01 2.72846073e-01 -3.84967983e-01
1.58619136e-01 3.05301607e-01 3.89378339e-01 -7.23193884e-01
-9.69396383e-02 -7.21721470e-01 -1.24961293e+00 -1.03692695e-01
-1.56577229e-01 -3.99407506e-01 7.25164413e-01 7.78791249e-01
6.44449711e-01 3.23680118e-02 9.43490028e-01 -1.37872779e+00
-3.65904242e-01 -3.96707982e-01 -3.54674131e-01 2.48157710e-01
5.46079218e-01 -9.23567355e-01 -5.03465652e-01 -1.96518958e-01] | [7.990467548370361, -1.6969341039657593] |
86b6ee18-6bb8-4c32-9129-66eef3daf6e5 | simpleview-neighborhood-views-for-point-cloud | null | null | https://ieeexplore.ieee.org/document/9874679 | https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=9874679 | SimpleView++: Neighborhood Views for Point Cloud Classification | Existing multi-view-based point cloud classification methods only utilize multiple views of point clouds and discard the point clouds from further processing. Among these methods, the Simple View model demonstrates that features from six orthogonal perspective projections of a point cloud achieved comparable 3D classification. However, points on the local structures overlap in these projections resulting in the loss of structural information. Also, we found that the performance of Simple View degrades at lower projection resolutions. We propose the use of neighbor projections along with object projections to learn finer local structural information. In this paper, we introduce SimpleView++ to concatenate features from the combined orthogonal perspective projections at object and neighbor levels with encoded features from the point cloud. We evaluated Simple-View++ using ModelNet40 and ScanObjectNN benchmark datasets. Our idea to use neighbor views broadly applies to other existing view-based methods. For example, neighbor views improve the results of MVTN by 2% on the hardest variant of ScanObjectNN. Our experiments also validate that neighbor view features considerably enhance classification accuracy at lower resolutions. | ['Chandra Kambhamettu', 'Shivanand Venkanna Sheshappanavar'] | 2022-09-08 | null | null | null | ieee-5th-international-conference-on | ['3d-point-cloud-classification', '3d-classification', 'point-cloud-classification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.89441964e-01 -3.99482727e-01 -2.33395964e-01 -5.51771164e-01
-5.66763341e-01 -8.86189401e-01 8.16735268e-01 -6.68211747e-03
2.01725289e-01 6.48485171e-03 -3.00730448e-02 -3.88418473e-02
-2.17974380e-01 -9.65224624e-01 -6.51764572e-01 -5.30375838e-01
1.60108685e-01 7.71262348e-01 6.56574547e-01 -1.65659830e-01
6.13935471e-01 1.04950547e+00 -2.03318048e+00 1.03556263e+00
3.95748258e-01 1.05804336e+00 2.34814972e-01 3.30981880e-01
-3.61521155e-01 2.69145072e-02 -2.36163512e-01 -1.34422496e-01
7.40858555e-01 5.42480528e-01 -4.07300681e-01 8.04701298e-02
1.32541895e+00 -3.09490234e-01 3.09073944e-02 6.96885586e-01
3.00255388e-01 -1.55843064e-01 4.82160866e-01 -1.27562010e+00
-4.28075671e-01 6.12210706e-02 -9.17732477e-01 3.57610136e-02
3.98917019e-01 1.09863408e-01 1.16060209e+00 -1.73520112e+00
8.64382029e-01 1.32002723e+00 9.86134052e-01 1.87828615e-01
-1.35815215e+00 -6.55896783e-01 3.37664425e-01 1.95141017e-01
-1.28010476e+00 -2.53784537e-01 9.32557881e-01 -5.22204638e-01
1.30827379e+00 7.93993354e-01 8.91727507e-01 7.97298253e-01
3.49934310e-01 3.88874084e-01 1.38466382e+00 3.04890443e-02
-2.97291260e-02 2.27278307e-01 4.30164516e-01 6.36932433e-01
2.72274047e-01 3.03615212e-01 -6.46258771e-01 -6.79137170e-01
6.97165668e-01 6.02986991e-01 -3.30621392e-01 -1.17782986e+00
-1.39845622e+00 7.65317559e-01 5.56037784e-01 -5.53464666e-02
-1.91544220e-01 -1.22771345e-01 2.08793446e-01 3.19145352e-01
7.04687476e-01 3.01293850e-01 -6.25284493e-01 1.46630898e-01
-7.63082445e-01 1.48414150e-01 5.34676373e-01 1.27524102e+00
9.87154305e-01 -3.12845975e-01 3.55928391e-01 8.21714640e-01
2.64090896e-01 7.52501607e-01 -2.12323174e-01 -1.16056025e+00
8.17651749e-01 9.34439540e-01 -3.59615326e-01 -1.06006050e+00
-4.07379866e-01 -5.04642785e-01 -5.95260918e-01 7.97902346e-01
5.04727103e-02 6.16228640e-01 -9.49973822e-01 7.43597746e-01
3.77334088e-01 -7.54541978e-02 -1.65293753e-01 7.86035895e-01
1.03407598e+00 5.32887280e-01 -6.49668038e-01 3.23554844e-01
1.25878108e+00 -9.77727592e-01 3.00532989e-02 5.41652180e-02
3.71064484e-01 -1.04645216e+00 1.09584999e+00 7.10757613e-01
-1.00851810e+00 -7.60774672e-01 -1.06424558e+00 1.91559102e-02
-3.95817339e-01 -4.67591658e-02 7.05184340e-01 4.54184711e-01
-1.03356576e+00 7.84490645e-01 -9.05379772e-01 -3.28766942e-01
4.40826923e-01 4.40330505e-01 -8.45567644e-01 -4.18602496e-01
-2.29812726e-01 6.84308112e-01 -1.02886111e-01 -2.04492092e-01
-3.10285330e-01 -1.17444777e+00 -5.50960958e-01 -4.48488817e-02
1.67728990e-01 -8.64998817e-01 7.21490443e-01 -2.37736389e-01
-8.83620620e-01 8.64740014e-01 -3.57004732e-01 1.33364439e-01
4.75728124e-01 -2.34369397e-01 -1.46096051e-01 2.38705710e-01
1.97553352e-01 6.86896205e-01 6.43776894e-01 -1.96917725e+00
-8.93543661e-01 -8.47642958e-01 1.20967098e-01 4.71313089e-01
-2.20451932e-02 -4.33616310e-01 -6.10428154e-01 -1.64570317e-01
1.16549122e+00 -1.13217509e+00 1.00417351e-02 3.73507708e-01
-2.43639722e-01 -1.72400594e-01 1.47569942e+00 -5.30977659e-02
3.95504028e-01 -2.22656274e+00 -8.63829851e-02 4.14134443e-01
6.09844327e-01 -2.61791110e-01 8.76710340e-02 6.47843838e-01
-2.94684798e-01 1.18295863e-01 3.01045328e-01 -2.79931337e-01
-3.98016274e-01 4.00798202e-01 -3.79023403e-01 5.57072699e-01
-1.45618379e-01 5.13858378e-01 -4.96237069e-01 -2.01528430e-01
5.75195432e-01 4.84148055e-01 -7.80767679e-01 -2.88414299e-01
6.72772452e-02 2.39933550e-01 -2.16982380e-01 8.27475369e-01
1.27503920e+00 -3.20011020e-01 -3.60033423e-01 -4.97734219e-01
-3.20059240e-01 3.25549603e-01 -1.04875576e+00 1.83461893e+00
-5.27711034e-01 4.48254466e-01 -7.24165365e-02 -3.90796840e-01
8.53862882e-01 8.52062479e-02 7.78844237e-01 -1.59321681e-01
-5.18009067e-01 7.39999861e-02 -1.23701699e-01 -1.67505324e-01
3.58415693e-01 3.00737899e-02 3.77098620e-01 3.00842226e-01
-1.51247665e-01 -5.15199602e-01 -4.39787984e-01 -2.77825631e-02
8.66530657e-01 1.71941057e-01 3.17809075e-01 -1.86105296e-01
3.28637749e-01 2.21189156e-01 2.70034939e-01 5.69932044e-01
2.89478600e-01 1.11502159e+00 1.20713353e-01 -8.51781666e-01
-9.73106444e-01 -1.51893544e+00 -5.03328323e-01 7.64969826e-01
3.07547659e-01 -7.20491350e-01 -2.92579886e-02 -9.17821825e-01
4.39422011e-01 6.78088605e-01 -3.00806135e-01 1.93439737e-01
-7.62011945e-01 -3.44448268e-01 -4.40622531e-02 6.76208436e-01
2.43244871e-01 -2.60985672e-01 -5.89907944e-01 -2.85255045e-01
1.09726079e-01 -1.12400723e+00 -7.99614936e-02 2.07316026e-01
-1.52171421e+00 -9.76425588e-01 -4.03940260e-01 -4.27354932e-01
7.64081180e-01 1.04781902e+00 1.28871429e+00 -7.37294788e-03
1.91422239e-01 6.26771092e-01 -3.36919695e-01 -4.22179013e-01
-2.40158774e-02 4.97252606e-02 7.05914125e-02 -3.69868815e-01
5.60330868e-01 -1.03054547e+00 -5.34363508e-01 7.23492086e-01
-3.32511485e-01 3.50669503e-01 3.53611410e-01 7.05115139e-01
1.02910101e+00 -3.55053067e-01 -2.95793653e-01 -9.16564107e-01
-1.96472451e-01 -3.07734728e-01 -5.68206728e-01 -6.41717091e-02
-6.03180647e-01 -2.37807795e-01 3.87626290e-01 1.60992015e-02
-6.83943331e-01 2.99409658e-01 -5.07935807e-02 -1.15277743e+00
-3.54628295e-01 -4.42605093e-03 -1.73129290e-01 -4.36856002e-01
5.59997976e-01 2.75401920e-01 -8.68769884e-02 -7.40051568e-01
3.15176815e-01 3.62717509e-01 -1.52583905e-02 -4.17260885e-01
1.01515877e+00 1.18079674e+00 2.77321160e-01 -8.12738001e-01
-5.10625780e-01 -8.91072810e-01 -1.03985894e+00 -2.61921406e-01
5.16500533e-01 -1.01989627e+00 -5.53392828e-01 -1.74570337e-01
-1.32374585e+00 4.86668110e-01 -3.50362480e-01 4.83516008e-01
-6.19065821e-01 3.28120857e-01 -2.19180673e-01 -3.86675537e-01
-1.59726724e-01 -1.22534049e+00 1.62404084e+00 -4.07383502e-01
5.66714397e-03 -6.28391147e-01 -7.02990890e-02 3.99142087e-01
4.36997116e-02 9.13159028e-02 9.87005353e-01 -5.59960544e-01
-1.05200827e+00 -1.04665108e-01 -1.12127036e-01 1.83014289e-01
1.36394337e-01 1.27710342e-01 -1.35430300e+00 -3.66045415e-01
5.10244846e-01 2.56189376e-01 9.97324526e-01 1.80371910e-01
1.14404166e+00 6.49366826e-02 -8.14266443e-01 1.05789828e+00
1.68665576e+00 7.75500834e-02 3.00082415e-01 2.06428915e-01
1.08642495e+00 6.26499653e-01 6.05380774e-01 2.22437814e-01
2.00262368e-01 8.91834438e-01 7.76853919e-01 -5.63092530e-02
-6.00563399e-02 -1.58790961e-01 6.43533468e-02 1.06562448e+00
-4.76328582e-01 4.08119671e-02 -1.20076406e+00 1.37429357e-01
-1.38792288e+00 -1.00054729e+00 -7.15635896e-01 2.20050144e+00
-1.33830056e-01 2.20841989e-01 -1.52791619e-01 1.43393710e-01
5.24507582e-01 2.41255537e-01 -4.03970599e-01 -6.77914768e-02
-5.26747331e-02 3.48973013e-02 6.95521832e-01 2.04397961e-01
-1.01032531e+00 5.14841437e-01 5.98347664e+00 6.33951783e-01
-1.25663042e+00 2.23835886e-01 1.38712868e-01 -4.41415101e-01
-4.65646118e-01 1.98082492e-01 -1.12179518e+00 1.52785361e-01
4.57262486e-01 3.07075709e-01 5.34168258e-02 1.21407652e+00
-5.00037409e-02 8.33753124e-02 -1.43117940e+00 1.21286535e+00
2.62763649e-01 -1.51130402e+00 4.11455184e-01 4.55382794e-01
7.43490577e-01 6.76433504e-01 1.30373880e-01 -6.81453273e-02
-1.18758783e-01 -6.08707130e-01 6.94387674e-01 4.45036560e-01
7.16868579e-01 -6.69975758e-01 4.86779422e-01 6.34996414e-01
-1.36794758e+00 9.37934741e-02 -6.71332955e-01 7.87831992e-02
6.63232356e-02 6.68708444e-01 -9.91730571e-01 1.00041735e+00
8.90165925e-01 1.00203800e+00 -8.06111991e-01 9.68715370e-01
4.40439969e-01 1.72934413e-01 -6.25744343e-01 4.08015966e-01
1.21625006e-01 -3.15450430e-01 9.30382371e-01 7.30849087e-01
5.62463641e-01 -4.69683781e-02 2.77728230e-01 7.56494880e-01
3.76631618e-01 -1.33540258e-01 -1.24703455e+00 7.52804458e-01
3.11682075e-01 1.18170774e+00 -8.02428961e-01 -3.41794163e-01
-9.03748870e-01 6.94321990e-01 1.68305457e-01 -2.77848821e-02
-4.74106610e-01 2.18207642e-01 7.28162348e-01 6.27678871e-01
5.59138238e-01 -3.83054137e-01 -8.10346365e-01 -1.26629925e+00
2.46363014e-01 -6.64897740e-01 1.55784711e-01 -8.43182266e-01
-1.44462538e+00 6.43781126e-01 3.73348475e-01 -2.11482191e+00
4.97981161e-02 -7.38146365e-01 -2.96387047e-01 9.06260908e-01
-1.20720100e+00 -1.33324683e+00 -5.69654465e-01 4.09580201e-01
9.33347642e-01 -1.18303761e-01 8.77805352e-01 1.03604645e-01
2.73123354e-01 -1.68234874e-02 2.49654010e-01 -3.51105720e-01
5.19168198e-01 -1.28482485e+00 7.27113545e-01 1.56851649e-01
5.65697908e-01 7.04910219e-01 3.67512226e-01 -6.64220810e-01
-1.47521007e+00 -9.59262013e-01 3.89430314e-01 -1.08940792e+00
8.80575553e-02 -5.82307756e-01 -9.67713118e-01 6.90942526e-01
5.96424192e-02 4.25766110e-01 6.48688316e-01 1.68062016e-01
-7.00638413e-01 -1.99947476e-01 -1.14749694e+00 2.46057749e-01
1.38070440e+00 -4.98834640e-01 -6.22171283e-01 3.16673428e-01
5.53016603e-01 -6.14941180e-01 -1.03498161e+00 8.22090149e-01
8.17929566e-01 -1.57605410e+00 1.45648718e+00 -3.48426968e-01
3.14936846e-01 -3.56321037e-01 -5.67611814e-01 -1.37392676e+00
-4.53307748e-01 2.72285968e-01 1.19460016e-01 9.01534438e-01
3.38140935e-01 -8.92463028e-01 1.13884795e+00 -6.53343573e-02
-4.54300642e-01 -1.01317501e+00 -9.37679827e-01 -8.85415614e-01
-3.14675048e-02 -5.04871368e-01 6.51084542e-01 1.03793502e+00
-4.36078221e-01 2.26130888e-01 2.85307378e-01 5.98691940e-01
6.89481378e-01 7.68028259e-01 1.05005455e+00 -1.81955922e+00
-2.79126525e-01 -1.60476536e-01 -7.48987079e-01 -1.11228490e+00
-1.79345012e-01 -1.17892885e+00 -4.34693843e-01 -1.50070930e+00
2.90801764e-01 -7.83895552e-01 4.15080320e-03 1.20326981e-01
2.03569785e-01 2.86160350e-01 4.56746727e-01 7.51538575e-01
-6.85831383e-02 1.92278340e-01 1.44641769e+00 -1.58807099e-01
4.88120802e-02 2.18846172e-01 -1.77715033e-01 1.15348399e+00
5.89668632e-01 -5.90280592e-01 -4.92449760e-01 -7.26377487e-01
1.63609684e-02 -4.31616179e-04 4.96995062e-01 -1.23969936e+00
9.81166139e-02 -1.10495083e-01 7.77617276e-01 -1.57485330e+00
1.12535727e+00 -1.23961520e+00 3.35563660e-01 3.05565029e-01
2.06602573e-01 6.13767982e-01 1.39722750e-01 6.55497372e-01
-7.21271709e-02 4.41107936e-02 5.61110616e-01 -5.66743791e-01
-4.67536509e-01 2.49553829e-01 1.77636161e-01 -5.66079199e-01
1.03087783e+00 -7.72281051e-01 -4.18826669e-01 -9.50892270e-02
-8.53801250e-01 -1.50619615e-02 1.02952754e+00 4.69126195e-01
9.17525232e-01 -1.29797661e+00 -4.67791200e-01 6.35584295e-01
4.49420869e-01 2.74282247e-01 3.37508991e-02 8.14594865e-01
-6.24444366e-01 5.97566009e-01 -3.78181040e-01 -1.61957932e+00
-1.80252934e+00 5.42123020e-01 1.05723999e-01 9.69785228e-02
-1.05242217e+00 7.60939419e-01 6.58224940e-01 -7.99640357e-01
-1.63327619e-01 -6.28625393e-01 -1.69099435e-01 -6.75309449e-02
1.32977515e-01 4.94173288e-01 5.93028367e-01 -7.43031442e-01
-3.87498230e-01 1.13562131e+00 -2.01491594e-01 -5.05016223e-02
1.52877891e+00 7.16184452e-02 -8.17977861e-02 7.36149490e-01
1.38870811e+00 3.92053932e-01 -1.01062357e+00 -6.25422895e-02
-1.82701677e-01 -9.68104303e-01 1.36088124e-02 -4.29301322e-01
-1.12364388e+00 1.15840650e+00 8.26157808e-01 2.15784162e-01
7.61920035e-01 2.36200914e-01 3.84791553e-01 4.78367358e-01
9.56934154e-01 -3.71419758e-01 -3.45029414e-01 4.21937734e-01
1.11395049e+00 -1.11177647e+00 3.95641714e-01 -1.16051459e+00
-2.75654197e-01 1.36347628e+00 8.14289570e-01 -2.23774329e-01
7.40302205e-01 2.63791531e-01 -6.35862816e-03 -7.98039496e-01
-1.02908039e+00 1.76947400e-01 4.23842609e-01 6.99881375e-01
9.43607986e-02 -7.80096054e-02 1.78272858e-01 -1.41011178e-01
-4.50478166e-01 -5.30251145e-01 2.29146332e-01 1.07215071e+00
-3.84821087e-01 -9.73006248e-01 -7.68714666e-01 8.34771693e-01
-8.54067430e-02 -4.36392464e-02 -4.78747278e-01 9.61575687e-01
3.20682853e-01 4.65078861e-01 5.79181790e-01 -6.95413828e-01
7.02838600e-01 2.20913743e-03 5.07351160e-01 -9.74845111e-01
-5.80638945e-01 9.31113213e-02 2.15358082e-02 -8.97592127e-01
-3.38235736e-01 -9.83911574e-01 -9.11442995e-01 -2.87963033e-01
-5.32685816e-01 -1.21405102e-01 8.60209227e-01 4.31711704e-01
5.59269309e-01 1.15796715e-01 7.95603216e-01 -1.27347326e+00
-6.15390122e-01 -5.99724174e-01 -3.38788539e-01 2.52060533e-01
2.55898595e-01 -8.48936558e-01 -5.25882483e-01 -1.99833930e-01] | [8.149969100952148, -3.345304250717163] |
f570d4a3-d575-4382-b6c6-4440a694d246 | continuous-diagnosis-and-prognosis-by | 2210.02719 | null | https://arxiv.org/abs/2210.02719v1 | https://arxiv.org/pdf/2210.02719v1.pdf | Continuous Diagnosis and Prognosis by Controlling the Update Process of Deep Neural Networks | Continuous diagnosis and prognosis are essential for intensive care patients. It can provide more opportunities for timely treatment and rational resource allocation, especially for sepsis, a main cause of death in ICU, and COVID-19, a new worldwide epidemic. Although deep learning methods have shown their great superiority in many medical tasks, they tend to catastrophically forget, over fit, and get results too late when performing diagnosis and prognosis in the continuous mode. In this work, we summarized the three requirements of this task, proposed a new concept, continuous classification of time series (CCTS), and designed a novel model training method, restricted update strategy of neural networks (RU). In the context of continuous prognosis, our method outperformed all baselines and achieved the average accuracy of 90%, 97%, and 85% on sepsis prognosis, COVID-19 mortality prediction, and eight diseases classification. Superiorly, our method can also endow deep learning with interpretability, having the potential to explore disease mechanisms and provide a new horizon for medical research. We have achieved disease staging for sepsis and COVID-19, discovering four stages and three stages with their typical biomarkers respectively. Further, our method is a data-agnostic and model-agnostic plug-in, it can be used to continuously prognose other diseases with staging and even implement CCTS in other fields. | ['Shenda Hong', 'Baofeng Zhang', 'Derun Cai', 'Moxian Song', 'Hongyan Li', 'Chenxi Sun'] | 2022-10-06 | null | null | null | null | ['mortality-prediction'] | ['medical'] | [-2.12306097e-01 -5.28965533e-01 -1.81380421e-01 -1.56065673e-01
3.83008318e-03 7.34338686e-02 1.61323890e-01 3.21283728e-01
-3.67598355e-01 8.69538724e-01 7.43441433e-02 -6.32622659e-01
-5.02087891e-01 -6.19937181e-01 -3.68829444e-02 -8.03179801e-01
-5.29387712e-01 9.13770914e-01 -1.55490220e-01 -8.81035626e-02
-6.72557801e-02 6.44427419e-01 -9.62928295e-01 2.75251508e-01
7.47309685e-01 1.27109015e+00 -5.35115674e-02 5.50402761e-01
-2.02224310e-02 1.01299846e+00 -4.13353235e-01 4.49287472e-03
-7.54756778e-02 -3.56617659e-01 -6.57002687e-01 -5.28426707e-01
-5.30741453e-01 -3.93884808e-01 -2.56556958e-01 3.15846205e-01
9.20069814e-01 -1.30131021e-01 7.67484307e-01 -1.21397150e+00
-2.19487607e-01 3.11812013e-01 -4.27441269e-01 3.49957794e-01
-1.44082149e-02 2.77305901e-01 2.45907307e-01 -5.34685314e-01
3.78090926e-02 1.07759404e+00 1.28059137e+00 7.82052457e-01
-7.31026888e-01 -6.44978762e-01 -1.05392270e-01 2.71388233e-01
-9.59857106e-01 -1.33728096e-02 3.60775471e-01 -7.88860559e-01
8.22101057e-01 2.28246063e-01 8.38925004e-01 1.39994335e+00
6.75453424e-01 3.84819806e-01 1.01382363e+00 6.04836680e-02
2.06626862e-01 -1.70668751e-01 1.79011852e-01 6.49730921e-01
2.74072498e-01 3.50424826e-01 -2.28971839e-01 -1.76321179e-01
8.93966556e-01 1.05587423e+00 -4.83818322e-01 1.28205717e-01
-1.78123450e+00 6.26642168e-01 1.00798599e-01 2.76343793e-01
-7.06392527e-01 9.02862027e-02 8.46154213e-01 3.53910983e-01
4.01480347e-01 6.19146645e-01 -1.16579020e+00 -3.57961595e-01
-5.85287154e-01 -1.82455987e-01 7.53599882e-01 3.87586266e-01
4.45639975e-02 2.04044238e-01 -4.31393743e-01 7.25611925e-01
-2.16386169e-01 5.28170168e-01 1.11766791e+00 -6.80023432e-01
-8.13939609e-03 7.07984209e-01 2.68869668e-01 -8.38418305e-01
-1.32451582e+00 -8.17279160e-01 -1.85769534e+00 -1.93484619e-01
-3.59519906e-02 -4.08819228e-01 -6.89915359e-01 1.50366306e+00
1.07484274e-01 5.55984080e-01 1.00960463e-01 7.14406133e-01
7.56859660e-01 3.97243351e-01 2.04190403e-01 -6.28976226e-01
1.45665216e+00 -8.74002278e-01 -7.79569924e-01 2.27741227e-01
7.96781361e-01 -2.63200819e-01 7.05830932e-01 7.35662580e-01
-8.16884458e-01 -4.52026576e-01 -6.09642565e-01 3.94809872e-01
-3.09542179e-01 1.86356947e-01 8.18334043e-01 5.98272271e-02
-9.84296560e-01 8.88755739e-01 -1.19512141e+00 -4.91385639e-01
4.47864324e-01 3.25905055e-01 -1.65333301e-01 9.11602899e-02
-1.34162867e+00 1.08573484e+00 3.45894068e-01 1.77025884e-01
-9.61172938e-01 -1.06152928e+00 -3.73801857e-01 9.99186262e-02
-8.40896145e-02 -1.50971854e+00 9.82374966e-01 -1.91763952e-01
-1.32703173e+00 5.88627696e-01 9.17922258e-02 -6.66275442e-01
6.41445518e-01 -3.82268906e-01 -4.84544247e-01 4.32765372e-02
-2.82800168e-01 2.03680962e-01 6.27531290e-01 -5.27609944e-01
-5.14092863e-01 -2.87043124e-01 -4.02228981e-01 -1.15446039e-01
-4.75195169e-01 -3.26598100e-02 2.49214396e-01 -6.34151042e-01
3.22091244e-02 -4.63253915e-01 -3.35055918e-01 -1.89181924e-01
-1.11393422e-01 -1.52760163e-01 7.64833331e-01 -6.34746671e-01
1.53789985e+00 -2.06658268e+00 1.15842372e-01 -3.45188648e-01
6.12684250e-01 4.42512274e-01 1.42299548e-01 6.47375047e-01
-1.42480969e-01 2.74086952e-01 -1.61014199e-01 -2.74534851e-01
-4.19679910e-01 3.12108219e-01 -3.03842247e-01 4.30107415e-01
2.51085550e-01 1.00706649e+00 -1.13900673e+00 -3.33731949e-01
2.96526581e-01 4.27143604e-01 -3.86574984e-01 6.81695402e-01
1.45300794e-02 9.32559133e-01 -4.27124083e-01 5.20346522e-01
2.97032386e-01 -6.44808054e-01 -1.05182603e-01 -1.08065806e-01
-8.52241123e-04 -3.17906499e-01 -5.57545006e-01 1.17747045e+00
-5.26094973e-01 2.36686423e-01 -3.36754411e-01 -1.45184410e+00
9.42648649e-01 7.25014031e-01 1.02794564e+00 -3.84324402e-01
3.42212379e-01 1.89130738e-01 -5.47377542e-02 -1.22970378e+00
-3.93600285e-01 -4.32418168e-01 1.27719879e-01 3.49861443e-01
-3.59072268e-01 2.27903128e-01 -2.14205116e-01 -2.49402240e-01
1.24840546e+00 -2.36100197e-01 4.36147213e-01 -1.66079491e-01
5.38823783e-01 -6.75770268e-02 7.26318061e-01 6.65724695e-01
-2.25759044e-01 4.81029302e-01 5.15782535e-01 -1.13079751e+00
-7.31090903e-01 -8.24206948e-01 -3.22705239e-01 7.36744583e-01
-2.29708984e-01 1.86835930e-01 -1.87207878e-01 -5.03499389e-01
1.31852940e-01 3.15354437e-01 -8.09087873e-01 -5.23000479e-01
-5.88052034e-01 -1.27016509e+00 6.24080896e-01 6.30409658e-01
4.03932959e-01 -9.84052122e-01 -8.74063551e-01 5.75954914e-01
-2.87427813e-01 -7.55751133e-01 -2.51913872e-02 3.59445661e-01
-1.33263218e+00 -1.32736874e+00 -9.43513274e-01 -6.97000682e-01
4.43661690e-01 -1.45710576e-02 1.17120707e+00 2.82954842e-01
-4.15455252e-01 -3.09275892e-02 -2.93546408e-01 -6.30900264e-01
-3.84456336e-01 -1.58905163e-02 5.20053148e-01 -3.60104702e-02
1.61532402e-01 -6.57213628e-01 -1.09166884e+00 3.11105311e-01
-7.71453321e-01 2.38632068e-01 5.11722863e-01 1.32267642e+00
3.24874282e-01 -3.43895964e-02 1.20417488e+00 -8.66201937e-01
6.79483473e-01 -9.30014670e-01 -4.75414097e-02 4.19376157e-02
-1.15712714e+00 -3.22086573e-01 1.12230396e+00 -3.78181577e-01
-5.73897898e-01 -4.59392309e-01 -2.15252653e-01 -7.42665827e-01
-1.01740316e-01 6.69018269e-01 6.36052728e-01 5.15061259e-01
4.45122451e-01 3.28176498e-01 3.37408483e-01 -7.72692382e-01
-2.89500654e-01 6.39211714e-01 3.59145105e-01 -2.82066077e-01
1.53853402e-01 2.95932353e-01 3.89708966e-01 -2.42143482e-01
-9.71391976e-01 -6.81129634e-01 -4.85099524e-01 8.94508511e-02
8.75477195e-01 -1.02672446e+00 -9.20437455e-01 8.66667867e-01
-1.11245346e+00 -2.70992368e-01 -1.30208254e-01 6.88860059e-01
-3.58212739e-01 2.51445353e-01 -8.72008026e-01 -5.66924512e-01
-8.73704433e-01 -8.15659106e-01 7.38143504e-01 -6.12181984e-02
-2.05382273e-01 -1.51251721e+00 2.63178587e-01 -1.39369503e-01
8.89435232e-01 7.29938507e-01 1.28519773e+00 -1.05042481e+00
2.07652867e-01 -3.23994875e-01 -2.88374245e-01 6.24088883e-01
4.00402039e-01 -2.42843349e-02 -6.75969601e-01 -5.35470307e-01
1.97661489e-01 -2.90264547e-01 7.23018110e-01 3.76518905e-01
1.59649551e+00 -3.98114771e-01 -5.19197941e-01 9.50699210e-01
1.26799119e+00 6.16437852e-01 3.23970318e-01 1.74212843e-01
5.62416494e-01 2.51480788e-01 2.88443178e-01 9.17992353e-01
4.06608850e-01 1.69034213e-01 4.58497047e-01 -5.10015786e-01
1.06002644e-01 3.49899352e-01 -1.44101428e-02 1.37199903e+00
-2.39687681e-01 -1.47047237e-01 -1.31697989e+00 4.29745048e-01
-1.87932730e+00 -8.09962809e-01 -3.71519536e-01 1.94077134e+00
9.16467369e-01 -1.54475480e-01 3.33622657e-02 1.87479272e-01
4.60884809e-01 -3.38458329e-01 -6.13762259e-01 -4.24377143e-01
3.27106677e-02 3.23155791e-01 -4.56921346e-02 -1.03354789e-01
-9.75577474e-01 3.76990348e-01 6.81507063e+00 2.93113202e-01
-1.54549015e+00 2.31887102e-01 9.51736212e-01 2.50251275e-02
3.64928037e-01 -5.35544753e-01 -3.78135115e-01 7.44996130e-01
1.08732951e+00 -7.85004124e-02 1.50530130e-01 7.65065312e-01
6.45728111e-01 3.23153973e-01 -1.35735166e+00 1.08945799e+00
-4.00596619e-01 -1.24227631e+00 -6.48298860e-02 -3.15633953e-01
6.46629393e-01 1.59527436e-01 1.78737752e-02 5.81801355e-01
8.37663561e-02 -1.21305037e+00 -1.67759016e-01 9.78648603e-01
1.11320722e+00 -3.49708855e-01 1.36939096e+00 4.78479713e-01
-6.51694059e-01 -4.01470542e-01 -2.99724221e-01 -4.61264968e-01
3.95052806e-02 8.03012252e-01 -8.52907598e-01 6.24971032e-01
6.22083783e-01 9.91284907e-01 -3.42641234e-01 1.21845508e+00
2.25318566e-01 6.93300784e-01 -9.51127261e-02 -9.05065835e-02
4.71333973e-02 1.11365773e-01 2.21928507e-01 1.23037410e+00
4.64132547e-01 2.02398822e-01 2.44743124e-01 2.22498059e-01
2.25789800e-01 -5.44043072e-03 -4.62162942e-01 9.69766602e-02
3.18398803e-01 1.00577068e+00 -4.10596848e-01 -6.36222422e-01
-6.08036369e-02 7.45174885e-01 5.85544668e-02 3.41765463e-01
-9.69798267e-01 -3.22835743e-01 4.08082992e-01 4.00338229e-03
-2.69679964e-01 8.95505548e-02 -5.54584146e-01 -1.40510499e+00
-3.72782081e-01 -8.44777226e-01 5.59168100e-01 -6.42193317e-01
-1.49366164e+00 6.00017071e-01 -2.07550451e-01 -1.49443269e+00
-2.36604840e-01 -7.88402677e-01 -8.47382367e-01 7.73837149e-01
-1.71654475e+00 -6.03430450e-01 -6.20469868e-01 5.37185907e-01
3.65501404e-01 -2.27124348e-01 1.18534887e+00 5.13463199e-01
-9.28629935e-01 2.61853844e-01 5.11928022e-01 8.74474868e-02
7.52777696e-01 -1.02435660e+00 -9.92221981e-02 2.57190794e-01
-8.17212224e-01 6.37056112e-01 3.49860132e-01 -3.91683877e-01
-1.23535132e+00 -1.42624903e+00 6.61180735e-01 -3.59294891e-01
6.75604939e-01 3.09273869e-01 -8.79087150e-01 2.63789207e-01
-1.15166627e-01 -1.01303905e-02 7.96987116e-01 -3.35556231e-02
1.17552698e-01 -4.10526365e-01 -1.01686776e+00 2.09398389e-01
9.15984571e-01 4.24594618e-02 -5.61701000e-01 6.18069589e-01
8.69137824e-01 -2.97987759e-01 -1.29973435e+00 1.00243104e+00
6.14489436e-01 -8.86785507e-01 9.39092934e-01 -1.18358660e+00
6.61260486e-01 2.29753688e-01 3.76292586e-01 -1.36555314e+00
-6.95228159e-01 -4.21433955e-01 -5.37564337e-01 6.72067583e-01
4.08313461e-02 -1.07142985e+00 3.71648461e-01 1.55674815e-01
-4.13887829e-01 -1.62020338e+00 -8.01460981e-01 -7.00187147e-01
1.61800489e-01 -1.90596893e-01 7.85436392e-01 1.29844236e+00
-8.80941302e-02 1.14748947e-01 -6.13111138e-01 -1.14401393e-01
1.94311455e-01 -6.80712238e-02 4.57225263e-01 -1.53698909e+00
-2.68030554e-01 -7.84849167e-01 -1.72991589e-01 -5.02924204e-01
-3.72886240e-01 -6.11141980e-01 -3.35394681e-01 -1.82623708e+00
9.78911296e-02 -7.30856597e-01 -1.14665866e+00 7.53663540e-01
-2.85318524e-01 -6.55230358e-02 -3.82920772e-01 5.39440632e-01
-3.53386551e-01 4.22282964e-01 1.40275156e+00 -5.10855168e-02
-1.73400626e-01 2.51296043e-01 -4.38001752e-01 6.47974491e-01
9.64624763e-01 -4.15082455e-01 -2.92536557e-01 -3.79565716e-01
1.93807811e-01 6.20818973e-01 3.38480443e-01 -1.09565985e+00
1.69227254e-02 -5.17099380e-01 5.53037584e-01 -4.80480492e-01
-2.60006599e-02 -8.21399510e-01 4.30063784e-01 1.26057398e+00
4.58215276e-04 4.74062055e-01 1.82418093e-01 4.42636341e-01
-1.90031499e-01 1.57141268e-01 7.10703969e-01 -2.16818199e-01
-4.46943432e-01 7.68732131e-01 -4.58228946e-01 2.69892782e-01
1.09198689e+00 5.10943048e-02 -3.82263690e-01 5.49363345e-02
-7.64627159e-01 5.75986862e-01 8.73929560e-02 3.79109651e-01
7.57325888e-01 -1.09587324e+00 -8.72148812e-01 2.44396314e-01
2.53906082e-02 2.37910911e-01 3.56462240e-01 1.55550563e+00
-9.06243086e-01 5.35185516e-01 -1.97648332e-01 -7.98576772e-01
-6.77424848e-01 8.37785065e-01 5.67403555e-01 -5.36366224e-01
-7.51848102e-01 7.05922186e-01 3.34636301e-01 -1.39772564e-01
4.61874813e-01 -3.40207785e-01 -3.69054765e-01 7.60086104e-02
5.11352658e-01 5.98827541e-01 1.74393773e-01 5.08576073e-02
-4.71115887e-01 4.27518696e-01 1.10464785e-02 7.32173741e-01
1.53696704e+00 1.11003928e-01 -4.40376729e-01 6.81987524e-01
1.11908722e+00 -6.44011974e-01 -8.06235552e-01 -7.01654982e-03
-1.73337623e-01 1.14162732e-02 -1.86540186e-01 -1.14553857e+00
-9.74321485e-01 1.17990315e+00 7.57153392e-01 3.73546004e-01
1.34034479e+00 -3.87289077e-01 1.06094837e+00 4.32552129e-01
2.80180812e-01 -4.54726726e-01 1.30794346e-01 6.27274990e-01
7.25732327e-01 -1.21377027e+00 -1.43003583e-01 2.95283020e-01
-4.68801230e-01 1.32160127e+00 4.40246433e-01 1.54155225e-01
8.09751689e-01 1.57742292e-01 3.18135440e-01 -1.98505402e-01
-1.18246639e+00 3.56061220e-01 -7.92003702e-03 5.75332940e-01
4.37674046e-01 2.04779908e-01 -4.10758346e-01 9.28181052e-01
1.62854612e-01 5.11681080e-01 2.92088091e-01 6.73418045e-01
-3.69413614e-01 -7.60861278e-01 -5.44120334e-02 9.87628341e-01
-8.16700816e-01 -2.69749939e-01 2.24808171e-01 6.00386024e-01
2.42831737e-01 6.64631188e-01 -1.88710272e-01 -5.55106878e-01
1.94286674e-01 3.35038632e-01 3.27195078e-02 -4.71064538e-01
-6.96967602e-01 -3.29987198e-01 -1.85568601e-01 -3.09854329e-01
-1.40544608e-01 -3.34544480e-01 -1.10242772e+00 -3.53083193e-01
-1.13336399e-01 2.66877800e-01 3.95919085e-01 8.46175373e-01
7.44237304e-01 9.83147383e-01 9.63408887e-01 -3.80270272e-01
-8.95477831e-01 -1.16478896e+00 -2.96310484e-01 3.41502041e-01
6.42117739e-01 -7.51670003e-01 -4.00931716e-01 4.19313610e-02] | [7.957047462463379, 6.179295063018799] |
fa8029d3-cdbd-4b91-a13d-b835047dea5d | a-mixed-quantization-network-for | 2203.06504 | null | https://arxiv.org/abs/2203.06504v1 | https://arxiv.org/pdf/2203.06504v1.pdf | A Mixed Quantization Network for Computationally Efficient Mobile Inverse Tone Mapping | Recovering a high dynamic range (HDR) image from a single low dynamic range (LDR) image, namely inverse tone mapping (ITM), is challenging due to the lack of information in over- and under-exposed regions. Current methods focus exclusively on training high-performing but computationally inefficient ITM models, which in turn hinder deployment of the ITM models in resource-constrained environments with limited computing power such as edge and mobile device applications. To this end, we propose combining efficient operations of deep neural networks with a novel mixed quantization scheme to construct a well-performing but computationally efficient mixed quantization network (MQN) which can perform single image ITM on mobile platforms. In the ablation studies, we explore the effect of using different attention mechanisms, quantization schemes, and loss functions on the performance of MQN in ITM tasks. In the comparative analyses, ITM models trained using MQN perform on par with the state-of-the-art methods on benchmark datasets. MQN models provide up to 10 times improvement on latency and 25 times improvement on memory consumption. | ['Paul Wisbey', 'Frederik Laboyrie', 'Mete Ozay', 'Juan Borrego-Carazo'] | 2022-03-12 | null | null | null | null | ['tone-mapping', 'inverse-tone-mapping'] | ['computer-vision', 'computer-vision'] | [ 5.84921360e-01 -1.94505110e-01 -2.44577661e-01 -2.43855521e-01
-1.02650523e+00 -2.01977894e-01 4.94277179e-01 -5.78618884e-01
-4.84213412e-01 2.96810955e-01 2.70642012e-01 -3.53210866e-01
-1.82386823e-02 -8.18610072e-01 -1.01367104e+00 -2.80309469e-01
1.07713595e-01 1.49357557e-01 2.24373385e-01 -2.50110537e-01
2.92925090e-01 3.83287877e-01 -1.84251857e+00 4.38865066e-01
7.10315585e-01 1.18241823e+00 5.03752768e-01 8.71377110e-01
5.45996167e-02 9.10195291e-01 -4.87979621e-01 -2.94882536e-01
4.33409989e-01 -2.85056919e-01 -7.29802549e-01 -3.18196297e-01
7.51460254e-01 -8.82955670e-01 -5.56150854e-01 8.72323096e-01
8.11302483e-01 9.86349881e-02 3.71546626e-01 -1.01448679e+00
-9.29868162e-01 3.42993110e-01 -6.88627660e-01 4.13801700e-01
3.17437276e-02 1.38981849e-01 6.94329798e-01 -9.63431180e-01
6.83234513e-01 1.10152340e+00 7.53827631e-01 5.06573498e-01
-1.15310037e+00 -6.23111248e-01 -2.41652742e-01 3.92173588e-01
-1.51402307e+00 -7.92598903e-01 4.52239215e-01 2.45641798e-01
1.52118576e+00 3.45286340e-01 1.36930391e-01 7.36641526e-01
3.21495444e-01 5.20763159e-01 1.04807162e+00 -3.57364953e-01
-5.97235374e-02 -1.83168739e-01 -3.49564165e-01 4.96694356e-01
-1.59252718e-01 1.16663091e-01 -9.77471769e-01 2.78459221e-01
8.15093577e-01 -1.70133680e-01 -2.86216348e-01 1.33290634e-01
-1.14749444e+00 6.37975395e-01 5.11110604e-01 1.35400414e-01
-2.86646605e-01 5.67627490e-01 3.08834791e-01 4.60688829e-01
5.52977979e-01 3.48199308e-01 -4.00278658e-01 -5.12052059e-01
-1.22626615e+00 -1.84713826e-01 3.55431080e-01 9.49044824e-01
7.41609633e-01 -8.76902714e-02 -1.38086051e-01 9.45635021e-01
6.48501441e-02 4.92532492e-01 3.41950715e-01 -1.24832284e+00
7.13874757e-01 1.44216150e-01 -5.12151308e-02 -7.10520566e-01
-3.09729367e-01 -4.35899884e-01 -7.62251675e-01 2.47271672e-01
2.26155482e-02 2.03689739e-01 -1.19674492e+00 1.57170165e+00
-4.00557220e-02 2.95426458e-01 -5.59809059e-02 1.20191634e+00
8.28686297e-01 8.23852956e-01 -1.41158119e-01 1.16662085e-01
1.13166904e+00 -1.16248989e+00 -7.03456938e-01 -2.30649173e-01
6.24802828e-01 -9.53402996e-01 1.52027452e+00 3.79165918e-01
-1.29637992e+00 -7.93760657e-01 -1.39618683e+00 -7.74174094e-01
-3.47888023e-01 8.30250084e-02 4.96285319e-01 5.77506483e-01
-1.40754104e+00 8.14720929e-01 -7.31168032e-01 -1.11269884e-01
5.82102120e-01 5.16456723e-01 -2.05554560e-01 -5.19788444e-01
-1.26473582e+00 9.15720224e-01 3.34052853e-02 1.48807734e-01
-8.21551323e-01 -9.98768628e-01 -7.17704892e-01 6.78217262e-02
2.00677723e-01 -5.98109543e-01 1.19770455e+00 -8.30141068e-01
-1.64265978e+00 8.12063992e-01 8.21093544e-02 -5.22718549e-01
3.96651983e-01 -4.40151751e-01 -3.48745793e-01 3.04284602e-01
-4.16538306e-02 1.17674208e+00 9.02006507e-01 -8.81530762e-01
-4.75672543e-01 -1.51614755e-01 9.02462974e-02 4.08638448e-01
-3.94328833e-01 -4.59000207e-02 -8.52584541e-01 -4.93685722e-01
2.36110538e-02 -8.81245553e-01 1.18258670e-01 2.26435184e-01
-1.21043757e-01 3.72502446e-01 1.15334845e+00 -8.67277920e-01
1.23772919e+00 -2.38169074e+00 -3.60544659e-02 -1.67623490e-01
2.33219424e-03 4.17474478e-01 -4.21679735e-01 2.43820678e-02
1.52081281e-01 2.14167029e-01 -5.23235165e-02 -6.43250883e-01
1.74578577e-02 2.55490988e-01 -3.41579318e-01 3.32319528e-01
1.07545897e-01 9.61247385e-01 -4.19191152e-01 -4.60584670e-01
4.22966033e-01 9.07845557e-01 -6.42220140e-01 1.43547505e-01
-1.39388710e-01 1.49782524e-01 2.00833172e-01 7.37736762e-01
8.31768036e-01 -4.29750562e-01 2.82301307e-02 -5.13097525e-01
-8.75060707e-02 3.54463816e-01 -7.78752983e-01 2.04031992e+00
-1.10764349e+00 9.65921879e-01 -2.17273869e-02 -4.96069282e-01
7.06752598e-01 1.19511373e-01 1.37519330e-01 -1.69639969e+00
8.92176293e-03 3.33149761e-01 -2.32342869e-01 -2.76375979e-01
8.88959467e-01 1.66387528e-01 1.80139408e-01 4.99554485e-01
-7.85422996e-02 1.83158651e-01 -1.13945685e-01 -5.97113185e-02
1.00987017e+00 2.88621098e-01 -9.40015987e-02 -1.03921950e-01
9.10023004e-02 -2.72226214e-01 2.08975181e-01 7.94016957e-01
-1.72876686e-01 9.67816830e-01 2.32218847e-01 -4.99132395e-01
-1.26050925e+00 -1.07610738e+00 -1.92699060e-01 1.43054938e+00
3.79883915e-01 -3.73153806e-01 -6.16289258e-01 -1.77082151e-01
-4.44334060e-01 5.02574861e-01 -4.21158940e-01 -2.86384374e-01
-7.49158204e-01 -9.62259829e-01 6.20360374e-01 5.75706303e-01
1.08781064e+00 -9.63335931e-01 -9.86676157e-01 7.65537098e-02
-4.33727711e-01 -1.39485180e+00 -6.18438184e-01 2.54458547e-01
-7.99162507e-01 -4.85993654e-01 -7.88829029e-01 -6.13417923e-01
9.59210470e-02 4.73344594e-01 1.57891679e+00 2.55099144e-02
-3.17123473e-01 -9.48289484e-02 -3.18089485e-01 -2.11193070e-01
-2.04609334e-01 4.34825838e-01 -3.32174093e-01 -1.73458859e-01
1.13284819e-01 -5.93739808e-01 -9.85157073e-01 3.97754192e-01
-1.14504695e+00 3.39045405e-01 9.17251706e-01 6.69202626e-01
8.59375298e-01 2.60120213e-01 4.41758484e-01 -7.48494983e-01
3.18119496e-01 -2.67692894e-01 -4.82599050e-01 1.94745898e-01
-7.25817144e-01 1.03504322e-01 5.14012754e-01 -5.39057732e-01
-1.08912933e+00 -9.09202993e-02 -2.39812285e-01 -6.09718621e-01
2.91008115e-01 2.10098088e-01 -1.52497247e-01 -2.59724379e-01
6.45227015e-01 2.71758229e-01 -1.10378459e-01 -3.64348561e-01
4.64733452e-01 6.23235404e-01 7.34792829e-01 -1.60503298e-01
5.04027665e-01 6.22541964e-01 4.81493697e-02 -5.69692135e-01
-7.22242475e-01 -1.70869276e-01 -1.68799102e-01 -1.30824864e-01
8.68375480e-01 -1.14452672e+00 -4.18460995e-01 5.46253383e-01
-8.14113557e-01 -8.56327415e-01 -2.38766335e-02 2.67570615e-01
-5.77293694e-01 1.22869328e-01 -8.11292529e-01 -3.31084311e-01
-6.14585102e-01 -1.29494691e+00 1.50059915e+00 1.75798357e-01
6.82651475e-02 -6.97957098e-01 -2.63805956e-01 5.12348413e-01
1.06906319e+00 -1.42132565e-01 8.16887379e-01 1.96386531e-01
-1.01732635e+00 1.96028590e-01 -7.23650634e-01 2.18350455e-01
-1.98584720e-01 -3.33091974e-01 -1.23209715e+00 -3.34425271e-01
-1.36560947e-01 -4.55919504e-01 1.15489662e+00 6.28891349e-01
1.47034943e+00 -6.92524612e-02 -9.81074721e-02 1.16999340e+00
1.69565856e+00 -8.12525824e-02 1.26203728e+00 6.30506814e-01
7.69711733e-01 2.26324379e-01 7.91648269e-01 2.87436277e-01
3.66925985e-01 9.45454419e-01 6.84128344e-01 -4.22897786e-01
-5.67273974e-01 -5.46111241e-02 4.36751217e-01 4.16982770e-01
1.54469848e-01 -3.53133380e-01 -7.21688688e-01 3.92656356e-01
-1.52126646e+00 -6.36035979e-01 1.97312698e-01 2.12635589e+00
8.68458807e-01 3.00949782e-01 -2.35530213e-01 1.56343535e-01
4.77420092e-01 4.01094824e-01 -6.04212105e-01 -6.13396764e-01
-1.88691616e-01 3.95411104e-01 8.55848610e-01 4.80069160e-01
-1.01700175e+00 7.08284557e-01 6.32671404e+00 1.15562451e+00
-1.50273931e+00 3.69847268e-01 1.01874018e+00 -4.68477428e-01
-2.71705419e-01 -2.70789921e-01 -7.38208771e-01 5.67365050e-01
1.35207868e+00 4.79644150e-01 8.52651119e-01 7.00539231e-01
1.34473145e-01 -1.79499567e-01 -9.33606446e-01 1.42655075e+00
1.56947896e-01 -1.44620717e+00 -1.14664242e-01 2.59956628e-01
8.14749122e-01 4.34459865e-01 5.99115849e-01 3.44781607e-01
-2.16220066e-01 -1.30438626e+00 7.18818188e-01 3.72123539e-01
1.50779355e+00 -8.68001580e-01 7.48241007e-01 5.70102669e-02
-1.19009864e+00 -6.19980060e-02 -4.12035167e-01 1.47530446e-02
6.02531694e-02 6.41804039e-01 -5.17458081e-01 2.70465881e-01
1.21261013e+00 4.42241430e-01 -6.94943726e-01 4.53708053e-01
2.98765510e-01 4.31815475e-01 -3.51851523e-01 4.68226880e-01
1.30031958e-01 1.53908774e-01 4.59050648e-02 1.09202898e+00
3.79842252e-01 -1.50261268e-01 -2.88556337e-01 7.78118134e-01
-4.93669689e-01 -2.64977276e-01 -4.70776916e-01 1.62508696e-01
5.26334107e-01 1.18300378e+00 -7.77272522e-01 -1.96925640e-01
-5.93401432e-01 1.31382442e+00 2.74770141e-01 2.25953057e-01
-1.16063142e+00 -3.56060356e-01 6.05158567e-01 2.36117095e-01
3.95073891e-01 -1.06605984e-01 -4.50967461e-01 -8.65106046e-01
2.95894474e-01 -8.09412122e-01 1.89217940e-01 -1.06033683e+00
-9.03338313e-01 7.00845659e-01 -2.78038085e-01 -1.26535964e+00
-8.99478495e-02 -4.53173995e-01 -2.84540743e-01 9.58928645e-01
-2.08905149e+00 -1.09402442e+00 -6.32660329e-01 4.64140385e-01
6.01615429e-01 3.99521962e-02 6.03198588e-01 8.40943336e-01
-3.91758293e-01 8.15367103e-01 9.54990089e-02 -2.14882955e-01
8.05906177e-01 -1.05624044e+00 7.57927299e-01 9.59190905e-01
9.99085698e-03 4.34086621e-01 4.21746999e-01 -2.05761850e-01
-1.48668718e+00 -1.38549364e+00 6.75561905e-01 -2.66921759e-01
3.57389487e-02 -4.68260914e-01 -8.95458639e-01 3.35247219e-01
5.87373972e-02 2.74843007e-01 4.70049351e-01 -1.78582773e-01
-4.30041194e-01 -1.99769363e-01 -1.06712234e+00 6.65107191e-01
1.10686994e+00 -9.08381522e-01 2.16273233e-01 1.59658000e-01
1.05662990e+00 -5.98000884e-01 -1.00699735e+00 4.34635401e-01
6.15033865e-01 -1.23742938e+00 1.26579499e+00 1.96868926e-01
7.15268135e-01 -2.88177490e-01 -5.99277139e-01 -7.17302203e-01
-1.64823249e-01 -6.35999084e-01 -3.50316435e-01 1.02600956e+00
2.24419892e-01 -3.40848476e-01 8.19758594e-01 3.75534654e-01
-2.07446530e-01 -1.02718651e+00 -1.01385009e+00 -5.89747787e-01
-2.11965814e-01 -5.00467122e-01 5.48159778e-01 5.78867972e-01
-7.96419621e-01 2.64492333e-01 -7.05529749e-01 9.02764723e-02
4.17979479e-01 1.08895600e-01 4.83126491e-01 -5.26922941e-01
-4.15365219e-01 -2.07299098e-01 -2.71873295e-01 -1.12267494e+00
-9.21059325e-02 -5.51160991e-01 8.56731907e-02 -1.48911071e+00
1.66886806e-01 -4.65962321e-01 -2.08596319e-01 3.05932671e-01
-4.83265221e-02 1.09624088e+00 2.88867980e-01 3.29760551e-01
-8.40542197e-01 4.92540330e-01 1.17187881e+00 -2.11299607e-03
-1.56592175e-01 -4.86562222e-01 -6.67647004e-01 3.85496110e-01
7.57153392e-01 -3.04875493e-01 -6.58844709e-01 -9.46767330e-01
3.50733012e-01 1.12997264e-01 4.06761169e-01 -1.20766747e+00
5.83028309e-02 1.96122617e-01 5.41231632e-01 -6.64120197e-01
5.33060372e-01 -6.30585372e-01 2.58761287e-01 2.61730283e-01
-2.90238529e-01 1.36103272e-01 2.35591099e-01 2.54748523e-01
-2.13936165e-01 5.08092865e-02 8.76297891e-01 5.55477291e-03
-1.13383949e+00 3.28088641e-01 1.16806459e-02 -1.00489117e-01
4.95077044e-01 -4.25862014e-01 -5.12089968e-01 -4.63239789e-01
-2.88052142e-01 -1.98793367e-01 6.31648421e-01 5.52315116e-01
7.58232176e-01 -1.27989638e+00 -5.02904654e-01 2.15746894e-01
-1.10630840e-01 6.29849359e-02 6.74175024e-01 7.49494374e-01
-8.58483434e-01 3.45289826e-01 -3.93040597e-01 -6.29818738e-01
-1.17980897e+00 5.24488568e-01 5.06489813e-01 -3.67757916e-01
-6.24311328e-01 6.19802117e-01 2.07380071e-01 -3.06096822e-01
2.23867238e-01 -2.78293699e-01 1.29640177e-01 -3.48504961e-01
5.88134825e-01 4.34084386e-01 4.80161726e-01 -5.08032858e-01
-2.66606033e-01 6.24067903e-01 -1.45531923e-01 -1.70290485e-01
1.10617149e+00 -4.56682801e-01 1.24964796e-01 1.05548956e-01
1.67408633e+00 -4.83622134e-01 -1.43396747e+00 -1.96507722e-01
-4.19722289e-01 -6.37666464e-01 5.35721064e-01 -7.78552473e-01
-1.26080310e+00 9.21517730e-01 1.16982126e+00 -4.16076839e-01
1.84088969e+00 -1.43539488e-01 1.11927593e+00 2.73597956e-01
2.77928919e-01 -1.23610461e+00 2.49989256e-01 2.99479038e-01
7.06943333e-01 -1.36317432e+00 4.98702042e-02 -3.23140323e-01
-3.75287473e-01 8.70456934e-01 6.30691469e-01 9.58939940e-02
5.82885027e-01 4.87733632e-01 2.05569506e-01 -2.38569975e-02
-8.99948776e-01 -9.59371626e-02 1.54727384e-01 6.04462624e-01
3.82429332e-01 -2.13278294e-01 4.09906134e-02 3.78156304e-02
-1.59513921e-01 1.69300258e-01 4.95860755e-01 8.69578183e-01
-1.59900755e-01 -7.91535735e-01 -1.48610115e-01 4.81468320e-01
-6.47315800e-01 -4.12175000e-01 2.14881793e-01 8.30615222e-01
2.31250674e-01 8.16465497e-01 3.49527985e-01 -6.41143680e-01
1.90971091e-01 -4.67553556e-01 3.69466573e-01 -1.62712023e-01
-3.55450273e-01 -1.98747646e-02 4.86619063e-02 -9.88534987e-01
-4.91139114e-01 -2.34133676e-01 -1.02122915e+00 -6.26633346e-01
-1.21879518e-01 -6.89402163e-01 8.49074841e-01 7.33254075e-01
7.38195717e-01 7.09961474e-01 4.00266677e-01 -1.00146687e+00
-3.53966355e-01 -8.04210901e-01 -3.58296663e-01 3.90808731e-01
2.95638829e-01 -4.63650852e-01 4.78596613e-02 3.74998078e-02] | [10.9269380569458, -2.117189407348633] |
781c2496-09b2-489a-a143-3e263a2139dc | ask-me-anything-in-your-native-language | null | null | https://openreview.net/forum?id=OJYPlFa4rmX | https://openreview.net/pdf?id=OJYPlFa4rmX | Ask Me Anything in Your Native Language | Cross-lingual question answering is a thriving field in the modern world, helping people to search information on the web more efficiently. One of the important scenarios is to give an answer even there is no answer in the language a person asks a question with. We present a novel approach to achieving a new state of the art in both retrieval and end-to-end tasks on the XOR TyDi dataset outperforming the previous results up to 10\% on several languages. We find that our approach can be generalized to more than 100 languages in zero-shot approach and outperform all previous models by 8\%. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['cross-lingual-question-answering'] | ['natural-language-processing'] | [-3.46601725e-01 -2.65626818e-01 -1.31600529e-01 -2.80187786e-01
-1.68950522e+00 -8.84539068e-01 5.33334732e-01 4.01668817e-01
-8.41203749e-01 7.78220594e-01 3.36015999e-01 -3.47471118e-01
-2.97508389e-01 -7.94770896e-01 -3.19408745e-01 1.84901163e-01
2.52783298e-01 1.07316959e+00 8.25289011e-01 -1.00087893e+00
3.05005908e-01 -2.47020632e-01 -1.37919116e+00 5.58913350e-01
7.53455281e-01 9.59200084e-01 -2.61635184e-02 8.94302249e-01
-8.23717654e-01 8.47034097e-01 -4.05059636e-01 -7.72906184e-01
1.11054599e-01 -2.25539401e-01 -1.55117476e+00 -7.17225492e-01
9.94067669e-01 -7.22013339e-02 -1.84192896e-01 8.69054377e-01
7.06836224e-01 2.60876894e-01 4.13987041e-01 -6.08327925e-01
-1.28098071e+00 2.93951541e-01 -2.06015304e-01 6.18698359e-01
8.71635973e-01 -2.51881689e-01 1.42683887e+00 -1.03025842e+00
6.84917271e-01 1.38687551e+00 1.54979825e-01 5.78027129e-01
-7.79971898e-01 -4.33928728e-01 -1.86490759e-01 6.64105177e-01
-1.35416067e+00 -4.37340021e-01 3.05153072e-01 -6.10111691e-02
1.36197591e+00 4.12920833e-01 -1.68673739e-01 4.41042125e-01
-1.20753944e-01 7.00726330e-01 8.14788520e-01 -6.45721495e-01
2.00750912e-03 1.81260943e-01 7.10504413e-01 7.52814591e-01
-2.02868190e-02 -5.74706733e-01 -6.52520776e-01 -3.85012209e-01
-1.95892632e-01 -3.76206040e-01 5.25397360e-02 1.68200761e-01
-8.70989561e-01 1.00040841e+00 3.30264360e-01 5.05144119e-01
-3.18725072e-02 1.15009569e-01 6.74081922e-01 8.99024367e-01
5.24071336e-01 8.08567107e-01 -6.53448701e-01 -1.23572432e-01
-8.63837063e-01 6.49517715e-01 1.41881883e+00 8.34223568e-01
6.74540758e-01 -4.67316955e-01 -5.62288404e-01 1.06711650e+00
3.40439677e-02 5.82750857e-01 6.67139828e-01 -1.27166772e+00
6.10021651e-01 6.64191008e-01 2.85619885e-01 -5.01973867e-01
9.51339304e-02 -3.62286747e-01 -2.73548633e-01 -5.12625337e-01
4.70343530e-01 -1.78101167e-01 -6.36309206e-01 1.40365791e+00
3.42262030e-01 -4.06483144e-01 1.78738385e-01 8.05321515e-01
1.37715113e+00 8.26651931e-01 3.20442021e-02 1.60679519e-02
1.91976392e+00 -1.19275856e+00 -8.77075255e-01 -5.44520497e-01
7.63155878e-01 -1.23343599e+00 1.22604203e+00 7.10885450e-02
-9.84752178e-01 -4.71574694e-01 -7.32896686e-01 -8.60411882e-01
-6.78020954e-01 -5.74056022e-02 4.16158408e-01 5.62032938e-01
-1.23819339e+00 8.26928020e-02 2.34299526e-02 -8.06022763e-01
-1.35015026e-01 -9.24563408e-03 -1.99663028e-01 -6.27777100e-01
-1.68230689e+00 1.21395969e+00 2.30810300e-01 -3.74819934e-01
-5.16193330e-01 -6.45034552e-01 -5.45132697e-01 3.56172398e-02
6.97864115e-01 -6.76109552e-01 1.59422147e+00 -2.98566729e-01
-1.06521893e+00 1.20734763e+00 -7.21183360e-01 -4.20976490e-01
7.69910589e-02 -5.01424670e-01 -5.42034090e-01 1.31508946e-01
4.12982732e-01 5.15435040e-01 4.24753815e-01 -3.81010205e-01
-6.02064729e-01 -4.89918292e-01 4.90642339e-01 1.69577941e-01
-2.97608137e-01 5.30578971e-01 -7.92754352e-01 -1.51746005e-01
-7.46663064e-02 -6.66289926e-01 1.14712924e-01 2.11625218e-01
1.60643086e-01 -1.16571116e+00 9.26982641e-01 -9.50613618e-01
1.28971052e+00 -1.66509449e+00 -1.54261783e-01 -4.19098347e-01
6.58457354e-02 5.66488147e-01 -2.66428053e-01 9.01860118e-01
3.18060726e-01 1.93545938e-01 -5.56405187e-02 -1.93742216e-01
2.90962756e-01 2.70181686e-01 -5.07465363e-01 6.72560243e-04
-1.54411614e-01 1.22571385e+00 -9.81898427e-01 -6.85779512e-01
-3.72809827e-01 1.61265537e-01 -4.55880314e-01 2.63027579e-01
-6.27748132e-01 -2.85117239e-01 -5.94709098e-01 6.47364855e-01
1.24032147e-01 -3.58342171e-01 -2.47075409e-01 3.37845802e-01
1.52124450e-01 5.39998412e-01 -7.78191626e-01 1.88207912e+00
-5.32821655e-01 5.88175535e-01 2.33461382e-03 -6.22157812e-01
8.92341018e-01 5.80890298e-01 1.02457389e-01 -1.07939553e+00
-2.47571275e-01 5.68400145e-01 -5.41518033e-02 -7.89669096e-01
5.23412943e-01 -1.52023688e-01 -2.56733268e-01 5.73144555e-01
9.85815674e-02 -2.68439472e-01 7.63527334e-01 4.58189398e-01
1.30783546e+00 -4.99969237e-02 2.70303845e-01 -5.75482905e-01
9.93681014e-01 1.70413941e-01 -1.69908069e-02 9.27081048e-01
-3.33251148e-01 3.57752621e-01 2.27189124e-01 -5.36448121e-01
-1.00162470e+00 -8.84915113e-01 1.11854831e-02 1.61227846e+00
-1.43867299e-01 -5.17864347e-01 -5.99899769e-01 -4.78892118e-01
-1.13753015e-02 8.19355547e-01 -3.83573294e-01 4.97812964e-02
-4.08084214e-01 -1.16149880e-01 6.79961026e-01 1.70498803e-01
7.62789428e-01 -8.47906351e-01 -3.97294253e-01 3.02969366e-01
-7.01606870e-01 -1.39906538e+00 -7.72579610e-01 -8.26134384e-02
-5.91156781e-01 -7.77836978e-01 -1.04478872e+00 -1.04977417e+00
-1.71003327e-01 4.31051522e-01 1.72112751e+00 2.06253633e-01
-4.43069816e-01 6.10552669e-01 -5.67558467e-01 -3.28320861e-01
9.86400545e-02 5.23304939e-01 -2.30265498e-01 -4.12307620e-01
1.00982749e+00 -8.82063732e-02 -7.82866061e-01 1.92921031e-02
-8.39583457e-01 -7.26990819e-01 -3.01803909e-02 4.66806710e-01
3.11784714e-01 -3.90852183e-01 9.32446420e-01 -1.02079117e+00
1.12751806e+00 -7.32968628e-01 -3.90058875e-01 8.56697738e-01
-5.89259028e-01 4.88110602e-01 5.76959968e-01 -3.54535109e-03
-9.41415191e-01 -3.43383491e-01 -4.59502697e-01 8.05667713e-02
9.67969671e-02 5.90955913e-01 4.04064983e-01 -3.00390087e-02
9.32764769e-01 3.38771269e-02 -4.24960375e-01 -7.40181029e-01
6.85976684e-01 5.31628191e-01 2.78096318e-01 -6.15785301e-01
5.65991580e-01 3.59403759e-01 -3.49226773e-01 -8.81142735e-01
-1.25075710e+00 -1.28537309e+00 -3.27317923e-01 -3.33051160e-02
9.24237072e-01 -7.04675794e-01 -7.24017322e-01 -9.00549591e-02
-1.47138155e+00 4.25957263e-01 -1.09793387e-01 -1.03006467e-01
-1.90733954e-01 4.89927709e-01 -5.79169571e-01 -8.92117202e-01
-8.94573390e-01 -6.54003799e-01 1.13838553e+00 1.73954234e-01
-3.67575824e-01 -1.03186989e+00 5.14193416e-01 8.34520161e-01
7.74160385e-01 -3.70754838e-01 1.35903180e+00 -1.01307356e+00
-4.11376417e-01 -5.08061349e-01 -3.21214169e-01 1.96523651e-01
-1.45188749e-01 -6.70569718e-01 -7.54667640e-01 -1.98730007e-01
1.86455205e-01 -8.63143086e-01 9.31343675e-01 -3.39045018e-01
8.04976881e-01 -2.59976566e-01 3.71224135e-02 -3.56981009e-01
1.69290388e+00 -1.07715972e-01 4.49966311e-01 8.37471858e-02
1.42604068e-01 9.48314309e-01 2.16288701e-01 -5.94895631e-02
7.10165501e-01 7.66538918e-01 4.90308516e-02 4.84973997e-01
-4.24426079e-01 -3.30657691e-01 8.09530020e-02 9.46859598e-01
4.37357992e-01 -3.13826770e-01 -1.06707418e+00 1.01539683e+00
-1.77697444e+00 -9.67781126e-01 -3.09562504e-01 2.23463488e+00
8.85573506e-01 -1.56333312e-01 -1.16749249e-01 -2.75505185e-01
3.79308790e-01 2.26619259e-01 -5.72488189e-01 -6.90858781e-01
-5.73583767e-02 8.57874870e-01 2.19509333e-01 9.16385889e-01
-6.62044585e-01 1.20243120e+00 7.19813395e+00 8.59919727e-01
-6.05571508e-01 5.10445893e-01 4.31728750e-01 1.38776135e-02
-3.95099163e-01 1.02347821e-01 -1.11270702e+00 6.41838536e-02
1.01561391e+00 -7.16625631e-01 4.72022474e-01 7.63197660e-01
-3.10301155e-01 -3.18112493e-01 -1.01972854e+00 8.96360338e-01
6.18569970e-01 -1.13545120e+00 2.96195336e-02 -4.65710461e-01
6.52757108e-01 2.73611873e-01 -1.67124301e-01 7.38631546e-01
4.65035290e-01 -1.00392723e+00 9.40211937e-02 5.05706131e-01
5.56887090e-01 -6.42494082e-01 6.84330821e-01 7.79464424e-01
-1.07590902e+00 7.68016130e-02 -4.67903852e-01 -2.02515990e-01
3.19092393e-01 2.37468258e-01 -6.41361475e-01 4.37231839e-01
8.79698813e-01 -1.34143397e-01 -7.70661831e-01 1.08379197e+00
-1.13445811e-01 4.87250745e-01 -4.22834665e-01 -6.06799781e-01
5.75119376e-01 1.47491604e-01 3.17011803e-01 1.01950645e+00
3.21419299e-01 4.02576983e-01 8.29360113e-02 4.68376964e-01
-3.69051516e-01 5.34653306e-01 -6.86530352e-01 -5.62024638e-02
1.77407935e-01 9.58848536e-01 -9.63229388e-02 -4.97688800e-01
-6.36711538e-01 1.25956833e+00 6.04109704e-01 3.55096787e-01
-2.44636297e-01 -7.94214964e-01 2.23925009e-01 9.61179361e-02
2.21761331e-01 -3.77775222e-01 2.82430053e-01 -1.20385742e+00
4.06912357e-01 -1.11833823e+00 9.93611038e-01 -7.74643540e-01
-1.64309716e+00 6.92904055e-01 -2.05419838e-01 -5.95899165e-01
-7.61624992e-01 -6.66704774e-01 -2.36714542e-01 1.27119231e+00
-1.96004903e+00 -9.44483519e-01 -3.57159637e-02 5.49898803e-01
9.62295651e-01 -9.10814106e-02 1.15670395e+00 5.96739352e-01
7.00578615e-02 4.08046514e-01 2.28717953e-01 1.88846812e-02
1.06813920e+00 -1.10843074e+00 4.71564502e-01 6.84529603e-01
5.88783801e-01 5.77159762e-01 6.87068105e-01 -3.07347178e-01
-1.72199404e+00 -5.46187758e-01 2.08940601e+00 -8.54004920e-01
8.82477522e-01 -1.27179310e-01 -1.11774147e+00 2.71556944e-01
7.96088099e-01 -9.87534076e-02 7.49260366e-01 4.25350904e-01
-7.28456736e-01 -3.33524942e-01 -9.50050831e-01 4.66542602e-01
6.78120971e-01 -1.07915485e+00 -1.07682192e+00 7.77590811e-01
1.13601387e+00 -1.96341053e-01 -4.72853154e-01 1.40715927e-01
3.45105827e-01 -4.94367510e-01 9.54761088e-01 -9.59359884e-01
4.13543493e-01 -1.51977330e-01 -4.74379987e-01 -8.23190510e-01
-1.72725037e-01 -5.47412813e-01 -2.09375486e-01 1.04327750e+00
7.18792140e-01 -4.94729728e-01 6.26891136e-01 9.23918843e-01
4.13049221e-01 -6.07160509e-01 -1.35725439e+00 -7.46865094e-01
6.83904648e-01 -5.44164479e-01 2.96842039e-01 5.22575080e-01
2.33605672e-02 8.33433092e-01 -2.56264269e-01 -2.08770141e-01
5.34947753e-01 1.80883203e-02 3.68372053e-01 -9.75408137e-01
-2.36768365e-01 -2.03535333e-01 -1.78095847e-02 -1.47802460e+00
3.47250491e-01 -1.13726079e+00 -1.29489437e-01 -1.88276219e+00
1.87168047e-01 5.34766018e-02 -2.69195378e-01 2.29645997e-01
-2.48045504e-01 4.73648280e-01 9.14672762e-02 6.57216981e-02
-1.09205008e+00 3.34631056e-01 1.07564998e+00 -2.82084763e-01
5.10619760e-01 -1.07929945e-01 -9.02060449e-01 4.12412673e-01
3.94050837e-01 -4.37050641e-01 -4.86003846e-01 -9.79019284e-01
8.36225152e-01 1.19431466e-01 8.42428952e-02 -6.53052509e-01
7.71081448e-01 2.75891006e-01 -3.04625720e-01 -5.22892356e-01
4.16635036e-01 -7.20196784e-01 -7.02812195e-01 3.55697930e-01
-6.01241946e-01 3.66454780e-01 8.97755325e-02 4.22111422e-01
-5.15754759e-01 -6.84318423e-01 3.69977027e-01 -5.41765869e-01
-1.03071797e+00 4.09333766e-01 -1.22068167e-01 9.30683792e-01
6.46766186e-01 4.40525711e-01 -5.09398341e-01 -9.42508936e-01
-2.38818482e-01 8.29137385e-01 -3.78765374e-01 9.05135691e-01
5.86885393e-01 -1.37142599e+00 -1.03941607e+00 -4.88301218e-01
6.19439781e-01 -5.66177905e-01 1.05010450e-01 2.98220873e-01
-7.76763320e-01 1.23710966e+00 4.98100311e-01 -1.88648686e-01
-1.29473853e+00 4.06761169e-01 3.99088383e-01 -5.57311296e-01
-1.21176101e-01 1.06111479e+00 -2.63383031e-01 -6.01686180e-01
2.81874001e-01 8.26443955e-02 -3.83919418e-01 2.00822502e-01
9.73699212e-01 3.89490068e-01 3.36334735e-01 -3.73367131e-01
-2.71114796e-01 6.68299079e-01 -3.66920471e-01 -4.22352403e-01
9.83319819e-01 -2.33094171e-01 -5.84519804e-01 5.96531630e-01
1.47817802e+00 -2.32235402e-01 -1.43979102e-01 -8.37446272e-01
4.63786751e-01 -3.81136268e-01 -1.18412033e-01 -1.07571137e+00
-4.34182912e-01 9.80353892e-01 5.65009832e-01 3.75096619e-01
8.01263750e-01 5.44103086e-01 1.32151353e+00 1.23625970e+00
5.50687909e-01 -1.19414902e+00 1.03623830e-01 1.05497313e+00
1.01263475e+00 -1.45250618e+00 -1.96724340e-01 3.21027227e-02
-4.55089390e-01 8.77806842e-01 3.83046448e-01 -1.89874768e-01
6.47392750e-01 -2.26139650e-01 2.10826442e-01 -4.19468641e-01
-1.08577275e+00 -5.90322375e-01 7.45213509e-01 2.81282365e-01
9.51531887e-01 -2.93248206e-01 -8.24429095e-01 3.34489942e-01
-2.08702058e-01 -5.25311846e-03 -6.89245015e-02 7.93612242e-01
-8.07061255e-01 -1.47169781e+00 2.65408512e-02 3.38622242e-01
-9.44802761e-01 -6.05273128e-01 -5.89536548e-01 4.34581667e-01
-3.10908943e-01 1.33355153e+00 -2.70038158e-01 1.49016455e-01
4.34074461e-01 6.35223329e-01 1.75491244e-01 -9.57110643e-01
-7.20541358e-01 -5.68585336e-01 3.88770491e-01 -3.37266743e-01
-2.11902469e-01 -3.03243250e-01 -8.99026811e-01 -2.79944807e-01
-1.92708999e-01 5.53828299e-01 3.98241699e-01 1.18999052e+00
3.34683061e-01 -9.55506712e-02 2.21470699e-01 3.97776693e-01
-7.63725162e-01 -9.53686595e-01 -1.15333281e-01 2.24774912e-01
3.53853524e-01 4.26169410e-02 -2.60781735e-01 -1.52804911e-01] | [11.427189826965332, 7.931922435760498] |
b1cc2b3e-a9e8-4e03-8f6c-5b681c5b26c5 | primary-object-segmentation-in-videos-via | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Jang_Primary_Object_Segmentation_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Jang_Primary_Object_Segmentation_CVPR_2016_paper.pdf | Primary Object Segmentation in Videos via Alternate Convex Optimization of Foreground and Background Distributions | An unsupervised video object segmentation algorithm, which discovers a primary object in a video sequence automatically, is proposed in this work. We introduce three energies in terms of foreground and background probability distributions: Markov, spatiotemporal, and antagonistic energies. Then, we minimize a hybrid of the three energies to separate a primary object from its background. However, the hybrid energy is nonconvex. Therefore, we develop the alternate convex optimization (ACO) scheme, which decomposes the nonconvex optimization into two quadratic programs. Moreover, we propose the forward-backward strategy, which performs the segmentation sequentially from the first to the last frames and then vice versa, to exploit temporal correlations. Experimental results on extensive datasets demonstrate that the proposed ACO algorithm outperforms the state-of-the-art techniques significantly. | ['Chang-Su Kim', 'Won-Dong Jang', 'Chulwoo Lee'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['unsupervised-video-object-segmentation'] | ['computer-vision'] | [ 2.35932037e-01 -2.21657947e-01 -1.86113641e-01 -8.60076919e-02
-4.88973141e-01 -4.13509041e-01 1.99840993e-01 -1.96268618e-01
-4.67625052e-01 6.03800476e-01 -2.60025561e-01 -1.25323147e-01
-6.40647486e-02 -3.75371009e-01 -6.56255186e-01 -1.23188472e+00
1.45052418e-01 3.59491892e-02 6.09890521e-01 3.46855551e-01
2.16582805e-01 2.41007298e-01 -1.10572970e+00 -7.65981078e-02
1.18587554e+00 1.10916996e+00 3.46437395e-01 3.18597168e-01
-4.11648490e-02 9.69263434e-01 -2.60511726e-01 -1.48002818e-01
3.17561567e-01 -6.68216109e-01 -6.30008221e-01 7.26731360e-01
-1.55539084e-02 -3.47667307e-01 -2.46059880e-01 1.34099305e+00
8.42834339e-02 3.59271377e-01 1.96094617e-01 -1.29557824e+00
-4.62615579e-01 2.07166523e-01 -9.85257864e-01 4.94414806e-01
8.60357285e-02 1.85842231e-01 7.18906641e-01 -9.08033729e-01
4.39554065e-01 1.01650226e+00 2.53309608e-01 2.34631374e-01
-1.08421850e+00 -3.82632881e-01 7.37485766e-01 2.34833449e-01
-1.51247120e+00 -2.22900510e-01 9.82299626e-01 -5.94726562e-01
2.96185464e-01 2.86543459e-01 9.44576740e-01 5.41805148e-01
-7.26813152e-02 1.21335518e+00 1.09389198e+00 -1.15613669e-01
2.02467352e-01 -7.76306540e-02 1.47624329e-01 8.65740478e-01
2.60026366e-01 -1.30256057e-01 -2.35379979e-01 -8.89935642e-02
5.34325957e-01 1.07898600e-01 -5.82875073e-01 -3.95947903e-01
-1.23459470e+00 5.09494901e-01 3.25782001e-02 8.70864317e-02
-6.10171795e-01 -4.59073633e-02 1.17089100e-01 -3.35963875e-01
3.97405386e-01 -1.24105603e-01 -3.43375921e-01 -4.53725532e-02
-1.13715255e+00 1.89284354e-01 6.51743352e-01 1.07381928e+00
6.72790110e-01 -2.20741034e-02 -2.11911514e-01 6.20492101e-01
7.71925926e-01 2.55169928e-01 1.00677505e-01 -9.20879483e-01
5.12784600e-01 7.23388672e-01 2.51132131e-01 -1.20908296e+00
2.23500282e-02 -2.99197555e-01 -6.36018515e-01 -2.25111961e-01
2.08986670e-01 -3.23434114e-01 -9.32019651e-01 1.37898171e+00
7.23550439e-01 7.39025354e-01 -6.74004778e-02 1.29355407e+00
6.22508526e-01 7.93146908e-01 1.20828167e-01 -9.26916242e-01
8.80828083e-01 -1.21293855e+00 -9.69456196e-01 -1.12965122e-01
-7.55499303e-02 -6.01393282e-01 5.01074851e-01 3.05235595e-01
-1.32300138e+00 -3.79735976e-01 -8.98325384e-01 2.58733124e-01
7.38724172e-02 2.25278124e-01 4.19841558e-01 3.70762855e-01
-6.56415105e-01 4.24277812e-01 -1.03169370e+00 5.30663878e-02
5.23089051e-01 4.22934026e-01 1.90073088e-01 1.92617923e-01
-8.43780518e-01 3.71183574e-01 6.63368642e-01 5.59656084e-01
-9.73971426e-01 -4.21183258e-01 -7.34286785e-01 -1.16747968e-01
9.82986271e-01 -4.72347558e-01 1.00844681e+00 -1.27574825e+00
-1.67181122e+00 7.23587632e-01 -3.75017166e-01 -1.95088774e-01
7.36813188e-01 -1.88376546e-01 -1.81867361e-01 2.91221410e-01
-5.47313206e-02 3.38346273e-01 8.59680355e-01 -1.64278746e+00
-9.78046894e-01 -9.23441872e-02 1.14510246e-01 2.19000593e-01
-2.81314611e-01 2.07914725e-01 -1.33089709e+00 -7.39099920e-01
3.32566589e-01 -9.09167647e-01 -4.67265099e-01 -2.32868612e-01
-6.25184953e-01 -3.60202044e-02 9.67504144e-01 -8.91804397e-01
1.68643188e+00 -2.28282356e+00 6.37430847e-01 2.85124809e-01
1.96766749e-01 1.18448108e-01 3.68666768e-01 -1.17735140e-01
1.15858950e-01 7.79936835e-02 -6.73355997e-01 -2.09794685e-01
-2.38530874e-01 2.31552362e-01 5.94346710e-02 6.73413515e-01
1.62839353e-01 6.35128975e-01 -1.04398811e+00 -9.51043963e-01
1.42872244e-01 1.96487844e-01 -5.04343688e-01 3.05867255e-01
-7.07970038e-02 5.89268804e-01 -7.28277147e-01 8.30848992e-01
8.83567154e-01 -2.96860844e-01 2.77653903e-01 -2.98023969e-01
-4.96104419e-01 -1.97840735e-01 -1.41885185e+00 1.41384697e+00
2.75072724e-01 2.26145580e-01 4.57908243e-01 -1.12614310e+00
4.84731436e-01 1.92544594e-01 1.06155694e+00 -1.73428312e-01
3.30551147e-01 1.44327447e-01 -1.55618340e-02 -7.24212229e-01
3.17950279e-01 -7.06212446e-02 2.94716775e-01 -1.16740763e-01
-2.76388168e-01 4.36509460e-01 6.37480617e-01 2.42231227e-02
6.65838122e-01 3.93258870e-01 4.30907160e-02 -3.42969269e-01
9.84298289e-01 6.03458621e-02 1.22676826e+00 3.94224137e-01
-4.91459668e-01 5.78122020e-01 5.76016545e-01 -1.10972390e-01
-4.75584477e-01 -1.11486459e+00 1.36494949e-01 7.17559874e-01
1.00202096e+00 -1.65757328e-01 -9.18674588e-01 -8.53703022e-01
-2.06748232e-01 2.88376749e-01 -3.85763675e-01 6.26400039e-02
-9.23771977e-01 -9.80759144e-01 -7.93510899e-02 4.13680494e-01
6.78853333e-01 -8.61212790e-01 -7.38218248e-01 2.70590901e-01
-4.53181624e-01 -1.23311257e+00 -9.25772667e-01 -1.54815793e-01
-8.39037180e-01 -1.09714866e+00 -9.01998103e-01 -1.07384670e+00
8.82442355e-01 4.66304868e-01 7.33358443e-01 2.68869936e-01
-1.86691154e-02 4.13315803e-01 -2.82690823e-01 -1.20573923e-01
6.62163422e-02 -2.25485578e-01 -1.79040041e-02 6.33465469e-01
-3.36087048e-02 -2.68374860e-01 -7.99282849e-01 4.32760388e-01
-8.92204881e-01 2.49870300e-01 6.12441957e-01 6.04148448e-01
1.25716102e+00 4.32513714e-01 1.90254256e-01 -6.48663044e-01
2.05389023e-01 -6.06839120e-01 -8.58902097e-01 3.92954648e-01
-2.76090145e-01 -4.60895151e-01 3.93037170e-01 -6.05890989e-01
-1.22572422e+00 4.24218357e-01 3.31485242e-01 -8.30868244e-01
9.03879404e-02 5.97162247e-01 -4.27454889e-01 -1.36684686e-01
-3.50835234e-01 6.92842066e-01 -3.75192851e-01 -3.52726310e-01
1.83272481e-01 3.70402455e-01 6.75591469e-01 -5.15729666e-01
9.02987540e-01 7.02285767e-01 -2.81725258e-01 -8.33734751e-01
-7.44543493e-01 -7.28506267e-01 -6.97745323e-01 -5.64275324e-01
1.30294156e+00 -6.93639338e-01 -7.29819536e-01 5.19234836e-01
-1.10385263e+00 -1.51064426e-01 -8.61870646e-02 6.21700943e-01
-4.45244819e-01 6.49134099e-01 -4.59548056e-01 -1.06267297e+00
5.99871483e-03 -1.27552092e+00 8.68468344e-01 6.91039741e-01
3.68479550e-01 -6.84442163e-01 -1.29159108e-01 4.32413936e-01
-2.49231786e-01 5.60512960e-01 4.21494007e-01 -2.83925027e-01
-1.10646975e+00 1.97302550e-01 -3.32086891e-01 3.90947819e-01
2.14424863e-01 4.00650322e-01 -3.87255788e-01 -2.48287842e-01
2.30571300e-01 2.59334922e-01 7.75688410e-01 4.59599853e-01
1.28978705e+00 -4.70172495e-01 -4.87510592e-01 8.26591671e-01
1.46082997e+00 9.03047442e-01 4.76154178e-01 4.21721131e-01
9.31428552e-01 4.66765761e-01 8.27110291e-01 5.09226263e-01
4.57323551e-01 5.37346005e-01 4.74759996e-01 -2.24957839e-01
4.20235068e-01 3.02512702e-02 4.14309442e-01 1.10938632e+00
-3.47273856e-01 -3.01067889e-01 -7.14780927e-01 6.28362715e-01
-2.09027719e+00 -8.75280261e-01 -2.57483572e-01 2.00073004e+00
6.53441966e-01 2.03352764e-01 2.14735523e-01 -8.30488726e-02
1.02119052e+00 1.93986848e-01 -6.32140458e-01 1.00355014e-01
-9.74192247e-02 -2.68281043e-01 5.95228076e-01 1.96690947e-01
-1.46937811e+00 7.49413788e-01 6.04842758e+00 6.75091505e-01
-1.01943183e+00 6.34280741e-02 6.57413602e-01 -4.12388369e-02
-6.12568259e-02 1.66488424e-01 -5.81147373e-01 8.98301065e-01
3.01607966e-01 -2.80455232e-01 6.34357333e-01 7.30462074e-01
5.41057885e-01 -1.56517759e-01 -8.03947270e-01 9.03411388e-01
5.70805697e-03 -1.05605352e+00 -9.34200734e-02 8.66727680e-02
9.54659283e-01 -4.01388228e-01 -1.78609177e-01 1.07364386e-01
-2.50299945e-02 -4.66241747e-01 1.12936401e+00 6.25557899e-01
1.53304070e-01 -7.78245389e-01 4.57387090e-01 3.46227705e-01
-1.57500029e+00 -9.36358720e-02 9.34567451e-02 3.26067477e-01
5.71907222e-01 3.38440031e-01 9.29518491e-02 6.64667428e-01
8.38860929e-01 8.63478065e-01 -2.67883003e-01 1.25790024e+00
-1.12389684e-01 5.94801962e-01 -1.92040920e-01 1.03538379e-01
4.14343566e-01 -8.51129413e-01 9.28115845e-01 1.13464344e+00
-2.07267832e-02 6.73138082e-01 8.44786644e-01 8.36213708e-01
2.07133591e-02 1.83749318e-01 1.05196021e-01 -1.56615917e-02
2.47044176e-01 1.28905320e+00 -1.14237511e+00 -3.36625993e-01
-5.85251153e-01 1.12828469e+00 1.73357073e-02 6.64336443e-01
-1.28432059e+00 -2.58118331e-01 3.33873540e-01 -1.97146758e-01
5.51638901e-01 -2.74557590e-01 -8.57664868e-02 -1.21851921e+00
3.38134438e-01 -7.67659247e-01 4.16759253e-01 -4.89578873e-01
-9.62039292e-01 3.19762647e-01 1.62812054e-01 -1.25211906e+00
4.34419721e-01 -3.22392613e-01 -7.39441276e-01 5.38206935e-01
-1.58689785e+00 -9.02772665e-01 -4.06823426e-01 5.58133304e-01
7.76244044e-01 3.20611924e-01 -2.13762790e-01 5.71524322e-01
-1.18032598e+00 1.12845309e-01 1.44286811e-01 2.72344351e-01
7.53075927e-02 -1.13710773e+00 -3.00893158e-01 1.18410671e+00
-1.52749151e-01 5.36995888e-01 5.10975182e-01 -7.79839158e-01
-1.44174206e+00 -1.11097598e+00 4.02471632e-01 1.56598881e-01
6.12706125e-01 1.49610229e-02 -9.97852504e-01 5.90900421e-01
2.56449819e-01 6.29697964e-02 4.50705171e-01 -6.05331898e-01
3.43433022e-01 -4.96119484e-02 -9.60285842e-01 6.56002402e-01
9.25447524e-01 3.29308957e-02 -4.39268738e-01 2.81346887e-01
7.34464169e-01 -6.50991023e-01 -8.34398925e-01 6.06184125e-01
4.25054759e-01 -8.07331145e-01 9.43471968e-01 -4.07263279e-01
3.38549972e-01 -7.93585062e-01 -1.15251895e-02 -8.08366477e-01
-2.67876357e-01 -8.88169527e-01 -5.95030248e-01 1.47632527e+00
1.18322216e-01 -4.90566880e-01 6.83512688e-01 4.16629195e-01
-1.20558009e-01 -1.18950522e+00 -7.93439269e-01 -9.52405691e-01
-4.32234347e-01 -1.65732279e-01 2.12018088e-01 8.54039609e-01
-3.49970639e-01 -6.06242828e-02 -3.87167186e-01 4.71133977e-01
7.78728187e-01 3.45343173e-01 4.68664616e-01 -8.39389503e-01
-3.30617607e-01 -5.42562544e-01 -1.51238054e-01 -1.38165069e+00
1.10352501e-01 -5.29346347e-01 5.29617786e-01 -1.35614192e+00
5.26455045e-01 -2.73454517e-01 -6.23355746e-01 9.91795510e-02
-6.09928727e-01 -3.82248200e-02 2.67931461e-01 3.99387687e-01
-9.25457001e-01 6.71989441e-01 1.36588967e+00 -3.04498196e-01
-4.83560652e-01 1.10407090e-02 -5.52490890e-01 1.11969721e+00
5.77120304e-01 -5.36868930e-01 -2.46356279e-01 -3.66990954e-01
-4.18587506e-01 9.21789706e-02 2.38692120e-01 -7.88224816e-01
1.40824690e-01 -6.49499178e-01 1.06300488e-01 -8.34411204e-01
1.36667043e-01 -9.95278001e-01 1.75026104e-01 4.73135442e-01
-3.37725542e-02 -8.98846760e-02 6.20552078e-02 8.26515794e-01
-3.30686718e-01 -1.43715665e-01 9.03659999e-01 -3.67816947e-02
-7.45996594e-01 6.21360958e-01 -3.08851212e-01 2.76597202e-01
1.50064147e+00 -3.36745471e-01 1.06942959e-01 7.25842118e-02
-7.90157616e-01 6.66587472e-01 2.89183170e-01 3.57084394e-01
5.40102661e-01 -1.37952852e+00 -5.32320082e-01 -2.39295423e-01
-3.61718893e-01 1.69890016e-01 2.65760690e-01 1.47404087e+00
-6.60086632e-01 1.72003403e-01 2.24428028e-01 -8.05160940e-01
-1.38039124e+00 8.24525952e-01 3.39102417e-01 -2.40832746e-01
-4.58615363e-01 7.09651589e-01 3.25300664e-01 1.79909244e-01
1.59924701e-01 -2.55364984e-01 -4.03650939e-01 1.26221150e-01
1.78462759e-01 7.43523061e-01 -5.15577853e-01 -9.96920884e-01
-7.01514661e-01 7.85829127e-01 1.34481192e-01 2.05097925e-02
1.11475098e+00 -4.14302737e-01 -3.68415505e-01 2.87502527e-01
1.24832129e+00 1.64255581e-03 -1.53227603e+00 -3.42090018e-02
1.69745669e-01 -7.25247681e-01 -2.35411152e-03 -1.59171343e-01
-1.51161110e+00 3.18562150e-01 3.81740510e-01 4.84428108e-01
1.56645298e+00 -1.71489954e-01 9.95769203e-01 -4.49482165e-02
1.79646656e-01 -1.16889405e+00 -5.08968625e-03 3.63801569e-01
5.37622035e-01 -1.06774139e+00 1.05248600e-01 -8.08202386e-01
-6.68084323e-01 7.64074981e-01 8.53159189e-01 -1.10783726e-01
6.73328698e-01 1.37553811e-02 -2.14271590e-01 -3.75932232e-02
-3.95931184e-01 -3.14887345e-01 4.50633675e-01 1.18453540e-01
7.85866007e-02 -1.77628562e-01 -6.84144914e-01 6.40450776e-01
5.20578623e-01 1.30350605e-01 2.51149118e-01 1.20007110e+00
-3.01341653e-01 -6.71234787e-01 -4.85840619e-01 2.41796494e-01
-7.67855108e-01 9.88509655e-02 -3.55759293e-01 6.12171292e-01
5.94835460e-01 1.13018036e+00 -1.00695342e-01 -2.32476801e-01
3.48670423e-01 -2.82207459e-01 2.16201678e-01 -3.76696974e-01
-3.37386608e-01 5.96189439e-01 -3.45788479e-01 -5.03341138e-01
-8.99368107e-01 -8.93958867e-01 -1.42789471e+00 2.13849530e-01
-7.03387737e-01 1.72282562e-01 2.41933241e-01 1.02375948e+00
1.21662067e-02 7.35935986e-01 8.24768066e-01 -9.89232183e-01
-1.23440608e-01 -4.13667828e-01 -3.83412182e-01 1.72154665e-01
2.74567038e-01 -7.10091114e-01 -3.51401001e-01 5.14778018e-01] | [9.129173278808594, -0.4710143804550171] |
99ebfbd5-95cc-4a05-ae49-ecd9ea6ea2ac | holistic-3d-human-and-scene-mesh-estimation | 2012.01591 | null | https://arxiv.org/abs/2012.01591v2 | https://arxiv.org/pdf/2012.01591v2.pdf | Holistic 3D Human and Scene Mesh Estimation from Single View Images | The 3D world limits the human body pose and the human body pose conveys information about the surrounding objects. Indeed, from a single image of a person placed in an indoor scene, we as humans are adept at resolving ambiguities of the human pose and room layout through our knowledge of the physical laws and prior perception of the plausible object and human poses. However, few computer vision models fully leverage this fact. In this work, we propose an end-to-end trainable model that perceives the 3D scene from a single RGB image, estimates the camera pose and the room layout, and reconstructs both human body and object meshes. By imposing a set of comprehensive and sophisticated losses on all aspects of the estimations, we show that our model outperforms existing human body mesh methods and indoor scene reconstruction methods. To the best of our knowledge, this is the first model that outputs both object and human predictions at the mesh level, and performs joint optimization on the scene and human poses. | ['Serena Yeung', 'Zhenzhen Weng'] | 2020-12-02 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Weng_Holistic_3D_Human_and_Scene_Mesh_Estimation_From_Single_View_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Weng_Holistic_3D_Human_and_Scene_Mesh_Estimation_From_Single_View_CVPR_2021_paper.pdf | cvpr-2021-1 | ['indoor-scene-reconstruction'] | ['computer-vision'] | [ 1.14783615e-01 4.82902527e-01 2.70591617e-01 -3.08673322e-01
-3.50969583e-01 -3.72181356e-01 3.67810220e-01 -6.56403005e-02
-2.41435125e-01 2.25787818e-01 1.99194983e-01 2.21680254e-01
1.53908566e-01 -5.30895174e-01 -1.14221573e+00 -2.46464297e-01
2.90044218e-01 1.06805766e+00 2.47121260e-01 -7.90422261e-02
-5.44669591e-02 2.21494868e-01 -1.52498031e+00 -1.89522699e-01
5.09253561e-01 9.64765191e-01 4.09231633e-01 9.88642335e-01
4.98373717e-01 7.29579926e-01 -1.30373806e-01 -4.54238117e-01
4.46545333e-01 -1.17348939e-01 -6.26613855e-01 7.63457775e-01
8.24096620e-01 -5.94238460e-01 -5.04618704e-01 8.13407719e-01
4.29009795e-01 1.26052454e-01 4.72545624e-01 -8.90299916e-01
-2.35878423e-01 -8.04410782e-03 -4.72013563e-01 -4.29280013e-01
9.36749578e-01 2.62202054e-01 7.26997614e-01 -1.00469220e+00
7.23456025e-01 1.35811794e+00 9.70735788e-01 4.07204658e-01
-1.20811617e+00 -1.95873559e-01 5.47977626e-01 -1.02266662e-01
-1.49314988e+00 -4.46749955e-01 8.11571479e-01 -6.68626428e-01
6.56205893e-01 1.39383584e-01 1.15803075e+00 1.14569092e+00
1.33585736e-01 5.31696320e-01 9.26945925e-01 -4.76297826e-01
2.45634407e-01 4.04964946e-03 -3.35888684e-01 1.20921445e+00
1.65349826e-01 -1.52295724e-01 -7.70003974e-01 4.45805155e-02
1.21677768e+00 1.39346063e-01 -3.17808568e-01 -1.12251186e+00
-1.38904905e+00 1.55234620e-01 4.71360892e-01 -3.02348733e-01
-3.90585870e-01 3.37419599e-01 -2.63695687e-01 -3.34030241e-01
2.69575477e-01 1.92932174e-01 -4.60190356e-01 8.18608887e-03
-6.32040679e-01 3.56406569e-01 8.55233788e-01 1.10850465e+00
6.53173089e-01 -3.84930968e-01 2.19128907e-01 2.74189770e-01
7.31977105e-01 8.68654251e-01 -3.55125725e-01 -1.38853574e+00
5.56059241e-01 4.72705156e-01 4.50259030e-01 -1.06864130e+00
-6.01325095e-01 -2.48346716e-01 -5.03146231e-01 3.35991494e-02
5.85932791e-01 1.07364086e-02 -1.12189698e+00 1.58460665e+00
7.92939603e-01 2.72216678e-01 -4.58301574e-01 1.31866658e+00
6.23112857e-01 2.25481898e-01 -1.93162858e-01 2.50654042e-01
1.37675059e+00 -8.02012324e-01 -5.45677662e-01 -8.83841097e-01
-1.39526904e-01 -7.35278606e-01 9.51596379e-01 5.34571826e-01
-1.43766475e+00 -7.72578895e-01 -8.96017909e-01 -5.13554394e-01
7.17852712e-02 1.55577064e-01 5.62298298e-01 4.82949615e-01
-7.31649160e-01 4.43565756e-01 -1.26629877e+00 -6.97625339e-01
7.52821937e-02 4.48023349e-01 -3.33106697e-01 -2.00528488e-01
-5.44083536e-01 1.22374904e+00 -1.19336918e-01 5.44299424e-01
-1.00061023e+00 -4.39637899e-01 -1.09163761e+00 -3.61323476e-01
6.54125810e-01 -1.49255037e+00 1.37507904e+00 -5.93000710e-01
-1.54263365e+00 1.06410456e+00 -1.14071257e-01 -1.55256674e-01
7.34147966e-01 -6.83333457e-01 3.01147133e-01 7.97957256e-02
-2.27817491e-01 5.47331095e-01 8.76116514e-01 -1.59082270e+00
-1.71266407e-01 -8.91893148e-01 1.29071251e-01 5.49780250e-01
3.49354625e-01 -6.30893350e-01 -9.06839013e-01 -2.50301510e-01
5.49756944e-01 -1.08579445e+00 -3.90567482e-01 5.39889932e-01
-6.51065826e-01 4.75958645e-01 1.06847502e-01 -8.27169657e-01
4.01065975e-01 -1.81580532e+00 8.10245335e-01 4.15411413e-01
3.02317262e-01 -4.41891491e-01 3.84209275e-01 1.64654851e-01
5.76928794e-01 -2.98250318e-01 -6.22124001e-02 -1.08271968e+00
2.64124423e-01 4.50114071e-01 -1.28707383e-02 9.11278486e-01
-2.70563960e-01 7.52401888e-01 -8.20502460e-01 -5.18449306e-01
6.76441967e-01 9.34591651e-01 -7.41670072e-01 5.68416357e-01
-2.89005548e-01 9.25719440e-01 -4.52154934e-01 5.95807314e-01
6.35534525e-01 -3.31044227e-01 4.38680083e-01 -4.37753856e-01
1.69649556e-01 2.66087025e-01 -1.37500691e+00 2.30981326e+00
-5.00644207e-01 8.44907388e-02 4.10636067e-01 -3.97860557e-01
3.37725282e-01 3.10655415e-01 4.71149057e-01 -4.22106624e-01
2.03730434e-01 -1.57530531e-02 -4.78680730e-01 -4.90690261e-01
2.71551818e-01 -2.96574920e-01 -8.02348405e-02 -3.08960434e-02
-1.60382148e-02 -6.08400881e-01 -5.08416653e-01 -5.28275780e-02
8.23295891e-01 7.31709599e-01 2.62039453e-01 1.37326971e-01
4.47529033e-02 -2.75538146e-01 3.20292205e-01 5.91346502e-01
1.48513123e-01 9.44669008e-01 -5.00404276e-02 -6.10316873e-01
-1.19589829e+00 -1.59678495e+00 2.00799525e-01 8.68695080e-01
4.65553015e-01 -4.16410416e-01 -7.74207771e-01 -1.13292195e-01
1.52069986e-01 3.91469449e-01 -6.21179461e-01 7.77221173e-02
-6.12013757e-01 -1.46109492e-01 -3.75595316e-02 5.60100853e-01
2.76565313e-01 -5.13475657e-01 -1.09135127e+00 -6.46909326e-02
-5.45504332e-01 -1.57769775e+00 -5.18316031e-01 -5.50638773e-02
-7.23169088e-01 -1.20500338e+00 -3.74464393e-01 -6.05650723e-01
9.10091639e-01 -8.61139894e-02 1.32288098e+00 3.61895561e-02
-5.33994853e-01 9.32564974e-01 -8.79353732e-02 -3.91337305e-01
-2.25936621e-02 -3.94370615e-01 2.57829756e-01 -3.67129082e-03
-3.00482512e-01 -7.55359590e-01 -7.61675715e-01 3.11046094e-01
-8.21540654e-02 5.65648198e-01 2.97584623e-01 2.64933169e-01
8.52124631e-01 -6.60455599e-02 -7.38544941e-01 -6.92717433e-01
-3.31949800e-01 -3.32408212e-02 -5.43700695e-01 5.93401566e-02
-1.11578681e-01 -1.17228955e-01 2.53105730e-01 -2.82876700e-01
-1.11073554e+00 7.03965783e-01 -8.45006201e-03 -5.80045938e-01
-4.14435804e-01 -5.43367900e-02 -4.96572256e-01 6.61826432e-02
4.89805877e-01 -1.39121890e-01 -2.63551950e-01 -7.03277588e-01
4.38150585e-01 -5.20814024e-02 9.76913750e-01 -7.60565281e-01
1.03223765e+00 8.43424678e-01 3.17035645e-01 -7.05887318e-01
-1.23229694e+00 -5.56585610e-01 -1.29862607e+00 -4.19974983e-01
1.24910510e+00 -1.14452302e+00 -1.05246615e+00 3.82674485e-01
-1.25716650e+00 -5.48742652e-01 -3.05559456e-01 5.00649750e-01
-9.68929589e-01 2.76270330e-01 -5.02562642e-01 -1.01662016e+00
1.36323094e-01 -9.98738110e-01 1.78333867e+00 -2.73765117e-01
-4.26056057e-01 -6.84980035e-01 -1.38400331e-01 6.91872418e-01
-2.08876804e-01 6.28618777e-01 3.80515188e-01 3.22662950e-01
-9.78063345e-01 -9.08256173e-02 3.24506223e-01 -1.65532112e-01
-3.91903780e-02 -3.15215826e-01 -1.09968448e+00 -1.65874943e-01
1.58028137e-02 -2.78983235e-01 4.43296432e-01 6.23311520e-01
1.02266872e+00 -4.27556217e-01 -3.08833569e-01 9.22158659e-01
1.23901093e+00 -4.43395078e-01 3.61284256e-01 1.91562653e-01
1.20935810e+00 6.94219828e-01 3.67396772e-01 7.18193352e-01
8.69969308e-01 8.35864425e-01 8.06037426e-01 -2.38306656e-01
-9.59658846e-02 -6.67536855e-01 9.81887057e-02 6.26932502e-01
-4.41529095e-01 -1.19281806e-01 -1.04580736e+00 1.83595315e-01
-1.83240235e+00 -6.20662868e-01 -3.63003239e-02 2.34371662e+00
6.10061467e-01 2.31204912e-01 1.36584416e-01 -1.19281463e-01
2.85492986e-01 -4.61396836e-02 -7.95082390e-01 1.78951442e-01
2.45208278e-01 3.01247817e-02 5.78972876e-01 8.06273818e-01
-1.05653417e+00 7.50574052e-01 6.17953253e+00 -1.71572819e-01
-4.72599298e-01 1.38692986e-02 2.18320325e-01 -3.09502840e-01
-4.04687896e-02 6.17044158e-02 -7.35257924e-01 4.67891991e-02
4.73027498e-01 4.40348119e-01 7.38478184e-01 7.13617802e-01
-2.32815854e-02 -1.57897681e-01 -1.60152984e+00 1.15369737e+00
3.02965343e-01 -7.85744488e-01 -2.69107372e-01 2.84354120e-01
4.33381587e-01 -1.39182866e-01 5.39954230e-02 -1.19741291e-01
3.29601258e-01 -1.03893077e+00 1.47762847e+00 1.07247698e+00
4.07221913e-01 -3.68000478e-01 4.30711985e-01 8.04736495e-01
-1.20266831e+00 -5.95229270e-04 -1.09929591e-01 -3.43864977e-01
4.52326655e-01 4.04728472e-01 -7.75339127e-01 2.26945058e-01
7.48273909e-01 5.99535286e-01 -5.99472761e-01 8.09649289e-01
-3.06924701e-01 1.52473643e-01 -5.63972354e-01 6.00903392e-01
-4.45245624e-01 -1.08403690e-01 4.62914109e-01 8.05124521e-01
1.79685056e-01 4.46696848e-01 8.05784285e-01 1.02902913e+00
9.63326469e-02 -3.18071961e-01 -3.76476049e-01 4.57490623e-01
1.25515342e-01 9.47673023e-01 -5.97515881e-01 -1.96920038e-04
-1.08226679e-01 1.18612814e+00 3.23183924e-01 3.94852042e-01
-8.54660809e-01 4.86412913e-01 4.61010426e-01 6.02636695e-01
2.91146785e-01 -6.05948746e-01 -3.79604101e-01 -1.33495700e+00
3.41410667e-01 -6.72022700e-01 2.85546966e-02 -1.21295726e+00
-9.44319487e-01 1.74025729e-01 1.12291453e-02 -9.13882434e-01
-1.23556025e-01 -6.87994540e-01 4.67407852e-02 6.63728356e-01
-9.06162202e-01 -1.48610079e+00 -6.06082141e-01 5.82421362e-01
5.91198027e-01 6.35188222e-01 7.55922377e-01 -1.22249916e-01
-2.36967385e-01 1.46584615e-01 -5.32611549e-01 1.89129263e-02
4.95952696e-01 -1.42512286e+00 4.74777907e-01 6.39255464e-01
1.72361985e-01 7.34948754e-01 1.12628961e+00 -6.67583227e-01
-1.98668826e+00 -7.50754178e-01 5.80815375e-01 -1.20204008e+00
8.03370029e-02 -9.08652484e-01 -3.91026676e-01 1.09141958e+00
-2.32879683e-01 1.44822881e-01 2.43408754e-01 2.74115920e-01
-2.83924848e-01 7.04750419e-02 -9.94135678e-01 5.91573477e-01
1.49417806e+00 -5.95662355e-01 -5.45314610e-01 4.71078634e-01
6.10906005e-01 -1.28130162e+00 -9.36153889e-01 4.20023471e-01
8.94701719e-01 -1.01701069e+00 1.60501862e+00 -1.85111105e-01
2.11566865e-01 -4.08677071e-01 -5.60816109e-01 -9.34968114e-01
-3.18393886e-01 -3.52100968e-01 -4.30699766e-01 6.32423520e-01
1.79954097e-02 1.47487987e-02 8.78548443e-01 1.05008590e+00
-8.33404437e-02 -5.47988415e-01 -9.47984159e-01 -3.58026594e-01
-4.34220970e-01 -4.85183179e-01 3.49982619e-01 4.38330740e-01
-3.75386387e-01 2.52798438e-01 -7.31375813e-01 8.67202938e-01
1.12436998e+00 2.41056859e-01 1.26855350e+00 -1.46432316e+00
-7.06329525e-01 3.60717624e-02 -3.46956551e-01 -1.35097992e+00
1.14937887e-01 -2.80169487e-01 4.69225913e-01 -1.78164947e+00
2.90053576e-01 -5.18293716e-02 3.27258825e-01 2.72340953e-01
-2.48365458e-02 1.75611883e-01 3.06126684e-01 5.84078319e-02
-8.59305501e-01 3.38591963e-01 1.36934876e+00 8.52784738e-02
1.49795031e-02 6.95164949e-02 -2.88642973e-01 1.39646804e+00
3.68769526e-01 -1.19090229e-01 -2.68727064e-01 -8.28103125e-01
5.07571936e-01 1.55465081e-01 1.00070691e+00 -1.04052365e+00
3.64133149e-01 -2.61660129e-01 1.00762844e+00 -7.10426331e-01
1.03707790e+00 -1.06491673e+00 4.48881805e-01 3.75146896e-01
-1.44002438e-01 -1.82149321e-01 2.15999465e-02 8.16435754e-01
5.96766055e-01 4.19564754e-01 5.65183163e-01 -5.62814415e-01
-4.59702760e-01 3.80062491e-01 6.20797910e-02 -2.46969312e-02
5.33124208e-01 -4.56180811e-01 2.11976349e-01 -3.93355489e-01
-9.68068361e-01 2.56938070e-01 1.04096663e+00 3.23886961e-01
7.44569182e-01 -1.02224958e+00 -3.35457295e-01 3.74375939e-01
-4.50094938e-02 6.61035895e-01 2.42093384e-01 7.78417051e-01
-5.63858271e-01 1.59772307e-01 4.66495082e-02 -8.78181458e-01
-1.32194829e+00 5.99844873e-01 6.59689128e-01 1.21206187e-01
-8.71140301e-01 8.60503078e-01 4.62891996e-01 -9.17312264e-01
5.16792178e-01 -5.33995926e-01 4.12858814e-01 -6.27194643e-01
1.24571808e-01 4.63361442e-01 -1.34698465e-01 -1.08398509e+00
-4.23015147e-01 1.24000645e+00 5.17018199e-01 -2.08799034e-01
1.08053327e+00 -6.11685276e-01 -3.34145938e-04 6.78640902e-01
8.51879656e-01 2.26284981e-01 -1.68519413e+00 -2.02349544e-01
-5.50043106e-01 -6.04923785e-01 -1.28670290e-01 -7.81758487e-01
-5.59571862e-01 9.12164867e-01 3.59527469e-01 -5.35247326e-01
9.23083365e-01 4.03246999e-01 7.39715576e-01 4.67744231e-01
9.24817979e-01 -1.10146010e+00 2.69782752e-01 3.67767900e-01
8.13889563e-01 -1.09194660e+00 3.64585489e-01 -8.18301976e-01
-4.50618476e-01 6.94633782e-01 8.00577402e-01 -4.76938337e-02
5.93016028e-01 3.02713394e-01 -2.32829571e-01 -3.51114243e-01
-3.87505710e-01 -2.00370669e-01 6.17577136e-01 7.10674345e-01
1.19131543e-01 2.51793832e-01 6.38881087e-01 2.87931889e-01
-7.44354248e-01 -1.45471707e-01 8.45407471e-02 8.62497807e-01
-5.64877629e-01 -6.43332720e-01 -7.54847646e-01 -1.02733843e-01
-2.61269331e-01 3.09452951e-01 -5.04117787e-01 5.53019762e-01
6.05659783e-01 6.81554854e-01 -1.54467225e-02 -2.42714599e-01
8.82080793e-01 -9.51587409e-02 1.20283902e+00 -7.58317173e-01
-1.24409445e-01 2.13068262e-01 -2.16995496e-02 -8.60233784e-01
-4.27280039e-01 -5.85429966e-01 -1.38869965e+00 -1.79252073e-01
-1.52832091e-01 -4.72647935e-01 6.85541391e-01 1.00374913e+00
3.39176953e-02 4.05851126e-01 1.48807317e-01 -1.70564914e+00
-2.93588787e-01 -5.73086917e-01 -6.75605237e-01 6.91622853e-01
5.26152730e-01 -8.14110398e-01 -2.31283456e-01 4.24478263e-01] | [7.036870956420898, -1.1905986070632935] |
0561bcba-6f89-42bd-aadf-1b65a9c413d5 | pushing-the-envelope-from-discrete-to | 2004.13477 | null | https://arxiv.org/abs/2004.13477v1 | https://arxiv.org/pdf/2004.13477v1.pdf | Pushing the Envelope: From Discrete to Continuous Movements in Multi-Agent Path Finding via Lazy Encodings | Multi-agent path finding in continuous space and time with geometric agents MAPF$^\mathcal{R}$ is addressed in this paper. The task is to navigate agents that move smoothly between predefined positions to their individual goals so that they do not collide. We introduce a novel solving approach for obtaining makespan optimal solutions called SMT-CBS$^\mathcal{R}$ based on {\em satisfiability modulo theories} (SMT). The new algorithm combines collision resolution known from conflict-based search (CBS) with previous generation of incomplete SAT encodings on top of a novel scheme for selecting decision variables in a potentially uncountable search space. We experimentally compare SMT-CBS$^\mathcal{R}$ and previous CCBS algorithm for MAPF$^\mathcal{R}$. | ['Pavel Surynek'] | 2020-04-25 | null | null | null | null | ['multi-agent-path-finding'] | ['playing-games'] | [ 3.45173687e-01 3.54662120e-01 1.05753034e-01 6.19560182e-02
-7.68189907e-01 -8.06297362e-01 6.29404634e-02 4.43028897e-01
-5.86673200e-01 1.52379417e+00 -7.08579242e-01 -6.66435838e-01
-1.18566275e+00 -1.30741370e+00 -5.93125641e-01 -6.49382889e-01
-6.48628831e-01 1.24330795e+00 5.59602499e-01 -6.11536980e-01
3.74273211e-01 4.31487143e-01 -1.33158743e+00 -1.06056012e-01
9.68171954e-01 7.56838858e-01 3.29933196e-01 8.15889716e-01
-1.79644629e-01 4.22125131e-01 -6.21581972e-01 -2.00925246e-01
5.35806239e-01 -5.07079482e-01 -1.29323304e+00 -1.84207365e-01
-2.68506259e-01 1.28586357e-02 9.71601531e-02 1.24077964e+00
-3.39919305e-03 4.21281010e-01 5.48779368e-01 -1.97695625e+00
1.29267648e-01 6.53726935e-01 -6.21354640e-01 2.45314628e-01
6.40129507e-01 1.15605533e-01 8.17557335e-01 8.76273885e-02
8.82362127e-01 1.13155353e+00 3.23250651e-01 2.97015220e-01
-1.05618751e+00 -5.58865488e-01 3.74601424e-01 4.85380054e-01
-1.79277432e+00 1.81183040e-01 2.39875928e-01 -6.70777559e-02
1.61260915e+00 7.84680128e-01 6.44258380e-01 8.03013816e-02
4.95717049e-01 1.34362981e-01 1.13748968e+00 -6.56379580e-01
6.50659263e-01 -3.44429433e-01 -9.42006633e-02 7.41137385e-01
7.36621678e-01 3.13394129e-01 -1.24126256e-01 -7.15379119e-02
6.81679249e-01 -4.25176352e-01 4.81973439e-02 -1.05875701e-01
-9.80007827e-01 1.10103047e+00 7.81397149e-03 2.16148466e-01
-1.89491287e-01 4.32757944e-01 2.72351392e-02 4.26505744e-01
-1.72233447e-01 7.31148183e-01 -4.33915555e-01 -1.50580168e-01
-7.08087981e-01 8.47090065e-01 7.57538080e-01 1.52511966e+00
6.49078190e-01 -1.66862637e-01 3.04898113e-01 -3.01498622e-02
-2.06002258e-02 7.85713434e-01 -4.74966764e-01 -1.38874114e+00
7.48361230e-01 6.57151937e-01 5.31936705e-01 -7.96014965e-01
-8.27050030e-01 -2.12563440e-01 -3.61364186e-01 7.42769539e-01
4.92457718e-01 -3.19555879e-01 -6.59031332e-01 1.62462795e+00
4.23268437e-01 -1.60008892e-01 2.58705854e-01 6.72303557e-01
1.46131709e-01 8.96411896e-01 -3.44194174e-01 -8.31342578e-01
9.82891619e-01 -9.29204762e-01 -2.99045533e-01 -1.35409340e-01
8.01287770e-01 -4.04337436e-01 1.37827083e-01 7.65281379e-01
-1.70311093e+00 1.96027637e-01 -1.17566240e+00 4.93191838e-01
-3.46831650e-01 -8.01414371e-01 5.99366546e-01 7.11637139e-01
-1.19946134e+00 1.98607534e-01 -6.54643178e-01 -8.25829506e-02
9.58808959e-02 9.26730275e-01 -1.22504406e-01 -4.26153451e-01
-9.42112505e-01 8.75160158e-01 5.66038430e-01 -1.51710570e-01
-8.15733075e-01 5.48084117e-02 -1.01106691e+00 -2.02372357e-01
1.33590174e+00 -5.41369379e-01 1.19368231e+00 -2.69330800e-01
-1.17376745e+00 4.14079785e-01 -3.93120021e-01 -5.66999316e-01
4.69845414e-01 6.50188744e-01 -1.88971564e-01 3.47955734e-01
3.83725524e-01 3.91016096e-01 3.22166495e-02 -1.26660883e+00
-1.17458379e+00 -3.90186816e-01 7.40329623e-01 1.58384219e-01
5.98182023e-01 1.53838620e-01 -1.04918346e-01 4.57181334e-02
4.58070874e-01 -1.07564437e+00 -9.33111668e-01 -6.34818435e-01
-3.97229582e-01 -2.26581126e-01 1.24048598e-01 -3.94794606e-02
1.32914710e+00 -1.19445956e+00 6.35953009e-01 8.93195093e-01
8.67950767e-02 -1.36173472e-01 -2.70959765e-01 8.79761875e-01
2.95657098e-01 2.10646167e-01 -6.47835508e-02 1.48780167e-01
1.30306140e-01 5.70346117e-01 1.58213258e-01 3.57391745e-01
-1.67221949e-01 6.01697385e-01 -9.41767335e-01 -5.31282365e-01
1.29078105e-01 -5.33573985e-01 -6.79273307e-01 -4.10106182e-01
-8.76445115e-01 2.59806275e-01 -7.32858658e-01 6.30213857e-01
1.13250339e+00 1.65298522e-01 4.21031475e-01 7.31628239e-01
-7.68980443e-01 4.64852229e-02 -1.68654144e+00 1.45927632e+00
9.28829908e-02 -1.39522359e-01 5.36217809e-01 -9.34829772e-01
6.37093008e-01 -8.21174085e-02 6.40739322e-01 -1.07164216e+00
5.23157835e-01 3.83264959e-01 -2.52982676e-02 -1.58186778e-01
6.32494986e-01 -2.72107750e-01 -6.40251219e-01 5.67197084e-01
-5.99474609e-01 -4.92359459e-01 7.74948359e-01 2.16591522e-01
1.45510161e+00 -1.59554273e-01 2.52718419e-01 -4.48988378e-01
6.66982114e-01 8.41334879e-01 7.26644695e-01 1.04274750e+00
-1.22698702e-01 -1.53988808e-01 6.11930609e-01 -5.40705144e-01
-8.69564593e-01 -8.70134473e-01 3.34399283e-01 6.46266520e-01
9.02019262e-01 -5.00506341e-01 -8.02953124e-01 -2.55465686e-01
-3.49433333e-01 9.73989248e-01 -4.44101751e-01 8.93763378e-02
-8.31201613e-01 -3.36848348e-01 3.56499612e-01 1.85794502e-01
2.65653670e-01 -9.04988825e-01 -1.21172380e+00 4.79214609e-01
-1.90569401e-01 -1.02561295e+00 -5.22326445e-03 4.70451146e-01
-3.49159837e-01 -1.26165652e+00 -5.70374401e-03 -5.91888845e-01
9.21488762e-01 4.88395989e-01 6.49304092e-01 3.48929077e-01
-3.87073487e-01 8.40693787e-02 -6.55677438e-01 -5.63918293e-01
-2.06740186e-01 8.84568226e-03 -4.11998220e-02 -1.00545835e+00
2.44777307e-01 -3.64927262e-01 -2.99903870e-01 5.27062297e-01
-6.99665606e-01 1.39893726e-01 5.37943095e-02 2.92765498e-01
9.41873074e-01 1.08959401e+00 2.66230583e-01 -3.28417838e-01
5.08682430e-01 -3.09993237e-01 -1.32413757e+00 2.24540666e-01
-3.68795812e-01 6.18329085e-02 7.03355432e-01 2.78436076e-02
-5.16872108e-01 -9.04839933e-02 2.71260083e-01 -3.87446284e-02
-8.93758908e-02 5.75965822e-01 -2.73622833e-02 -9.95298102e-02
3.21298808e-01 1.34873226e-01 -2.61503816e-01 1.71193779e-01
4.65241186e-02 1.78730357e-02 4.57650721e-01 -9.20471549e-01
7.80553997e-01 3.56209785e-01 8.35348487e-01 -6.81620613e-02
-4.76750955e-02 -7.00600445e-02 -2.58634269e-01 -9.10846740e-02
6.19578481e-01 -1.93014935e-01 -1.42782593e+00 -4.49103303e-02
-9.18003738e-01 -6.62012994e-01 -2.67209530e-01 3.37100685e-01
-9.79347527e-01 1.55209646e-01 -1.78534284e-01 -1.52006841e+00
9.97469723e-02 -1.29328156e+00 7.93438852e-01 3.00626040e-01
-2.63143986e-01 -6.66067958e-01 5.77686839e-02 3.12898397e-01
2.34182760e-01 6.89463079e-01 9.89093244e-01 -4.34984148e-01
-1.03088832e+00 -9.68767926e-02 -1.52920917e-01 -6.47972167e-01
-1.67757049e-01 -2.38144293e-01 1.55047685e-01 -5.56310356e-01
-2.77615547e-01 1.40037358e-01 8.64349455e-02 6.33348584e-01
6.28165364e-01 -5.92358708e-01 -7.57616758e-01 8.45968630e-03
1.77718675e+00 1.03725064e+00 6.33899212e-01 9.36018586e-01
-3.72623324e-01 5.42662919e-01 9.76097167e-01 6.11497343e-01
7.08420873e-01 7.61632860e-01 9.15900230e-01 4.75054353e-01
5.27453184e-01 3.97245198e-01 3.47322412e-02 -5.87234274e-02
-4.32250261e-01 -7.48108268e-01 -9.87322152e-01 6.56379402e-01
-1.79931605e+00 -1.07568681e+00 -3.20620358e-01 2.15938401e+00
5.93771100e-01 5.02196968e-01 3.17042053e-01 4.76021439e-01
8.00529540e-01 -3.65210354e-01 -2.75021553e-01 -1.06138170e+00
-5.43689951e-02 6.83383703e-01 9.80336905e-01 1.08728909e+00
-5.00753582e-01 9.59946811e-01 4.98849916e+00 7.74399161e-01
-5.27611196e-01 -5.99726215e-02 -2.79080626e-02 -3.65013033e-01
-3.22395146e-01 3.64401102e-01 -6.45140111e-01 2.16387242e-01
8.05475831e-01 -4.10796285e-01 9.12640512e-01 5.01813769e-01
3.32086831e-01 -7.51847029e-01 -8.59073877e-01 4.40079361e-01
-2.54598022e-01 -1.33112943e+00 -4.42046016e-01 2.29029179e-01
1.00386345e+00 -4.82799411e-01 -7.50622526e-03 -3.18763149e-03
9.39892650e-01 -1.19455242e+00 9.36848521e-01 -1.44612610e-01
7.50990033e-01 -1.22127998e+00 6.99725449e-01 6.07492626e-01
-1.54443872e+00 -3.93053263e-01 -1.80436179e-01 -5.47172844e-01
6.81199729e-01 -1.06985718e-01 -8.21782172e-01 1.17577779e+00
6.03340089e-01 -2.79474318e-01 3.56625557e-01 1.19324875e+00
-1.64878383e-01 -3.26152921e-01 -6.29546940e-01 -4.02812660e-01
7.14895904e-01 -2.93233871e-01 9.07826126e-01 6.77558661e-01
5.12353420e-01 9.12310481e-01 5.59777081e-01 9.09889877e-01
7.61951268e-01 -2.77150005e-01 -2.11543530e-01 1.59228787e-01
7.29777098e-01 6.68664217e-01 -1.31729722e+00 2.77847610e-02
1.83575049e-01 4.29967791e-01 -1.66332901e-01 1.63876534e-01
-9.82642472e-01 -5.74329019e-01 5.99900007e-01 1.82259958e-02
3.01150471e-01 -5.55701673e-01 -3.89378458e-01 -2.94363111e-01
-9.80101600e-02 -7.19886184e-01 6.73333883e-01 -6.95590436e-01
-4.41298515e-01 7.92633593e-01 5.81532478e-01 -8.85846555e-01
-4.04025823e-01 -4.01874691e-01 -4.64949578e-01 6.78859174e-01
-1.49635077e+00 -6.85364127e-01 -2.01431364e-01 7.93089509e-01
4.73364413e-01 -5.83608560e-02 8.35028529e-01 -3.31095964e-01
-3.43333215e-01 3.99865031e-01 -1.77428931e-01 -5.83041131e-01
-4.29158956e-01 -8.74962270e-01 8.52028430e-02 9.06992376e-01
-6.49526179e-01 4.20892209e-01 1.04360163e+00 -8.03346574e-01
-1.65446115e+00 -7.99154997e-01 6.46789014e-01 6.71510994e-02
3.34619701e-01 1.84952810e-01 2.07790756e-03 6.65727854e-01
1.22274980e-01 -6.93549156e-01 2.26533547e-01 -5.02256334e-01
2.22964361e-02 6.90037757e-02 -1.71860814e+00 7.05268264e-01
1.23494864e+00 3.49190652e-01 -3.20978612e-01 2.81924695e-01
7.15553224e-01 -7.21825361e-01 -4.31135982e-01 4.36601490e-01
2.20444389e-02 -6.85332298e-01 7.28979111e-01 -5.18259227e-01
-1.81533098e-02 -7.43679345e-01 -4.03533041e-01 -1.06744957e+00
-2.76330769e-01 -1.07494593e+00 5.22770882e-01 4.00721908e-01
5.56457937e-01 -8.77744317e-01 8.53115022e-01 7.04676867e-01
-4.41032976e-01 -7.67142415e-01 -1.68812811e+00 -1.14476037e+00
4.06974316e-01 -4.81465250e-01 9.47080910e-01 6.86966658e-01
6.51815355e-01 -2.75852799e-01 -7.81215727e-02 4.65022534e-01
8.00178766e-01 2.94582278e-01 4.82500821e-01 -1.07705796e+00
-3.47906262e-01 -6.31175458e-01 -2.04889148e-01 -4.55019653e-01
-2.51796842e-02 -6.22786880e-01 6.92692697e-02 -2.06428409e+00
5.45474701e-02 -1.05705392e+00 8.55856091e-02 7.19949961e-01
7.15856254e-01 -1.54552579e-01 2.52705038e-01 -3.42288017e-01
-1.06088555e+00 -8.48061368e-02 1.48709404e+00 -1.68400958e-01
-2.48557299e-01 3.27243144e-03 -6.11487269e-01 4.31470573e-01
8.45395148e-01 -5.60120225e-01 -6.50394082e-01 -2.92503238e-01
8.37839305e-01 1.03284454e+00 2.44490504e-02 -8.41036379e-01
5.38240612e-01 -1.25592661e+00 -5.39315224e-01 -6.32574797e-01
5.52396715e-01 -9.67313409e-01 7.57058203e-01 8.74148250e-01
-9.34099108e-02 5.16106665e-01 5.04648387e-01 2.59493709e-01
3.04378122e-01 -5.77315748e-01 3.94676745e-01 -3.08851629e-01
-6.07605875e-01 -1.07644230e-01 -6.95098758e-01 -1.84915945e-01
1.90853333e+00 -6.74431443e-01 -7.59087861e-01 -4.66224134e-01
-7.12886691e-01 9.00654376e-01 3.97300541e-01 -8.10410827e-02
6.46749616e-01 -7.96990871e-01 -5.44182301e-01 -1.55324638e-01
-1.76388398e-01 4.69995588e-01 3.59899700e-01 9.18624103e-01
-8.51508260e-01 8.84656608e-01 -4.40609038e-01 1.40683334e-02
-1.22171223e+00 7.72058129e-01 3.37826669e-01 -4.73410755e-01
-2.22254694e-01 1.11660075e+00 -4.88200516e-01 -5.55504486e-02
2.70101782e-02 -4.48827237e-01 6.56783730e-02 -4.82875437e-01
4.30732429e-01 8.45370591e-01 -5.62422797e-02 -5.45388281e-01
-6.77066147e-01 5.32868564e-01 9.04990733e-02 -5.27394891e-01
1.33845794e+00 -4.23078120e-01 -5.12501895e-01 -4.40186679e-01
5.41772664e-01 3.75072509e-02 -6.43364310e-01 2.18155965e-01
-1.55621260e-01 -5.03144920e-01 -4.15536523e-01 -9.94005561e-01
-7.33933568e-01 5.37063293e-02 9.47320275e-03 5.07510483e-01
1.10045457e+00 4.22433354e-02 6.45601034e-01 3.83273661e-01
1.59740782e+00 -1.26874745e+00 -1.64129019e-01 6.98995173e-01
4.51979518e-01 -4.01453197e-01 2.15415642e-01 -7.31174469e-01
-4.34923083e-01 9.59267616e-01 6.29880130e-01 -1.09691091e-01
3.34908180e-02 7.11182415e-01 -5.22765577e-01 -1.91237032e-01
-8.05794001e-01 -4.82585758e-01 -6.17562294e-01 6.97120786e-01
-6.74644709e-01 1.79065615e-01 -8.69280696e-01 7.07559049e-01
-6.63470030e-01 -5.68056665e-02 9.41563964e-01 1.55903506e+00
-9.17649448e-01 -1.39009142e+00 -6.75402939e-01 -8.70029554e-02
-7.25306720e-02 3.72819752e-01 -3.67728114e-01 8.71380031e-01
5.44706464e-01 1.58964860e+00 -5.76233193e-02 -2.94472128e-01
2.28614718e-01 -3.42323959e-01 8.94063234e-01 -2.73176163e-01
-3.66345912e-01 1.63337439e-01 6.10482991e-01 -5.22157669e-01
-3.04467976e-01 -7.96252847e-01 -2.07447791e+00 -6.07712626e-01
-3.16458851e-01 7.87353337e-01 3.67514849e-01 1.08037186e+00
-1.47385880e-01 4.63630855e-01 4.25941557e-01 -5.55899382e-01
-3.65181953e-01 -1.23775415e-01 -4.32848006e-01 -4.61229652e-01
-9.70057100e-02 -9.04732406e-01 -5.74160740e-02 -6.14279449e-01] | [4.979506015777588, 1.847761869430542] |
ff259b90-cd9e-4ab1-b17e-3629dffd10b8 | relative-and-incomplete-time-expression | null | null | https://aclanthology.org/2020.clinicalnlp-1.14 | https://aclanthology.org/2020.clinicalnlp-1.14.pdf | Relative and Incomplete Time Expression Anchoring for Clinical Text | Extracting and modeling temporal information in clinical text is an important element for developing timelines and disease trajectories. Time information in written text varies in preciseness and explicitness, posing challenges for NLP approaches that aim to accurately anchor temporal information on a timeline. Relative and incomplete time expressions (RI-Timexes) are expressions that require additional information for their temporal anchor to be resolved, but few studies have addressed this challenge specifically. In this study, we aimed to reproduce and verify a classification approach for identifying anchor dates and relations in clinical text, and propose a novel relation classification approach for this task. | ['Sumithra Velupillai', 'Hegler Tissot', 'Nicol Bergou', 'Louise Dupuis'] | null | null | null | null | emnlp-clinicalnlp-2020-11 | ['relation-classification'] | ['natural-language-processing'] | [ 2.05230191e-01 3.38394940e-02 -9.13698018e-01 -6.81936443e-01
-4.74304438e-01 -7.30380476e-01 5.75474918e-01 1.28743255e+00
-4.90863681e-01 9.99620676e-01 5.72922111e-01 -5.43136954e-01
-8.62953126e-01 -5.22378981e-01 -1.36362268e-02 -2.66261727e-01
-6.12330675e-01 6.43344522e-01 -1.07803345e-01 6.26356900e-02
-9.70835462e-02 6.72043324e-01 -7.35264719e-01 3.97195220e-01
7.63909936e-01 6.06104314e-01 -4.67109978e-01 5.37043154e-01
-4.61398244e-01 1.15398991e+00 -8.20988178e-01 -2.50931412e-01
-3.72514486e-01 -4.19132739e-01 -1.08529544e+00 -5.03972530e-01
-5.43895423e-01 9.73089263e-02 -4.13117915e-01 5.03915906e-01
1.43948391e-01 -9.36962366e-02 8.88251662e-01 -1.50044858e+00
-1.95659325e-01 9.45464849e-01 -2.91887760e-01 6.13944530e-01
7.44781554e-01 -5.03690124e-01 8.89171064e-01 -1.49681047e-01
1.26893556e+00 6.69737041e-01 1.01334739e+00 5.57998061e-01
-1.01855946e+00 -5.56381464e-01 3.21079224e-01 1.19291179e-01
-1.53767717e+00 -2.43064165e-01 4.92879897e-01 -6.62781477e-01
1.22556293e+00 6.81347430e-01 7.63058662e-01 1.00111604e+00
6.45099699e-01 4.49405342e-01 8.68249536e-01 -3.52436364e-01
1.21539913e-01 -2.70476520e-01 3.84555668e-01 9.53814238e-02
2.16869507e-02 4.37356681e-02 -6.20524645e-01 -6.33518100e-01
4.02159095e-01 2.78670222e-01 -1.43490061e-01 4.73251909e-01
-1.36135685e+00 5.34500360e-01 1.49315596e-01 6.81236327e-01
-5.05380392e-01 -3.08476508e-01 7.92346001e-01 1.95747226e-01
8.92963290e-01 4.75446492e-01 -7.53692865e-01 -4.81157720e-01
-1.14376628e+00 5.97930960e-02 9.99372721e-01 1.04754519e+00
-1.40014783e-01 -6.79368198e-01 -5.40180326e-01 4.68946815e-01
-1.20616592e-01 -1.30513042e-01 4.61698115e-01 -4.18930292e-01
3.57945442e-01 7.97770321e-01 2.73705900e-01 -1.08958292e+00
-1.07453370e+00 -1.67928517e-01 -7.55630970e-01 -7.30592430e-01
4.88104701e-01 8.32379535e-02 -9.01249111e-01 1.73133814e+00
5.29050887e-01 1.29021183e-01 2.84645081e-01 4.08523679e-01
9.44180191e-01 4.42810208e-01 7.35129476e-01 -1.00916469e+00
1.75593686e+00 -2.06619963e-01 -1.58819211e+00 1.94034696e-01
1.04189277e+00 -6.62735105e-01 2.11238131e-01 2.20558085e-02
-7.36693501e-01 3.27663749e-01 -6.91468358e-01 -1.66685551e-01
-7.56235242e-01 -1.59025133e-01 9.39420998e-01 1.22306839e-01
-4.94594753e-01 8.29254270e-01 -1.22896564e+00 -4.20418024e-01
3.24686140e-01 3.30991715e-01 -4.63023990e-01 2.31685638e-01
-1.68599367e+00 1.11455297e+00 5.22850752e-01 2.88855761e-01
-3.17952186e-02 -1.16933620e+00 -9.28928912e-01 -3.44615608e-01
1.56380355e-01 -6.49766684e-01 1.32743859e+00 -1.38946056e-01
-9.08745229e-01 8.69115889e-01 -5.56671500e-01 -6.03729546e-01
4.86043274e-01 8.02244470e-02 -1.16224766e+00 1.67064946e-02
1.92225948e-01 1.08640566e-01 2.64741451e-01 -3.74659956e-01
-6.27269685e-01 -4.16262537e-01 -1.86065719e-01 -1.58142760e-01
-1.89264208e-01 6.13644719e-01 -3.17261785e-01 -1.17944634e+00
1.62311897e-01 -5.67642808e-01 -6.37071192e-01 -2.52320051e-01
-4.18728411e-01 -6.70948684e-01 5.17388284e-01 -8.97998035e-01
2.24028945e+00 -1.63932288e+00 -1.54391974e-01 8.36910754e-02
6.02701604e-01 -1.00977838e-01 4.34182078e-01 1.08385408e+00
-4.67223138e-01 2.44246453e-01 -2.88069427e-01 -7.18635097e-02
-2.86124349e-01 5.54003656e-01 -4.02171820e-01 5.30607641e-01
2.12075129e-01 1.08447385e+00 -1.21090353e+00 -9.45036829e-01
-1.98446020e-01 4.26463813e-01 -4.79681455e-02 1.98171288e-02
-3.66169930e-01 4.42904413e-01 -5.73950648e-01 8.30234051e-01
2.65414983e-01 -4.29498821e-01 4.10219818e-01 -2.16161907e-01
-3.72830510e-01 5.68017185e-01 -6.87836647e-01 1.37373626e+00
-1.83039129e-01 5.03304183e-01 -6.05436027e-01 -5.45700133e-01
7.12222338e-01 7.11996019e-01 1.39795721e+00 -5.98016679e-01
2.84823477e-02 1.11144431e-01 -2.36911118e-01 -7.38847554e-01
4.80446845e-01 -9.20781791e-02 -3.22288215e-01 4.27344710e-01
-4.28770036e-01 2.77346164e-01 3.65746766e-01 1.59416512e-01
1.38578093e+00 -7.80645236e-02 9.34520900e-01 -2.67939698e-02
3.56811374e-01 5.41309178e-01 9.40957546e-01 2.02375337e-01
-2.47483075e-01 2.32380718e-01 7.94403970e-01 -8.27350676e-01
-7.54001260e-01 -6.13491893e-01 -6.00228548e-01 4.74250376e-01
-4.27092433e-01 -1.10726190e+00 -2.52992868e-01 -8.46014023e-01
2.55771470e-03 5.66615760e-01 -9.34520543e-01 7.10394010e-02
-6.94841564e-01 -8.14697862e-01 7.62031496e-01 6.51393354e-01
-4.51938838e-01 -6.64174497e-01 -6.94556594e-01 7.71863759e-01
-5.49829483e-01 -1.24954438e+00 -6.16801202e-01 3.90578121e-01
-8.67468536e-01 -1.45090783e+00 -5.93708694e-01 -4.53657687e-01
8.15771103e-01 -4.93911862e-01 1.17033494e+00 -3.55612695e-01
-3.82870585e-01 2.97697596e-02 -4.66061145e-01 -8.64102483e-01
-3.60895395e-01 2.10356832e-01 -3.73771936e-02 -3.79678816e-01
6.75354123e-01 -2.62440890e-01 -6.42971277e-01 3.48997295e-01
-9.95882988e-01 -4.12450507e-02 3.47367562e-02 5.92521369e-01
6.96054220e-01 -2.27322616e-03 6.25384688e-01 -1.12807953e+00
8.23103786e-01 -7.69644618e-01 -5.65431654e-01 3.02076697e-01
-9.97952938e-01 3.32264951e-03 1.95557132e-01 -5.18696070e-01
-6.51656270e-01 4.91230600e-02 -2.36968160e-01 3.19253989e-02
4.42885384e-02 1.22543323e+00 3.46521944e-01 4.23279107e-01
6.09360099e-01 -8.25938284e-02 -1.65621534e-01 -3.69268537e-01
2.33349964e-01 7.15884686e-01 4.16281074e-01 -3.62345636e-01
1.48181885e-01 6.16368592e-01 3.18792969e-01 -3.86990547e-01
-8.86234939e-01 -8.83826196e-01 -9.43459868e-01 -2.58711097e-03
7.80060709e-01 -6.41603231e-01 -6.51954472e-01 2.92687397e-02
-1.34912193e+00 -4.16218974e-02 -3.87538016e-01 6.98647380e-01
-3.43740880e-01 1.36215955e-01 -6.80222690e-01 -6.10013366e-01
-3.88757527e-01 -5.80316365e-01 7.74284720e-01 -1.99596405e-01
-1.23046982e+00 -1.41123724e+00 4.90866423e-01 -3.57960910e-01
1.91038296e-01 8.32619250e-01 8.51047575e-01 -9.46284711e-01
3.03987842e-02 -4.18763846e-01 -5.34839705e-02 -7.71946728e-01
7.00864494e-01 1.71109632e-01 -5.71184397e-01 2.52150178e-01
5.94896488e-02 4.74422604e-01 4.35928166e-01 5.85670710e-01
1.07109690e+00 -9.88253534e-01 -1.14833796e+00 5.87596357e-01
1.05315661e+00 6.11046374e-01 4.11309659e-01 3.54916334e-01
3.71682495e-01 9.25101161e-01 1.08663762e+00 6.14731789e-01
6.27927840e-01 7.54445910e-01 -1.42116264e-01 1.83067762e-03
2.79343814e-01 -7.61012882e-02 -3.50450844e-01 8.12656283e-01
-1.93454325e-02 -3.05689692e-01 -1.29594421e+00 8.11368346e-01
-1.94466102e+00 -8.31582189e-01 -5.18507600e-01 1.89725602e+00
1.56076157e+00 6.75564958e-03 9.40412804e-02 3.10259938e-01
3.65847021e-01 9.12180394e-02 -3.48782837e-01 -3.69166434e-01
-6.85923398e-02 3.13418661e-03 5.71545362e-01 4.02031451e-01
-1.11123526e+00 7.10589051e-01 7.50384760e+00 6.23501241e-01
-9.97227967e-01 -6.74279267e-03 4.52071905e-01 1.42273277e-01
-3.31540614e-01 6.28223121e-02 -8.23917150e-01 4.11156476e-01
1.45426285e+00 -5.73624134e-01 -2.35317662e-01 2.50869721e-01
7.21341491e-01 1.03102244e-01 -1.38297248e+00 8.24647665e-01
-3.88994604e-01 -1.41880834e+00 -2.28756681e-01 3.26805040e-02
4.59891677e-01 -4.08873111e-01 -3.10666084e-01 -2.33516783e-01
2.30368719e-01 -1.07149172e+00 3.84157330e-01 8.69098723e-01
7.95472622e-01 -4.18632299e-01 6.69611514e-01 8.19354132e-03
-1.38198566e+00 2.25924075e-01 2.72992194e-01 -2.59599090e-02
5.25221109e-01 9.88105118e-01 -1.26287174e+00 1.11319888e+00
4.64164436e-01 1.14396083e+00 -1.71265572e-01 1.03877592e+00
-1.76154315e-01 5.43390691e-01 -4.05831426e-01 1.36455214e-02
7.00897872e-02 1.27535850e-01 4.00886685e-01 1.40976989e+00
2.62929767e-01 5.33517420e-01 -1.43345416e-01 5.21384001e-01
1.62300050e-01 2.66798824e-01 -6.18286192e-01 -7.21869349e-01
5.53229868e-01 8.54827464e-01 -1.09410369e+00 -2.39616916e-01
-2.68654853e-01 6.52902842e-01 8.33842605e-02 3.35923165e-01
-9.55549181e-01 -2.37209171e-01 7.94329286e-01 2.20663965e-01
-3.41674298e-01 -4.67211336e-01 -2.84238487e-01 -8.55423331e-01
-1.59264896e-02 -4.59388018e-01 1.47652137e+00 -4.12067384e-01
-1.30976117e+00 8.05504382e-01 2.47996226e-01 -1.63386703e+00
-6.08047366e-01 -2.68474609e-01 -3.97269167e-02 7.83211708e-01
-1.31084239e+00 -1.27690077e+00 9.42612588e-02 5.42666435e-01
1.92541301e-01 4.54700679e-01 1.08736610e+00 4.50966716e-01
-3.48713487e-01 4.96407181e-01 -1.71144828e-01 1.03904985e-01
9.80477571e-01 -1.07971525e+00 2.07588375e-01 3.97817940e-01
-1.16503797e-01 9.43894148e-01 8.95827830e-01 -1.12236524e+00
-1.13912475e+00 -1.19232774e+00 2.00378346e+00 -8.33691120e-01
1.00546503e+00 -3.63759361e-02 -9.56510603e-01 7.89080262e-01
-1.11298420e-01 -1.54677451e-01 1.11911631e+00 3.22034717e-01
-4.45047021e-01 4.86564785e-02 -1.03290915e+00 5.52502155e-01
1.04276407e+00 -6.76345408e-01 -6.73709512e-01 6.79840446e-01
9.09489930e-01 -7.99167395e-01 -1.56015527e+00 5.33176959e-01
5.25243998e-01 4.30667624e-02 8.63229036e-01 -9.27864492e-01
2.99482167e-01 -2.97404170e-01 4.60889131e-01 -9.78358328e-01
-9.32104141e-03 -9.96530831e-01 -4.53573138e-01 1.22951150e+00
5.73030293e-01 -7.26640880e-01 6.89651489e-01 9.05714691e-01
-1.73683800e-02 -7.98256099e-01 -1.04325652e+00 -8.38243186e-01
-1.05208062e-01 -5.65253973e-01 9.11774158e-01 1.54346919e+00
7.47073591e-01 -5.51380068e-02 -2.09578410e-01 1.82869509e-01
5.44271842e-02 9.43601355e-02 7.34487595e-03 -1.24767470e+00
4.99091655e-01 -4.65459466e-01 -4.25932080e-01 -2.14804053e-01
7.52132237e-02 -8.13058853e-01 -3.98262322e-01 -1.95483220e+00
3.05068642e-02 -7.01725185e-01 -1.95972472e-01 7.70404279e-01
-5.81637695e-02 -3.04927856e-01 -4.79983062e-01 5.42741895e-01
-4.19941247e-01 2.03981280e-01 7.70560801e-01 -2.09350735e-01
-5.08303165e-01 2.22917318e-01 -2.62414843e-01 4.85254765e-01
6.68460310e-01 -1.03095591e+00 -4.52057838e-01 -5.80030829e-02
6.06768489e-01 6.42940998e-01 -7.82979373e-03 -4.00652140e-01
8.27286661e-01 -5.44959724e-01 5.87509088e-02 -1.12104809e+00
-3.05390544e-02 -1.06010342e+00 7.97744751e-01 5.77775061e-01
-5.39894283e-01 5.82111657e-01 4.81070161e-01 5.74224830e-01
-4.52293277e-01 -6.16152212e-02 -4.15954813e-02 1.28302142e-01
-4.11262631e-01 4.93227392e-01 -5.33862650e-01 7.99673144e-03
1.21556759e+00 -4.70151231e-02 -3.34435970e-01 -7.30443448e-02
-1.17250109e+00 4.35744733e-01 8.55122358e-02 4.74372953e-01
5.62413394e-01 -1.31304419e+00 -5.86511254e-01 -4.22738045e-01
4.54337388e-01 -1.14705168e-01 2.18456984e-01 1.22417951e+00
-5.03777921e-01 7.96711802e-01 1.42385900e-01 -4.39565837e-01
-1.49464405e+00 9.32250500e-01 2.34602138e-01 -8.31105053e-01
-6.70048535e-01 3.08214277e-01 -3.86009425e-01 -2.32298262e-02
2.24762648e-01 -7.75047362e-01 -6.00849390e-01 7.60134816e-01
6.68011129e-01 8.91483724e-02 2.18116373e-01 -4.54479963e-01
-7.57526278e-01 3.39573950e-01 -2.03172237e-01 -3.19624878e-02
1.36652851e+00 -6.56087250e-02 -5.87908626e-01 6.21311963e-01
1.04717243e+00 8.29261728e-03 -5.56576967e-01 -3.30936432e-01
8.27534258e-01 -2.61816233e-02 -3.15008849e-01 -7.83055782e-01
-4.08200324e-01 1.59869179e-01 7.39336908e-02 3.92006546e-01
1.13135338e+00 1.56406179e-01 6.39700472e-01 -1.57170370e-01
1.87405288e-01 -8.08055997e-01 -5.53803205e-01 5.44757366e-01
9.20637190e-01 -8.44978333e-01 2.38927081e-01 -8.45768332e-01
-3.11200023e-01 1.17266107e+00 5.75163700e-02 8.64292324e-01
9.89469945e-01 6.52819872e-01 1.71910033e-01 -5.86515903e-01
-9.05309916e-01 -1.64573073e-01 5.61995208e-01 5.27252793e-01
8.24586213e-01 1.37113914e-01 -1.06869328e+00 8.87423992e-01
-9.16967317e-02 4.10471559e-01 9.30790529e-02 1.24387038e+00
5.35598516e-01 -1.33600652e+00 -1.73959032e-01 5.00524580e-01
-9.80073392e-01 -1.16176419e-01 -3.31841201e-01 7.91251719e-01
-1.70304343e-01 1.01920903e+00 3.63205746e-02 -1.98598653e-01
4.57437694e-01 2.58175023e-02 8.85822624e-02 -3.99230719e-01
-7.19728410e-01 -3.53539269e-03 6.22098982e-01 -5.89519620e-01
-6.44413948e-01 -7.83980608e-01 -1.65126836e+00 4.17215079e-02
-1.25968292e-01 5.63000917e-01 5.58671415e-01 9.09415543e-01
4.09143984e-01 8.80262971e-01 3.37535828e-01 1.13864064e-01
1.39008462e-02 -6.79374516e-01 -2.26156473e-01 2.60633886e-01
3.42224628e-01 -4.58673924e-01 -7.96664953e-02 3.57181340e-01] | [8.663333892822266, 9.076717376708984] |
87113742-6785-4549-9668-48fc58d4b1e4 | point2sequence-learning-the-shape | 1811.02565 | null | http://arxiv.org/abs/1811.02565v2 | http://arxiv.org/pdf/1811.02565v2.pdf | Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-based Sequence to Sequence Network | Exploring contextual information in the local region is important for shape
understanding and analysis. Existing studies often employ hand-crafted or
explicit ways to encode contextual information of local regions. However, it is
hard to capture fine-grained contextual information in hand-crafted or explicit
manners, such as the correlation between different areas in a local region,
which limits the discriminative ability of learned features. To resolve this
issue, we propose a novel deep learning model for 3D point clouds, named
Point2Sequence, to learn 3D shape features by capturing fine-grained contextual
information in a novel implicit way. Point2Sequence employs a novel sequence
learning model for point clouds to capture the correlations by aggregating
multi-scale areas of each local region with attention. Specifically,
Point2Sequence first learns the feature of each area scale in a local region.
Then, it captures the correlation between area scales in the process of
aggregating all area scales using a recurrent neural network (RNN) based
encoder-decoder structure, where an attention mechanism is proposed to
highlight the importance of different area scales. Experimental results show
that Point2Sequence achieves state-of-the-art performance in shape
classification and segmentation tasks. | ['Yu-Shen Liu', 'Xinhai Liu', 'Matthias Zwicker', 'Zhizhong Han'] | 2018-11-06 | null | null | null | null | ['3d-part-segmentation', 'shape-representation-of-3d-point-clouds'] | ['computer-vision', 'computer-vision'] | [ 4.16439101e-02 -2.46178955e-01 -9.58000869e-02 -6.02599084e-01
-7.66802311e-01 -4.03441399e-01 4.95042384e-01 2.55906552e-01
-1.98200196e-01 1.78602487e-01 3.30350846e-01 -9.15665403e-02
-6.42594397e-02 -1.02508688e+00 -8.76414955e-01 -8.03183377e-01
1.44732058e-01 4.08487171e-01 3.19333255e-01 -1.27399176e-01
3.45906287e-01 8.57104897e-01 -1.21974301e+00 3.89384657e-01
5.70624232e-01 1.06090856e+00 6.97682738e-01 3.98349136e-01
-6.42301142e-01 2.32529208e-01 -4.57218945e-01 7.26763308e-02
2.14199632e-01 -1.97468251e-01 -5.30109107e-01 2.70957291e-01
4.32968676e-01 -3.08418512e-01 -2.00919598e-01 9.05367374e-01
3.19262981e-01 -4.59026396e-02 6.11047864e-01 -7.30929911e-01
-8.70485783e-01 2.12507829e-01 -7.06021965e-01 2.09426969e-01
-1.92144021e-01 2.56671250e-01 1.18732798e+00 -9.70621407e-01
3.41017008e-01 1.35271609e+00 4.65923756e-01 1.67528093e-01
-9.68319297e-01 -5.18607378e-01 5.36146641e-01 -2.42814012e-02
-1.56650424e+00 1.25158191e-01 1.17695153e+00 -4.24147487e-01
9.58415926e-01 -6.01052493e-02 8.42699885e-01 6.76874936e-01
7.08116144e-02 1.05778456e+00 7.75388479e-01 -8.02198350e-02
1.88277543e-01 -4.47375536e-01 -1.89445347e-01 6.94635928e-01
-1.82679608e-01 1.50224924e-01 -2.63118356e-01 5.47962375e-02
1.43146002e+00 5.07844627e-01 1.69063717e-01 -4.48903978e-01
-1.10971892e+00 8.36883307e-01 1.07003117e+00 5.43327272e-01
-4.59066361e-01 2.55468726e-01 3.29622567e-01 5.98080270e-02
6.31149173e-01 2.48211026e-01 -5.69509983e-01 -8.91380310e-02
-8.87957513e-01 1.16615919e-02 1.94626421e-01 1.04220986e+00
1.10845912e+00 4.51332517e-02 -1.89653262e-01 8.69800866e-01
4.81836528e-01 5.84542572e-01 2.80483335e-01 -7.59973645e-01
6.43488944e-01 1.11481261e+00 -1.58870265e-01 -9.47889388e-01
-1.99701428e-01 -4.72221851e-01 -8.87523830e-01 1.29285187e-01
-2.83092503e-02 2.77335346e-01 -1.04411268e+00 1.51408219e+00
3.53633881e-01 4.63436872e-01 -2.39992782e-01 1.09633708e+00
8.11316311e-01 7.63359308e-01 5.11133857e-03 2.15559632e-01
1.35401320e+00 -8.43143702e-01 1.97025388e-02 -1.82084098e-01
3.28789324e-01 -5.39659321e-01 1.18518996e+00 -2.22889885e-01
-8.95197868e-01 -6.54570758e-01 -7.52477884e-01 -3.70682329e-01
-3.62321079e-01 1.60695836e-01 3.40534717e-01 -1.29201431e-02
-8.06010365e-01 4.23921227e-01 -9.08740699e-01 -6.09919354e-02
5.97631216e-01 2.38137230e-01 -1.89335302e-01 5.45974448e-02
-8.32658529e-01 5.14324486e-01 8.17856714e-02 2.05769360e-01
-7.33720481e-01 -7.36128092e-01 -1.10827303e+00 2.80146539e-01
2.39747718e-01 -6.80590749e-01 1.12317610e+00 -7.84370780e-01
-1.23842311e+00 7.86530435e-01 -4.00655895e-01 -2.37335950e-01
6.39508590e-02 -2.55930662e-01 3.70161608e-02 1.98416054e-01
2.52726376e-01 7.31458068e-01 9.48111951e-01 -1.17741299e+00
-4.70651090e-01 -5.06248295e-01 -1.06661230e-01 1.84438422e-01
4.62785810e-02 -1.63682520e-01 -7.14767933e-01 -8.74820173e-01
3.53776336e-01 -6.22081459e-01 -4.19678390e-01 1.35112718e-01
-1.36941403e-01 -5.50250709e-01 9.85450447e-01 -4.02607828e-01
9.60541666e-01 -2.41776800e+00 6.32466227e-02 2.34769583e-01
1.57251328e-01 3.48011464e-01 -3.89877886e-01 3.72865826e-01
1.85420364e-01 1.47292688e-01 -5.23532391e-01 -2.33104557e-01
2.02932060e-02 5.29119670e-01 -4.03479218e-01 7.90126696e-02
7.37572312e-01 1.25488460e+00 -9.64923680e-01 -3.27235073e-01
4.68272448e-01 7.48227835e-01 -4.41678971e-01 2.18628824e-01
-4.35104519e-01 4.76073414e-01 -1.03805399e+00 7.12055147e-01
6.66591167e-01 -4.15370703e-01 -3.87572944e-01 -2.97240078e-01
-2.90895551e-01 3.17512512e-01 -8.26139510e-01 1.94459438e+00
-7.04899669e-01 4.04116154e-01 -1.40948340e-01 -9.52319860e-01
1.43156314e+00 1.44366995e-01 5.12815833e-01 -7.85904646e-01
-1.54986426e-01 3.16247106e-01 -3.09252590e-01 -2.73095638e-01
2.77398705e-01 -2.83732563e-01 -1.83311403e-01 2.84167022e-01
-1.83454797e-01 -5.48784912e-01 -4.39302027e-01 2.37614959e-02
1.06824231e+00 4.44200598e-02 2.89700091e-01 4.36437950e-02
5.11224031e-01 -4.34735894e-01 5.72277904e-01 5.11448801e-01
-1.38297319e-01 9.74695504e-01 1.70564070e-01 -7.79419959e-01
-1.41427088e+00 -9.71708477e-01 5.44869639e-02 8.24236751e-01
2.26349026e-01 -4.00651097e-01 -2.85604686e-01 -7.30879724e-01
2.86140561e-01 3.10829282e-01 -7.27812052e-01 -6.70621470e-02
-7.26559103e-01 -4.97777425e-02 1.39160246e-01 9.48844910e-01
6.96852684e-01 -1.35507858e+00 -6.97746634e-01 3.33172530e-01
2.26608098e-01 -1.08732557e+00 -7.51638770e-01 7.06583261e-02
-1.20109594e+00 -7.25726545e-01 -7.37585723e-01 -7.97587395e-01
7.67045021e-01 5.73781788e-01 1.11485946e+00 2.25589320e-01
-1.19679376e-01 1.51220232e-01 -4.48278904e-01 -2.07592279e-01
1.96554095e-01 1.91429004e-01 -5.75594485e-01 1.35893077e-01
5.16226232e-01 -7.64986634e-01 -7.56697476e-01 3.34780812e-01
-8.48933756e-01 7.80727640e-02 9.73251164e-01 9.37267125e-01
9.91571009e-01 -2.66989380e-01 3.78645867e-01 -6.09197319e-01
2.89131671e-01 -3.30424577e-01 -5.07024705e-01 1.52321994e-01
-1.74201783e-02 3.83119792e-01 5.56042731e-01 -2.00563326e-01
-7.61823177e-01 1.03901088e-01 -4.34008121e-01 -7.76473403e-01
-4.06309158e-01 4.44233388e-01 -3.01085860e-01 8.72540772e-02
6.05680756e-02 5.58787882e-01 -2.66130179e-01 -6.68547511e-01
3.86287510e-01 5.50815642e-01 4.78455216e-01 -8.21983039e-01
7.30739713e-01 5.22975087e-01 -3.20388451e-02 -7.24513769e-01
-9.85734999e-01 -5.70289195e-01 -9.92143333e-01 6.18170276e-02
9.07287896e-01 -9.53342855e-01 -5.48628390e-01 2.77752221e-01
-1.16593599e+00 -3.35527986e-01 -4.98250693e-01 2.38793045e-01
-5.63519776e-01 1.54992908e-01 -5.80834210e-01 -6.22134268e-01
-3.66354167e-01 -1.20433092e+00 1.91469681e+00 3.12012970e-01
-2.80031804e-02 -8.93041670e-01 -6.16620742e-02 1.67656258e-01
2.39713416e-01 2.66932935e-01 9.74263072e-01 -3.24698120e-01
-8.90000641e-01 -6.80917725e-02 -6.14196837e-01 1.43933252e-01
3.44484597e-01 -1.54285282e-01 -5.61809599e-01 -1.48190051e-01
-9.40453932e-02 -3.37092602e-03 9.39844251e-01 2.64697403e-01
1.56861794e+00 -2.04627559e-01 -2.77025998e-01 5.72848856e-01
1.30826974e+00 1.42168880e-01 4.35264885e-01 1.52532637e-01
9.24804807e-01 2.17025742e-01 5.02521396e-01 3.99710178e-01
6.14483833e-01 6.85014844e-01 5.60670674e-01 -1.41015053e-01
3.79412039e-03 -5.31052768e-01 1.95524935e-02 9.87868249e-01
-7.29746446e-02 3.34463775e-01 -8.86363029e-01 8.31996441e-01
-1.81679595e+00 -7.85020173e-01 2.81511575e-01 1.85312486e+00
7.15070367e-01 1.24475598e-01 -8.26152787e-02 -2.31164113e-01
7.10410655e-01 4.59385484e-01 -8.14847171e-01 -3.80897373e-01
-3.04337498e-02 1.04219623e-01 1.98647767e-01 3.88508439e-01
-9.19104397e-01 1.14796388e+00 5.14743710e+00 9.30119097e-01
-1.47343814e+00 -7.33825266e-02 5.05081296e-01 5.48663829e-03
-5.82240522e-01 -3.73415500e-02 -6.90842271e-01 6.18070364e-01
4.69827563e-01 3.25690299e-01 1.45300910e-01 7.38955915e-01
3.25882763e-01 3.17091346e-01 -9.08551753e-01 1.00499427e+00
-9.46439058e-02 -1.35213530e+00 2.45909587e-01 1.09831810e-01
7.26760089e-01 3.21604311e-01 5.17338701e-02 1.91770166e-01
2.71766156e-01 -8.70630920e-01 6.46815658e-01 6.78832889e-01
7.49200165e-01 -7.75886416e-01 5.09180605e-01 5.28669238e-01
-1.64227986e+00 -3.23757194e-02 -5.39600551e-01 -7.61591000e-05
6.77797198e-02 6.65286183e-01 -9.04943109e-01 5.50996244e-01
5.92721522e-01 1.04075456e+00 -3.39351445e-01 9.70751286e-01
-4.32710469e-01 5.84006250e-01 -3.96331847e-01 -5.34857437e-02
7.06836402e-01 -2.99739569e-01 5.15999615e-01 1.18084395e+00
5.47836542e-01 4.90373611e-01 1.72288045e-01 1.00348091e+00
-9.38683152e-02 -1.26045346e-01 -7.64357269e-01 6.89736381e-02
5.64912558e-01 9.59627986e-01 -6.69854164e-01 -3.94927144e-01
-5.70045471e-01 8.02141249e-01 5.13252974e-01 3.76342475e-01
-3.68399113e-01 -2.72587180e-01 6.80886626e-01 1.00453265e-01
9.55822825e-01 -6.97067380e-01 -5.48964977e-01 -1.02486598e+00
-1.16115306e-02 -3.53032976e-01 3.73890728e-01 -9.82525945e-01
-1.42162204e+00 4.80291128e-01 -2.96771705e-01 -1.30870712e+00
1.52062885e-02 -3.80104244e-01 -8.03762853e-01 9.51194406e-01
-1.57452607e+00 -1.35924816e+00 -2.10910484e-01 5.45312583e-01
9.13292050e-01 -8.68061110e-02 5.76805294e-01 -1.77128296e-02
-1.71154618e-01 3.35301340e-01 -1.99816991e-02 3.20921600e-01
2.61127174e-01 -1.21936607e+00 7.85782337e-01 4.80075717e-01
4.07593638e-01 6.19612038e-01 1.52163774e-01 -7.06381381e-01
-1.20205534e+00 -1.30357265e+00 9.60387230e-01 -2.13478848e-01
4.08175915e-01 -4.07792777e-01 -1.21634579e+00 4.94267642e-01
-1.05096646e-01 4.50195819e-01 4.46373791e-01 7.26502836e-02
-5.41544020e-01 -1.27644956e-01 -8.53684187e-01 4.63998109e-01
1.22875071e+00 -1.04235482e+00 -9.28384244e-01 -1.07951172e-01
1.00025201e+00 -3.83540869e-01 -8.35710466e-01 5.13083577e-01
2.85992473e-01 -5.21337569e-01 1.11279523e+00 -3.75596166e-01
6.41156554e-01 -5.02755165e-01 -2.30960786e-01 -1.37780058e+00
-4.47352260e-01 -1.36308283e-01 -1.86037913e-01 9.83729124e-01
2.33653903e-01 -4.19532925e-01 9.29511130e-01 3.99958760e-01
-3.97775233e-01 -1.24703300e+00 -9.35616016e-01 -5.82457364e-01
2.80638963e-01 -4.42520082e-01 1.13465047e+00 6.73131704e-01
-4.65464234e-01 4.09883708e-01 -7.06567317e-02 3.59966338e-01
3.29636365e-01 9.29161012e-01 6.63407683e-01 -1.10981226e+00
5.24545982e-02 -5.49346507e-01 -6.93719983e-01 -1.63146043e+00
1.19454309e-01 -8.20257723e-01 -2.88415458e-02 -1.64450252e+00
1.48683771e-01 -6.32082403e-01 -4.47159708e-01 4.84891713e-01
-3.11834782e-01 1.77376956e-01 4.50763255e-01 3.43834043e-01
-6.61713660e-01 8.98651659e-01 1.70542002e+00 -3.71900052e-01
-6.01347946e-02 2.59500835e-02 -6.01159751e-01 6.05547905e-01
5.61831176e-01 -2.58335382e-01 -2.46610612e-01 -7.49560356e-01
1.15193218e-01 -3.54484804e-02 6.58416748e-01 -8.25540245e-01
3.96873385e-01 -2.63236314e-01 5.62219262e-01 -1.17837942e+00
3.50955158e-01 -9.90413189e-01 -2.43627116e-01 1.04653805e-01
-2.81211227e-01 6.97719306e-03 1.24860756e-01 6.67022347e-01
-5.86823046e-01 -5.13515957e-02 3.48117590e-01 -3.92421722e-01
-8.86229515e-01 7.46116281e-01 1.43540442e-01 -6.69467896e-02
7.58384168e-01 -2.83543706e-01 2.42472038e-01 -1.68546587e-01
-6.59123838e-01 2.59086639e-01 4.58413184e-01 5.49892545e-01
9.63870049e-01 -1.51667678e+00 -6.59603357e-01 6.02200389e-01
1.81809053e-01 6.95354700e-01 3.69956404e-01 3.53203416e-01
-1.73983261e-01 4.03203815e-01 -2.50074983e-01 -1.02954781e+00
-1.09591329e+00 4.65657532e-01 2.72037953e-01 -2.01675102e-01
-9.32256579e-01 9.85429585e-01 6.57106936e-01 -5.35942078e-01
-1.67950824e-01 -7.08357036e-01 -1.09659255e-01 -2.03032836e-01
3.03224593e-01 -2.65104562e-01 -2.44551376e-02 -7.49151111e-01
-2.07204044e-01 1.30844760e+00 -5.41314706e-02 1.78769976e-01
1.60556626e+00 -1.60678357e-01 3.41958851e-02 5.54430604e-01
1.43533552e+00 -1.77021518e-01 -1.88309324e+00 -7.31446445e-01
-9.57764313e-02 -6.07706010e-01 -3.00636049e-03 -6.50742412e-01
-1.29960656e+00 1.28662050e+00 1.37116447e-01 -1.55356973e-01
1.12769032e+00 3.27109158e-01 1.08103597e+00 1.07467167e-01
3.82897913e-01 -8.31761837e-01 1.98868543e-01 8.93689871e-01
1.04740036e+00 -1.23800933e+00 -2.40767032e-01 -2.93995768e-01
-6.29694819e-01 1.14304185e+00 5.70924222e-01 -3.90731633e-01
8.04957509e-01 1.08834408e-01 -1.02771178e-01 -4.80443537e-01
-8.12976956e-01 -2.92173535e-01 4.91114527e-01 2.65531957e-01
2.83098161e-01 2.71106839e-01 1.50190532e-01 6.76335931e-01
-6.67883530e-02 -2.98039913e-01 -2.04236150e-01 8.38431954e-01
-5.68768084e-01 -1.08230162e+00 -1.61394179e-01 4.34394777e-01
-5.21128299e-04 -7.04463720e-02 -3.83055449e-01 5.25757611e-01
2.77492285e-01 3.13458562e-01 4.82584119e-01 -2.99834758e-01
4.15275693e-01 -9.36735868e-02 2.27960303e-01 -7.08048463e-01
-5.06670237e-01 1.72726616e-01 -5.75367153e-01 -4.70160127e-01
-2.78294951e-01 -7.63851106e-01 -1.50561285e+00 3.60102840e-02
-6.25035912e-02 9.78726596e-02 5.33470988e-01 1.15504467e+00
7.16888487e-01 5.05669475e-01 7.92753160e-01 -9.11294639e-01
-2.32878014e-01 -7.25418210e-01 -4.66112792e-01 5.25499761e-01
5.71262658e-01 -6.59657836e-01 1.03360049e-01 -2.21785247e-01] | [8.027195930480957, -3.5167465209960938] |
c551ce3d-c0ab-4c21-8b19-410740a37740 | deep-amortized-variational-inference-for | null | null | https://openreview.net/forum?id=H1xXYy3VKr | https://openreview.net/pdf?id=H1xXYy3VKr | Deep Amortized Variational Inference for Multivariate Time Series Imputation with Latent Gaussian Process Models | Multivariate time series with missing values are common in areas such as healthcare and finance, and have grown in number and complexity over the years. This raises the question whether deep learning methodologies can outperform classical data imputation methods in this domain. However, naive applications of deep learning fall short in giving reliable confidence estimates and lack interpretability. We propose a new deep sequential latent variable model for dimensionality reduction and data imputation. Our modeling assumption is simple and interpretable: the high dimensional time series has a lower-dimensional representation which evolves smoothly in time according to a Gaussian process. The non-linear dimensionality reduction in the presence of missing data is achieved using a VAE approach with a novel structured variational approximation. We demonstrate that our approach outperforms several classical and deep learning-based data imputation methods on high-dimensional data from the domains of computer vision and healthcare, while additionally improving the smoothness of the imputations and providing interpretable uncertainty estimates. | ['Stephan Mandt', 'Gunnar Rätsch', 'Dmitry Baranchuk', 'Vincent Fortuin'] | 2019-10-16 | null | null | null | pproximateinference-aabi-symposium-2019-12 | ['multivariate-time-series-imputation'] | ['time-series'] | [ 9.98777524e-02 2.20598504e-01 -1.04870729e-01 -6.54840648e-01
-1.00463986e+00 -1.36188507e-01 5.61909854e-01 4.81281988e-02
-2.46486604e-01 1.08423197e+00 5.33493698e-01 -1.25647232e-01
-5.39737403e-01 -6.33354127e-01 -8.95341218e-01 -8.95697653e-01
1.07065403e-04 8.73687327e-01 -8.49407732e-01 2.56957293e-01
-1.78103179e-01 1.20010853e-01 -1.24332249e+00 -3.55568379e-02
1.10085297e+00 8.65310490e-01 -4.24001098e-01 2.44218469e-01
-1.18336745e-01 5.98141670e-01 -1.18851759e-01 -4.53605860e-01
1.38057247e-01 -3.18999738e-01 -3.68313938e-01 -2.18761027e-01
4.34024334e-02 -4.13951844e-01 -2.91083544e-01 7.43492186e-01
3.96714687e-01 6.92568943e-02 9.86676991e-01 -1.44268405e+00
-7.33852327e-01 3.94745976e-01 -6.69019043e-01 -3.79077226e-01
-1.13181826e-02 -1.75609946e-01 5.36413014e-01 -6.86111629e-01
5.60102224e-01 1.24647236e+00 1.26369262e+00 5.65113604e-01
-1.69931936e+00 -4.58434641e-01 -1.39960229e-01 -3.42037827e-02
-1.02478302e+00 -4.55637842e-01 7.99965024e-01 -8.86489034e-01
6.59826994e-01 7.88863599e-02 2.16529727e-01 1.68876588e+00
3.78865182e-01 7.02030480e-01 8.89892757e-01 -1.39140218e-01
6.29140735e-01 -3.55794132e-01 1.12024896e-01 1.34911373e-01
2.14128986e-01 3.18367094e-01 -5.03102124e-01 -6.25055432e-01
6.86482608e-01 7.06925154e-01 1.25092179e-01 -6.85161829e-01
-1.46179497e+00 1.06576574e+00 -1.83947384e-01 -2.14606285e-01
-8.68338943e-01 3.25907558e-01 4.99518305e-01 2.20807716e-01
8.89275193e-01 -7.61746895e-03 -6.81166649e-01 -5.01739025e-01
-1.21868420e+00 3.61945271e-01 7.06228316e-01 7.41229236e-01
3.68205637e-01 2.35697404e-01 -3.51037771e-01 5.36272347e-01
2.28378996e-01 7.15176165e-01 1.69753700e-01 -1.54093659e+00
5.03893077e-01 2.77387828e-01 5.70187807e-01 -8.10014367e-01
-5.62002957e-01 -2.86481470e-01 -1.71208000e+00 2.90615380e-01
6.64640784e-01 -5.23274839e-01 -9.54542994e-01 2.05424833e+00
3.15789521e-01 4.09802198e-01 1.74142763e-01 6.52046084e-01
3.72557431e-01 3.60456318e-01 1.34570852e-01 -7.01149285e-01
9.64466870e-01 -2.82959729e-01 -1.27433038e+00 2.76846975e-01
3.85564536e-01 -3.43985558e-01 7.13051975e-01 4.78780270e-01
-1.12648845e+00 -3.24583977e-01 -4.62764919e-01 -3.71455222e-01
1.23266816e-01 -1.40502870e-01 6.17387831e-01 5.51838815e-01
-6.05806828e-01 8.01571488e-01 -1.38991630e+00 1.15936518e-01
7.20005035e-01 3.46420377e-01 -3.95841986e-01 -2.86589283e-02
-9.68167782e-01 7.01678216e-01 -2.91907489e-01 3.29913825e-01
-6.57778382e-01 -1.06617200e+00 -9.53222811e-01 6.37794882e-02
-3.01779285e-02 -1.19612730e+00 1.03892827e+00 -7.93215513e-01
-1.17016780e+00 4.57218409e-01 -6.45324945e-01 -6.48834646e-01
1.10670483e+00 -2.17326760e-01 -2.70966202e-01 -4.90180433e-01
1.05059288e-01 1.40329674e-01 1.02534819e+00 -9.27377403e-01
-1.36944532e-01 -8.21291208e-01 -7.06689119e-01 -2.52943367e-01
1.49590299e-01 -3.50540727e-01 9.56958458e-02 -8.83420348e-01
2.77775258e-01 -7.27265716e-01 -4.91905779e-01 8.64982381e-02
-2.64307082e-01 -9.15140510e-02 4.19632792e-01 -1.06782520e+00
1.00000691e+00 -2.02644730e+00 5.21898568e-01 2.92947646e-02
3.24730903e-01 -5.48797190e-01 1.12515844e-01 3.93923134e-01
-5.42487688e-02 -5.27702831e-02 -8.63768518e-01 -1.05697155e+00
3.43912780e-01 5.20686150e-01 -5.04460514e-01 8.18744421e-01
2.06443807e-03 1.06513250e+00 -7.46879637e-01 -2.55306959e-01
2.24922940e-01 8.26927245e-01 -4.02847201e-01 1.44634649e-01
-2.48317942e-01 9.35833573e-01 -1.64495647e-01 5.61759174e-01
9.75858271e-01 -2.62291282e-01 1.36699200e-01 6.38725981e-02
2.08905023e-02 -1.58017904e-01 -1.21366918e+00 1.97365558e+00
-3.71478349e-01 4.58233893e-01 -2.15960331e-02 -1.18665957e+00
7.61079192e-01 4.44026440e-01 8.80216360e-01 -6.07757330e-01
-2.88421139e-02 3.88831608e-02 -3.72622550e-01 -5.70553243e-01
1.10902756e-01 -4.73434806e-01 -2.66535282e-01 3.18507820e-01
-1.91754237e-01 2.79520363e-01 -4.10364091e-01 -2.12520555e-01
9.09238935e-01 4.87638414e-01 -7.88657367e-03 -5.88730425e-02
-5.61317429e-02 -1.33935779e-01 1.04296231e+00 8.13736379e-01
1.27730295e-01 8.21279764e-01 6.77602649e-01 -6.40553236e-01
-1.35203016e+00 -1.43073356e+00 -6.32416904e-01 5.21564364e-01
-4.76474583e-01 5.28210215e-02 -6.50487781e-01 -2.60377675e-01
3.91327262e-01 9.10688400e-01 -9.64012265e-01 8.82874131e-02
-3.61929983e-01 -1.19290972e+00 2.66131580e-01 6.33066237e-01
-6.79501072e-02 -8.71921778e-01 -3.23610097e-01 5.27644455e-01
-4.81206954e-01 -7.37917185e-01 -1.11409247e-01 2.10033268e-01
-1.27893019e+00 -7.55387485e-01 -8.29339087e-01 -2.19000369e-01
5.81461251e-01 -5.48014522e-01 1.06366539e+00 -6.36815608e-01
-1.02247812e-01 3.99310142e-01 4.26032208e-02 -6.25253975e-01
-3.84063184e-01 -1.73112750e-01 3.53087723e-01 1.79459214e-01
7.52383828e-01 -7.31335640e-01 -6.08434379e-01 -1.67505562e-01
-7.70075858e-01 2.43392922e-02 1.17451355e-01 1.15389752e+00
8.33128095e-01 -1.83899909e-01 8.88141513e-01 -8.24919164e-01
5.03240824e-01 -9.54991162e-01 -7.27434456e-01 1.14520594e-01
-8.02162290e-01 4.02663946e-01 5.10442972e-01 -3.58174801e-01
-1.11324513e+00 1.94185704e-01 -1.08051054e-01 -5.64714491e-01
-2.80464143e-01 5.20175815e-01 7.12204203e-02 6.65118754e-01
3.73790115e-01 1.42127890e-02 5.37944198e-01 -8.87147963e-01
2.94282168e-01 5.44051409e-01 5.60341954e-01 -4.52982485e-01
5.91799200e-01 9.09122229e-01 4.64723408e-01 -5.27803719e-01
-5.93450010e-01 -8.08947459e-02 -7.58130133e-01 1.74840212e-01
9.60706890e-01 -1.16010690e+00 -9.69820201e-01 5.69646120e-01
-1.15639400e+00 -3.06481242e-01 -6.79934919e-01 7.01914668e-01
-1.00182927e+00 3.96266490e-01 -5.96552312e-01 -1.12508178e+00
-3.06509286e-01 -9.32607293e-01 1.24638426e+00 -3.15959215e-01
-4.86860007e-01 -1.22337353e+00 3.72738659e-01 2.37412676e-01
2.45106936e-01 8.70807469e-01 1.06663001e+00 -2.63365954e-01
-3.23371917e-01 -1.78471580e-01 4.56497148e-02 2.08624378e-01
1.27764240e-01 -2.79423356e-01 -8.61487985e-01 -1.26006827e-01
3.38312805e-01 -1.37723759e-02 9.12225008e-01 1.23158336e+00
1.47918773e+00 -3.23432267e-01 -1.91759139e-01 8.79145443e-01
1.35221398e+00 -1.30183354e-01 5.81432343e-01 7.94562027e-02
5.69897711e-01 7.60557294e-01 1.35804445e-01 8.74242663e-01
5.79304934e-01 5.13160169e-01 3.68542314e-01 -1.13945134e-01
4.76824373e-01 -1.53989211e-01 5.60879260e-02 6.29522443e-01
-1.11265577e-01 9.99598727e-02 -1.04727387e+00 8.08186114e-01
-2.48067141e+00 -1.13180888e+00 -6.70439243e-01 2.35419655e+00
7.59262979e-01 -3.47487658e-01 1.26578793e-01 -5.51732592e-02
3.92991960e-01 -2.61799753e-01 -1.00815916e+00 -3.37236762e-01
-3.68071526e-01 9.03243944e-02 4.87060964e-01 3.85197818e-01
-1.04584825e+00 9.82611701e-02 6.99805212e+00 2.99846381e-01
-3.97231311e-01 4.05079871e-01 7.64843106e-01 -2.55740941e-01
-6.57704592e-01 -3.75884265e-01 -3.61403883e-01 7.56381452e-01
1.23079240e+00 1.23847136e-02 3.33786726e-01 5.57932496e-01
6.51945233e-01 2.82034725e-01 -1.29410017e+00 1.15154552e+00
-2.64014572e-01 -1.54425550e+00 -3.87977898e-01 4.48550731e-01
1.04874182e+00 1.94984227e-01 3.11218977e-01 2.16538012e-01
4.74799573e-01 -1.30713141e+00 4.39033538e-01 1.32354355e+00
7.03919113e-01 -8.93115580e-01 7.67603695e-01 4.69708085e-01
-3.24157715e-01 -1.64549381e-01 -3.17295313e-01 -2.58655041e-01
5.30959070e-01 8.61131966e-01 -6.63027614e-02 4.65808719e-01
7.52745926e-01 7.22863495e-01 1.08453430e-01 8.70404959e-01
2.54884958e-01 6.16139531e-01 -3.43127728e-01 6.13512874e-01
-1.18488170e-01 -7.27423847e-01 4.61540192e-01 6.48654282e-01
6.58412933e-01 -1.07196160e-01 -3.63971621e-01 1.22170019e+00
-1.44836642e-02 -3.52764964e-01 -5.63764632e-01 2.42433101e-01
1.37583539e-01 6.48732603e-01 2.06088033e-02 -1.07413933e-01
-5.72755456e-01 9.07530725e-01 6.59646392e-02 6.60435915e-01
-7.65146554e-01 1.91915810e-01 9.93061483e-01 7.24723609e-03
2.51486331e-01 -4.14942592e-01 -9.20476496e-01 -1.44825590e+00
2.35040903e-01 -7.31685817e-01 5.26373923e-01 -5.67819178e-01
-1.94996274e+00 2.01802820e-01 -7.04605430e-02 -1.12228262e+00
-8.25062037e-01 -1.86293274e-01 -2.54733413e-01 1.16629803e+00
-1.31353414e+00 -1.16711438e+00 2.97680255e-02 6.76694810e-01
2.40647539e-01 -2.11589560e-01 1.08920407e+00 1.52697891e-01
-4.96217787e-01 3.62961173e-01 1.25218558e+00 -9.12929401e-02
4.42249864e-01 -1.34698200e+00 5.40057778e-01 6.89077497e-01
-2.13869661e-01 5.36807239e-01 9.57511127e-01 -7.97666252e-01
-1.49064529e+00 -1.12135279e+00 9.97041941e-01 -9.15256619e-01
4.77717429e-01 -5.24141908e-01 -1.13407099e+00 8.44703794e-01
-6.24791123e-02 -1.39268652e-01 9.38911259e-01 4.48194265e-01
-2.55891204e-01 1.46941599e-02 -1.44680488e+00 2.89349049e-01
7.39353180e-01 -4.72791672e-01 -6.90211713e-01 4.08355385e-01
6.65763795e-01 -1.88990533e-01 -1.16800511e+00 3.90427202e-01
6.25084460e-01 -5.73804021e-01 1.02936959e+00 -1.02506781e+00
5.81727207e-01 3.37538086e-02 -2.90580958e-01 -1.05245078e+00
-2.46593952e-01 -7.15789139e-01 -7.41552055e-01 1.09124696e+00
1.26718298e-01 -5.26454866e-01 9.64630961e-01 1.26511061e+00
2.19287559e-01 -4.42123085e-01 -1.37071562e+00 -7.12922454e-01
6.19610906e-01 -7.52530992e-01 8.57784450e-01 1.12247229e+00
-3.38811070e-01 -1.05461091e-01 -9.87388313e-01 7.56084621e-02
1.46385515e+00 2.37969905e-01 6.80142522e-01 -1.79307544e+00
-5.15350178e-02 6.13467768e-02 -8.13310817e-02 -5.91057420e-01
4.89588886e-01 -6.09055579e-01 -3.12136501e-01 -1.64287007e+00
2.12674946e-01 -2.84144133e-01 -2.53776312e-01 2.96784610e-01
1.84626684e-01 -2.99928375e-02 -2.86639601e-01 2.66049474e-01
-1.67023957e-01 8.87358904e-01 7.26952314e-01 -1.93775892e-01
-2.45574728e-01 2.48561844e-01 -4.52456832e-01 6.94210231e-01
5.94450235e-01 -6.90523863e-01 -1.96991384e-01 -6.07537985e-01
4.56599206e-01 6.75159514e-01 7.17777073e-01 -4.57870722e-01
6.87406808e-02 -3.35459381e-01 7.65260100e-01 -7.87155986e-01
4.05190498e-01 -1.13085663e+00 7.90019989e-01 2.92720765e-01
-2.57683158e-01 -1.18825501e-02 8.93996581e-02 9.87694502e-01
-8.34633783e-02 1.65261716e-01 4.52536643e-01 1.31717160e-01
-2.01502696e-01 5.78037918e-01 -3.87974828e-01 7.11724162e-03
6.84459269e-01 -2.49933437e-01 9.28688049e-02 -5.23309827e-01
-1.29401743e+00 2.54066437e-01 3.93064499e-01 3.97675991e-01
5.63154459e-01 -1.51746261e+00 -9.79736507e-01 2.78676897e-01
-5.15328646e-02 1.13174498e-01 7.63665915e-01 9.91490901e-01
3.14752050e-02 1.67068362e-01 -1.75176367e-01 -7.33658910e-01
-6.12632930e-01 7.03343570e-01 3.23241591e-01 1.30606990e-04
-8.43555093e-01 1.71574309e-01 2.71454677e-02 -7.02532291e-01
3.49472404e-01 -4.13013190e-01 2.05581918e-01 1.94971815e-01
3.60413015e-01 7.04482615e-01 -1.92523953e-02 -5.20210445e-01
-2.02381030e-01 3.84571791e-01 3.83031249e-01 -1.08884815e-02
1.88709557e+00 -4.38509434e-01 -2.40016982e-01 8.27255011e-01
1.24950385e+00 -5.10249138e-01 -1.64165783e+00 -2.30838716e-01
1.10623084e-01 -3.09671760e-01 1.64968986e-02 -7.36111879e-01
-6.74561679e-01 9.76015985e-01 6.81585789e-01 -4.20250520e-02
9.82956827e-01 -2.32710674e-01 7.29433477e-01 6.74465373e-02
2.54150957e-01 -9.60394561e-01 -6.41675770e-01 3.33899289e-01
8.68533909e-01 -1.55178154e+00 -1.69523478e-01 3.43947083e-01
-5.09148538e-01 9.13937628e-01 8.01141839e-03 -3.43777277e-02
8.45451891e-01 3.11622143e-01 -8.77396017e-02 -3.75061929e-02
-7.40140498e-01 2.32545704e-01 3.00854087e-01 9.66587961e-01
3.12303156e-01 1.90177262e-01 -1.55498728e-01 9.55820441e-01
-1.10359646e-01 3.35040331e-01 5.83488643e-01 3.41523528e-01
3.40817928e-01 -1.12896633e+00 -4.84710366e-01 6.91242397e-01
-6.09586060e-01 6.48436248e-02 2.11147949e-01 4.98893410e-01
-7.81662762e-03 7.70191014e-01 4.86257821e-01 3.44636470e-01
1.93360358e-01 3.13922942e-01 1.78594500e-01 -4.19140756e-02
-8.92071724e-02 1.90133363e-01 -2.10168988e-01 -4.91886526e-01
-4.67875868e-01 -1.22430301e+00 -9.93087411e-01 -5.54617822e-01
1.30926386e-01 -1.26507916e-02 8.03820074e-01 1.02021468e+00
5.85656226e-01 4.09982055e-01 2.51518607e-01 -5.45392096e-01
-9.30881798e-01 -7.17599630e-01 -5.67417264e-01 6.47598982e-01
9.25793469e-01 -6.15334809e-01 -3.84978056e-01 1.60838693e-01] | [7.192180633544922, 3.592771053314209] |
a3874947-373f-42e6-9c18-18f32ec0e779 | evaluating-open-question-answering-evaluation | 2305.12421 | null | https://arxiv.org/abs/2305.12421v1 | https://arxiv.org/pdf/2305.12421v1.pdf | Evaluating Open Question Answering Evaluation | This study focuses on the evaluation of Open Question Answering (Open-QA) tasks, which have become vital in the realm of artificial intelligence. Current automatic evaluation methods have shown limitations, indicating that human evaluation still remains the most reliable approach. We introduce a new task, QA Evaluation (QA-Eval), designed to assess the accuracy of AI-generated answers in relation to standard answers within Open-QA. Our evaluation of these methods utilizes human-annotated results, and we employ accuracy and F1 score to measure their performance. Specifically, the work investigates methods that show high correlation with human evaluations, deeming them more reliable. We also discuss the pitfalls of current methods, such as their inability to accurately judge responses that contain excessive information. The dataset generated from this work is expected to facilitate the development of more effective automatic evaluation tools. We believe this new QA-Eval task and corresponding dataset will prove valuable for future research in this area. | ['Yue Zhang', 'Yidong Wang', 'Bowen Ding', 'Zhikun Xu', 'Sirui Cheng', 'Cunxiang Wang'] | 2023-05-21 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 2.11159229e-01 3.89359862e-01 2.93433994e-01 -6.33325577e-01
-1.54586935e+00 -9.64524746e-01 5.54613829e-01 4.51416701e-01
-5.45054555e-01 9.49592233e-01 4.38623220e-01 -4.92144793e-01
-1.39009759e-01 -6.92952454e-01 -2.74121463e-01 -9.23672616e-02
3.07030857e-01 8.93569767e-01 4.08076465e-01 -5.79863966e-01
6.16498113e-01 -8.61546397e-02 -1.70980239e+00 7.34531999e-01
1.27240562e+00 1.00930023e+00 -2.59533644e-01 8.12138796e-01
-2.07800627e-01 1.09574926e+00 -1.11104536e+00 -9.61070478e-01
-1.39072761e-01 -6.23530626e-01 -1.56882894e+00 -4.84524220e-01
7.01864779e-01 -1.62456751e-01 4.82511729e-01 8.08393180e-01
6.97530270e-01 1.29087195e-01 6.58810735e-01 -1.07018280e+00
-7.77687013e-01 2.42468163e-01 5.56289136e-01 2.71307945e-01
1.15073121e+00 2.71163285e-01 1.34041536e+00 -4.95747864e-01
7.88295925e-01 1.09559250e+00 4.93361354e-01 5.91267586e-01
-1.15342689e+00 -1.35947704e-01 -5.35432816e-01 6.12462878e-01
-9.99338746e-01 -4.04901147e-01 4.77980465e-01 -4.02535796e-01
8.85598958e-01 6.68262541e-01 -3.04550622e-02 9.24707770e-01
-2.00464502e-02 6.60619557e-01 1.67690039e+00 -7.65803993e-01
3.82212967e-01 4.11835521e-01 5.74927628e-01 4.95967299e-01
1.68154746e-01 -8.73421878e-02 -2.87160873e-01 -4.09016728e-01
9.45118293e-02 -8.30678523e-01 -5.55162549e-01 -1.13070324e-01
-1.13005567e+00 1.02522314e+00 3.61916929e-01 5.14428496e-01
-4.87282544e-01 -3.84620547e-01 4.74328160e-01 7.14741468e-01
3.96905065e-01 1.20624733e+00 -5.14895499e-01 -5.69243789e-01
-6.29143059e-01 6.32430613e-01 1.31427121e+00 3.42647344e-01
7.10632801e-01 -4.97478038e-01 -5.27354002e-01 1.03045321e+00
4.77997437e-02 3.16865355e-01 5.43237627e-01 -1.44335651e+00
4.54906642e-01 1.00295234e+00 5.85659266e-01 -9.60560620e-01
-2.57846564e-01 -1.29923940e-01 3.34823993e-03 -1.22404858e-01
6.96095288e-01 -6.18197396e-02 -3.97622854e-01 1.55716848e+00
3.65588278e-01 -7.08872914e-01 2.70164013e-01 8.54097664e-01
1.14038444e+00 5.49247146e-01 2.18954563e-01 -1.63605988e-01
1.57158601e+00 -8.49979401e-01 -9.73559439e-01 1.02389872e-01
9.71955538e-01 -8.28537881e-01 1.60824835e+00 3.55737388e-01
-1.08796370e+00 -4.65598136e-01 -8.21169376e-01 -2.34294295e-01
-6.04177952e-01 -1.83927685e-01 3.96036506e-01 8.68903756e-01
-1.02235472e+00 2.27239981e-01 -1.70869902e-01 -2.36483559e-01
-2.13108994e-02 1.41691610e-01 -2.24868998e-01 -1.68814808e-01
-1.53297794e+00 1.44284868e+00 1.28585532e-01 -1.46382913e-01
-4.83805418e-01 -3.70777935e-01 -8.82631361e-01 -3.62332165e-02
5.14531374e-01 -6.98407114e-01 1.89273274e+00 -9.74605083e-01
-1.15228426e+00 1.15519750e+00 -2.12035701e-01 -5.34830213e-01
3.02114338e-01 -1.66861147e-01 -5.63754320e-01 3.21459711e-01
4.60508823e-01 4.89785463e-01 2.97536790e-01 -1.09426451e+00
-4.06646132e-01 -4.89399761e-01 4.92839485e-01 1.35598913e-01
-3.61543328e-01 3.82392436e-01 2.00821832e-01 -2.78925210e-01
-1.59239158e-01 -6.83754265e-01 -2.86407676e-03 -5.66870511e-01
2.88015027e-02 -9.38966453e-01 2.75649756e-01 -8.35128367e-01
1.57225418e+00 -1.44638431e+00 -1.43367261e-01 -5.41644134e-02
1.91799358e-01 4.83505249e-01 -2.63417721e-01 7.99704373e-01
2.67580688e-01 2.71566808e-01 -3.01742792e-01 1.92346185e-01
2.25616097e-01 1.43873379e-01 -2.27919698e-01 -1.86415434e-01
2.58299053e-01 1.05158794e+00 -9.62735057e-01 -7.76712656e-01
-1.43230811e-01 -9.56003368e-03 -2.47372240e-01 4.47998732e-01
-6.69203818e-01 1.73168197e-01 -5.06931901e-01 5.51749229e-01
1.28938809e-01 -1.99864402e-01 -2.53530532e-01 -1.09880485e-01
-4.76072505e-02 7.51096129e-01 -7.95240343e-01 1.39892292e+00
-4.88318652e-01 4.95092124e-01 -2.44671553e-01 -6.32134855e-01
9.10578668e-01 5.65516353e-01 -8.32023248e-02 -1.07799971e+00
2.75400490e-01 5.22205651e-01 1.45002231e-01 -1.00458694e+00
6.77287221e-01 -1.06512969e-02 -6.89523071e-02 5.72120011e-01
1.03940114e-01 -6.16670310e-01 7.30114937e-01 2.16518462e-01
1.20305872e+00 2.68221106e-02 3.06597143e-01 -3.44933629e-01
9.00297463e-01 4.94683027e-01 1.47099406e-01 7.85857260e-01
-4.85020310e-01 5.09217799e-01 4.96232182e-01 -4.76652086e-01
-7.61136532e-01 -7.88235188e-01 -2.54259706e-01 1.01466751e+00
-2.06342191e-01 -4.49602991e-01 -1.15390599e+00 -7.89964795e-01
-4.80852753e-01 1.08456957e+00 -4.94131148e-01 1.33036211e-01
-3.73795033e-01 -3.28089297e-01 7.10062981e-01 2.37738416e-01
3.39923590e-01 -1.30491459e+00 -9.55795228e-01 2.58629054e-01
-1.02562523e+00 -1.01976037e+00 6.00283146e-02 -1.56071305e-01
-8.42862785e-01 -1.36524415e+00 -6.81198835e-01 -7.32383192e-01
2.38094166e-01 -8.16637874e-02 1.67386484e+00 3.39596003e-01
9.81045291e-02 7.46421754e-01 -8.49811435e-01 -4.84446943e-01
-7.87621856e-01 1.82128921e-01 -4.61536348e-01 -2.26344630e-01
8.17073584e-01 -2.72266474e-02 -4.77134556e-01 6.28488719e-01
-1.07267797e+00 -2.86137193e-01 2.92159885e-01 8.40152264e-01
1.71056911e-01 -3.49689066e-01 9.35793698e-01 -1.07070518e+00
1.49328518e+00 -3.61441851e-01 -2.74370790e-01 6.04135096e-01
-7.51821280e-01 1.19984485e-01 3.60112280e-01 -4.74016443e-02
-1.21577847e+00 -6.88912809e-01 -5.08787155e-01 5.14042735e-01
-4.22630101e-01 7.83674538e-01 1.62045479e-01 -2.28527725e-01
1.33727741e+00 -2.61149675e-01 -1.24500677e-01 -2.45690107e-01
1.44322351e-01 1.04603958e+00 6.30237043e-01 -7.38651037e-01
3.73330086e-01 -5.98250926e-02 -3.14725250e-01 -6.83770537e-01
-1.10556185e+00 -6.10019743e-01 9.56480205e-03 -4.57684427e-01
8.01100254e-01 -5.00374556e-01 -5.85601628e-01 -2.80633103e-02
-1.30420399e+00 -1.30028918e-01 -1.70840845e-01 2.77757049e-01
-6.21604383e-01 4.84847546e-01 -6.51485384e-01 -1.01291633e+00
-5.48111141e-01 -1.09126258e+00 1.00901747e+00 2.11109117e-01
-9.80116665e-01 -9.40630794e-01 5.41975319e-01 1.36285651e+00
5.24022102e-01 1.37834132e-01 8.81520987e-01 -1.06191385e+00
-1.35455102e-01 -3.52996975e-01 1.09680943e-01 6.08872533e-01
-3.40062678e-01 -9.95819047e-02 -1.18761587e+00 1.27445877e-01
1.68305397e-01 -1.17372799e+00 3.55420709e-01 -1.26018897e-01
7.81847239e-01 -1.82341978e-01 2.91846573e-01 -4.27934438e-01
1.16228247e+00 3.46015170e-02 9.02649581e-01 5.47867000e-01
-6.50353134e-02 1.15811992e+00 9.36227977e-01 -1.75779000e-01
5.94071686e-01 5.68327248e-01 2.49331743e-01 2.89228261e-01
-6.49264902e-02 -1.42447501e-01 2.53544450e-01 9.00350749e-01
-1.92771509e-01 -3.82318377e-01 -1.04154420e+00 7.14096785e-01
-1.53147733e+00 -8.11116755e-01 -5.50224662e-01 2.10639286e+00
8.28334272e-01 7.28180036e-02 -7.11631551e-02 4.39105690e-01
4.23701674e-01 -7.60757849e-02 -2.61602248e-03 -8.55796695e-01
-1.59074217e-01 6.07574463e-01 -4.69651967e-01 5.18973231e-01
-7.18267441e-01 6.40928328e-01 6.61395311e+00 6.23135746e-01
-3.13300490e-01 1.96203336e-01 5.89320421e-01 6.50827825e-01
-6.11391187e-01 5.31919599e-02 -3.55350643e-01 3.42426598e-01
1.35791028e+00 4.96672876e-02 1.49874538e-01 6.80858374e-01
-1.14616789e-01 -3.88917744e-01 -1.14350629e+00 4.63738084e-01
3.09585929e-01 -8.84880960e-01 2.66868789e-02 -3.36993486e-01
5.98777413e-01 -2.37262860e-01 -3.51120621e-01 7.41159916e-01
1.08668685e-01 -8.44885468e-01 3.32795322e-01 3.92138571e-01
3.43994051e-01 -3.75153601e-01 1.32605505e+00 4.05121863e-01
-3.93112808e-01 1.65658165e-02 -1.77196920e-01 -4.40850526e-01
2.62311935e-01 2.49406621e-01 -9.57807302e-01 5.50410867e-01
6.87484503e-01 -1.88898370e-01 -9.42745149e-01 9.32213843e-01
-5.37090540e-01 8.01698625e-01 -5.39629795e-02 -6.55296385e-01
1.31103158e-01 -1.91606075e-01 3.55852157e-01 8.36175025e-01
-4.38604318e-02 1.31210089e-01 -1.05276868e-01 7.55174577e-01
5.88787608e-02 4.93352503e-01 -6.11071587e-01 -2.48619273e-01
3.63595039e-01 1.05315351e+00 -4.42376494e-01 -4.31840181e-01
-5.02326906e-01 7.97434926e-01 5.01525760e-01 4.36140411e-02
-4.77779806e-01 -4.51280743e-01 1.31170541e-01 -7.48266354e-02
-2.73096621e-01 1.78626478e-01 -2.43685469e-01 -1.03511751e+00
3.13315809e-01 -1.40616667e+00 8.89289320e-01 -1.18113101e+00
-1.41982973e+00 9.89493728e-01 5.18033579e-02 -8.98866355e-01
-6.31050229e-01 -6.19936526e-01 -2.55693614e-01 6.86943591e-01
-1.31412268e+00 -6.78182364e-01 -4.32044953e-01 3.50250334e-01
6.05376184e-01 2.15388820e-01 1.26411796e+00 2.94073015e-01
-3.46496068e-02 5.00458539e-01 -3.51473570e-01 -2.71678977e-02
9.14181650e-01 -1.18055069e+00 1.29877523e-01 4.87204015e-01
2.66945511e-01 5.22849619e-01 9.25410628e-01 -4.83497769e-01
-8.65258992e-01 -2.70235628e-01 1.51409245e+00 -1.18243384e+00
5.91695547e-01 2.39117399e-01 -1.07264221e+00 3.32936823e-01
4.68679041e-01 -5.90681732e-01 9.76420045e-01 2.17880264e-01
-3.05675745e-01 1.57668203e-01 -1.19657612e+00 3.95427287e-01
5.82604885e-01 -6.92055821e-01 -1.43309724e+00 3.69809151e-01
8.35648239e-01 -3.25437963e-01 -9.43926871e-01 6.31776333e-01
3.15154463e-01 -1.06024849e+00 6.99548662e-01 -8.38782728e-01
5.25733650e-01 -1.68769762e-01 -1.97125450e-01 -1.26670527e+00
4.85355780e-02 -2.67424583e-01 1.14161193e-01 1.19604552e+00
7.64712870e-01 -7.04770982e-01 4.38958466e-01 9.44987237e-01
-1.58677235e-01 -6.00090563e-01 -8.01275373e-01 -5.79071581e-01
4.26000208e-02 -4.64014173e-01 4.73516047e-01 1.00807118e+00
1.46494687e-01 7.46061325e-01 1.31321087e-01 -1.03662498e-01
1.64788216e-01 1.49717527e-02 8.00760686e-01 -1.28926969e+00
-3.32024731e-02 -2.75738537e-01 -4.29647535e-01 -6.41112506e-01
-4.66349348e-03 -5.14634669e-01 7.32157603e-02 -1.58177221e+00
6.19095843e-03 -1.50575414e-01 3.47785801e-02 1.44245014e-01
-5.17549634e-01 4.16837186e-01 6.37635291e-02 -1.00832470e-01
-8.67099881e-01 4.91885006e-01 1.28013051e+00 -2.47129090e-02
6.22383319e-02 1.46519893e-03 -6.92348719e-01 5.19331157e-01
8.79759610e-01 -3.88263434e-01 -3.39511842e-01 -4.34314758e-01
3.98937017e-01 1.23233058e-01 2.17657462e-01 -1.03981841e+00
1.03824824e-01 1.47425081e-03 -1.42737970e-01 -2.69439816e-01
2.62841463e-01 -6.24515355e-01 -1.75277710e-01 3.29001039e-01
-5.96606314e-01 5.35409033e-01 -3.13266553e-02 3.63796949e-02
-6.20682180e-01 -8.78038704e-01 1.43402740e-01 -2.86979347e-01
-6.05753422e-01 -3.44446927e-01 -3.68474126e-01 6.38452470e-01
7.82254815e-01 -1.14772029e-01 -5.13458014e-01 -7.43450403e-01
-3.71609479e-01 4.63124394e-01 3.32205236e-01 3.85826975e-01
4.66499060e-01 -1.03543091e+00 -8.91979575e-01 -4.00695592e-01
6.27355576e-01 -3.03441405e-01 1.46943182e-01 6.29365504e-01
-8.51737797e-01 7.07566082e-01 -1.52191564e-01 -3.07980448e-01
-1.37932861e+00 4.08537567e-01 1.94025367e-01 -5.54279745e-01
5.07738292e-02 5.46788096e-01 -4.15520310e-01 -7.50857115e-01
2.84451455e-01 -6.22893646e-02 -6.45146608e-01 -8.34279656e-02
7.00220764e-01 6.30411267e-01 4.69261765e-01 -5.77754617e-01
-2.14004219e-02 6.66415095e-02 1.09043211e-01 -4.60038483e-01
8.99182916e-01 1.28269777e-01 -4.26390588e-01 5.10615766e-01
8.23373616e-01 3.97107974e-02 -1.38171107e-01 -1.40021726e-01
4.06947702e-01 -4.03170258e-01 -1.41188964e-01 -1.30368781e+00
-5.87823847e-03 8.59202564e-01 5.66219509e-01 7.34123766e-01
9.84239399e-01 -4.46138121e-02 7.51640260e-01 7.57028818e-01
5.28935790e-01 -1.23844731e+00 4.45689969e-02 5.27956545e-01
1.05989873e+00 -1.56312454e+00 -2.66581863e-01 -3.38859528e-01
-6.72701240e-01 1.04390812e+00 9.00780737e-01 4.93696183e-01
-1.27086982e-01 -4.88773257e-01 5.68054259e-01 -5.48222899e-01
-7.99913168e-01 -3.15078467e-01 4.19037610e-01 4.82326925e-01
8.60594451e-01 1.01764478e-01 -1.13231468e+00 5.43128908e-01
-4.20654029e-01 1.09095924e-01 4.63711202e-01 9.87290859e-01
-4.51592475e-01 -1.26953375e+00 -6.43244028e-01 5.14796436e-01
-6.17768705e-01 1.73724256e-02 -9.97455955e-01 6.73336923e-01
-2.16364369e-01 1.60996997e+00 -3.60372603e-01 -3.02993417e-01
5.67698658e-01 4.76189524e-01 4.77645487e-01 -5.31540871e-01
-9.89575326e-01 -7.58116066e-01 9.12051499e-01 -4.67287838e-01
-7.21148789e-01 -3.90123039e-01 -6.90416336e-01 -8.62717107e-02
-5.61707556e-01 8.79694641e-01 6.14632785e-01 8.87894332e-01
2.68492460e-01 1.73446655e-01 2.41933376e-01 -7.53995851e-02
-9.93745625e-01 -1.27699780e+00 2.75253244e-02 8.40214729e-01
-6.58807829e-02 -5.25640965e-01 -3.61393720e-01 -1.01320468e-01] | [11.416780471801758, 8.171406745910645] |
3524f274-d816-4249-ac5c-7dd867d9ac95 | near-optimal-high-probability-convergence-for | 2302.06032 | null | https://arxiv.org/abs/2302.06032v1 | https://arxiv.org/pdf/2302.06032v1.pdf | Near-Optimal High-Probability Convergence for Non-Convex Stochastic Optimization with Variance Reduction | Traditional analyses for non-convex stochastic optimization problems characterize convergence bounds in expectation, which is inadequate as it does not supply a useful performance guarantee on a single run. Motivated by its importance, an emerging line of literature has recently studied the high-probability convergence behavior of several algorithms, including the classic stochastic gradient descent (SGD). However, no high-probability results are established for optimization algorithms with variance reduction, which is known to accelerate the convergence process and has been the de facto algorithmic technique for stochastic optimization at large. To close this important gap, we introduce a new variance-reduced algorithm for non-convex stochastic optimization, which we call Generalized SignSTORM. We show that with probability at least $1-\delta$, our algorithm converges at the rate of $O(\log(dT/\delta)/T^{1/3})$ after $T$ iterations where $d$ is the problem dimension. This convergence guarantee matches the existing lower bound up to a log factor, and to our best knowledge, is the first high-probability minimax (near-)optimal result. Finally, we demonstrate the effectiveness of our algorithm through numerical experiments. | ['Zhengyuan Zhou', 'Srikanth Jagabathula', 'Perry Dong', 'Zijian Liu'] | 2023-02-13 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 1.18383961e-02 -7.95881078e-02 -9.60207134e-02 -2.02685699e-01
-1.13900554e+00 -5.84376395e-01 -3.22235003e-02 3.09266210e-01
-7.08272398e-01 9.16074574e-01 -1.67830274e-01 -6.68569744e-01
-4.60601419e-01 -4.76191461e-01 -8.79099250e-01 -1.12272012e+00
-1.75291911e-01 4.06122059e-01 2.21862830e-02 -1.73132896e-01
2.33850628e-01 2.57538199e-01 -1.19611549e+00 -5.41396201e-01
1.13485110e+00 1.20078349e+00 -3.93720977e-02 6.10930264e-01
1.57711059e-02 2.28068575e-01 -2.26486519e-01 -6.42868817e-01
5.64681768e-01 -7.91129708e-01 -6.46288276e-01 -1.92679968e-02
2.23961368e-01 -2.18386669e-02 2.17023250e-02 1.54304862e+00
6.25372827e-01 2.73066044e-01 3.09277385e-01 -1.07667243e+00
-2.44504064e-01 7.96620071e-01 -1.04324365e+00 3.06270689e-01
7.87123367e-02 -1.29485279e-01 1.13780344e+00 -7.60706484e-01
2.39946276e-01 9.32221174e-01 6.33373260e-01 3.68744403e-01
-9.24671710e-01 -6.91546023e-01 3.15525383e-01 -1.82622373e-01
-1.39414155e+00 -1.36973247e-01 6.26600802e-01 -2.08881512e-01
4.22717482e-01 4.94332045e-01 3.19341809e-01 4.78033841e-01
-1.27256170e-01 8.84831429e-01 1.26630723e+00 -5.86455286e-01
2.93476552e-01 1.42679721e-01 2.48945475e-01 9.85596240e-01
7.58300245e-01 -2.35028807e-02 -6.14644706e-01 -2.86507934e-01
2.88153648e-01 -2.68995702e-01 -5.48854291e-01 -3.22130144e-01
-8.54783952e-01 8.62695634e-01 -2.44999249e-02 2.34703794e-01
-1.98180184e-01 4.66209888e-01 2.71608889e-01 4.06573296e-01
6.92725480e-01 3.86304706e-02 -6.05354667e-01 -6.86993420e-01
-9.23698604e-01 2.87150055e-01 1.10023987e+00 8.09138834e-01
3.16636592e-01 1.33377925e-01 -1.64503865e-02 5.96352160e-01
1.84917063e-01 1.03726339e+00 -4.52952310e-02 -7.23790586e-01
8.56736958e-01 9.35481638e-02 5.21992326e-01 -8.26565146e-01
-2.50566989e-01 -7.73819864e-01 -8.82436872e-01 1.62411600e-01
9.43269849e-01 -4.90079433e-01 -2.31486022e-01 1.88001478e+00
3.31479639e-01 -4.77921851e-02 -2.66403139e-01 9.64117706e-01
-2.35590443e-01 5.97372413e-01 -5.65265357e-01 -8.55393350e-01
8.94536197e-01 -5.55586219e-01 -6.20398343e-01 -3.32332030e-02
6.89393759e-01 -6.96326077e-01 1.26710916e+00 5.56640387e-01
-1.20360053e+00 2.00124919e-01 -1.06871819e+00 5.21214664e-01
1.53664738e-01 -1.74759150e-01 6.62426353e-01 1.24342620e+00
-8.26810956e-01 5.19983888e-01 -8.57894659e-01 1.62078179e-02
3.36455464e-01 3.78931016e-01 1.12363376e-01 -5.34502091e-03
-6.08128488e-01 5.08544028e-01 -7.65461996e-02 3.01447123e-01
-5.22971570e-01 -8.50398123e-01 -5.03513932e-01 6.54064724e-03
8.39127362e-01 -6.11793935e-01 1.13232994e+00 -6.09438598e-01
-1.66673815e+00 6.15156412e-01 -5.87307334e-01 -6.34333909e-01
8.82568717e-01 -4.13822949e-01 1.31684169e-01 -2.72875596e-02
-7.78364688e-02 -4.68673557e-01 7.67528772e-01 -1.06989825e+00
-8.06434989e-01 -6.86642110e-01 8.84769186e-02 2.00174794e-01
-4.88636136e-01 2.39510372e-01 -3.55277568e-01 -5.13013244e-01
1.39071703e-01 -1.00451803e+00 -5.87217689e-01 -9.14621055e-02
-4.19848889e-01 -4.02089283e-02 1.48089245e-01 -3.63008499e-01
1.55867302e+00 -2.15125275e+00 1.80440079e-02 4.82436508e-01
2.72468120e-01 8.44667032e-02 4.28767592e-01 3.47617596e-01
1.65294513e-01 3.55187982e-01 -6.60247147e-01 -4.90785152e-01
2.27955908e-01 -2.05938742e-01 -3.28976840e-01 1.06504977e+00
-5.20203650e-01 5.05983710e-01 -1.09093630e+00 -1.72468260e-01
3.00457850e-02 1.89057216e-01 -5.88250220e-01 -2.87731975e-01
4.40266691e-02 1.07585294e-02 -5.49242198e-01 3.15573812e-01
8.83720160e-01 -5.17821848e-01 2.09135994e-01 1.78995371e-01
-1.12507649e-01 -7.30951130e-02 -1.63465655e+00 1.34404576e+00
-5.32831013e-01 4.76045519e-01 4.52874064e-01 -1.30110729e+00
5.06822884e-01 1.09295055e-01 6.31807625e-01 -2.26688549e-01
2.11973056e-01 6.06995523e-01 -2.05109373e-01 -1.13753490e-01
2.89233774e-01 -4.44689065e-01 -2.63415333e-02 4.99733150e-01
-4.69274700e-01 1.36652738e-01 4.16602790e-01 9.66470987e-02
9.85994220e-01 -3.25778127e-01 1.68342143e-01 -5.62254727e-01
5.01514554e-01 -3.54244709e-01 5.35316348e-01 1.16441953e+00
-2.24923581e-01 4.95294005e-01 6.76696777e-01 -2.75036693e-02
-7.71938264e-01 -1.01147330e+00 -1.21199526e-01 1.09233940e+00
3.35562706e-01 -3.52081239e-01 -7.63311923e-01 -4.79166985e-01
-5.12956716e-02 6.42534137e-01 -5.57466209e-01 2.94699352e-02
-4.29394215e-01 -1.43537474e+00 1.45479620e-01 1.90902412e-01
4.27511036e-01 -4.13165927e-01 -4.13583547e-01 3.03658098e-01
1.28686398e-01 -1.08067214e+00 -9.15001452e-01 2.17309505e-01
-1.01665568e+00 -1.01324379e+00 -9.00527060e-01 -5.46179354e-01
8.47619116e-01 3.11477482e-01 7.75125623e-01 -8.33879337e-02
7.85523579e-02 2.71919459e-01 -1.79262087e-01 -3.90475094e-01
-1.59223735e-01 -7.29183331e-02 3.44435155e-01 2.60644376e-01
4.14905045e-03 -5.71519554e-01 -6.09001160e-01 1.18893668e-01
-7.39585996e-01 -2.72318721e-01 4.66586649e-01 7.28278160e-01
7.29996800e-01 3.28089178e-01 2.19824240e-01 -8.07555914e-01
8.34468305e-01 -3.95796090e-01 -1.24503684e+00 2.08257347e-01
-9.26770806e-01 3.38565975e-01 1.00295520e+00 -1.95272610e-01
-8.20508420e-01 -7.40687400e-02 -5.14321215e-02 -1.87471479e-01
4.63282406e-01 7.18969226e-01 7.90768638e-02 -2.52592981e-01
4.90468591e-01 5.09022713e-01 -4.29858305e-02 -6.18438423e-01
3.54669631e-01 3.74362558e-01 4.22333926e-01 -7.46684372e-01
9.77331400e-01 8.97666812e-01 4.57100362e-01 -7.25257039e-01
-1.30143464e+00 -3.99963289e-01 -4.91929799e-02 -1.61337352e-03
3.77439767e-01 -4.21699792e-01 -1.21322608e+00 2.84952283e-01
-6.14702582e-01 -1.52770802e-01 -2.58364171e-01 6.76736295e-01
-6.64416313e-01 6.11330390e-01 -3.21189195e-01 -1.51044238e+00
-4.22281593e-01 -9.22799468e-01 5.06022155e-01 1.62918612e-01
2.67155468e-01 -9.67168033e-01 -1.36684448e-01 1.58173099e-01
4.56682295e-01 4.23118651e-01 4.64798152e-01 -3.88377368e-01
-3.87178391e-01 -3.15421671e-01 -9.56673175e-02 5.56493521e-01
-6.55489340e-02 -2.87541598e-01 -3.02643776e-01 -4.39279735e-01
5.09812891e-01 9.25613269e-02 8.17297339e-01 6.45827115e-01
1.12874603e+00 -5.28458416e-01 -3.06701660e-01 7.33493149e-01
1.70054436e+00 -2.95018833e-02 1.08246647e-01 1.44976437e-01
3.51186603e-01 8.41240957e-02 5.77923179e-01 9.98241246e-01
9.62352902e-02 4.26981032e-01 3.17426741e-01 2.02741817e-01
3.84923160e-01 -8.73222649e-02 4.96563584e-01 8.12536120e-01
-1.62608057e-01 -4.15413976e-01 -7.28279471e-01 4.84591901e-01
-1.75229061e+00 -8.57481897e-01 -2.63569146e-01 2.76693320e+00
9.16035235e-01 4.29176748e-01 1.86658084e-01 2.88460374e-01
8.15028906e-01 1.78839527e-02 -5.79455078e-01 -3.57995033e-01
-2.80922294e-01 5.10866702e-01 9.41479504e-01 7.81204045e-01
-8.33722472e-01 6.77355587e-01 6.06192636e+00 1.23462880e+00
-7.69536614e-01 2.03234866e-01 6.81122839e-01 -5.37617385e-01
-3.17513913e-01 7.79483542e-02 -8.72709036e-01 7.22360373e-01
8.72675836e-01 -6.32346988e-01 5.35123825e-01 9.87208962e-01
4.51739132e-01 -1.82075948e-01 -7.00263977e-01 1.29079652e+00
2.07761545e-02 -1.21788716e+00 -2.89674252e-01 4.44606751e-01
9.32264507e-01 -2.03105673e-01 2.08526805e-01 -7.79037625e-02
4.84926432e-01 -8.14214170e-01 5.15599370e-01 1.61591694e-01
4.40614492e-01 -1.19540811e+00 8.59657645e-01 4.09577727e-01
-1.09865999e+00 -1.64377794e-01 -2.50418544e-01 6.59777895e-02
6.07208490e-01 1.18693411e+00 -2.46963248e-01 5.02715290e-01
6.28699422e-01 1.40362486e-01 7.92530458e-03 1.32066858e+00
-1.22090146e-01 1.03044128e+00 -1.03865004e+00 -5.96463501e-01
3.66193652e-01 -5.61289966e-01 1.00415468e+00 1.00796175e+00
4.80179608e-01 1.79294840e-01 1.11995727e-01 3.71923059e-01
-3.19664329e-01 3.55415821e-01 -1.61075935e-01 -1.44600519e-03
4.40636933e-01 7.35263109e-01 -8.99487257e-01 -1.22995615e-01
-3.42607170e-01 9.90084231e-01 1.53358117e-01 2.46533677e-01
-9.63525593e-01 -5.90570211e-01 7.19815433e-01 -4.25651930e-02
5.19956768e-01 -5.69478095e-01 -5.63486218e-01 -1.05923116e+00
7.55473197e-01 -5.76072037e-01 4.88693655e-01 6.96091577e-02
-1.31954432e+00 2.02413082e-01 -7.78394938e-02 -9.40770924e-01
-1.42319854e-02 -4.29660171e-01 -1.13426253e-01 7.28408158e-01
-1.26060796e+00 -2.90246665e-01 1.74685955e-01 3.43375027e-01
2.01585025e-01 2.31638588e-02 3.66373122e-01 1.97496250e-01
-5.68025827e-01 9.24946487e-01 8.87256384e-01 -1.42899171e-01
2.42251426e-01 -1.37638175e+00 -1.31918848e-01 1.21761763e+00
3.74518409e-02 5.27362525e-01 1.12595117e+00 -2.07142457e-01
-1.83184826e+00 -4.48693097e-01 7.95042932e-01 -2.49571666e-01
8.90437543e-01 -2.31251746e-01 -5.61864495e-01 1.95811048e-01
-2.17009246e-01 1.41932433e-02 7.34483719e-01 2.63559818e-01
-1.07667133e-01 -3.89160007e-01 -1.07060373e+00 7.70018399e-01
1.18424857e+00 -2.19597057e-01 -1.83288261e-01 4.60359633e-01
5.06877959e-01 -5.44251502e-01 -7.36475050e-01 4.00635660e-01
3.61850917e-01 -9.98592675e-01 7.64184296e-01 -4.42237377e-01
1.15635194e-01 -2.78237492e-01 -3.97813976e-01 -9.60154116e-01
2.76261717e-01 -1.43299091e+00 -4.01187390e-01 8.91650438e-01
5.82477689e-01 -7.88590312e-01 9.86437201e-01 5.10739625e-01
1.73134245e-02 -1.18693650e+00 -1.25358307e+00 -1.36095965e+00
3.45858961e-01 -8.71966660e-01 1.94892779e-01 5.33053696e-01
-4.39289324e-02 -7.00787129e-03 -5.85087359e-01 2.53967732e-01
8.53598595e-01 3.27946782e-01 5.18862009e-01 -7.13336945e-01
-6.78093672e-01 -9.28619444e-01 -1.13030843e-01 -1.46893692e+00
-1.08092003e-01 -8.29360962e-01 9.73334014e-02 -1.14574945e+00
3.17810893e-01 -4.88114983e-01 -4.79719639e-01 -7.53735155e-02
-4.98344868e-01 -5.62202260e-02 1.08280435e-01 -1.07937939e-01
-7.34209836e-01 8.01127255e-01 1.12680829e+00 2.25562975e-01
-3.47413480e-01 4.02140290e-01 -1.09224081e+00 6.88430130e-01
8.43797982e-01 -6.83119953e-01 -2.84041464e-01 -2.91341007e-01
6.61049306e-01 2.83819661e-02 -1.79565236e-01 -9.35005963e-01
1.20691918e-01 -2.20764741e-01 -3.47584397e-01 -4.16905791e-01
9.59557071e-02 -5.51866770e-01 -2.86648303e-01 6.17877305e-01
-3.30024540e-01 -1.76590122e-02 -9.11415368e-02 8.36220145e-01
5.80600016e-02 -4.54813838e-01 9.66102540e-01 2.24543571e-01
-2.19305512e-02 5.87469399e-01 -2.62487411e-01 5.41927695e-01
8.58129621e-01 2.33421717e-02 1.03780583e-01 -6.59671247e-01
-3.23024184e-01 2.71284103e-01 1.37104347e-01 -1.23545408e-01
3.22565943e-01 -8.99225473e-01 -9.25551116e-01 -4.29199576e-01
-3.72138768e-01 -7.18419850e-02 1.74496487e-01 1.20154619e+00
-4.70140100e-01 3.41675311e-01 7.85797954e-01 -4.85445857e-01
-1.07222188e+00 5.38610339e-01 2.62195468e-01 -4.09890354e-01
-5.01521766e-01 1.19119656e+00 -1.20816514e-01 1.01534985e-01
2.62206346e-01 -2.65989572e-01 5.70916891e-01 -3.44418257e-01
6.25942230e-01 6.59147143e-01 -2.56104581e-02 -1.94516212e-01
-4.70448524e-01 5.31652868e-01 5.02650663e-02 -6.21928096e-01
1.22191000e+00 -3.22959125e-01 -1.77342221e-01 5.00993669e-01
1.47906792e+00 5.81558466e-01 -1.23420584e+00 -3.17285359e-01
3.99423316e-02 -5.80426157e-01 -6.87178299e-02 -3.73449266e-01
-1.16411948e+00 7.20084727e-01 5.90390801e-01 3.37940097e-01
1.21065688e+00 -1.13085888e-01 8.90147805e-01 5.98985791e-01
6.05176270e-01 -1.35357416e+00 -1.62813663e-01 7.07348824e-01
6.25299454e-01 -1.21286416e+00 1.11725718e-01 -4.71777469e-01
-2.81213284e-01 8.43352616e-01 9.54104960e-02 -1.68268427e-01
9.79019463e-01 2.99274802e-01 -3.91105980e-01 1.68689981e-01
-4.80470628e-01 -2.67113805e-01 2.10501198e-02 1.51551172e-01
3.41763228e-01 1.09013334e-01 -9.50322568e-01 6.71730220e-01
-5.78590930e-01 -8.57206360e-02 4.12773311e-01 8.27501535e-01
-5.19463241e-01 -1.00889277e+00 -2.75771260e-01 5.59357703e-01
-1.01020896e+00 -3.04942310e-01 7.67865404e-02 6.42580271e-01
-5.00857294e-01 1.12105048e+00 -4.38242525e-01 -1.06267862e-01
9.31876227e-02 -1.54764786e-01 5.41272700e-01 -9.22485068e-02
-3.89532119e-01 -1.82036385e-01 -6.56931549e-02 -4.57931817e-01
-1.46800086e-01 -7.11060703e-01 -1.16401517e+00 -6.59732461e-01
-3.93825501e-01 5.25086582e-01 7.95037210e-01 1.02961946e+00
1.91020384e-01 2.07903851e-02 1.00459504e+00 -2.27468982e-01
-1.06820118e+00 -5.84987640e-01 -7.01050580e-01 2.03455314e-01
1.75637528e-01 -3.39642406e-01 -8.57180715e-01 -2.73236215e-01] | [6.639564037322998, 4.442726135253906] |
c648e18c-bdd0-421b-a3b4-6fb98f0a2543 | a-multi-class-structured-dictionary-learning | 1812.01389 | null | http://arxiv.org/abs/1812.01389v1 | http://arxiv.org/pdf/1812.01389v1.pdf | A multi-class structured dictionary learning method using discriminant atom selection | In the last decade, traditional dictionary learning methods have been
successfully applied to various pattern classification tasks. Although these
methods produce sparse representations of signals which are robust against
distortions and missing data, such representations quite often turn out to be
unsuitable if the final objective is signal classification. In order to
overcome or at least to attenuate such a weakness, several new methods which
incorporate discriminative information into sparse-inducing models have emerged
in recent years. In particular, methods for discriminative dictionary learning
have shown to be more accurate (in terms of signal classification) than the
traditional ones, which are only focused on minimizing the total representation
error. In this work, we present both a novel multi-class discriminative measure
and an innovative dictionary learning method. For a given dictionary, this new
measure, which takes into account not only when a particular atom is used for
representing signals coming from a certain class and the magnitude of its
corresponding representation coefficient, but also the effect that such an atom
has in the total representation error, is capable of efficiently quantifying
the degree of discriminability of each one of the atoms. On the other hand, the
new dictionary construction method yields dictionaries which are highly
suitable for multi-class classification tasks. Our method was tested with a
widely used database for handwritten digit recognition and compared with three
state-of-the-art classification methods. The results show that our method
significantly outperforms the other three achieving good recognition rates and
additionally, reducing the computational cost of the classifier. | ['H. L. Rufiner', 'R. E. Rolón', 'L. E. Di Persia', 'R. D. Spies'] | 2018-12-04 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 3.38047296e-01 -3.63167137e-01 -1.31792486e-01 -1.87867448e-01
-5.18155158e-01 -3.08894128e-01 6.25253439e-01 6.63641393e-01
-4.38717365e-01 6.53073072e-01 7.91186932e-03 1.09193787e-01
-3.35421830e-01 -9.51347470e-01 -2.91353106e-01 -1.08572519e+00
1.39391735e-01 4.25742507e-01 8.10275003e-02 -2.51127809e-01
2.87008882e-01 7.79718995e-01 -1.79288971e+00 6.88439980e-02
7.07129061e-01 1.11469722e+00 2.15009794e-01 4.12864238e-02
7.62324333e-02 5.33073962e-01 -7.44555771e-01 -2.00145468e-01
3.53029408e-02 -5.31348407e-01 -2.09167928e-01 1.22444309e-01
3.96421462e-01 2.12754354e-01 -2.15680927e-01 1.10082603e+00
5.64222455e-01 5.42843528e-02 7.76122928e-01 -6.05051696e-01
-3.72249901e-01 3.91666621e-01 -2.98289835e-01 3.02976668e-01
4.14216787e-01 -2.67167270e-01 1.02275670e+00 -1.02300978e+00
4.65097725e-01 7.55715549e-01 5.03245413e-01 1.79135934e-01
-1.43213141e+00 -4.39679503e-01 -6.59953803e-02 2.84085602e-01
-1.47352552e+00 -5.69799125e-01 1.11294341e+00 -5.85765421e-01
5.84528983e-01 2.83376247e-01 5.21266282e-01 8.62998426e-01
2.56204069e-01 5.33269167e-01 1.22149765e+00 -6.06913745e-01
3.74601454e-01 2.70965278e-01 9.83037055e-02 5.52577376e-01
3.35064381e-01 7.59578943e-02 -4.82925773e-01 -1.38799503e-01
4.21817273e-01 -3.10434829e-02 -4.59380329e-01 -6.12134635e-01
-1.02237499e+00 9.61643994e-01 2.20027015e-01 1.16289473e+00
-5.02827883e-01 -2.45591834e-01 3.10685366e-01 2.94291764e-01
3.31772596e-01 4.20071483e-01 3.85018960e-02 4.69869794e-03
-1.02301717e+00 1.65467322e-01 7.62691200e-01 3.05249482e-01
7.24536538e-01 4.36374575e-01 -1.50415853e-01 1.01366651e+00
1.22692278e-02 4.46887612e-01 7.41712213e-01 -1.50706232e-01
2.37948880e-01 6.19151831e-01 -2.41645604e-01 -1.64320838e+00
-3.33928198e-01 -8.04644883e-01 -1.12016082e+00 2.43423149e-01
4.83196199e-01 3.88349593e-01 -5.73259592e-01 1.61009693e+00
8.45629871e-02 6.43273368e-02 1.00665629e-01 7.77731419e-01
4.75012988e-01 6.14011705e-01 -1.39042705e-01 -4.32451487e-01
1.07773399e+00 -3.61250460e-01 -6.74677432e-01 -9.61693823e-02
3.79794180e-01 -9.31902945e-01 7.39325345e-01 8.53997529e-01
-7.46511579e-01 -8.03239524e-01 -1.24288332e+00 4.05429631e-01
-2.43888065e-01 3.26058805e-01 3.87780488e-01 6.65487349e-01
-5.51160216e-01 5.88982582e-01 -5.27065217e-01 -4.59378846e-02
1.41694546e-01 2.74363130e-01 -4.58450049e-01 -1.54243395e-01
-1.00936413e+00 1.03556728e+00 3.29299361e-01 3.53351445e-03
-7.20240235e-01 -3.12882632e-01 -7.25956142e-01 2.39423335e-01
2.18158588e-01 -6.47625104e-02 4.93443429e-01 -1.15277469e+00
-1.30203295e+00 7.17701137e-01 1.65100291e-03 -5.21799684e-01
2.86762923e-01 1.81095883e-01 -6.38531208e-01 2.21164644e-01
-7.86903203e-02 8.49908069e-02 1.25102723e+00 -1.08225465e+00
-2.94498652e-01 -5.00808179e-01 -2.04358280e-01 -1.08485624e-01
-6.41740918e-01 -2.72676706e-01 -5.94437160e-02 -1.09774995e+00
3.42345297e-01 -7.30298162e-01 -6.09198399e-02 -9.08706561e-02
4.01943289e-02 -1.73078269e-01 7.44559944e-01 -4.40937459e-01
1.24632156e+00 -2.31880307e+00 6.47783160e-01 5.84926963e-01
-7.04885647e-02 6.16113722e-01 -8.08913708e-02 4.58290428e-01
-6.63208142e-02 -4.27672356e-01 -4.62344915e-01 -3.16757679e-01
-2.61079818e-01 3.72283816e-01 -2.31992096e-01 6.30998313e-01
1.08395740e-01 9.38058123e-02 -7.12846100e-01 -3.31476450e-01
5.69198608e-01 8.10847819e-01 -4.37533826e-01 7.91561604e-03
2.29542218e-02 5.29341757e-01 -3.00860584e-01 5.16749620e-01
4.59512115e-01 2.22560987e-01 3.39975655e-01 -2.39001110e-01
-8.76416489e-02 -5.31439446e-02 -1.62478995e+00 1.43806779e+00
-6.47706389e-01 4.93436694e-01 -1.94033936e-01 -1.59575975e+00
1.38742745e+00 4.02272999e-01 6.67166173e-01 -7.71042764e-01
1.03423730e-01 6.73211932e-01 1.50292277e-01 -2.08654061e-01
4.00716662e-01 -3.43258917e-01 6.73573390e-02 1.04163038e-02
2.65195727e-01 -7.62414113e-02 3.85724574e-01 -3.48804623e-01
7.73732543e-01 -3.69672209e-01 5.61986566e-01 -2.35425770e-01
1.08970726e+00 -3.58102500e-01 5.13725519e-01 4.82207775e-01
1.16525427e-01 4.48734164e-01 2.22266600e-01 -3.98750037e-01
-7.82258451e-01 -6.64375484e-01 -3.69954616e-01 4.74744201e-01
5.33855222e-02 -1.24370493e-01 -3.50018322e-01 -3.90525043e-01
1.86931431e-01 5.88979840e-01 -3.58124733e-01 -2.93231010e-01
-5.45754433e-01 -6.95868373e-01 5.34223557e-01 1.67070866e-01
3.21194619e-01 -9.84306216e-01 -5.77757418e-01 5.21973431e-01
6.06404571e-03 -9.87493038e-01 -1.63617536e-01 3.62303823e-01
-1.03621972e+00 -1.03030646e+00 -9.04552400e-01 -7.14852452e-01
5.70518851e-01 2.66953409e-01 7.88167536e-01 1.16220057e-01
-3.30086380e-01 2.04370335e-01 -7.33005285e-01 -1.22982629e-01
-5.96598387e-01 -5.85105792e-02 7.89415836e-02 7.16133773e-01
3.06624204e-01 -6.45895541e-01 -2.24097207e-01 3.41487765e-01
-1.14613545e+00 -5.36957145e-01 7.72916198e-01 1.00136685e+00
7.25921750e-01 2.45704994e-01 7.22655714e-01 -6.01918876e-01
5.29124975e-01 -4.28929865e-01 -6.05176449e-01 4.05814089e-02
-7.20146060e-01 6.44330829e-02 1.03332162e+00 -4.12876606e-01
-6.46345556e-01 8.97277668e-02 -4.59277332e-01 -2.41464689e-01
-1.47561744e-01 7.83582687e-01 -1.06934510e-01 -4.24572200e-01
7.65734553e-01 7.14041293e-01 1.29682571e-01 -6.96975052e-01
-1.43476114e-01 5.09714961e-01 3.31735671e-01 -3.83982718e-01
6.66698754e-01 3.19110781e-01 3.73012662e-01 -1.18657053e+00
-3.74290675e-01 -5.64214826e-01 -5.75052619e-01 -2.00275332e-01
4.80547488e-01 -6.37295365e-01 -2.70897895e-01 4.96130109e-01
-8.60511959e-01 3.43984783e-01 -3.90510261e-01 6.47713661e-01
-3.51539344e-01 6.95124269e-01 -7.23321140e-02 -7.32860506e-01
-5.24246693e-02 -1.26513958e+00 6.54359043e-01 1.01659223e-02
-6.11264817e-02 -1.04267013e+00 5.97144477e-02 3.06704827e-02
4.41160381e-01 3.12853843e-01 1.04540336e+00 -6.23406291e-01
-2.23661736e-01 -5.38781703e-01 1.64697021e-01 7.71553278e-01
2.76243299e-01 -3.31782341e-01 -7.54002392e-01 -5.01286745e-01
2.84037858e-01 -2.27818862e-02 8.88084948e-01 2.20195040e-01
9.19885397e-01 -8.40365067e-02 -2.00041294e-01 2.77871013e-01
1.77428567e+00 4.03406680e-01 6.66664541e-01 2.03480005e-01
3.36733729e-01 3.21941048e-01 4.87526149e-01 5.89655042e-01
-3.43770653e-01 1.27860057e+00 2.73342669e-01 -4.92669903e-02
-2.17143118e-01 1.10110298e-01 3.22910011e-01 8.83018374e-01
-1.55567437e-01 -2.08797768e-01 -6.42788589e-01 5.10094166e-01
-1.52925515e+00 -1.09577727e+00 -8.45368952e-02 2.36591601e+00
6.43326342e-01 2.50926495e-01 7.27276579e-02 9.34819877e-01
4.55517083e-01 3.49852741e-01 -2.16000155e-01 -3.98717463e-01
-4.94159758e-01 6.97636425e-01 2.40403369e-01 4.06947136e-01
-1.04831994e+00 3.62734795e-01 5.33241367e+00 9.13634956e-01
-1.25355613e+00 5.32183098e-03 5.79749867e-02 3.75393063e-01
-2.04607829e-01 -1.44411728e-01 -6.07354283e-01 4.84298080e-01
5.91733038e-01 -2.91152648e-03 2.76668519e-01 8.10353458e-01
1.04319984e-02 -2.49298096e-01 -9.49474990e-01 1.14735186e+00
4.18655068e-01 -1.09724450e+00 3.96906197e-01 1.45203426e-01
4.64295000e-01 -5.24223089e-01 1.09538436e-01 -8.97603333e-02
-6.78857744e-01 -9.09407079e-01 8.02169740e-01 6.00434959e-01
4.52811241e-01 -8.66221726e-01 9.93903399e-01 3.96240115e-01
-1.03951573e+00 -1.46219879e-01 -3.08855504e-01 -1.06960364e-01
-6.78048953e-02 1.12821877e+00 -3.74249369e-01 7.40127861e-01
1.83797181e-01 8.31413448e-01 -4.65853453e-01 1.06802845e+00
-1.76206425e-01 4.28834826e-01 -8.59716311e-02 -4.55595851e-02
2.44982213e-01 -3.73985857e-01 8.70439529e-01 1.21655858e+00
5.39246023e-01 -8.74045640e-02 1.15576729e-01 6.53519809e-01
2.35491112e-01 3.84311944e-01 -7.91038752e-01 -4.78641270e-03
2.69287527e-01 1.02496231e+00 -6.58923209e-01 -2.18222082e-01
-3.82743090e-01 7.99355090e-01 4.35383022e-02 -9.86867845e-02
-4.64647263e-01 -4.06461537e-01 6.40118420e-01 3.91380601e-02
5.25556982e-01 -4.49035794e-01 -1.87222421e-01 -1.10523605e+00
1.74615592e-01 -1.13524628e+00 2.19943687e-01 -1.43193290e-01
-1.07510543e+00 6.47589266e-01 -2.04548955e-01 -1.49512339e+00
-2.45110884e-01 -7.00737119e-01 -5.73121347e-02 8.10779989e-01
-1.55603254e+00 -8.26302648e-01 -1.98597446e-01 7.49706030e-01
4.61009026e-01 -5.54026961e-01 1.17096579e+00 9.00108039e-01
-3.52009147e-01 5.19537151e-01 3.34463835e-01 -3.25568356e-02
5.39726734e-01 -9.33668971e-01 -4.22472537e-01 9.30418789e-01
6.24993026e-01 4.26987231e-01 8.21992993e-01 -2.56248951e-01
-1.32854760e+00 -6.17228448e-01 8.88930619e-01 1.29126951e-01
3.75986218e-01 -6.29713759e-02 -9.47971284e-01 1.33780539e-01
-3.01684499e-01 -4.95070107e-02 7.62339056e-01 2.64233276e-02
-4.15138394e-01 -4.42066759e-01 -1.04158914e+00 2.12764703e-02
5.38790166e-01 -5.65359831e-01 -7.18708575e-01 9.48807597e-02
-1.97809562e-01 -1.64683983e-01 -9.21184897e-01 3.54686946e-01
5.73390067e-01 -1.18232942e+00 9.91069138e-01 -1.70309767e-01
2.39392012e-01 -4.67236668e-01 -3.74777079e-01 -1.32816541e+00
-3.03733408e-01 -2.03595739e-02 -1.12325653e-01 1.12447739e+00
1.58528060e-01 -6.55317664e-01 4.80529040e-01 -1.64189920e-01
-1.37120813e-01 -7.39911318e-01 -1.21935284e+00 -9.53866422e-01
-1.39315411e-01 -1.89718798e-01 2.17955053e-01 9.35000360e-01
-1.28535897e-01 2.35053286e-01 -5.34312725e-01 -7.14681298e-02
6.50638342e-01 2.89372712e-01 5.26728094e-01 -1.50752032e+00
-4.63423789e-01 -5.06466269e-01 -9.21471179e-01 -7.65868127e-01
1.25860423e-01 -1.08097732e+00 -2.13890016e-01 -1.04056001e+00
-2.05595359e-01 -6.68793201e-01 -3.88345897e-01 3.27300280e-01
7.58693740e-02 3.91663343e-01 2.83116102e-01 3.90296370e-01
1.32930964e-01 4.16542441e-01 9.39724743e-01 -4.14628625e-01
-1.28235415e-01 1.59286395e-01 -4.81545538e-01 6.00739777e-01
5.97413003e-01 -5.35630345e-01 -1.43245310e-01 -1.97086260e-01
-2.76133921e-02 -3.99780609e-02 4.68494356e-01 -1.46939540e+00
6.07711710e-02 3.01496834e-01 3.76087874e-01 -2.05912709e-01
5.02653480e-01 -1.11094093e+00 4.19695824e-01 7.66543567e-01
-1.53462783e-01 -2.47541353e-01 8.30280334e-02 6.00105703e-01
-8.13428044e-01 -5.76618135e-01 1.18392587e+00 3.79073210e-02
-7.88228333e-01 -1.18728757e-01 -4.98992354e-01 -4.06548023e-01
9.23214197e-01 -3.06839138e-01 2.06120163e-01 -1.28011540e-01
-8.38262320e-01 -5.08283317e-01 2.55442381e-01 2.95868158e-01
6.45087421e-01 -1.30880463e+00 -6.75381362e-01 4.17429715e-01
1.37824640e-01 -5.19798577e-01 8.85045007e-02 8.56625319e-01
-3.15531015e-01 5.13062596e-01 -4.79052514e-01 -5.29282808e-01
-1.39098144e+00 5.70851386e-01 2.30337635e-01 -3.84615898e-01
-5.15560031e-01 5.08384049e-01 -3.36742401e-01 1.54739499e-01
2.36105666e-01 -3.51421773e-01 -4.74572420e-01 5.31056881e-01
4.10905749e-01 3.81969184e-01 5.41949987e-01 -9.43415821e-01
-4.24974352e-01 9.47650313e-01 2.14627758e-01 1.74039990e-01
1.42943215e+00 3.44033659e-01 -1.10111408e-01 5.47692001e-01
1.14844465e+00 3.41809779e-01 -6.56150162e-01 -3.14744502e-01
6.93498105e-02 -6.59040868e-01 2.40154088e-01 -5.92330456e-01
-1.10879934e+00 8.56953025e-01 6.40362442e-01 4.72004145e-01
1.40378475e+00 -4.50813085e-01 5.28207660e-01 2.23137066e-01
7.87041306e-01 -9.82267380e-01 3.97447161e-02 2.79707968e-01
9.33257103e-01 -1.02744079e+00 2.14603290e-01 -3.49594533e-01
-2.13901192e-01 1.40634215e+00 -2.20699273e-02 -3.54167104e-01
6.74882948e-01 -1.72509626e-02 -1.82425976e-01 -2.39209905e-02
-4.45912927e-02 -2.51895607e-01 3.49263430e-01 4.88736600e-01
4.07879710e-01 4.01262790e-02 -8.85804594e-01 1.88896805e-01
1.45049557e-01 3.64000574e-02 3.14477354e-01 8.02564025e-01
-4.39812332e-01 -1.46144235e+00 -5.75037718e-01 2.82160342e-01
-3.55565518e-01 2.28103876e-01 -2.98271686e-01 7.27137029e-01
3.68501216e-01 9.58250880e-01 -2.29364604e-01 -3.77564937e-01
5.64913690e-01 -5.54394871e-02 6.26340091e-01 -5.10603607e-01
-5.67093194e-01 -5.73340841e-02 -1.15219876e-01 -2.92885065e-01
-7.76113212e-01 -8.30351591e-01 -8.64779413e-01 3.88552472e-02
-4.88175690e-01 2.86354542e-01 6.99802637e-01 8.93299758e-01
7.66366068e-03 4.72955823e-01 8.38720322e-01 -8.50995958e-01
-7.47715175e-01 -8.89108717e-01 -8.82123947e-01 4.32022721e-01
2.46995017e-01 -9.81826663e-01 -4.00036752e-01 -6.37154728e-02] | [12.248486518859863, 0.5902098417282104] |
74cd9409-bd38-40e8-971d-6e2b6965dd85 | depth-induced-multi-scale-recurrent-attention | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Piao_Depth-Induced_Multi-Scale_Recurrent_Attention_Network_for_Saliency_Detection_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Piao_Depth-Induced_Multi-Scale_Recurrent_Attention_Network_for_Saliency_Detection_ICCV_2019_paper.pdf | Depth-Induced Multi-Scale Recurrent Attention Network for Saliency Detection | In this work, we propose a novel depth-induced multi-scale recurrent attention network for saliency detection. It achieves dramatic performance especially in complex scenarios. There are three main contributions of our network that are experimentally demonstrated to have significant practical merits. First, we design an effective depth refinement block using residual connections to fully extract and fuse multi-level paired complementary cues from RGB and depth streams. Second, depth cues with abundant spatial information are innovatively combined with multi-scale context features for accurately locating salient objects. Third, we boost our model's performance by a novel recurrent attention module inspired by Internal Generative Mechanism of human brain. This module can generate more accurate saliency results via comprehensively learning the internal semantic relation of the fused feature and progressively optimizing local details with memory-oriented scene understanding. In addition, we create a large scale RGB-D dataset containing more complex scenarios, which can contribute to comprehensively evaluating saliency models. Extensive experiments on six public datasets and ours demonstrate that our method can accurately identify salient objects and achieve consistently superior performance over 16 state-of-the-art RGB and RGB-D approaches.
| [' Huchuan Lu', ' Miao Zhang', ' Jingjing Li', ' Wei Ji', 'Yongri Piao'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [ 2.75517553e-01 -1.06972501e-01 -1.96478754e-01 -4.64856595e-01
-8.15872192e-01 5.21457605e-02 2.72261322e-01 -8.72867629e-02
-1.50621876e-01 3.82259458e-01 6.82535827e-01 1.10700183e-01
5.60446642e-02 -6.88342035e-01 -6.34179592e-01 -5.96114576e-01
1.28475279e-01 -2.93723762e-01 8.21378589e-01 -4.64151233e-01
4.88799512e-01 4.22446996e-01 -1.88519561e+00 2.85165817e-01
9.22680974e-01 1.26225007e+00 7.34828532e-01 4.10550326e-01
7.89317712e-02 9.46058571e-01 -2.71579266e-01 1.94812194e-01
1.54552594e-01 -2.92279989e-01 -7.12461412e-01 1.23514950e-01
3.33804220e-01 -5.39023459e-01 -4.19444829e-01 8.62601280e-01
7.29042888e-01 9.65815410e-02 1.76015809e-01 -9.80306208e-01
-9.83445346e-01 3.34444553e-01 -8.95055473e-01 6.61937416e-01
3.69294792e-01 3.33567083e-01 1.02932096e+00 -1.12458313e+00
3.91629368e-01 1.27948093e+00 2.68350959e-01 3.52511406e-01
-6.89882278e-01 -6.37706757e-01 5.80421150e-01 3.85703295e-01
-1.05934465e+00 -2.70481110e-01 1.36795926e+00 1.92436814e-01
8.26168895e-01 1.81877717e-01 8.36627662e-01 8.02648187e-01
3.91229987e-03 1.52103317e+00 1.03544009e+00 -5.97290993e-02
3.01549397e-02 -2.16702625e-01 -5.48434556e-02 8.78999352e-01
-1.33280065e-02 7.26038544e-03 -9.20785606e-01 2.68519491e-01
1.19925606e+00 4.49762404e-01 -3.26707900e-01 -5.13956606e-01
-1.57632983e+00 7.28403270e-01 1.32916391e+00 3.63560945e-01
-6.39155924e-01 3.51334512e-01 -2.48747580e-02 -3.67850244e-01
6.23860717e-01 4.09895957e-01 -2.67176658e-01 2.59808034e-01
-9.13160741e-01 3.71989273e-02 -1.48861721e-01 1.05819583e+00
9.35669601e-01 9.32718590e-02 -4.75984812e-01 5.79716682e-01
5.31250834e-01 6.72529757e-01 6.52569950e-01 -9.24372733e-01
2.53163368e-01 7.83276200e-01 2.82434504e-02 -1.19782913e+00
-6.43106461e-01 -6.35641396e-01 -7.26498067e-01 -5.35053052e-02
-1.53729290e-01 2.72144616e-01 -1.00524306e+00 1.52934015e+00
2.98500329e-01 5.24980664e-01 -2.56881882e-02 1.42365968e+00
1.35580420e+00 4.12402570e-01 8.90025049e-02 1.75236434e-01
1.28631663e+00 -1.30163765e+00 -5.32305658e-01 -5.71552753e-01
-9.06660408e-02 -5.48543036e-01 1.14744377e+00 -1.67076901e-01
-1.36071301e+00 -7.66181171e-01 -1.12284684e+00 -5.93903124e-01
-2.73909599e-01 1.71611100e-01 1.16678643e+00 -3.00139934e-02
-1.29606986e+00 3.03555638e-01 -7.18655944e-01 -5.92332184e-02
8.10774148e-01 2.47689173e-01 1.59542590e-01 -7.59056211e-02
-1.26392329e+00 6.16735756e-01 1.39685065e-01 2.38853380e-01
-1.04325271e+00 -6.84242249e-01 -1.18260491e+00 -7.56298825e-02
1.89928561e-01 -8.68229568e-01 1.30092180e+00 -8.12879264e-01
-1.22835135e+00 8.83475661e-01 -6.85389161e-01 -2.25794524e-01
4.76468727e-02 -2.19324961e-01 -1.36754721e-01 5.46698511e-01
4.81174797e-01 1.11413121e+00 9.15179372e-01 -1.37539697e+00
-9.18776572e-01 -4.61965829e-01 7.30118752e-02 4.70301360e-01
-2.69058168e-01 -1.17921876e-02 -6.48027837e-01 -7.99647510e-01
7.40855753e-01 -3.29469889e-01 -3.92397344e-01 -1.16139846e-02
-3.30580294e-01 -1.62286997e-01 6.76519215e-01 -4.94495332e-01
9.65511739e-01 -2.01240468e+00 3.41603905e-01 -1.93981037e-01
6.59491420e-01 1.10889196e-01 -1.74608707e-01 -1.64982185e-01
1.05269276e-01 -8.56653303e-02 -1.83846325e-01 -5.74802756e-01
-1.89699009e-01 -1.29920959e-01 -6.83451951e-01 2.35901251e-01
5.88237882e-01 1.64312911e+00 -1.31602561e+00 -5.08659244e-01
3.26208562e-01 5.65015495e-01 -3.60487282e-01 2.13737950e-01
-1.13639727e-01 3.69248867e-01 -9.26225543e-01 1.14008427e+00
5.09239674e-01 -6.33339047e-01 -6.31736875e-01 -3.21194232e-01
-1.99598923e-01 4.50369745e-01 -6.94079578e-01 2.20652771e+00
-3.04445595e-01 6.03722155e-01 -4.01283234e-01 -6.18941367e-01
1.02737880e+00 -2.31459707e-01 2.50522286e-01 -1.10275459e+00
1.64270237e-01 1.57430753e-01 -5.40391147e-01 -2.86902994e-01
7.23514974e-01 1.27995834e-01 7.79190287e-03 3.90726835e-01
1.50836604e-02 -2.56135702e-01 -4.10529613e-01 2.81504691e-01
7.75965273e-01 1.65738031e-01 7.79852346e-02 -1.75023347e-01
5.79139650e-01 -2.42170244e-01 6.65799677e-01 5.64764917e-01
-5.70809007e-01 1.03711832e+00 1.77841350e-01 -3.99463564e-01
-7.35365570e-01 -1.16382539e+00 1.79356173e-01 1.12568605e+00
9.49885726e-01 -4.31686454e-02 -2.93308198e-01 -5.98256230e-01
-1.56982139e-01 2.88857251e-01 -9.77167666e-01 -3.65153015e-01
-4.91781086e-01 -6.57671511e-01 1.01140551e-01 9.15425241e-01
1.10429442e+00 -1.29568481e+00 -9.82327938e-01 1.11580685e-01
-3.75132143e-01 -9.70143855e-01 -4.99128908e-01 4.86674672e-03
-1.03851438e+00 -8.48294556e-01 -9.51958537e-01 -9.56520200e-01
5.89823246e-01 1.22909176e+00 1.20683670e+00 2.23561630e-01
-3.64180297e-01 2.89865881e-01 -3.45041335e-01 -5.04400909e-01
5.41746080e-01 1.64887056e-01 -3.08449954e-01 -1.52993113e-01
4.01009828e-01 -5.74643075e-01 -1.04286444e+00 2.72585958e-01
-9.03655469e-01 4.14814979e-01 9.21289921e-01 6.22084498e-01
7.99658358e-01 -3.79271567e-01 7.89898336e-01 -1.81022331e-01
4.02025878e-01 -4.30234283e-01 -2.02793583e-01 1.83812991e-01
-2.48966604e-01 2.11509075e-02 -3.08108870e-02 -4.29100990e-02
-1.16476715e+00 2.89113019e-02 -6.38948828e-02 -4.95303899e-01
-7.96134919e-02 1.09549038e-01 -2.10885629e-01 -1.26586288e-01
2.23157302e-01 5.10865510e-01 -2.97201455e-01 -3.35803300e-01
3.71191680e-01 3.57403815e-01 6.93349004e-01 -2.78616548e-01
7.22333908e-01 8.56960833e-01 -8.91354457e-02 -4.08966482e-01
-1.43947732e+00 -6.00638807e-01 -6.60486639e-01 -1.52888030e-01
8.56180191e-01 -1.36361587e+00 -6.02678895e-01 7.37674713e-01
-1.05087888e+00 -3.85801524e-01 -1.85532600e-01 2.14749217e-01
-4.44254458e-01 2.17396975e-01 -5.67070842e-01 -5.87596357e-01
-5.00597596e-01 -1.14844573e+00 1.77386606e+00 7.65164256e-01
3.44439775e-01 -8.30392718e-01 -1.52028695e-01 1.60600111e-01
6.00521624e-01 1.13164529e-01 3.19657117e-01 9.21716541e-02
-1.13706124e+00 2.90718704e-01 -8.92156959e-01 -1.06633529e-01
2.42887393e-01 -3.02546769e-01 -1.11554456e+00 -7.06772134e-02
2.49240007e-02 -3.25019628e-01 1.34101498e+00 4.84539926e-01
1.41378534e+00 2.25968972e-01 -4.25137788e-01 8.45355093e-01
1.26334846e+00 -2.18697146e-01 6.01660788e-01 4.89177018e-01
9.65272248e-01 4.38717604e-01 9.51452136e-01 4.09141481e-01
8.91042829e-01 4.48049009e-01 6.96433604e-01 -7.05225408e-01
-3.44820499e-01 -3.89488131e-01 8.79457518e-02 5.81067383e-01
-3.22347693e-02 3.23382437e-01 -7.71136403e-01 8.91777992e-01
-1.94453013e+00 -9.76364136e-01 6.55950531e-02 1.69314694e+00
8.40075612e-01 1.97050124e-01 1.81580871e-01 -9.49716717e-02
6.15153611e-01 5.01308799e-01 -9.08938348e-01 2.86206961e-01
-4.47769910e-01 2.13289127e-01 3.88885021e-01 3.45140189e-01
-1.21638381e+00 1.25370145e+00 6.15267515e+00 5.57331562e-01
-1.25434053e+00 2.90713217e-02 8.21540534e-01 -4.27780300e-01
-7.80010402e-01 -3.06297928e-01 -7.93247223e-01 4.69984204e-01
2.02043191e-01 -1.16573505e-01 7.64565244e-02 8.41669917e-01
9.22783688e-02 -2.78519690e-01 -6.12247646e-01 1.09899020e+00
4.82560068e-01 -1.43175793e+00 -5.90651073e-02 -1.95656225e-01
1.03784072e+00 3.13753545e-01 4.99598056e-01 8.71602669e-02
1.96497723e-01 -9.34336960e-01 8.12502444e-01 7.59097457e-01
3.94643724e-01 -8.71820450e-01 7.21689284e-01 -1.52142849e-02
-1.39696205e+00 -8.68952274e-02 -5.68887711e-01 -1.94818571e-01
1.13052927e-01 7.33869135e-01 -3.97168845e-01 5.22187412e-01
1.05890393e+00 1.51126206e+00 -1.04626930e+00 1.14157641e+00
-5.57646811e-01 5.85845187e-02 -9.55245122e-02 -1.77393943e-01
5.19168735e-01 3.57834190e-01 3.74358326e-01 9.50374305e-01
1.45480007e-01 2.79618025e-01 -1.33263767e-01 1.08915925e+00
1.22081853e-01 -2.37002268e-01 -4.21559006e-01 5.16592205e-01
4.43103939e-01 1.38869536e+00 -8.03345382e-01 -3.60155582e-01
-2.74991691e-01 1.14553237e+00 4.43083495e-01 4.46012318e-01
-9.28272784e-01 -3.03034782e-01 8.46044123e-01 -1.96169257e-01
6.00655258e-01 -1.25171661e-01 -5.24939954e-01 -1.24209952e+00
-8.00226815e-03 -4.05942500e-01 1.62690088e-01 -1.35049260e+00
-1.00742221e+00 6.15391910e-01 -3.42318028e-01 -1.32270586e+00
1.22413978e-01 -4.15827274e-01 -7.03440547e-01 1.03217256e+00
-2.35114884e+00 -1.48095644e+00 -8.29137862e-01 7.44029105e-01
6.01465225e-01 9.07083601e-02 3.32584321e-01 -1.05094044e-02
-5.25334060e-01 2.93877363e-01 -4.07536477e-01 -2.76216958e-02
6.25075340e-01 -1.21631646e+00 6.28115535e-01 1.04524970e+00
3.00050117e-02 5.70308685e-01 3.13297302e-01 -5.48322380e-01
-1.23415208e+00 -1.17696500e+00 6.10617697e-01 -4.53911036e-01
4.78725970e-01 -2.73545444e-01 -7.68709362e-01 5.32676041e-01
-2.81866789e-02 1.96560055e-01 3.28811586e-01 -2.19576329e-01
-2.25921780e-01 -4.13510017e-02 -8.33720088e-01 6.08597636e-01
1.38601875e+00 -7.33955920e-01 -8.93077314e-01 9.16075706e-02
1.34412766e+00 -6.88271224e-01 -5.47689259e-01 6.87269866e-01
3.67229223e-01 -1.35846198e+00 1.36825371e+00 -2.45138139e-01
8.94825637e-01 -5.36608756e-01 -5.06620072e-02 -9.87482667e-01
-4.47172314e-01 -3.14485222e-01 -2.83140957e-01 1.13359094e+00
1.52128413e-01 -4.16983068e-01 7.57090986e-01 3.76627386e-01
-3.91653508e-01 -1.18915749e+00 -5.49400926e-01 -1.34103298e-01
-3.95712405e-01 -4.15070593e-01 8.93359065e-01 5.65686703e-01
-1.73466831e-01 2.05643639e-01 -2.57232487e-01 2.95250893e-01
7.51415908e-01 7.90052772e-01 5.16242504e-01 -9.52746511e-01
3.51045169e-02 -6.51193738e-01 -5.13864100e-01 -1.67534089e+00
5.35615161e-02 -6.58026576e-01 1.93364546e-01 -1.71386683e+00
3.24330747e-01 -4.73977596e-01 -7.93955684e-01 4.69387263e-01
-6.89672828e-01 6.19729042e-01 7.08565190e-02 1.85767978e-01
-1.19126999e+00 9.92287755e-01 1.72253990e+00 -8.75248760e-02
-2.05901340e-01 -2.15354636e-01 -1.17760789e+00 8.14906240e-01
7.11091816e-01 4.43258323e-02 -3.39469284e-01 -6.33671165e-01
9.48358253e-02 -2.10230440e-01 8.48118305e-01 -9.91360962e-01
3.64461124e-01 -1.94485188e-01 8.29958856e-01 -1.07353520e+00
3.17316771e-01 -3.98771793e-01 -7.51615226e-01 2.62035370e-01
-3.38573188e-01 -6.92962483e-02 2.47917429e-01 6.30494177e-01
-4.11297560e-01 3.28631312e-01 5.71280241e-01 -1.83403268e-01
-1.40387857e+00 5.29470146e-01 2.61441648e-01 1.82204228e-02
9.50987518e-01 -1.63060680e-01 -3.41411859e-01 -3.65258634e-01
-3.58605117e-01 4.31877613e-01 6.08692050e-01 7.26786494e-01
1.19962788e+00 -1.48154533e+00 -4.48617786e-01 4.62561280e-01
2.16976061e-01 4.11041528e-01 4.01933640e-01 7.21567273e-01
-1.85193852e-01 5.75294912e-01 -4.09973174e-01 -8.05040777e-01
-8.81902397e-01 5.01791060e-01 1.24017403e-01 7.34789371e-02
-5.36863506e-01 1.33550262e+00 5.69010615e-01 -1.54852405e-01
2.04066381e-01 -6.30452454e-01 -2.18700245e-01 -1.72092482e-01
8.25077832e-01 8.09828639e-02 -2.45925501e-01 -6.73745453e-01
-5.27569115e-01 8.52721095e-01 -2.69617904e-02 8.32286924e-02
1.51420867e+00 -5.58643043e-01 -1.73730955e-01 4.23094511e-01
1.03502607e+00 -2.12200657e-01 -1.76866567e+00 -5.10822475e-01
-4.28556651e-01 -7.30389476e-01 4.02511895e-01 -5.95217168e-01
-1.37354362e+00 1.08717322e+00 5.83745897e-01 -1.35525793e-01
1.61462486e+00 3.08176517e-01 1.05608594e+00 4.40966152e-02
3.76393169e-01 -8.46217453e-01 6.85163975e-01 4.01257098e-01
1.00996423e+00 -1.64951611e+00 1.14382304e-01 -3.61952811e-01
-7.42357671e-01 7.21206546e-01 8.52930069e-01 -3.43927652e-01
4.34584558e-01 -2.09894389e-01 -3.05786300e-02 -4.44847018e-01
-5.23226440e-01 -7.81631768e-01 6.44547760e-01 5.85693419e-01
2.26732969e-01 -2.02103883e-01 2.43642077e-01 7.69297123e-01
-7.95067400e-02 -1.82488814e-01 1.59155175e-01 8.35117877e-01
-8.00618351e-01 -5.34013808e-01 -1.41848996e-01 2.28381053e-01
-6.73293471e-02 -3.97029132e-01 -2.63520271e-01 5.24960518e-01
-3.42289284e-02 7.73514867e-01 5.23840711e-02 -3.76175553e-01
1.76856011e-01 -4.41628188e-01 3.02192271e-01 -5.79673767e-01
-3.25714290e-01 5.43213412e-02 -5.70480168e-01 -9.97580409e-01
-7.93856740e-01 -6.85521126e-01 -1.38228202e+00 3.40049192e-02
-7.99378231e-02 -3.05114895e-01 5.06898761e-01 9.10161197e-01
6.43364072e-01 9.64399993e-01 7.68645167e-01 -1.43353164e+00
-9.34676174e-03 -7.31777191e-01 -4.34404314e-01 2.31041670e-01
6.54256165e-01 -9.26954150e-01 -3.06027949e-01 -2.05638140e-01] | [9.73343563079834, -0.6393444538116455] |
4384fcb4-2039-4353-b2b0-3b0988751e57 | ecg-qa-a-comprehensive-question-answering | 2306.15681 | null | https://arxiv.org/abs/2306.15681v1 | https://arxiv.org/pdf/2306.15681v1.pdf | ECG-QA: A Comprehensive Question Answering Dataset Combined With Electrocardiogram | Question answering (QA) in the field of healthcare has received much attention due to significant advancements in natural language processing. However, existing healthcare QA datasets primarily focus on medical images, clinical notes, or structured electronic health record tables. This leaves the vast potential of combining electrocardiogram (ECG) data with these systems largely untapped. To address this gap, we present ECG-QA, the first QA dataset specifically designed for ECG analysis. The dataset comprises a total of 70 question templates that cover a wide range of clinically relevant ECG topics, each validated by an ECG expert to ensure their clinical utility. As a result, our dataset includes diverse ECG interpretation questions, including those that require a comparative analysis of two different ECGs. In addition, we have conducted numerous experiments to provide valuable insights for future research directions. We believe that ECG-QA will serve as a valuable resource for the development of intelligent QA systems capable of assisting clinicians in ECG interpretations. | ['Edward Choi', 'Joon-Myoung Kwon', 'Gyubok Lee', 'Seongsu Bae', 'JungWoo Oh'] | 2023-06-21 | null | null | null | null | ['question-answering'] | ['natural-language-processing'] | [ 3.58090311e-01 1.84167847e-01 1.75763398e-01 -7.67997921e-01
-1.24256933e+00 -5.78486085e-01 -1.64479062e-01 7.91165650e-01
1.39371097e-01 6.01959407e-01 4.88635808e-01 -7.35880017e-01
-3.61516893e-01 -7.76197970e-01 -6.07417338e-02 -2.38903478e-01
-9.68036279e-02 6.01245522e-01 -1.12640433e-01 7.34018674e-03
1.33688658e-01 1.22382097e-01 -9.53722119e-01 7.18207955e-01
1.07462144e+00 1.04452562e+00 -2.41762444e-01 6.52604520e-01
1.41665399e-01 1.06662285e+00 -7.37875760e-01 -7.49684274e-01
-3.87441856e-03 -1.01800239e+00 -1.00182307e+00 5.81390336e-02
2.26706937e-01 -3.23230058e-01 -1.75376952e-01 8.25267553e-01
1.04190218e+00 -3.51413190e-01 3.50180745e-01 -8.99623513e-01
-6.80721283e-01 6.03144228e-01 5.95225357e-02 3.88386995e-01
6.72191441e-01 2.56938398e-01 9.86721814e-01 -4.93459344e-01
4.80993867e-01 7.14984298e-01 7.24664152e-01 5.78494728e-01
-6.78134680e-01 -2.46478572e-01 -4.77527410e-01 4.75496650e-01
-1.31492734e+00 -1.86583340e-01 7.09612966e-01 -2.07380831e-01
6.28426313e-01 6.97843552e-01 6.39675498e-01 7.20385492e-01
2.96716779e-01 4.84940827e-01 1.02379715e+00 -3.37982565e-01
4.78277445e-01 -9.89847407e-02 3.25469166e-01 5.91626942e-01
3.21136832e-01 -4.13124561e-01 -3.81904632e-01 -6.83499336e-01
6.43793702e-01 -7.30943680e-03 -5.71504414e-01 6.37757219e-03
-1.42371905e+00 4.72714573e-01 1.33212745e-01 1.99482739e-01
-5.15749335e-01 -3.70838195e-01 6.58025682e-01 4.87529963e-01
-1.32355794e-01 6.70350015e-01 -5.84186673e-01 -3.32472682e-01
-7.05876946e-01 4.98589158e-01 1.05640268e+00 8.57121170e-01
2.80941576e-01 -3.68276060e-01 -4.21929151e-01 5.22451282e-01
3.34111333e-01 5.31660855e-01 3.65003288e-01 -1.13917243e+00
5.50825536e-01 9.74675655e-01 6.39932528e-02 -1.02451885e+00
-3.86000216e-01 -5.28592348e-01 -1.00999475e+00 -4.76266623e-01
3.93023431e-01 -1.57579407e-01 -6.00427747e-01 1.26553774e+00
2.29129046e-01 5.22494279e-02 1.63907915e-01 9.40765858e-01
1.09865260e+00 1.60824597e-01 2.63490707e-01 -3.55238020e-01
2.02365685e+00 -3.65476638e-01 -9.54092622e-01 6.63573965e-02
5.98446727e-01 -6.43141448e-01 1.05747163e+00 4.23137009e-01
-1.15938902e+00 -4.17944014e-01 -7.84521759e-01 4.79897521e-02
7.31060514e-03 -3.69223952e-02 2.20196515e-01 7.93580234e-01
-8.66152287e-01 3.16790879e-01 -7.19961345e-01 -2.84640014e-01
5.68002939e-01 -1.20801702e-02 -9.40272957e-02 -2.66782999e-01
-1.53916061e+00 8.23002815e-01 5.13396831e-03 1.83696330e-01
-5.44858277e-01 -7.31570840e-01 -8.89517248e-01 1.41338691e-01
5.07238507e-01 -1.19332337e+00 1.45182633e+00 -2.81310409e-01
-8.69371831e-01 9.28041935e-01 -3.32316667e-01 -6.38137877e-01
4.59838301e-01 -5.50295189e-02 -6.47202671e-01 5.07151425e-01
2.32132643e-01 1.12184316e-01 3.73401582e-01 -8.34242761e-01
-4.01124358e-01 -7.48342574e-01 3.28299105e-02 1.22459993e-01
-1.83533713e-01 1.21047884e-01 -4.25692320e-01 -7.66970515e-01
1.22618683e-01 -6.91130638e-01 -4.48851615e-01 -5.55229187e-02
-3.60526830e-01 -6.63503557e-02 2.57348537e-01 -8.79990578e-01
1.87848639e+00 -1.98475301e+00 -1.87699124e-01 1.55940384e-01
5.70638239e-01 2.27260694e-01 1.64808244e-01 5.13239980e-01
3.57801728e-02 4.29721653e-01 -4.46870983e-01 5.89324608e-02
-3.20381194e-01 2.87035584e-01 -2.60763347e-01 5.46278767e-02
3.57468039e-01 1.29297638e+00 -8.39083910e-01 -8.56490850e-01
1.70784984e-02 2.71560162e-01 -2.79114336e-01 3.10144126e-01
-6.34936914e-02 7.39815116e-01 -8.09151828e-01 8.66699398e-01
2.95196950e-01 -7.59645164e-01 3.50916266e-01 -2.79953539e-01
4.28241223e-01 3.69264364e-01 -8.42915297e-01 1.61308181e+00
1.96268350e-01 1.20917782e-01 -2.99391687e-01 -7.56888807e-01
8.00832391e-01 7.36410081e-01 7.73146570e-01 -5.59159338e-01
1.90665200e-01 1.17898911e-01 1.71774060e-01 -1.03244150e+00
6.33478910e-02 -1.64244294e-01 -6.71511218e-02 5.07232249e-01
-4.28161383e-01 -9.46603343e-02 2.66044736e-01 2.31271639e-01
1.39584529e+00 -4.48479772e-01 5.89226902e-01 4.65270728e-02
7.40236521e-01 1.95856899e-01 7.33512044e-01 8.99732649e-01
-6.56551301e-01 1.07630992e+00 3.37473601e-01 -7.04508364e-01
-7.58395433e-01 -1.07363105e+00 -4.54332769e-01 1.23738311e-01
-8.31086412e-02 -9.28792238e-01 -6.84165657e-01 -6.21915460e-01
-2.64293611e-01 2.73981303e-01 -4.45619196e-01 -5.28407507e-02
-6.01749241e-01 -7.32495666e-01 8.97022665e-01 8.26955080e-01
4.43757951e-01 -1.29054344e+00 -1.02424788e+00 4.59081352e-01
-8.41679096e-01 -9.97816145e-01 -2.75643378e-01 -3.07621747e-01
-1.28622913e+00 -1.52712953e+00 -7.85511732e-01 -7.19881058e-01
3.95789593e-01 -1.13025509e-01 1.65625572e+00 3.01211387e-01
-4.44137245e-01 5.99797606e-01 -6.32434249e-01 -7.39986539e-01
-5.30365109e-01 6.71023056e-02 -3.68666083e-01 -1.61849335e-01
5.85917532e-01 -2.21157610e-01 -1.00258553e+00 2.73910642e-01
-9.69442248e-01 -3.00685138e-01 4.38663036e-01 6.00875437e-01
7.34881163e-01 -1.79065362e-01 1.06226933e+00 -1.23224330e+00
1.21482110e+00 -5.00772595e-01 5.35249291e-03 6.46217465e-01
-8.03605020e-01 -2.49227151e-01 4.94110852e-01 2.88013071e-01
-7.85425007e-01 -2.26290286e-01 -4.92632329e-01 2.87712701e-02
-4.17514980e-01 9.76545572e-01 -8.05384666e-02 4.01151091e-01
8.85451734e-01 1.19319454e-01 9.72372480e-03 -4.50855434e-01
3.26846205e-02 8.40555072e-01 6.66345775e-01 -3.31807792e-01
2.12432787e-01 1.91946134e-01 -5.27729131e-02 -3.57949942e-01
-8.49490345e-01 -5.50810516e-01 -2.71196455e-01 -2.05411449e-01
8.49663556e-01 -7.28417933e-01 -6.81551397e-01 1.34628460e-01
-8.68225098e-01 2.27391526e-01 -5.82546771e-01 2.44482547e-01
-1.98322415e-01 8.18136156e-01 -8.07483196e-01 -8.07373822e-01
-7.11433589e-01 -9.99944746e-01 1.05750024e+00 5.48425429e-02
-7.47385383e-01 -8.28621566e-01 1.15952939e-01 9.51466978e-01
2.44104668e-01 3.49934518e-01 1.23046315e+00 -5.24456024e-01
-4.52597558e-01 -2.62313128e-01 -1.06387939e-02 1.19591586e-01
5.12489259e-01 -3.82726431e-01 -6.02222502e-01 -2.33803112e-02
3.85559171e-01 -4.07360047e-01 4.81747687e-01 2.95213878e-01
1.43691421e+00 1.17803290e-01 5.71591035e-02 3.90034355e-02
1.09088838e+00 4.09816325e-01 8.26390445e-01 -9.85806137e-02
4.29801226e-01 3.97501439e-01 6.47587895e-01 6.06846273e-01
1.00274205e+00 1.49641752e-01 2.68529266e-01 -2.67839164e-01
2.18191653e-01 6.24471828e-02 -3.19654465e-01 1.18354392e+00
-1.21944197e-01 -2.29354024e-01 -1.44417548e+00 6.33896053e-01
-1.83715975e+00 -5.37704945e-01 -3.68612319e-01 1.86275291e+00
1.07040906e+00 -1.80416062e-01 -3.94049846e-02 6.18949831e-01
3.04647863e-01 -2.56833792e-01 -6.82878733e-01 -1.84108779e-01
-1.60780251e-02 3.46200526e-01 -2.82060325e-01 -8.62251967e-02
-8.99898529e-01 1.34549230e-01 6.84289360e+00 1.05940022e-01
-6.61431670e-01 -2.41233632e-01 9.42069650e-01 5.11917651e-01
-3.89840245e-01 -2.26794243e-01 -1.23553708e-01 3.68205160e-01
9.92459893e-01 -1.62034571e-01 -6.85628802e-02 4.41202700e-01
3.60922247e-01 -1.14430310e-02 -1.14754951e+00 1.15453815e+00
1.39833137e-01 -1.37968695e+00 2.30407581e-01 -1.68582514e-01
3.59376520e-01 -3.51048172e-01 -1.69068307e-01 8.27366486e-02
-3.07569534e-01 -1.01122808e+00 7.54821002e-02 7.67063260e-01
7.94561267e-01 -2.73592263e-01 1.22321951e+00 3.01841468e-01
-9.61397409e-01 1.07481945e-02 -1.34857073e-01 -1.60941064e-01
3.24397624e-01 7.00416446e-01 -8.82947803e-01 1.09794295e+00
8.16340268e-01 6.13090634e-01 -7.07541585e-01 1.14793718e+00
-3.95664349e-02 9.30740416e-01 1.54572666e-01 2.94491887e-01
-3.74818891e-01 -8.84941891e-02 1.67404413e-01 9.15243506e-01
1.65994048e-01 6.58749819e-01 2.75802791e-01 5.69768190e-01
-1.88607499e-01 4.76889491e-01 -4.00627822e-01 -2.02645153e-01
3.16944659e-01 1.02737927e+00 -6.71369731e-01 -6.55018866e-01
-5.85214019e-01 5.11808395e-01 -2.16935091e-02 2.33816102e-01
-7.89410412e-01 -3.02891046e-01 2.28756547e-01 1.95155472e-01
-3.02403897e-01 1.92268595e-01 -8.12969089e-01 -1.39955688e+00
1.30025312e-01 -1.43612850e+00 9.63989675e-01 -8.12948525e-01
-1.60451150e+00 7.78161347e-01 -3.31133604e-01 -1.22681546e+00
-4.60639954e-01 -2.21074328e-01 -1.91364542e-01 8.38568747e-01
-1.37956226e+00 -7.14993060e-01 -6.55128121e-01 5.22044957e-01
4.49836522e-01 -5.32533228e-03 1.31425750e+00 6.76470459e-01
-2.86586016e-01 4.55595374e-01 -6.35882139e-01 4.17903632e-01
8.46196175e-01 -1.18513334e+00 3.88854027e-01 4.16666955e-01
7.14197531e-02 8.63069475e-01 3.17030907e-01 -6.66146815e-01
-1.32474971e+00 -1.15954161e+00 1.25656974e+00 -8.30686390e-01
1.14967339e-01 3.15809727e-01 -1.24888647e+00 4.54765409e-01
1.57113343e-01 -7.84933269e-02 1.38642168e+00 -1.33599371e-01
3.19353603e-02 -2.24718422e-01 -1.13685286e+00 4.36273396e-01
7.83498168e-01 -6.92134440e-01 -9.70490456e-01 7.11996481e-02
4.01151478e-01 -4.40948218e-01 -1.46268177e+00 9.27042782e-01
5.34244537e-01 -6.57179296e-01 8.55515420e-01 -8.67231905e-01
5.30089796e-01 -4.63570654e-01 2.08410770e-01 -8.06866407e-01
-1.20653942e-01 -3.07496965e-01 -2.40179390e-01 9.18592215e-01
5.33800244e-01 -5.52865446e-01 7.37355351e-01 9.87207353e-01
8.61103162e-02 -1.07519686e+00 -5.64326525e-01 4.56163511e-02
-2.95986712e-01 -5.17051756e-01 7.35872746e-01 1.08549535e+00
3.18934172e-01 4.20957714e-01 -1.74902186e-01 1.17075719e-01
4.17729080e-01 4.46149021e-01 4.50850517e-01 -1.17429090e+00
-3.48302603e-01 1.71347111e-01 -4.05916452e-01 -8.19941401e-01
-6.01492047e-01 -7.70903409e-01 -1.27092868e-01 -1.92658627e+00
4.02722061e-01 -3.55300426e-01 -5.16214848e-01 4.84424204e-01
-6.53990090e-01 5.57127535e-01 -4.30978909e-02 1.99079737e-01
-7.25520909e-01 9.20333639e-02 1.41589963e+00 -5.13724536e-02
-4.55608219e-02 2.12722540e-01 -1.04689956e+00 6.83808267e-01
1.11558735e+00 -3.73289734e-01 -7.31854975e-01 -4.50004667e-01
4.13835138e-01 7.67071784e-01 1.95605978e-01 -9.02645230e-01
2.75128841e-01 -2.24861074e-02 2.09872067e-01 -5.44173479e-01
-1.39493167e-01 -7.40161061e-01 9.92946103e-02 4.93578285e-01
-4.69117969e-01 4.46799666e-01 -2.20192708e-02 4.78172034e-01
-7.31374383e-01 9.57746059e-02 2.81992644e-01 -3.87151659e-01
-3.07179630e-01 3.08739334e-01 -5.55608511e-01 6.32237554e-01
7.72762120e-01 -8.21034387e-02 -6.29352778e-02 -5.89555204e-01
-1.02906251e+00 7.01220214e-01 7.81305954e-02 2.16075882e-01
9.46558774e-01 -9.35252011e-01 -1.01306117e+00 1.46275789e-01
3.85293931e-01 1.24546014e-01 4.40286696e-01 7.29210675e-01
-8.03034306e-01 4.07817334e-01 -5.39276078e-02 -6.65247321e-01
-1.34273911e+00 3.95287365e-01 1.80643886e-01 -4.12326247e-01
-6.86973870e-01 2.40337253e-01 -1.71868116e-01 -1.90992564e-01
3.53847116e-01 -7.12892711e-01 -2.95525730e-01 -1.81750268e-01
7.52442062e-01 2.20464334e-01 2.82882482e-01 -1.65878773e-01
-5.11214972e-01 2.93449551e-01 9.28568617e-02 7.47623518e-02
9.82487202e-01 -3.43079329e-01 -2.26393729e-01 3.71395677e-01
5.27943790e-01 -7.35469013e-02 -4.35556114e-01 -7.29559585e-02
1.44273013e-01 -2.17565611e-01 -4.23980683e-01 -1.21657288e+00
-7.37519205e-01 9.49424684e-01 5.59763908e-01 3.61763924e-01
1.48309922e+00 -3.87647524e-02 9.45058048e-01 3.62915784e-01
3.70017081e-01 -6.24902964e-01 -2.93311905e-02 1.85769185e-01
7.06560075e-01 -1.29723418e+00 -6.81020692e-02 -5.63528955e-01
-9.82229173e-01 1.05094373e+00 3.40861768e-01 3.90618742e-01
6.69929445e-01 1.30468369e-01 7.18856394e-01 -5.24917483e-01
-6.87039435e-01 -8.83616731e-02 3.22874725e-01 3.16753596e-01
7.21160889e-01 7.55279213e-02 -6.65200412e-01 9.47080493e-01
1.12322280e-02 5.21016061e-01 3.80229622e-01 1.12713110e+00
-5.33597805e-02 -1.49080539e+00 -2.12530345e-01 7.36053884e-01
-9.12485480e-01 -2.76759356e-01 -5.88237822e-01 2.53631055e-01
1.53634071e-01 1.21706498e+00 -5.94843149e-01 -1.82428196e-01
5.73482871e-01 4.17588890e-01 4.70623463e-01 -6.91168845e-01
-9.29788947e-01 -2.45502874e-01 5.44986837e-02 -3.28259587e-01
-4.25086290e-01 -5.95176518e-01 -1.36382067e+00 2.25903258e-01
-3.98905501e-02 3.84574860e-01 1.71843350e-01 7.27012098e-01
6.79237187e-01 7.90213525e-01 1.53290093e-01 4.12322789e-01
-6.55587137e-01 -7.50335395e-01 -3.62121940e-01 6.03742242e-01
1.94438338e-01 -2.10070908e-02 1.63809285e-01 5.69201589e-01] | [8.721908569335938, 8.533791542053223] |
b7f1beb0-25b3-49a1-9475-86ac788df0e5 | a-correct-and-certify-approach-to-self | 2302.06019 | null | https://arxiv.org/abs/2302.06019v2 | https://arxiv.org/pdf/2302.06019v2.pdf | A Correct-and-Certify Approach to Self-Supervise Object Pose Estimators via Ensemble Self-Training | Real-world robotics applications demand object pose estimation methods that work reliably across a variety of scenarios. Modern learning-based approaches require large labeled datasets and tend to perform poorly outside the training domain. Our first contribution is to develop a robust corrector module that corrects pose estimates using depth information, thus enabling existing methods to better generalize to new test domains; the corrector operates on semantic keypoints (but is also applicable to other pose estimators) and is fully differentiable. Our second contribution is an ensemble self-training approach that simultaneously trains multiple pose estimators in a self-supervised manner. Our ensemble self-training architecture uses the robust corrector to refine the output of each pose estimator; then, it evaluates the quality of the outputs using observable correctness certificates; finally, it uses the observably correct outputs for further training, without requiring external supervision. As an additional contribution, we propose small improvements to a regression-based keypoint detection architecture, to enhance its robustness to outliers; these improvements include a robust pooling scheme and a robust centroid computation. Experiments on the YCBV and TLESS datasets show the proposed ensemble self-training outperforms fully supervised baselines while not requiring 3D annotations on real data. | ['Luca Carlone', 'Dominic Maggio', 'Rajat Talak', 'Jingnan Shi'] | 2023-02-12 | null | null | null | null | ['keypoint-detection'] | ['computer-vision'] | [ 1.12993713e-03 1.20263971e-01 -1.31482139e-01 -4.07930762e-01
-1.14090705e+00 -7.12551355e-01 5.25403857e-01 1.77525952e-01
-3.93416524e-01 3.91668886e-01 -2.62815356e-01 2.19751433e-01
-2.67237961e-01 -5.85432410e-01 -1.27635407e+00 -6.88585281e-01
-6.63553476e-02 7.29051590e-01 7.95426726e-01 -1.39705855e-02
5.62430084e-01 7.18013406e-01 -1.54926574e+00 -1.29848003e-01
8.49241972e-01 1.22856116e+00 1.68923825e-01 6.64187551e-01
3.75847042e-01 4.21210468e-01 -7.11831391e-01 -1.16578355e-01
5.28115690e-01 9.47096497e-02 -4.97238785e-01 -1.67797238e-01
9.27463114e-01 -5.02777696e-01 -7.04622120e-02 9.16721046e-01
3.94685864e-01 7.00993389e-02 7.69461334e-01 -1.41634858e+00
-5.40361106e-01 2.75549322e-01 -3.30829322e-01 -2.47588441e-01
2.71023929e-01 2.60840338e-02 9.01963413e-01 -1.09305811e+00
4.88389164e-01 1.11834586e+00 1.09040105e+00 3.32738131e-01
-1.15410233e+00 -7.89933741e-01 2.72120625e-01 1.08099118e-01
-1.40360594e+00 -2.82471538e-01 6.62177384e-01 -2.20695198e-01
9.08981323e-01 -1.25806063e-01 3.94427419e-01 9.25818264e-01
1.56461522e-01 8.95910919e-01 7.84243405e-01 -1.41561896e-01
5.02428889e-01 4.00493219e-02 -1.99355185e-01 7.30954111e-01
4.40153688e-01 5.15046827e-02 -6.16941333e-01 -6.95435256e-02
8.37399662e-01 5.77067137e-02 -3.29630971e-02 -1.13666296e+00
-1.57216501e+00 4.89641309e-01 8.17350268e-01 -9.07359943e-02
-2.10636556e-01 3.77985328e-01 1.23475894e-01 1.35311022e-01
4.37857062e-01 7.43625104e-01 -8.31269383e-01 9.83752385e-02
-8.01738799e-01 4.33698595e-01 7.44938612e-01 1.45000589e+00
9.20111716e-01 -3.31803381e-01 1.02099940e-01 6.25360370e-01
4.71534997e-01 8.18773925e-01 4.44995373e-01 -1.02575886e+00
4.59553212e-01 8.09787571e-01 2.09538653e-01 -9.27534223e-01
-6.34594619e-01 -5.20496428e-01 -3.07907790e-01 4.03850049e-01
3.59790862e-01 6.53014258e-02 -1.03563702e+00 1.47453487e+00
5.54454684e-01 4.19475347e-01 -1.22297660e-01 8.44294012e-01
5.87617099e-01 1.55126289e-01 -2.65972674e-01 2.27469698e-01
7.81735837e-01 -9.31936622e-01 -2.23366424e-01 -3.12747896e-01
6.38118505e-01 -5.85221291e-01 7.58593500e-01 6.86180711e-01
-6.63728356e-01 -6.28027618e-01 -1.45189071e+00 3.14949304e-02
-4.48673368e-01 5.62999487e-01 4.93607670e-01 3.61071229e-01
-9.60939229e-01 7.70187378e-01 -1.04670703e+00 -3.35167319e-01
4.98652756e-01 6.36797369e-01 -5.84914804e-01 2.08950981e-01
-4.68777239e-01 1.26502907e+00 6.72397971e-01 2.11663052e-01
-9.04261887e-01 -7.22743809e-01 -1.04707003e+00 -3.66789311e-01
5.83609819e-01 -4.88405555e-01 1.45274842e+00 -5.58454394e-01
-1.73531270e+00 4.37387764e-01 1.37906358e-01 -3.94903034e-01
7.59408593e-01 -6.13642097e-01 1.24819234e-01 5.90499304e-02
2.43936539e-01 8.05939078e-01 1.00433278e+00 -1.47221828e+00
-8.48920822e-01 -6.46885991e-01 -2.73144513e-01 1.53259039e-01
-1.44664109e-01 -6.08077765e-01 -5.02815127e-01 -6.48600161e-01
8.67871642e-01 -1.17445886e+00 -1.07105725e-01 2.39111230e-01
-3.86276126e-01 -3.82692665e-01 9.76759970e-01 -2.09999740e-01
4.89181131e-01 -2.09883952e+00 3.63455713e-01 4.28613722e-01
1.46468077e-02 -8.88232887e-02 -1.24025792e-01 1.16071656e-01
7.24673122e-02 -3.90818745e-01 -3.12649876e-01 -4.90455568e-01
2.83643492e-02 1.67207897e-01 -3.67897481e-01 8.72309685e-01
5.14858663e-01 7.47291267e-01 -9.24113572e-01 -2.57824391e-01
4.80355680e-01 1.59079269e-01 -7.30984628e-01 1.99941933e-01
-3.17352176e-01 3.99124891e-01 -3.19652617e-01 7.45204926e-01
6.26199067e-01 -2.92393081e-02 -2.75732845e-01 -3.49265784e-01
-4.80737165e-02 8.16710144e-02 -1.77771401e+00 2.00672626e+00
-3.03815335e-01 3.35687727e-01 -1.97942391e-01 -8.69570136e-01
1.19377434e+00 -7.22828284e-02 6.12179637e-01 -9.19674560e-02
1.47705466e-01 5.49027264e-01 -3.87059838e-01 -2.00635299e-01
6.26760602e-01 4.21045870e-01 -2.43470564e-01 3.23862195e-01
5.05153060e-01 -7.08873689e-01 -1.10728905e-01 1.80770203e-01
1.01522899e+00 7.01010764e-01 1.51108310e-01 1.13890907e-02
4.54395771e-01 -2.11224481e-02 5.46064198e-01 6.13196135e-01
-3.04164201e-01 8.66695046e-01 1.29192263e-01 -2.13223696e-01
-9.34501469e-01 -1.17363870e+00 -2.24592969e-01 1.12541020e+00
6.30005777e-01 -3.17839026e-01 -5.92938364e-01 -9.90791738e-01
6.12115502e-01 3.98450375e-01 -4.89425212e-01 -4.28264409e-01
-5.94192684e-01 -3.33839506e-01 5.28015792e-01 1.00172174e+00
4.20262396e-01 -6.41746104e-01 -5.82820833e-01 2.07521021e-01
1.28348395e-01 -1.10687709e+00 -2.10291520e-01 4.81520504e-01
-9.48890269e-01 -1.34379911e+00 -6.28104746e-01 -7.94525504e-01
8.20160806e-01 1.31618902e-01 5.54469526e-01 -2.16009125e-01
-8.95501226e-02 5.57325065e-01 -4.22609627e-01 -7.51931131e-01
-1.40248299e-01 2.27470294e-01 5.29338121e-01 -2.93587238e-01
2.85418719e-01 -4.08121496e-01 -3.86617720e-01 5.72189748e-01
-5.30085087e-01 -3.74020278e-01 7.55855560e-01 6.83854938e-01
7.32444823e-01 -4.36523706e-01 6.37490571e-01 -4.94984746e-01
1.11738175e-01 -2.14969471e-01 -8.37601840e-01 1.45463735e-01
-6.74963236e-01 3.94747555e-01 3.82291019e-01 -4.72157449e-01
-8.87977540e-01 5.98364830e-01 7.22580478e-02 -4.22729105e-01
-2.22335681e-01 -3.01544573e-02 -1.44930094e-01 -4.10249829e-01
8.72936308e-01 -5.28332964e-02 5.76581582e-02 -5.16082287e-01
5.26062012e-01 6.84963882e-01 9.77884829e-01 -6.57071531e-01
1.12900674e+00 4.58376616e-01 3.59143056e-02 -5.03913522e-01
-7.62853384e-01 -6.92980587e-01 -1.10114431e+00 -3.14774454e-01
4.11389470e-01 -1.00645566e+00 -8.78920913e-01 7.21324563e-01
-1.07616913e+00 -2.06139728e-01 -2.24704549e-01 7.54866183e-01
-8.92719209e-01 2.36080304e-01 -3.12679589e-01 -5.76928675e-01
-1.08603492e-01 -1.10179591e+00 1.69630551e+00 1.94588050e-01
-1.66854069e-01 -4.71994817e-01 1.44833773e-01 4.33224216e-02
1.45290215e-02 2.96360552e-01 2.16360018e-01 -9.13101017e-01
-7.76648462e-01 -4.96294141e-01 -1.10900655e-01 5.28480947e-01
4.03668694e-02 1.43827409e-01 -9.02386963e-01 -4.07639474e-01
-3.28218132e-01 -4.24438328e-01 8.42027068e-01 2.41380230e-01
1.03280067e+00 1.71982512e-01 -5.73267877e-01 6.65431976e-01
1.06940532e+00 -2.73102134e-01 1.77444011e-01 6.28414512e-01
6.86387420e-01 4.47993249e-01 8.76228571e-01 2.75620580e-01
5.77610373e-01 6.62304163e-01 7.59171784e-01 1.31727785e-01
6.40332093e-03 -4.09488678e-01 3.38460028e-01 6.58921838e-01
-5.91070438e-03 3.94248098e-01 -8.48628879e-01 4.16666687e-01
-2.11750436e+00 -5.56151807e-01 -1.16046682e-01 2.24334931e+00
7.60700583e-01 2.02379927e-01 9.16707218e-02 2.17708588e-01
5.04812837e-01 -3.50071490e-01 -9.64780688e-01 2.27890581e-01
7.36837462e-02 1.72574446e-01 9.29874301e-01 2.34071836e-01
-1.47326589e+00 1.04232490e+00 6.12297344e+00 2.42148474e-01
-1.00509620e+00 -1.50815740e-01 -1.35461017e-01 -5.73592335e-02
2.26388305e-01 -1.00737952e-01 -8.67650867e-01 1.43133000e-01
3.49174052e-01 1.00541525e-01 1.37942582e-01 1.47567523e+00
-2.54512042e-01 -3.47787082e-01 -1.50820112e+00 9.27284777e-01
4.03087348e-01 -1.05633402e+00 -3.80526274e-01 -2.55365759e-01
7.39386499e-01 2.40141958e-01 -5.92596550e-03 4.12372142e-01
3.91610920e-01 -5.79517186e-01 1.11768878e+00 5.31599462e-01
4.86158341e-01 -7.01019645e-01 8.18399429e-01 4.98321533e-01
-1.09593570e+00 -2.38193199e-01 -3.77821088e-01 1.07821874e-01
-1.59575313e-01 4.41049904e-01 -1.18075633e+00 5.25363743e-01
8.86382520e-01 1.00365269e+00 -1.07287896e+00 1.36658323e+00
-6.58066571e-01 1.08109482e-01 -6.57676339e-01 -1.54552549e-01
-1.65666878e-01 3.65813524e-01 7.51393378e-01 1.02004194e+00
3.21053684e-01 -3.12422693e-01 2.12276712e-01 6.82735384e-01
1.78891588e-02 -1.44884437e-01 -4.19656128e-01 5.82493365e-01
8.04631889e-01 1.25208521e+00 -7.54605472e-01 -1.32922143e-01
4.45366316e-02 9.90040004e-01 4.16387886e-01 1.72269061e-01
-6.40971899e-01 -5.59024274e-01 5.72062671e-01 -3.85065079e-01
4.74327713e-01 -4.75019634e-01 -5.94100595e-01 -1.18913043e+00
3.53121251e-01 -5.76705813e-01 2.30066270e-01 -7.66882598e-01
-1.29600573e+00 1.20131843e-01 1.24422215e-01 -1.49007535e+00
-4.02283370e-01 -1.05294383e+00 -1.74673498e-01 4.99273241e-01
-1.51084065e+00 -1.28367817e+00 -5.74253380e-01 4.24158186e-01
4.31537062e-01 -1.17893919e-01 7.34347701e-01 -3.20088714e-02
-3.47802162e-01 6.78018451e-01 6.19201027e-02 1.37117043e-01
1.19674993e+00 -1.54580081e+00 3.31201375e-01 7.79973328e-01
2.14910090e-01 6.86233878e-01 6.59929514e-01 -6.52720630e-01
-1.57868075e+00 -1.27749586e+00 2.10024267e-01 -9.42045748e-01
4.85639066e-01 -4.17038590e-01 -7.60804892e-01 8.71400952e-01
-5.45191050e-01 3.34935993e-01 1.49231926e-01 9.55323353e-02
-5.42412758e-01 -4.11845714e-01 -1.22732496e+00 2.55099684e-01
1.04023349e+00 -3.32942814e-01 -9.11190271e-01 2.58056164e-01
6.23444557e-01 -9.34465289e-01 -8.25328052e-01 7.19017267e-01
6.55772150e-01 -7.14168847e-01 9.24273074e-01 -1.53703421e-01
3.69083732e-02 -8.09751689e-01 -7.26368204e-02 -1.41335905e+00
-1.22448698e-01 -2.05289230e-01 -3.63702387e-01 1.04605460e+00
3.47612053e-01 -7.14259744e-01 8.48427415e-01 3.39414537e-01
-3.52441251e-01 -5.05762339e-01 -1.00945055e+00 -9.64440405e-01
-7.39984438e-02 -6.80387616e-01 6.76607370e-01 5.46683967e-01
-8.21201354e-02 -5.77875003e-02 6.49550483e-02 6.64022267e-01
7.98660517e-01 -7.47756511e-02 1.41674507e+00 -1.43149006e+00
-1.31239206e-01 -2.69982219e-01 -9.07362819e-01 -1.18014038e+00
2.27599829e-01 -6.86013401e-01 7.19269931e-01 -1.29354346e+00
-1.92937687e-01 -4.76634294e-01 -1.06773965e-01 6.46891475e-01
-1.24944910e-01 4.07865971e-01 1.01429619e-01 3.22647542e-01
-8.74787450e-01 6.33252859e-01 8.71687472e-01 -1.49907008e-01
-2.72263348e-01 1.94022715e-01 -3.56222928e-01 9.58766997e-01
4.76901561e-01 -4.51790810e-01 -1.18200302e-01 -3.32617074e-01
9.98351723e-02 -6.82615638e-01 7.09794223e-01 -1.40444243e+00
6.34259045e-01 1.28399789e-01 7.61740863e-01 -7.93093324e-01
2.01308444e-01 -9.94133234e-01 -2.93280751e-01 2.55551785e-01
-1.17118374e-01 -7.27110058e-02 1.82684645e-01 8.07209134e-01
8.74104872e-02 -1.59468830e-01 6.43278480e-01 1.25219986e-01
-9.08306718e-01 2.08047256e-01 9.28392932e-02 -2.29586348e-01
1.48102295e+00 -3.80110174e-01 -1.57633215e-01 -6.35079220e-02
-3.88069987e-01 3.56588423e-01 7.13883340e-01 6.80070281e-01
6.06078565e-01 -1.29768908e+00 -4.55743223e-01 3.34896445e-01
5.12298465e-01 7.20592916e-01 -3.09843332e-01 7.18889117e-01
-5.42572975e-01 1.71045557e-01 8.98543671e-02 -1.25498760e+00
-8.82204652e-01 2.40802750e-01 3.66371840e-01 4.01868224e-01
-5.32703936e-01 1.02568698e+00 -4.44473803e-01 -1.14656723e+00
7.00106680e-01 -7.75571644e-01 4.01246212e-02 -4.62016650e-02
3.85685802e-01 5.37583232e-01 4.93489534e-01 -6.17677689e-01
-5.99357486e-01 7.30755389e-01 1.01030812e-01 -1.35332244e-02
1.53775048e+00 1.29614368e-01 6.29588440e-02 5.07823169e-01
1.11306536e+00 -2.09770694e-01 -1.58763254e+00 -3.15812796e-01
3.79925251e-01 -3.76905441e-01 -9.01978984e-02 -6.81080163e-01
-6.65039778e-01 4.55195338e-01 5.83414733e-01 -9.81848240e-02
7.87709534e-01 1.88030317e-01 4.81295794e-01 8.05347264e-01
7.12571561e-01 -1.39709735e+00 3.59595269e-01 6.45784378e-01
9.68928158e-01 -1.38465810e+00 2.44551286e-01 -3.11399788e-01
-3.34227443e-01 1.40225863e+00 9.22034502e-01 -4.84681904e-01
4.30485845e-01 4.29305166e-01 5.71242310e-02 -4.50620092e-02
-3.00844252e-01 -8.30707252e-02 5.22557676e-01 7.23572314e-01
-2.96704993e-02 -2.15017006e-01 1.92065209e-01 4.92995590e-01
-3.90381962e-01 -9.97363329e-02 1.45430654e-01 1.01806378e+00
-5.22549987e-01 -8.12670708e-01 -6.01118743e-01 3.55749607e-01
2.10365012e-01 4.82860535e-01 -5.10065675e-01 7.02174783e-01
2.29067028e-01 7.22695053e-01 7.62587264e-02 -5.17715871e-01
8.07760775e-01 -8.75024572e-02 6.30430341e-01 -7.95758188e-01
-4.19176430e-01 -1.60336822e-01 -4.81789649e-01 -8.86320829e-01
-4.85724896e-01 -8.05751622e-01 -1.47166800e+00 1.39602527e-01
-7.97844291e-01 -1.29576385e-01 9.49715078e-01 1.06082797e+00
3.83081406e-01 4.78803366e-01 5.31944633e-01 -1.39596391e+00
-9.95850623e-01 -1.12776208e+00 -2.72498995e-01 1.68535754e-01
5.82378268e-01 -1.03061378e+00 -3.84552628e-01 -4.09398451e-02] | [7.51814079284668, -2.624617338180542] |
bce8cb0c-e8af-4e92-aed4-a2fb0d430a29 | uriel-and-lang2vec-representing-languages-as | null | null | https://aclanthology.org/E17-2002 | https://aclanthology.org/E17-2002.pdf | URIEL and lang2vec: Representing languages as typological, geographical, and phylogenetic vectors | We introduce the URIEL knowledge base for massively multilingual NLP and the lang2vec utility, which provides information-rich vector identifications of languages drawn from typological, geographical, and phylogenetic databases and normalized to have straightforward and consistent formats, naming, and semantics. The goal of URIEL and lang2vec is to enable multilingual NLP, especially on less-resourced languages and make possible types of experiments (especially but not exclusively related to NLP tasks) that are otherwise difficult or impossible due to the sparsity and incommensurability of the data sources. lang2vec vectors have been shown to reduce perplexity in multilingual language modeling, when compared to one-hot language identification vectors. | ['David R. Mortensen', 'Patrick Littell', 'Lori Levin', 'Ke Lin', 'Katherine Kairis', 'Carlisle Turner'] | 2017-04-01 | null | null | null | eacl-2017-4 | ['multilingual-nlp'] | ['natural-language-processing'] | [-5.73256731e-01 -3.69404852e-01 -4.58367795e-01 2.46598776e-02
-4.23720747e-01 -1.16104400e+00 7.86905885e-01 4.33192521e-01
-7.52507567e-01 1.05767429e+00 7.86116302e-01 -6.27557933e-01
1.95033342e-01 -5.16065180e-01 -3.79668564e-01 -1.62741229e-01
-5.24189249e-02 8.00581396e-01 -3.48973840e-01 -3.46005887e-01
-2.21746728e-01 4.62343395e-01 -1.10858023e+00 -1.43196136e-01
7.95306921e-01 3.87539119e-01 3.37753922e-01 4.15029019e-01
-7.04525888e-01 5.42447150e-01 -1.14974402e-01 -5.97604096e-01
6.11921288e-02 1.64677948e-01 -8.23387802e-01 -4.64189857e-01
3.67559761e-01 6.76208884e-02 -4.25545216e-01 1.04744077e+00
2.83330619e-01 -9.94393602e-02 8.52981389e-01 -9.14777756e-01
-9.34991896e-01 8.54485989e-01 -2.98007369e-01 2.62040615e-01
4.69019026e-01 -4.46616337e-02 1.30153024e+00 -8.36259663e-01
1.28445935e+00 1.46236598e+00 5.59697270e-01 3.72919701e-02
-1.15611589e+00 -5.25886178e-01 -1.34610739e-02 1.95218906e-01
-1.59955585e+00 -3.83621603e-01 3.28562647e-01 -6.68392360e-01
1.05994427e+00 1.33921564e-01 3.17631185e-01 1.41091180e+00
3.98721881e-02 6.23723626e-01 8.13042939e-01 -4.36613113e-01
-1.85278893e-01 5.14121354e-01 3.29265505e-01 6.25366628e-01
5.02646565e-01 -1.58871263e-01 -3.76090378e-01 -3.73187691e-01
8.53559494e-01 -3.52499545e-01 -2.46804252e-01 -3.05506557e-01
-1.70241928e+00 1.10257089e+00 1.84159055e-01 8.53864133e-01
-5.55571020e-01 -1.72187179e-01 6.79590583e-01 4.42524523e-01
4.33601052e-01 8.27349126e-01 -7.61917114e-01 -3.20480704e-01
-5.88682592e-01 -1.63589288e-02 9.61902797e-01 9.21702325e-01
6.77546382e-01 4.92208511e-01 3.28999072e-01 1.19888496e+00
2.21539438e-01 9.03539777e-01 7.27131486e-01 -5.07296860e-01
4.80623841e-01 4.32304114e-01 -6.68472275e-02 -1.08727932e+00
-5.48352480e-01 -1.44802675e-01 -7.56677389e-01 -5.83393872e-01
3.81898999e-01 -1.77642763e-01 -4.93960083e-01 1.75240111e+00
1.55886665e-01 -2.92258590e-01 2.68308938e-01 5.80081701e-01
9.11449909e-01 8.36568415e-01 3.18288535e-01 -1.86197311e-01
1.45259225e+00 -3.58595967e-01 -5.48723817e-01 -2.11712152e-01
7.84984231e-01 -9.96285856e-01 1.05264926e+00 -4.29448746e-02
-4.33766276e-01 -3.03459495e-01 -6.46794438e-01 -4.37522501e-01
-7.93057084e-01 -1.18098013e-01 1.09592569e+00 4.99339104e-01
-8.79584134e-01 1.86814144e-01 -6.06257141e-01 -6.66067898e-01
1.60414279e-02 -8.69180039e-02 -1.07991695e+00 -2.27053374e-01
-1.39997303e+00 1.21394658e+00 7.03814089e-01 -2.48702005e-01
-3.74848455e-01 -6.80860221e-01 -1.27164590e+00 -4.54332083e-02
1.20389573e-01 -2.81983167e-01 5.39846301e-01 -6.39446795e-01
-9.05756295e-01 6.87159657e-01 -1.06098875e-03 -3.45258385e-01
1.91222295e-01 7.09162578e-02 -5.74126661e-01 -4.08442736e-01
2.97689974e-01 5.63905299e-01 1.04614310e-01 -8.80554616e-01
-3.25359792e-01 -2.63980806e-01 -1.79697409e-01 4.26886529e-01
-5.57552993e-01 3.36897433e-01 -7.14529872e-01 -7.53375351e-01
-2.19567910e-01 -6.50901973e-01 -3.63106608e-01 -3.36596698e-01
-3.79516602e-01 -6.86177388e-02 3.12231392e-01 -1.44462097e+00
9.58100498e-01 -1.91917062e+00 5.76486409e-01 1.52109489e-01
8.58553126e-02 1.40102446e-01 -5.19067109e-01 6.90565884e-01
1.70071766e-01 4.62602317e-01 2.04411045e-01 5.33024855e-02
1.09999202e-01 5.56455195e-01 -2.38366842e-01 5.67472041e-01
-2.13482901e-01 8.32929969e-01 -9.50600207e-01 -4.80996847e-01
4.03030962e-01 6.98934972e-01 -4.16801661e-01 -2.42708489e-01
-2.14245752e-01 2.14500964e-01 -1.11900486e-01 7.13743210e-01
2.45968863e-01 1.79285631e-01 7.65914261e-01 -2.07689792e-01
-3.47509116e-01 4.13888991e-01 -1.02958512e+00 1.45789778e+00
-8.35650563e-01 9.15400684e-01 -1.20701604e-01 -5.12236893e-01
6.60979688e-01 2.77397692e-01 4.08652753e-01 -5.52407265e-01
2.15478972e-01 2.90925086e-01 -8.33210573e-02 -2.06516042e-01
9.54731464e-01 1.83752123e-02 -5.72984993e-01 3.80463570e-01
6.19152009e-01 3.86980362e-02 4.16909933e-01 1.95669442e-01
4.86275494e-01 -3.68093550e-02 7.69815505e-01 -5.27652502e-01
3.40568006e-01 1.66722149e-01 7.04999328e-01 4.26686227e-01
2.30880138e-02 -2.34902110e-02 5.28694987e-01 -5.31634331e-01
-1.32964826e+00 -1.13029087e+00 -5.16133964e-01 1.06094623e+00
-3.11792284e-01 -5.35354197e-01 -2.90100873e-02 -5.35951614e-01
3.18720669e-01 6.95761979e-01 -3.42566133e-01 5.12595892e-01
-3.23659956e-01 -7.07507432e-01 8.57805669e-01 3.18469077e-01
-3.36612344e-01 -7.34887660e-01 2.37583026e-01 2.59815007e-01
-2.57410169e-01 -1.38254285e+00 -3.59162420e-01 1.01990074e-01
-2.25240648e-01 -1.08802140e+00 -8.93786073e-01 -8.71281803e-01
1.92376554e-01 -2.17058018e-01 1.09075117e+00 -5.54279447e-01
-1.52180344e-01 1.89358041e-01 -3.64788800e-01 -2.51732618e-01
-4.38326389e-01 2.74860173e-01 7.34958589e-01 -2.65793055e-01
4.43010092e-01 -3.23473781e-01 1.98976457e-01 -3.24714091e-03
-7.51881242e-01 -2.76505262e-01 7.40922615e-02 9.29029584e-01
6.90823376e-01 -3.12865883e-01 3.27107728e-01 -6.64279342e-01
6.38670683e-01 -7.07968056e-01 -7.49750912e-01 5.72560132e-01
-2.47011468e-01 2.82874048e-01 7.53262520e-01 -5.57342172e-01
-5.10324478e-01 -4.35598254e-01 -2.28647307e-01 -3.35078597e-01
-2.18626373e-02 8.36053431e-01 -8.23628530e-02 -1.99875310e-01
5.02872229e-01 3.83448154e-01 -3.73715758e-01 -7.33091354e-01
8.83573472e-01 9.15626407e-01 3.65543127e-01 -4.99690950e-01
5.56280434e-01 -3.36861387e-02 -3.47113073e-01 -1.56653309e+00
-1.00837611e-01 -5.20116389e-01 -8.84290814e-01 2.33569756e-01
8.45272958e-01 -1.16815817e+00 -5.20673811e-01 2.65697360e-01
-1.11695063e+00 7.38750622e-02 -1.06364943e-01 7.75528789e-01
-5.24847992e-02 5.94261229e-01 -7.60730505e-01 -5.18200994e-01
-2.38045260e-01 -1.12860346e+00 4.07868803e-01 -1.53475195e-01
-5.38813055e-01 -1.47966170e+00 3.74407947e-01 1.71370417e-01
4.46111560e-01 1.82563767e-01 1.30588245e+00 -1.05950701e+00
-3.20136249e-02 -3.64296325e-02 -2.76172698e-01 2.06524327e-01
2.09661379e-01 7.35341012e-02 -6.49048865e-01 -2.72069007e-01
-8.48615229e-01 -3.55923563e-01 6.10634983e-01 1.08897075e-01
3.81242216e-01 -4.20118153e-01 -1.69166729e-01 7.92085886e-01
1.47422099e+00 -5.56608625e-02 6.28385618e-02 2.38116801e-01
9.78589833e-01 5.04515290e-01 -1.24401145e-01 1.83469534e-01
7.80119717e-01 6.96955621e-01 -1.42449737e-01 1.55202048e-02
2.78085489e-02 -5.84332108e-01 1.79085255e-01 1.48922813e+00
1.80179313e-01 -2.65267044e-01 -1.42865717e+00 8.73436093e-01
-1.38699639e+00 -6.96206629e-01 -2.59475689e-02 2.10660744e+00
8.79108369e-01 -5.21577120e-01 -2.28233896e-02 -4.08115566e-01
2.98340172e-01 3.32692206e-01 -1.94294870e-01 -7.15827525e-01
-6.25257075e-01 1.78869307e-01 8.37470293e-01 6.95716381e-01
-9.33202267e-01 1.49078834e+00 6.68282700e+00 7.52340972e-01
-1.25584817e+00 3.05866599e-01 1.08340092e-01 1.05698325e-01
-7.95326233e-01 -7.85683766e-02 -7.38380134e-01 4.24953669e-01
9.62523997e-01 -5.09310603e-01 7.27348745e-01 6.03560150e-01
-2.43194640e-01 3.01782846e-01 -8.52536201e-01 1.09751725e+00
4.19928171e-02 -1.42889047e+00 2.70262003e-01 1.75731167e-01
5.31201005e-01 7.40662575e-01 -1.60966665e-01 5.45651257e-01
9.19992864e-01 -1.04502201e+00 5.04517257e-01 3.57917883e-02
1.18406892e+00 -8.81919742e-01 6.39248788e-01 1.10127397e-01
-1.21799254e+00 1.94719478e-01 -6.76386476e-01 1.58202648e-01
3.27081501e-01 4.18913305e-01 -9.11233962e-01 7.06568360e-01
5.04967928e-01 7.02610075e-01 -3.27085912e-01 6.63545966e-01
-1.41969144e-01 4.30090457e-01 -4.87445563e-01 -2.22664669e-01
5.08259654e-01 -1.41177341e-01 6.77967489e-01 1.40977871e+00
2.82057792e-01 -6.44649267e-01 3.60580295e-01 4.83834833e-01
-3.23222160e-01 9.48168635e-01 -9.24002945e-01 -7.01588690e-01
7.69433320e-01 1.05631816e+00 -4.13928777e-01 -1.61438063e-01
-4.88937259e-01 8.64700198e-01 6.60037577e-01 4.15980786e-01
-3.40920269e-01 -2.94707716e-01 1.22999918e+00 -1.81074619e-01
5.66959530e-02 -7.67178655e-01 -1.03474885e-01 -1.71554780e+00
-4.00411665e-01 -9.62370336e-01 4.45420474e-01 -4.35238302e-01
-1.42861557e+00 7.79907644e-01 -1.44909173e-01 -6.58683896e-01
-5.71486235e-01 -8.65614831e-01 1.12081565e-01 1.23432636e+00
-1.23731995e+00 -1.35373616e+00 5.55222809e-01 5.35976052e-01
3.61370623e-01 -6.18103683e-01 1.21016836e+00 2.37215534e-01
-5.14615119e-01 6.66605592e-01 6.73655391e-01 4.62965250e-01
6.18229806e-01 -1.06661451e+00 5.41427851e-01 6.11900866e-01
5.94589949e-01 8.38482559e-01 4.35562432e-01 -7.68358886e-01
-1.75756788e+00 -1.02000880e+00 1.65518463e+00 -3.77519011e-01
1.27628779e+00 -6.50005758e-01 -7.63318598e-01 9.27267611e-01
1.71948761e-01 -3.52328718e-01 9.64903831e-01 8.59538615e-01
-8.89628351e-01 3.12876195e-01 -8.65221858e-01 9.50444043e-01
6.82527304e-01 -9.58598316e-01 -7.20771790e-01 5.25195837e-01
5.42042792e-01 -7.40581900e-02 -1.21901417e+00 -2.91591212e-02
7.58520007e-01 1.61509383e-02 1.05418265e+00 -8.69366169e-01
8.52972940e-02 -5.29338382e-02 -7.02553630e-01 -1.55369639e+00
-4.46911007e-01 -7.76503682e-02 5.08070230e-01 1.38742363e+00
6.81570470e-01 -8.02607417e-01 3.52014825e-02 2.53764745e-02
3.03610504e-01 3.84578556e-02 -1.18183994e+00 -8.75047803e-01
3.82889211e-01 -5.76791942e-01 6.01037979e-01 1.61403835e+00
1.36366025e-01 4.89269137e-01 -6.87319458e-01 1.39344618e-01
4.45205331e-01 -1.15658015e-01 6.37939036e-01 -1.39499545e+00
-1.26736268e-01 -4.53293383e-01 -7.29064822e-01 -5.14964163e-01
6.23713255e-01 -1.42278016e+00 -2.96785742e-01 -1.41251826e+00
1.42419171e-02 -5.76872230e-01 -1.66035995e-01 7.06819296e-01
8.65009129e-02 1.74871117e-01 3.21063101e-01 1.76653028e-01
-1.48208126e-01 4.75954175e-01 6.75439417e-01 -2.63137937e-01
-2.79296279e-01 -9.00529742e-01 -7.06903934e-01 5.58629155e-01
3.71070027e-01 -1.93262607e-01 -1.26436993e-01 -7.90425122e-01
4.23995852e-01 1.03741601e-01 -1.76963672e-01 -6.05394959e-01
-1.58104807e-01 -3.16667020e-01 4.10850614e-01 -3.11557293e-01
3.06225121e-01 -5.74776828e-01 2.98349977e-01 2.56528348e-01
-1.61996901e-01 3.05234253e-01 3.25544327e-01 1.96903333e-01
-2.70096600e-01 1.06260851e-01 3.86124432e-01 -2.65074283e-01
-1.05197453e+00 3.73698115e-01 -6.24483943e-01 2.34577596e-01
7.27557898e-01 3.20234627e-01 -2.82148123e-01 -2.70853072e-01
-4.25962418e-01 2.36370921e-01 8.12899113e-01 8.47486377e-01
8.01098645e-02 -1.44326329e+00 -9.60286677e-01 4.03675824e-01
2.72874385e-01 -8.05975378e-01 1.96327761e-01 3.52221280e-01
-7.28542745e-01 8.25698972e-01 -6.06171906e-01 -2.64851570e-01
-1.05161965e+00 6.13188624e-01 7.98373856e-03 -3.96393925e-01
-4.46294606e-01 5.19055486e-01 2.71607824e-02 -1.11146879e+00
1.73600800e-02 3.26913558e-02 -5.62995076e-01 4.95944083e-01
4.05144483e-01 1.68521777e-01 -8.56495798e-02 -1.29085338e+00
-6.39407933e-01 4.20450866e-01 -8.80186111e-02 -2.93031871e-01
1.30009198e+00 -2.66876128e-02 -2.93238878e-01 6.84471011e-01
1.10419619e+00 3.34535897e-01 -2.28515357e-01 -3.94548386e-01
1.70499384e-01 -1.51056066e-01 1.17682680e-01 -5.82638502e-01
-6.84355915e-01 7.53852308e-01 2.47377813e-01 -1.76319182e-01
2.59781033e-01 -5.27478643e-02 7.38503218e-01 3.90039772e-01
6.68130338e-01 -1.05143380e+00 -9.34083939e-01 1.15455377e+00
7.04824328e-01 -9.53234017e-01 -1.69369236e-01 -1.01918131e-01
-6.98960423e-01 8.72993350e-01 1.23243622e-01 4.22509074e-01
6.36869431e-01 2.90645778e-01 2.89283693e-01 8.20303783e-02
-6.32870734e-01 -2.53319412e-01 3.53894889e-01 8.65472674e-01
7.13290095e-01 4.99636412e-01 -5.60478151e-01 5.40843129e-01
-3.52954656e-01 -3.68479162e-01 3.38639557e-01 4.04770017e-01
-3.29678655e-02 -1.38163936e+00 -3.75910550e-01 3.68837386e-01
-4.48793650e-01 -6.29626572e-01 -3.78372163e-01 1.05000079e+00
-3.42569090e-02 6.43040776e-01 2.41399869e-01 -4.07750458e-01
-2.52378476e-03 3.26114506e-01 2.06729650e-01 -3.29641700e-01
-5.67817926e-01 1.93304883e-03 6.19374812e-01 -4.01044160e-01
1.50074229e-01 -6.65814698e-01 -6.72515512e-01 -5.98086655e-01
1.09186366e-01 2.22140893e-01 9.82481241e-01 7.00433493e-01
2.22322851e-01 1.56153992e-01 3.85578036e-01 -6.31824136e-01
-2.53055006e-01 -9.93575811e-01 -7.38340318e-01 3.59073997e-01
-4.63140942e-02 -4.31177586e-01 -1.48867488e-01 -3.59203190e-01] | [10.862445831298828, 9.86462116241455] |
7df98c64-6402-409b-9fea-e4a57c6d5eeb | towards-end-to-end-semi-supervised-learning | 2302.11299 | null | https://arxiv.org/abs/2302.11299v1 | https://arxiv.org/pdf/2302.11299v1.pdf | Towards End-to-end Semi-supervised Learning for One-stage Object Detection | Semi-supervised object detection (SSOD) is a research hot spot in computer vision, which can greatly reduce the requirement for expensive bounding-box annotations. Despite great success, existing progress mainly focuses on two-stage detection networks like FasterRCNN, while the research on one-stage detectors is often ignored. In this paper, we focus on the semi-supervised learning for the advanced and popular one-stage detection network YOLOv5. Compared with Faster-RCNN, the implementation of YOLOv5 is much more complex, and the various training techniques used in YOLOv5 can also reduce the benefit of SSOD. In addition to this challenge, we also reveal two key issues in one-stage SSOD, which are low-quality pseudo-labeling and multi-task optimization conflict, respectively. To address these issues, we propose a novel teacher-student learning recipe called OneTeacher with two innovative designs, namely Multi-view Pseudo-label Refinement (MPR) and Decoupled Semi-supervised Optimization (DSO). In particular, MPR improves the quality of pseudo-labels via augmented-view refinement and global-view filtering, and DSO handles the joint optimization conflicts via structure tweaks and task-specific pseudo-labeling. In addition, we also carefully revise the implementation of YOLOv5 to maximize the benefits of SSOD, which is also shared with the existing SSOD methods for fair comparison. To validate OneTeacher, we conduct extensive experiments on COCO and Pascal VOC. The extensive experiments show that OneTeacher can not only achieve superior performance than the compared methods, e.g., 15.0% relative AP gains over Unbiased Teacher, but also well handle the key issues in one-stage SSOD. Our source code is available at: https://github.com/luogen1996/OneTeacher. | ['Rongrong Ji', 'Xiaoshuai Sun', 'Lei Jin', 'Yiyi Zhou', 'Gen Luo'] | 2023-02-22 | null | null | null | null | ['semi-supervised-object-detection'] | ['computer-vision'] | [-1.67295858e-01 -9.38856900e-02 -2.94752359e-01 -3.59794408e-01
-9.03423429e-01 -2.96133518e-01 2.44354367e-01 -3.77779067e-01
-6.00558519e-01 3.12116295e-01 -1.84090644e-01 -6.98396489e-02
3.31763625e-01 -4.13218915e-01 -5.56433022e-01 -8.88001859e-01
5.21306872e-01 2.53483891e-01 7.13184118e-01 -1.02131777e-02
-1.48696944e-01 3.31046313e-01 -1.47651780e+00 5.38006127e-02
9.74755883e-01 8.84361207e-01 3.81188512e-01 5.35299897e-01
1.59761772e-01 5.59533417e-01 -5.68290532e-01 -4.60470110e-01
2.19272047e-01 -1.59574285e-01 -5.80167174e-01 1.42473891e-01
7.76669621e-01 -4.98076230e-01 -2.83823490e-01 1.44171894e+00
7.84724236e-01 1.01841241e-01 4.69715983e-01 -1.37767506e+00
-4.94564950e-01 5.18896878e-01 -1.07323813e+00 1.32175133e-01
-1.84628755e-01 2.57803947e-01 1.04586077e+00 -1.33730471e+00
4.18444723e-01 1.37535727e+00 7.67184854e-01 6.33295178e-01
-1.14364147e+00 -1.05107331e+00 3.93826634e-01 2.18981072e-01
-1.52028704e+00 -3.67203563e-01 6.58894062e-01 -4.09377605e-01
4.63905334e-01 1.06236525e-01 4.92649645e-01 9.52856004e-01
-1.56102329e-01 1.32454777e+00 1.19685626e+00 -2.70270199e-01
-1.81267738e-01 3.92186016e-01 2.83837438e-01 1.09240270e+00
3.45482320e-01 1.52035475e-01 -2.82304883e-01 1.51988700e-01
6.82176709e-01 1.58201411e-01 -2.82950222e-01 -5.36136091e-01
-1.03157806e+00 7.73838401e-01 5.46908736e-01 -1.19846843e-01
2.65345603e-01 -1.35128181e-02 4.68100548e-01 -1.29686698e-01
5.41442931e-01 2.39092201e-01 -5.01653731e-01 2.01150581e-01
-8.75453115e-01 -1.17843589e-02 5.38906634e-01 1.27677417e+00
8.33540499e-01 1.52772471e-01 -2.52826780e-01 9.93966222e-01
5.88298857e-01 5.55124044e-01 2.62094796e-01 -9.41191196e-01
4.33984637e-01 7.02426553e-01 -1.61072895e-01 -7.49054134e-01
-3.80645961e-01 -7.33569741e-01 -8.07908475e-01 1.69011474e-01
4.74191785e-01 -1.57938316e-01 -1.04409432e+00 1.59802675e+00
7.62917697e-01 1.70167953e-01 -7.32321143e-02 1.06472433e+00
1.28902197e+00 5.17816663e-01 2.95283329e-02 -1.98815912e-01
1.55556273e+00 -1.70570409e+00 -5.43675005e-01 -3.66452187e-01
8.12353313e-01 -9.70948219e-01 1.07724202e+00 3.09597105e-01
-7.26673543e-01 -7.53819883e-01 -1.00837731e+00 -2.67263114e-01
-1.08020231e-01 7.51345813e-01 4.99656081e-01 5.64162433e-01
-6.79532170e-01 2.05182359e-01 -7.78354406e-01 -1.46372169e-01
6.54928982e-01 1.92005560e-01 -2.02814639e-01 -1.91628665e-01
-9.75962520e-01 7.21446753e-01 3.72604698e-01 3.10966194e-01
-1.16061509e+00 -6.87917829e-01 -7.85021901e-01 -6.25090674e-02
1.05502117e+00 -3.17630857e-01 1.48647034e+00 -6.56211019e-01
-1.36629117e+00 1.04578543e+00 -5.93672357e-02 -1.86372958e-02
7.41400063e-01 -3.53298098e-01 -1.90894440e-01 5.32970997e-03
3.86447370e-01 8.05347204e-01 7.86142886e-01 -1.16599011e+00
-9.40991640e-01 -2.97214895e-01 8.65079369e-03 3.89071643e-01
-3.47868204e-01 1.61320969e-01 -9.95754242e-01 -6.77681327e-01
1.03212908e-01 -1.03862059e+00 -2.71669745e-01 2.48332664e-01
-6.28994524e-01 -5.99016428e-01 9.78937984e-01 -3.04764003e-01
1.13081968e+00 -2.34261727e+00 -8.06784704e-02 -3.43483686e-01
5.77912271e-01 6.52557313e-01 -6.47794306e-02 -1.49180144e-01
5.75482100e-03 -4.01616581e-02 -2.40311205e-01 -7.18003035e-01
-9.96441245e-02 1.40352935e-01 -5.85999005e-02 6.32622898e-01
1.40746370e-01 7.78258145e-01 -9.51256871e-01 -8.31935048e-01
1.39716566e-01 3.13090146e-01 -4.07817513e-01 3.17003250e-01
-2.27758840e-01 3.21758121e-01 -6.06971204e-01 9.06682193e-01
8.28233302e-01 -4.39695656e-01 -1.08383797e-01 -4.07190740e-01
-3.71210307e-01 6.86624646e-02 -1.45720446e+00 1.40634942e+00
-2.00686038e-01 5.12090743e-01 2.60693282e-01 -8.40714395e-01
7.64342606e-01 8.22091475e-02 2.55565554e-01 -4.26765233e-01
2.93778330e-01 1.16385005e-01 -2.62942493e-01 -3.30870718e-01
3.00874084e-01 1.95199102e-01 1.50761813e-01 3.32548171e-01
2.16227263e-01 2.69397395e-03 2.65858680e-01 3.41985404e-01
5.54146290e-01 3.15796375e-01 3.61522406e-01 -2.60981202e-01
5.86783826e-01 5.04229292e-02 1.11763299e+00 6.63210034e-01
-6.59547269e-01 7.06475139e-01 4.34929013e-01 -2.37422764e-01
-6.01929128e-01 -8.40421677e-01 -2.81226575e-01 1.35702860e+00
5.24453461e-01 -4.51606274e-01 -8.35427344e-01 -1.09200621e+00
-8.56245160e-02 3.91544670e-01 -3.94649833e-01 -4.96399216e-02
-5.83400369e-01 -9.86821592e-01 6.36869788e-01 6.22982621e-01
8.07656944e-01 -8.73065472e-01 -3.31480145e-01 -5.78834563e-02
-5.71338311e-02 -1.31838095e+00 -8.96089911e-01 1.48939639e-01
-7.32208550e-01 -1.28455794e+00 -8.79409075e-01 -9.30839360e-01
7.41045892e-01 8.48287046e-01 8.90787840e-01 1.45541415e-01
-3.57465535e-01 1.76505104e-01 -3.90748441e-01 -5.13159096e-01
-1.24610044e-01 2.17410326e-01 2.26564094e-01 3.73743773e-02
3.33611548e-01 -1.26011252e-01 -6.51522756e-01 8.25273514e-01
-5.81546783e-01 2.00132325e-01 7.08766282e-01 8.63102138e-01
9.07094657e-01 -2.59221911e-01 3.34352076e-01 -1.20180106e+00
7.41939321e-02 -1.24613211e-01 -9.96690333e-01 3.35199177e-01
-1.04221463e+00 -2.13218465e-01 5.75544536e-01 -5.45921922e-01
-1.22121191e+00 3.17141622e-01 -1.99006170e-01 -7.20903993e-01
-9.71345752e-02 -8.59820768e-02 -3.45022500e-01 -3.47359866e-01
5.18229842e-01 1.57742038e-01 -2.02533379e-01 -6.77081764e-01
4.38042015e-01 6.22599006e-01 3.93550694e-01 -3.78912121e-01
8.53883982e-01 4.71664578e-01 -3.51123869e-01 -6.17769182e-01
-1.45220494e+00 -7.91432083e-01 -5.58252037e-01 -1.99057639e-01
8.90397787e-01 -1.29618680e+00 -6.97396159e-01 7.40969360e-01
-9.53670323e-01 -2.71881014e-01 -3.78948078e-02 5.74233830e-01
-1.91963241e-01 4.63289440e-01 -6.52538657e-01 -5.68374574e-01
-4.16846484e-01 -1.53362536e+00 1.03315759e+00 6.20234132e-01
3.65544975e-01 -7.67896295e-01 -1.42480925e-01 6.69266641e-01
4.42324998e-03 -1.94071606e-01 3.17346066e-01 -6.07698441e-01
-7.26293921e-01 1.41713902e-01 -6.97754622e-01 5.45664608e-01
-2.02508941e-01 1.51568741e-01 -1.21211016e+00 -6.26463711e-01
2.30001239e-03 -6.26656115e-01 1.02877605e+00 3.51842284e-01
1.09989583e+00 1.05869293e-01 -4.28839326e-01 1.04397011e+00
1.15919757e+00 -5.98936789e-02 1.49106354e-01 3.64048898e-01
1.18189633e+00 4.74427313e-01 1.07390189e+00 9.55350995e-02
6.58293545e-01 8.17456305e-01 4.79521036e-01 -4.14099723e-01
-4.44373965e-01 -4.29738879e-01 4.43299979e-01 9.58554864e-01
3.41818966e-02 1.00482300e-01 -7.82233536e-01 2.97371477e-01
-1.83473122e+00 -6.41764998e-01 -4.64638799e-01 1.85249758e+00
7.66898274e-01 1.90051734e-01 2.04095945e-01 -2.24408954e-01
8.37776840e-01 3.75698745e-01 -7.41318762e-01 1.73612699e-01
-8.19258671e-03 -2.73102373e-01 6.29934251e-01 3.76909941e-01
-1.39688635e+00 1.18517196e+00 5.35883999e+00 1.26049387e+00
-9.58895624e-01 4.45208251e-01 6.05896890e-01 -1.84959266e-02
3.08990210e-01 -2.37141103e-02 -1.54789662e+00 3.19123954e-01
1.62240461e-01 1.59122378e-01 9.12557691e-02 1.45727122e+00
5.59023321e-02 -1.15359880e-01 -1.01750267e+00 1.12634385e+00
2.39157453e-01 -1.06521380e+00 -2.60531992e-01 -1.72679245e-01
7.94149160e-01 2.55829334e-01 1.02683073e-02 6.29593790e-01
3.00884277e-01 -5.96661866e-01 6.98225439e-01 -1.03479795e-01
9.32788610e-01 -5.24839282e-01 6.96148515e-01 5.76876879e-01
-1.50496459e+00 2.23096088e-02 -5.38121462e-01 2.51853228e-01
3.73492076e-04 5.80029309e-01 -5.02156973e-01 4.39759344e-01
1.02862263e+00 8.54634166e-01 -7.83545852e-01 1.04905236e+00
-7.52527535e-01 7.30034292e-01 -2.87808001e-01 -2.43966021e-02
2.94437319e-01 -2.66768456e-01 5.76163471e-01 1.31956661e+00
-1.43547237e-01 5.23991622e-02 5.22015452e-01 8.00232649e-01
-1.65392309e-01 2.89742672e-03 -2.47954279e-01 4.55301017e-01
5.25146902e-01 1.69587529e+00 -7.24380076e-01 -4.20696646e-01
-4.96050984e-01 6.94990039e-01 5.20575225e-01 3.48617166e-01
-1.00385022e+00 -2.01110676e-01 4.72435832e-01 -1.17294930e-01
2.95758903e-01 -3.88824269e-02 -1.13899738e-01 -1.44741666e+00
-9.93760973e-02 -1.02862179e+00 5.36708772e-01 -5.23563802e-01
-1.15134072e+00 4.52481896e-01 1.16850566e-02 -1.22475779e+00
4.70975548e-01 -5.58230996e-01 -6.69437408e-01 5.96069038e-01
-1.70935988e+00 -1.14258265e+00 -4.23682660e-01 4.55251575e-01
9.35745478e-01 -1.36982212e-02 3.25375289e-01 5.89812040e-01
-1.33721268e+00 9.29360092e-01 5.84637523e-02 3.57501090e-01
1.02519357e+00 -1.27655447e+00 3.46680433e-01 1.03789949e+00
2.96401978e-01 4.07247335e-01 3.24654698e-01 -3.87464821e-01
-1.06426156e+00 -1.38385344e+00 4.51549351e-01 -4.07015502e-01
5.01343191e-01 -5.18575549e-01 -9.47785676e-01 6.91467762e-01
-1.59705132e-01 3.32277477e-01 3.65481257e-01 4.13379557e-02
-3.74827594e-01 -2.55269825e-01 -8.29422295e-01 5.92731476e-01
1.14203882e+00 -3.65290046e-01 -4.42040503e-01 4.21722680e-01
9.49472487e-01 -6.65564239e-01 -6.48709893e-01 6.01825953e-01
3.69473189e-01 -1.08834171e+00 8.92885149e-01 -1.92175955e-01
1.14643089e-01 -6.96680367e-01 1.08986802e-01 -1.08384013e+00
-2.90782630e-01 -4.94626820e-01 -1.47168890e-01 1.46958292e+00
3.34319830e-01 -6.45088613e-01 1.02262771e+00 2.37260446e-01
-2.68511176e-01 -1.13028443e+00 -7.44388819e-01 -6.99649274e-01
-2.92896837e-01 -2.50064939e-01 1.34863675e-01 9.52656984e-01
-7.11339653e-01 5.95136821e-01 -4.74937648e-01 3.66277456e-01
8.70220065e-01 2.39850283e-01 1.01655054e+00 -9.14840937e-01
-3.18135709e-01 -3.16399187e-01 -4.51872461e-02 -1.50822532e+00
-1.14952095e-01 -8.14057350e-01 2.10586682e-01 -1.25930882e+00
5.21935761e-01 -6.60709977e-01 -1.89482808e-01 5.17455995e-01
-5.28198600e-01 3.61613244e-01 3.03879410e-01 4.75411296e-01
-9.86687779e-01 7.09746540e-01 1.64169133e+00 -4.33601364e-02
-1.55055135e-01 2.10755169e-01 -5.96223950e-01 1.16280425e+00
6.35288477e-01 -8.69537115e-01 -2.90571213e-01 -4.76886243e-01
-2.23029211e-01 -2.82777488e-01 3.45337719e-01 -7.13397861e-01
3.99513692e-01 -9.99111757e-02 2.45528832e-01 -7.95127988e-01
1.42669797e-01 -5.20906866e-01 -4.04331475e-01 5.51616251e-01
-2.23777462e-02 -6.14926890e-02 -9.61306095e-02 6.28064394e-01
-2.22630411e-01 -3.14505965e-01 1.15172935e+00 -5.05744889e-02
-9.23906446e-01 5.76464295e-01 2.18175761e-02 3.18166852e-01
1.13953233e+00 -1.52467772e-01 -4.29936260e-01 3.99250761e-02
-5.53783059e-01 8.52206886e-01 3.10540885e-01 3.09684366e-01
4.89918828e-01 -1.11164832e+00 -5.93669295e-01 6.28585592e-02
1.97009295e-01 6.49746895e-01 2.11112693e-01 1.11000144e+00
-4.76390094e-01 1.55665368e-01 2.59843588e-01 -8.37507963e-01
-1.47419429e+00 4.07828242e-01 3.75616342e-01 -3.66904706e-01
-5.76591194e-01 1.25503469e+00 5.54611444e-01 -6.69139087e-01
6.16351724e-01 -6.84908926e-02 -3.38022292e-01 2.67108649e-01
4.18182105e-01 5.33974826e-01 -2.17517540e-01 -4.82914746e-01
-3.81592393e-01 7.73280740e-01 -5.17881572e-01 3.78051847e-01
1.01894093e+00 4.98623261e-03 2.69634444e-02 2.33396679e-01
1.08502018e+00 1.43407460e-03 -1.61098552e+00 -4.59683895e-01
-1.41642138e-01 -3.96657825e-01 1.24363355e-01 -4.26157773e-01
-1.28806615e+00 1.02831352e+00 6.25069320e-01 -1.36777475e-01
9.99070585e-01 4.59094830e-02 7.98075199e-01 4.03153598e-01
2.44633302e-01 -1.17239952e+00 3.54549229e-01 5.14066339e-01
5.05339622e-01 -1.55232084e+00 3.65614116e-01 -7.99005985e-01
-7.47052491e-01 8.05687547e-01 1.32232392e+00 -7.83268213e-02
4.74189490e-01 1.76741496e-01 3.66061598e-01 -1.26899421e-01
-5.51139057e-01 -3.80925685e-01 3.03231329e-01 3.53350163e-01
2.03079566e-01 -6.11322522e-02 -1.65392116e-01 5.26292920e-01
1.55303881e-01 -3.60775650e-01 2.86383510e-01 5.97121000e-01
-5.34599781e-01 -9.26761806e-01 -4.55493867e-01 3.65999699e-01
-4.12126273e-01 -3.69696170e-02 -2.32948229e-01 1.02728450e+00
4.31880146e-01 7.07199037e-01 -3.52855235e-01 -3.65576148e-01
4.53886509e-01 -3.58027637e-01 9.57438573e-02 -9.17325616e-01
-5.26186943e-01 3.83140057e-01 -1.04081459e-01 -6.03232682e-01
-4.68561769e-01 -5.06816864e-01 -1.17629576e+00 -1.00888155e-01
-1.06284165e+00 1.27813732e-02 3.92752230e-01 8.43433619e-01
1.04300633e-01 5.97202122e-01 6.52417243e-01 -7.79777467e-01
-8.30933332e-01 -8.38223279e-01 -4.39949393e-01 7.00879022e-02
1.98897362e-01 -9.76693213e-01 -4.54790235e-01 -1.67932078e-01] | [9.21894645690918, 1.337202548980713] |
f55d155f-f735-405e-a9e9-0eaca93c3867 | gptips-2-an-open-source-software-platform-for | 1412.4690 | null | http://arxiv.org/abs/1412.4690v2 | http://arxiv.org/pdf/1412.4690v2.pdf | GPTIPS 2: an open-source software platform for symbolic data mining | GPTIPS is a free, open source MATLAB based software platform for symbolic
data mining (SDM). It uses a multigene variant of the biologically inspired
machine learning method of genetic programming (MGGP) as the engine that drives
the automatic model discovery process. Symbolic data mining is the process of
extracting hidden, meaningful relationships from data in the form of symbolic
equations. In contrast to other data-mining methods, the structural
transparency of the generated predictive equations can give new insights into
the physical systems or processes that generated the data. Furthermore, this
transparency makes the models very easy to deploy outside of MATLAB. The
rationale behind GPTIPS is to reduce the technical barriers to using,
understanding, visualising and deploying GP based symbolic models of data,
whilst at the same time remaining highly customisable and delivering robust
numerical performance for power users. In this chapter, notable new features of
the latest version of the software are discussed with these aims in mind.
Additionally, a simplified variant of the MGGP high level gene crossover
mechanism is proposed. It is demonstrated that the new functionality of GPTIPS
2 (a) facilitates the discovery of compact symbolic relationships from data
using multiple approaches, e.g. using novel gene-centric visualisation analysis
to mitigate horizontal bloat and reduce complexity in multigene symbolic
regression models (b) provides numerous methods for visualising the properties
of symbolic models (c) emphasises the generation of graphically navigable
libraries of models that are optimal in terms of the Pareto trade off surface
of model performance and complexity and (d) expedites real world applications
by the simple, rapid and robust deployment of symbolic models outside the
software environment they were developed in. | ['Dominic P. Searson'] | 2014-12-15 | null | null | null | null | ['model-discovery'] | ['miscellaneous'] | [ 1.59243226e-01 1.05616651e-01 1.57369122e-01 1.01002455e-01
9.79170576e-03 -4.61218596e-01 4.13572103e-01 5.43217719e-01
1.92100495e-01 8.06533933e-01 -4.92524385e-01 -8.62997651e-01
-8.29580963e-01 -6.93017960e-01 -4.07106847e-01 -8.74452949e-01
-6.99244797e-01 4.26249653e-01 5.37607931e-02 -4.48747844e-01
4.59464103e-01 8.57561767e-01 -1.93658352e+00 2.17862730e-03
9.61389363e-01 6.75586522e-01 1.42881021e-01 7.76003420e-01
-1.36627769e-02 5.47191560e-01 -6.01423383e-01 5.44634871e-02
7.99687132e-02 -6.81840956e-01 -4.10184771e-01 -5.44891357e-01
-5.32053888e-01 4.49982226e-01 2.97166824e-01 5.42955160e-01
4.15847003e-01 -1.03261024e-01 4.88705248e-01 -1.43287885e+00
-2.06617519e-01 8.74067917e-02 -3.21324199e-01 1.32664531e-01
4.76701051e-01 3.91972750e-01 3.11400592e-01 -3.29829335e-01
6.72728837e-01 1.36808741e+00 7.05213666e-01 3.70543338e-02
-1.65931404e+00 -5.93500853e-01 -2.62313753e-01 -2.65602916e-01
-1.68028772e+00 -2.28830084e-01 3.79572809e-01 -5.42130053e-01
1.49111187e+00 8.45567822e-01 1.03807938e+00 5.09519994e-01
6.84469461e-01 -9.07603279e-02 1.10929537e+00 -7.47248888e-01
4.85709637e-01 1.41275868e-01 -1.26911044e-01 8.36102843e-01
4.79554027e-01 4.82955694e-01 -6.66235566e-01 -6.39872909e-01
7.76690304e-01 -3.14875036e-01 -2.32035778e-02 -5.01049161e-01
-7.75791764e-01 6.18919492e-01 1.21615469e-01 2.24847183e-01
-1.85060188e-01 3.84252220e-01 3.91614616e-01 1.84084892e-01
1.47640303e-01 6.96727455e-01 -4.72675234e-01 -4.71493363e-01
-7.65629530e-01 4.04587567e-01 7.38614559e-01 6.32312417e-01
5.07133961e-01 2.63380229e-01 5.01314759e-01 3.75523746e-01
6.58719122e-01 7.85096735e-02 4.72348541e-01 -5.78023851e-01
-1.41837478e-01 1.05866659e+00 -2.11528063e-01 -1.21241391e+00
-5.07220328e-01 -3.78886044e-01 -1.64355829e-01 7.07561433e-01
2.56713539e-01 -6.00881949e-02 -7.16800213e-01 1.30253220e+00
5.31516075e-01 -8.42304751e-02 1.68964222e-01 4.13744330e-01
5.55828333e-01 7.48534739e-01 3.73683095e-01 -1.10832021e-01
1.22516143e+00 -1.49176717e-01 -2.48621687e-01 2.67332226e-01
1.01560986e+00 -3.56341749e-01 6.95126295e-01 5.96216023e-01
-8.13149273e-01 -2.69753158e-01 -1.32684088e+00 5.73111847e-02
-9.45156574e-01 -3.49841863e-01 1.02726734e+00 9.79470015e-01
-1.01824057e+00 8.98195505e-01 -1.10359859e+00 -3.65066022e-01
3.59175563e-01 7.43511558e-01 -1.49367929e-01 3.00391793e-01
-9.38703835e-01 1.01678872e+00 6.77694440e-01 4.83481139e-02
-3.55003685e-01 -9.27209795e-01 -7.18664587e-01 1.20145418e-01
1.60182789e-01 -5.20935893e-01 5.20341754e-01 -8.02510142e-01
-1.29038548e+00 5.82720101e-01 -6.28699139e-02 -3.16085875e-01
2.63951182e-01 3.92128646e-01 -4.50481921e-01 1.16273984e-02
-2.08985731e-01 5.67057133e-01 2.32189417e-01 -1.15091693e+00
-3.71047080e-01 -1.99031070e-01 -4.09599274e-01 -2.07513526e-01
2.76829034e-01 2.45683476e-01 -7.98224881e-02 -5.63167512e-01
6.32094517e-02 -6.99488163e-01 -2.73472458e-01 1.27123045e-02
-2.03219205e-01 1.26051709e-01 9.33548927e-01 -7.61924505e-01
1.17926550e+00 -1.92809176e+00 2.07246140e-01 7.97666192e-01
1.06343076e-01 2.60108739e-01 5.04228830e-01 9.77581441e-01
-4.59145486e-01 4.36814666e-01 -1.75554037e-01 4.87131000e-01
-7.78216496e-02 2.38454446e-01 -8.06356966e-02 2.08179638e-01
3.62987131e-01 8.39837611e-01 -8.90012026e-01 -3.78738374e-01
2.60467649e-01 5.63781083e-01 -9.83344615e-02 -3.90102565e-01
-2.99094707e-01 3.06861460e-01 -2.90028840e-01 8.57817233e-01
3.23408008e-01 4.17171046e-03 4.80144829e-01 4.54347819e-01
-4.92052436e-01 1.04347050e-01 -1.05901647e+00 1.31023479e+00
-6.98055699e-02 5.94045460e-01 -7.81694651e-02 -7.85001755e-01
1.39107609e+00 2.98107624e-01 2.02118084e-01 -5.31067252e-01
1.63158730e-01 2.58506119e-01 2.36278117e-01 -3.70162040e-01
5.16710244e-02 -1.37930438e-01 1.76682368e-01 5.57189584e-01
-1.12439014e-01 -7.37505034e-02 3.01116735e-01 -6.58563077e-02
8.80933702e-01 6.78649127e-01 5.78728080e-01 -6.05817795e-01
2.74853975e-01 4.06469047e-01 2.90698946e-01 2.55445451e-01
3.22531760e-01 -4.64032888e-02 9.67739582e-01 -3.51680636e-01
-7.84711242e-01 -7.47840762e-01 -3.25972438e-01 5.18878162e-01
-4.89179194e-02 -7.00933993e-01 -4.04365212e-01 9.28767323e-02
2.46890917e-01 8.33856046e-01 -7.18676746e-01 -2.71425873e-01
-1.51530847e-01 -9.70022321e-01 6.79513514e-01 9.57119167e-02
-1.14609249e-01 -8.26555312e-01 -1.32313013e+00 1.89695165e-01
5.11663318e-01 -2.21780255e-01 7.42851317e-01 5.74109674e-01
-1.25556159e+00 -1.03403628e+00 -9.54903886e-02 -4.59678531e-01
6.24287844e-01 -1.02544867e-01 7.34880269e-01 4.40291852e-01
-1.05051208e+00 -9.03954655e-02 -2.30905771e-01 -9.47956920e-01
-7.08300471e-01 -4.00001913e-01 -3.41651201e-01 -6.44704282e-01
2.64964730e-01 -8.47541988e-01 -4.83720824e-02 1.94593370e-01
-7.21754551e-01 2.99631119e-01 2.40301907e-01 6.12989664e-01
6.46272719e-01 4.65670258e-01 5.98648310e-01 -3.95492643e-01
6.89384282e-01 -6.90279424e-01 -8.58875155e-01 3.95933777e-01
-8.58453751e-01 9.84866768e-02 3.67169917e-01 -8.66388306e-02
-5.22248745e-01 -3.83724878e-03 3.81755352e-01 -1.55727327e-01
-8.09421167e-02 8.19586396e-01 2.75938399e-02 -4.01051283e-01
6.50869846e-01 1.02944598e-01 6.14756048e-01 -4.25118923e-01
-1.02340065e-01 3.42805088e-01 1.60281315e-01 -5.43978393e-01
4.15723711e-01 1.96301609e-01 6.07600868e-01 -8.52148771e-01
7.12405980e-01 -1.59287043e-02 -6.33575082e-01 -3.43227535e-01
3.46818209e-01 -3.29796135e-01 -9.51308191e-01 1.45943269e-01
-5.77658474e-01 -2.45716855e-01 -7.49745220e-02 8.81066546e-02
-5.57107151e-01 1.24600105e-01 1.43025085e-01 -1.29854572e+00
-1.36111155e-01 -1.02047145e+00 6.12735331e-01 2.49659568e-01
-8.11705828e-01 -1.07914674e+00 3.87560502e-02 -1.43275097e-01
2.05082044e-01 1.32669461e+00 1.60792756e+00 -4.56954002e-01
-3.66509587e-01 -4.01960731e-01 2.64289409e-01 -3.90087038e-01
1.99850425e-01 7.79849827e-01 -7.73920894e-01 8.47934559e-02
-3.16345543e-01 2.03872398e-02 9.90986601e-02 1.39253572e-01
7.68936217e-01 -1.04480378e-01 -7.91731000e-01 5.98978937e-01
1.48770404e+00 8.52578700e-01 7.86983311e-01 6.43388450e-01
2.21026868e-01 9.91566122e-01 7.36008346e-01 1.58330798e-01
-1.03790693e-01 8.07777822e-01 4.35138196e-01 -3.10377300e-01
3.89071107e-01 -1.40424535e-01 1.36064455e-01 -1.33512830e-02
-1.44630373e-01 7.41699189e-02 -1.29333365e+00 2.01306194e-01
-1.87239754e+00 -7.29205906e-01 -3.67539734e-01 2.31682110e+00
6.62769556e-01 7.22183064e-02 3.98525894e-01 4.79537457e-01
5.72725356e-01 -6.96616769e-01 -4.36476082e-01 -1.26334345e+00
4.01126780e-02 5.36108553e-01 3.69317800e-01 2.66050071e-01
-4.57558930e-01 5.43646455e-01 6.23957968e+00 6.39245272e-01
-1.28358424e+00 -4.74297643e-01 5.74664950e-01 -4.71328884e-01
6.43143356e-02 2.17719465e-01 -5.13413131e-01 4.63226587e-01
1.42828012e+00 -6.13671541e-01 4.64080632e-01 6.73249900e-01
8.11785519e-01 -6.23396814e-01 -9.31824565e-01 6.20919883e-01
-5.61254859e-01 -1.56245756e+00 -1.07931241e-01 4.02730703e-01
1.24133229e-01 -3.88040274e-01 -2.55427092e-01 -2.59762734e-01
1.82837382e-01 -1.53757632e+00 8.76514673e-01 5.59179366e-01
7.34911978e-01 -1.22841227e+00 3.06484967e-01 2.09715083e-01
-7.12821960e-01 -1.40193120e-01 -7.44450837e-02 -3.00525695e-01
7.21253380e-02 2.95484900e-01 -1.03938687e+00 9.68244970e-01
8.23005021e-01 2.98263639e-01 -8.32344055e-01 1.12414885e+00
1.40887886e-01 4.75269496e-01 -5.55134773e-01 -1.48541734e-01
1.01301990e-01 -2.20395729e-01 4.91233110e-01 1.22325194e+00
3.41484964e-01 -1.13798477e-01 -5.33353329e-01 1.26188242e+00
1.02356386e+00 5.82054164e-03 -6.77870393e-01 -5.27002871e-01
4.35837656e-01 9.82339442e-01 -1.33116305e+00 1.55826941e-01
1.20108418e-01 1.59998342e-01 -1.92849502e-01 1.10982969e-01
-6.05472207e-01 -6.44256711e-01 6.25379324e-01 2.89987326e-01
3.02075386e-01 -2.94397473e-01 -6.30788743e-01 -1.52080864e-01
-2.90750265e-01 -1.03336525e+00 2.74488539e-01 -9.96030748e-01
-3.89631331e-01 2.26544484e-01 3.14210474e-01 -7.98653066e-01
-4.38985586e-01 -7.50220120e-01 -6.73732400e-01 1.35600269e+00
-7.25085676e-01 -9.01612401e-01 -9.60007310e-03 1.42338807e-02
-1.44942507e-01 -5.26424199e-02 1.23412454e+00 -4.16401535e-01
-6.87451839e-01 2.21948728e-01 3.76705021e-01 -6.58743322e-01
-3.71560431e-03 -9.94992375e-01 5.02676666e-01 5.99158466e-01
-1.83903322e-01 1.07659328e+00 9.33621883e-01 -1.03356051e+00
-1.46852720e+00 -5.81732750e-01 5.97731948e-01 -1.76197171e-01
5.89705765e-01 -5.02524436e-01 -9.26188946e-01 1.83684051e-01
-2.89489210e-01 -6.22534096e-01 9.57477033e-01 -1.14497468e-01
1.83223575e-01 1.67333499e-01 -1.40444350e+00 6.52130425e-01
5.33837140e-01 1.51486605e-01 -2.41239324e-01 1.03653871e-01
2.66807616e-01 -3.00623208e-01 -1.01974797e+00 -3.09483707e-02
7.57091463e-01 -1.11463630e+00 9.65917230e-01 -6.12135887e-01
2.62615055e-01 -6.16575956e-01 3.27587813e-01 -8.64172697e-01
-1.08988710e-01 -1.11681056e+00 -5.01235612e-02 1.15841484e+00
5.37213922e-01 -1.02776086e+00 4.50381845e-01 7.70377398e-01
-5.67597337e-02 -1.18984866e+00 -1.06846213e+00 -9.27596033e-01
-1.18856810e-01 -2.35109329e-01 8.81242633e-01 8.12160790e-01
5.03608584e-01 -4.05121118e-01 2.62908608e-01 5.57210296e-02
2.53823012e-01 -4.06777114e-02 6.07960343e-01 -1.47519016e+00
-4.72524494e-01 -4.66615945e-01 -1.01318288e+00 2.63310283e-01
-5.16718686e-01 -8.81673396e-01 -4.63441312e-01 -1.20206189e+00
-3.39249760e-01 -6.65394425e-01 5.11054806e-02 6.47352099e-01
2.95633733e-01 -3.62882428e-02 4.80934652e-03 2.72130460e-01
3.19976717e-01 -1.07210912e-01 7.05013812e-01 4.86180186e-01
-5.09331286e-01 -3.52326483e-01 -9.92463708e-01 3.35483342e-01
8.91955554e-01 -7.51742780e-01 -6.91633701e-01 3.03366929e-01
4.42359447e-01 -1.63201034e-01 6.87525690e-01 -8.02872539e-01
-1.07053638e-01 -3.12386155e-01 6.01082981e-01 -4.46959496e-01
1.49360910e-01 -5.65323114e-01 1.06244862e+00 1.02962470e+00
1.13353051e-01 2.22731113e-01 9.55002785e-01 5.34222245e-01
2.08278149e-01 -2.37879723e-01 5.75647056e-01 -1.91996545e-01
-4.97671932e-01 -6.26415610e-01 -6.55814230e-01 -6.40469909e-01
1.53829789e+00 -1.03657937e+00 -2.82768607e-01 -7.02746361e-02
-6.13634050e-01 2.15269521e-01 8.46477270e-01 2.56526679e-01
2.38220990e-01 -7.19147205e-01 -1.62881598e-01 3.16533804e-01
3.91180515e-02 -2.96477508e-02 -7.41969347e-02 8.60172451e-01
-1.12893128e+00 6.93512678e-01 -5.69130480e-01 -5.03022969e-01
-1.65751839e+00 6.04505241e-01 4.57690865e-01 1.43396124e-01
-6.82373047e-01 6.46693408e-01 -1.84692398e-01 -9.47214663e-02
-1.62639633e-01 -2.33287141e-01 -8.68577957e-02 -8.00176114e-02
4.40328360e-01 7.80937731e-01 2.50709862e-01 -4.00115430e-01
-5.28093874e-01 3.03284317e-01 2.55810678e-01 -4.62342799e-03
1.62428606e+00 9.67254192e-02 -4.12217587e-01 5.56456804e-01
6.11224890e-01 -2.17309877e-01 -1.06511056e+00 7.43170977e-01
2.74396062e-01 -2.66912073e-01 -4.52025374e-03 -1.26939809e+00
-3.78283262e-01 5.47553241e-01 5.21118462e-01 3.49554509e-01
1.06714094e+00 -2.58658737e-01 -5.15813529e-02 8.33256990e-02
1.38452336e-01 -8.42068255e-01 -6.68116748e-01 -1.76095679e-01
9.37329650e-01 -3.19049835e-01 4.59845155e-01 -6.51790917e-01
-2.62310803e-01 1.55572784e+00 1.70259371e-01 2.10215226e-01
2.91675121e-01 5.83878577e-01 -1.05021432e-01 -7.47807622e-01
-8.22077274e-01 2.84964055e-01 -3.75717804e-02 1.08434951e+00
3.51757705e-01 -2.47406274e-01 -3.68938774e-01 4.62364137e-01
-3.08502972e-01 3.10491413e-01 3.05543900e-01 1.46004987e+00
-2.99187899e-01 -1.20551503e+00 -8.04041088e-01 2.68755555e-01
-1.40664026e-01 2.45910250e-02 -8.82641733e-01 1.25455010e+00
1.37912661e-01 8.70054722e-01 -1.22957066e-01 -1.01044923e-01
-1.11804001e-01 3.11882913e-01 3.38122576e-01 -4.15388614e-01
-8.14233601e-01 -8.82515535e-02 2.67647684e-01 -3.62848341e-01
7.47178495e-02 -8.00121427e-01 -1.43846059e+00 -5.36877513e-01
-3.23980421e-01 2.64724702e-01 1.17643583e+00 7.32493699e-01
8.91405523e-01 6.48429453e-01 7.65682478e-03 -8.27057362e-01
1.64379850e-01 -4.22761828e-01 -5.36603153e-01 -5.18105447e-01
-1.86113924e-01 -9.32741642e-01 -1.36264086e-01 1.19933210e-01] | [6.969601631164551, 4.218203067779541] |
564a3f14-b21d-42a1-834a-bd9043d40c8f | earthformer-exploring-space-time-transformers | 2207.05833 | null | https://arxiv.org/abs/2207.05833v2 | https://arxiv.org/pdf/2207.05833v2.pdf | Earthformer: Exploring Space-Time Transformers for Earth System Forecasting | Conventionally, Earth system (e.g., weather and climate) forecasting relies on numerical simulation with complex physical models and are hence both expensive in computation and demanding on domain expertise. With the explosive growth of the spatiotemporal Earth observation data in the past decade, data-driven models that apply Deep Learning (DL) are demonstrating impressive potential for various Earth system forecasting tasks. The Transformer as an emerging DL architecture, despite its broad success in other domains, has limited adoption in this area. In this paper, we propose Earthformer, a space-time Transformer for Earth system forecasting. Earthformer is based on a generic, flexible and efficient space-time attention block, named Cuboid Attention. The idea is to decompose the data into cuboids and apply cuboid-level self-attention in parallel. These cuboids are further connected with a collection of global vectors. We conduct experiments on the MovingMNIST dataset and a newly proposed chaotic N-body MNIST dataset to verify the effectiveness of cuboid attention and figure out the best design of Earthformer. Experiments on two real-world benchmarks about precipitation nowcasting and El Nino/Southern Oscillation (ENSO) forecasting show Earthformer achieves state-of-the-art performance. Code is available: https://github.com/amazon-science/earth-forecasting-transformer . | ['Dit-yan Yeung', 'Mu Li', 'Yuyang Wang', 'Yi Zhu', 'Hao Wang', 'Xingjian Shi', 'Zhihan Gao'] | 2022-07-12 | null | null | null | null | ['weather-forecasting', 'earth-surface-forecasting'] | ['miscellaneous', 'time-series'] | [-6.58082306e-01 -5.10052741e-01 1.36333510e-01 -3.59402955e-01
-2.01666474e-01 -4.53547210e-01 8.90160084e-01 -1.25927553e-01
-9.00094807e-02 4.96057540e-01 1.79117054e-01 -6.55315399e-01
-9.51733589e-02 -8.18773508e-01 -5.61266720e-01 -9.73895669e-01
-2.94241488e-01 4.49107826e-01 -8.07139874e-02 -5.71338892e-01
1.68935895e-01 5.99625349e-01 -1.45338678e+00 -2.39141390e-01
1.11127305e+00 1.14834893e+00 4.12073016e-01 7.75520325e-01
-4.28275056e-02 7.16573238e-01 -3.22535723e-01 1.66611671e-01
2.78878570e-01 -1.69907421e-01 -4.07653689e-01 -3.94195825e-01
1.17216289e-01 -1.67421296e-01 -4.99104530e-01 8.17290843e-01
7.66461134e-01 4.87265378e-01 4.89371479e-01 -1.23425150e+00
-8.12661052e-01 2.25346699e-01 -5.47667027e-01 8.19429278e-01
-1.64636627e-01 2.57441700e-01 7.95615613e-01 -1.13060641e+00
1.95615128e-01 1.22928238e+00 7.66454458e-01 1.61309659e-01
-8.08986664e-01 -8.55850935e-01 3.03660661e-01 3.04090232e-01
-1.38406479e+00 -3.27045709e-01 5.43118954e-01 -6.67077422e-01
1.32506585e+00 3.05857986e-01 6.82219505e-01 6.37428105e-01
7.45861173e-01 4.76446986e-01 9.79602695e-01 -3.02107483e-02
2.67277479e-01 -1.67872801e-01 9.57282782e-02 4.17035192e-01
6.20121695e-02 3.12643886e-01 -2.47281730e-01 -1.17306441e-01
5.72883248e-01 2.67451406e-01 -4.66339260e-01 2.35568117e-02
-1.11542070e+00 1.02632642e+00 9.17087078e-01 2.60220140e-01
-6.70536399e-01 3.30356807e-01 3.23741823e-01 1.29324168e-01
1.00596797e+00 3.45599711e-01 -5.23757517e-01 6.34757653e-02
-1.05680728e+00 4.84202027e-01 6.34302139e-01 4.91410583e-01
3.93892944e-01 6.35764718e-01 2.37796366e-01 5.94111741e-01
3.66143286e-01 1.05784047e+00 7.70153046e-01 -4.24908280e-01
2.12995842e-01 3.84077102e-01 2.17753962e-01 -1.31425726e+00
-7.38332510e-01 -7.02922106e-01 -1.55220318e+00 1.64035693e-01
-1.53435946e-01 -4.31152672e-01 -9.19995785e-01 1.48271561e+00
6.39125764e-01 6.51348948e-01 1.02044784e-01 1.29379404e+00
8.09678376e-01 1.41464639e+00 -4.80317250e-02 1.32048696e-01
1.32060385e+00 -9.12954271e-01 -7.02393353e-01 -5.31474911e-02
4.67961222e-01 -4.76464272e-01 8.08828890e-01 1.31322876e-01
-7.03208506e-01 -7.42770135e-01 -9.08762753e-01 5.36959395e-02
-6.81953132e-01 1.14746287e-01 7.08141625e-01 2.18452126e-01
-1.04662800e+00 5.24359703e-01 -1.15992928e+00 -3.27748150e-01
-1.22811168e-01 2.22593278e-01 -8.56359452e-02 3.85597110e-01
-1.64958358e+00 8.71712446e-01 1.45166487e-01 4.41794068e-01
-1.04953766e+00 -9.33834612e-01 -8.09167266e-01 4.05876875e-01
-1.43799871e-01 -6.39820635e-01 1.06641889e+00 -7.02249110e-01
-1.50236356e+00 1.84984028e-01 2.48955563e-02 -7.91990221e-01
1.22814432e-01 -3.90372217e-01 -6.80966616e-01 -2.34115705e-01
5.99663593e-02 4.68679488e-01 7.66557038e-01 -7.95098305e-01
-6.05940759e-01 -3.01336408e-01 -2.14119345e-01 3.09151411e-01
-1.76234499e-01 -9.06952322e-02 -8.82895142e-02 -8.83828044e-01
8.67277756e-02 -1.20036983e+00 -3.77481788e-01 -2.54962891e-01
-1.39283314e-01 -2.95652211e-01 9.79010046e-01 -8.85851324e-01
1.27237666e+00 -1.86796248e+00 1.81806728e-01 3.27722076e-03
2.50031590e-01 3.02474320e-01 4.20712456e-02 5.70091724e-01
-3.27632338e-01 3.13719660e-02 -3.46919209e-01 -1.73664153e-01
-1.38874259e-03 1.07085668e-01 -9.24180329e-01 7.74408102e-01
5.05798904e-04 7.19142079e-01 -7.47811556e-01 1.34176482e-02
4.90919024e-01 7.86289930e-01 -5.31688035e-01 1.91200376e-01
-2.28173330e-01 5.70886850e-01 -4.48740512e-01 4.93509233e-01
8.93518806e-01 -4.69790280e-01 -2.11804941e-01 1.17717952e-01
-5.37159860e-01 9.12128612e-02 -1.00893581e+00 1.20061374e+00
-6.29425168e-01 8.25458288e-01 -4.43801545e-02 -8.92615080e-01
1.01750147e+00 3.79338443e-01 3.90925556e-01 -8.57278049e-01
7.37999976e-02 1.35858759e-01 9.88018662e-02 -5.50016761e-01
5.96287072e-01 -1.56780779e-01 -6.86954260e-02 2.89439470e-01
-1.19597971e-01 -1.56203330e-01 -1.46812528e-01 1.99016348e-01
5.48689485e-01 5.85416891e-02 4.33606505e-01 -8.15191090e-01
4.60321307e-01 -1.06922174e-02 5.93278766e-01 3.46421361e-01
9.22036022e-02 3.96656096e-01 7.07159415e-02 -1.09352183e+00
-9.28142428e-01 -5.48317671e-01 -2.63435185e-01 1.14452100e+00
-1.50859198e-02 -2.49909610e-01 -4.29744750e-01 -1.45361245e-01
2.15291083e-01 7.17936397e-01 -8.39119256e-01 1.28122941e-02
-3.19671065e-01 -1.24006486e+00 5.13341546e-01 4.79351401e-01
5.65305948e-01 -1.07589340e+00 -1.01539958e+00 3.94551992e-01
-2.70239890e-01 -7.64526606e-01 -2.90475339e-01 1.07232749e-01
-7.14265585e-01 -5.60489655e-01 -9.76848543e-01 -2.09396079e-01
2.96459109e-01 3.28768343e-01 1.21662116e+00 -3.36003676e-02
-4.42824438e-02 8.62424225e-02 -2.74915069e-01 -5.54669917e-01
2.73816288e-01 2.02826142e-01 5.24900019e-01 1.96657166e-01
-5.63190207e-02 -6.39058590e-01 -9.06920969e-01 2.26597175e-01
-7.30154097e-01 1.57213677e-02 2.41291687e-01 8.54574740e-01
2.85888791e-01 -3.31034921e-02 5.87315679e-01 -3.20907384e-01
4.87857580e-01 -1.14278376e+00 -1.09615290e+00 3.72327790e-02
-4.22837406e-01 5.58250062e-02 8.30950141e-01 -2.00191185e-01
-1.01711428e+00 -2.59043843e-01 -1.72618181e-01 -7.04231083e-01
8.95021930e-02 1.06539738e+00 2.63948083e-01 -1.45887351e-02
6.12196445e-01 4.15610135e-01 -3.95282626e-01 -6.03607595e-01
1.34706885e-01 6.76478148e-01 5.36650479e-01 -3.72539729e-01
5.71280658e-01 5.92408240e-01 -1.38411105e-01 -1.12768352e+00
-6.58810139e-01 -3.77988786e-01 -1.41708344e-01 -2.56764948e-01
9.10889208e-01 -1.21825862e+00 -9.18615818e-01 6.17875397e-01
-1.16692626e+00 -5.86102903e-01 1.54464826e-01 6.63473606e-01
1.11028980e-02 -5.86341741e-03 -3.46920222e-01 -8.78211260e-01
-7.57206500e-01 -9.66780484e-01 1.05118847e+00 3.46218616e-01
1.43958211e-01 -1.16448152e+00 6.41373754e-01 -2.19621196e-01
8.96459877e-01 2.33645365e-01 4.17781502e-01 -3.11781764e-01
-4.17112350e-01 6.75712377e-02 -2.85391599e-01 2.27961410e-02
-4.39493209e-02 9.25525501e-02 -9.18624341e-01 -5.14845669e-01
1.64568692e-01 -6.19264655e-02 1.02916551e+00 6.42832398e-01
1.15018034e+00 -4.20593977e-01 -3.95770341e-01 1.12966585e+00
1.47451472e+00 3.52066398e-01 3.86680007e-01 2.55761147e-01
7.98602641e-01 3.16389978e-01 3.26318949e-01 7.58977354e-01
7.09369838e-01 5.86221099e-01 4.82197434e-01 -2.70954281e-01
2.63891846e-01 2.11082771e-01 2.90945411e-01 1.21418011e+00
-2.03964397e-01 -6.21783495e-01 -1.60124600e+00 7.55579710e-01
-1.79798388e+00 -1.03494823e+00 -1.27833903e-01 1.84924889e+00
2.73614019e-01 -3.50723922e-01 -2.95748383e-01 -2.05790922e-01
4.14433748e-01 4.95402992e-01 -8.92237782e-01 -3.74236107e-01
-1.22463755e-01 4.78483029e-02 6.28822684e-01 5.53164065e-01
-1.28693330e+00 9.14538503e-01 4.98162651e+00 5.72011471e-01
-1.93336010e+00 1.55074462e-01 7.35625982e-01 1.92728732e-02
-1.94540441e-01 -1.98100299e-01 -8.85579109e-01 8.02129626e-01
1.26509738e+00 -3.87507528e-01 3.59350890e-01 8.51122260e-01
3.90944541e-01 1.46404132e-01 -4.59191412e-01 9.42368805e-01
-1.89963475e-01 -1.50079501e+00 -3.09834089e-02 -1.37444630e-01
8.23107004e-01 8.84045601e-01 1.23679057e-01 4.44754601e-01
4.76663530e-01 -1.04732871e+00 7.26153076e-01 6.71605825e-01
7.15898275e-01 -7.08517075e-01 8.15646410e-01 3.75824749e-01
-1.55636895e+00 -7.13658035e-02 -5.41028202e-01 -2.80407697e-01
1.36846781e-01 7.29981244e-01 -2.51726598e-01 7.51155078e-01
1.29418051e+00 9.37741995e-01 -1.79391310e-01 9.76955116e-01
1.69318706e-01 8.64071727e-01 -7.02360809e-01 1.04722954e-01
7.46095359e-01 -3.87994319e-01 6.55168295e-01 1.21861017e+00
7.68539429e-01 6.21159315e-01 1.67062208e-01 5.76262295e-01
9.25452635e-02 8.02745018e-03 -5.76965511e-01 1.90675333e-01
3.37490678e-01 1.30105925e+00 -7.25027144e-01 -5.38663268e-01
-3.37216973e-01 6.23121738e-01 1.15941055e-01 4.38269556e-01
-1.12919629e+00 -3.43866348e-01 9.55754161e-01 -7.62923956e-02
5.26863158e-01 -3.45101506e-01 -2.02650547e-01 -1.32098460e+00
-2.98144430e-01 -8.39932799e-01 2.12563336e-01 -9.56316113e-01
-1.18653107e+00 1.07446194e+00 -8.56156573e-02 -1.36442268e+00
-8.33483711e-02 -2.89530784e-01 -1.01768994e+00 1.04569030e+00
-1.64006829e+00 -8.22052836e-01 -6.36139154e-01 5.01198769e-01
5.42424738e-01 -1.31979987e-01 7.50411034e-01 4.12327468e-01
-7.14159608e-01 1.04530156e-01 7.30875731e-01 -1.53019458e-01
4.48086381e-01 -1.16556036e+00 1.06504083e+00 6.79397881e-01
-7.51751512e-02 5.21208525e-01 8.91627491e-01 -6.33386075e-01
-1.44242895e+00 -1.37108397e+00 8.36870313e-01 -8.34946632e-02
8.29913437e-01 -4.84913856e-01 -1.15230453e+00 7.02307940e-01
4.30503011e-01 4.77013707e-01 2.56126255e-01 -1.33869067e-01
-1.41390592e-01 -2.04654455e-01 -7.79098153e-01 3.91659766e-01
4.30770606e-01 -2.83093095e-01 -4.36490446e-01 5.44413507e-01
5.76872528e-01 -5.32877266e-01 -8.26418936e-01 4.66720074e-01
5.13493419e-01 -7.59860873e-01 8.31027806e-01 -5.13407052e-01
2.78634816e-01 -4.71353859e-01 -2.28494331e-01 -1.73118472e+00
-5.89096487e-01 -5.36126554e-01 -1.42075956e-01 8.09202313e-01
1.65975660e-01 -1.07651854e+00 4.66255665e-01 1.56628296e-01
-2.50920892e-01 -8.70881617e-01 -9.25950229e-01 -8.61560822e-01
3.50128949e-01 -1.46550715e-01 8.28646719e-01 1.35058618e+00
-2.46991158e-01 3.54705662e-01 -6.73094332e-01 5.43818235e-01
4.47104275e-01 5.14153242e-01 6.14446282e-01 -1.39749682e+00
-4.57363613e-02 -3.64871085e-01 -2.13368401e-01 -1.06496596e+00
6.97506592e-02 -7.50297070e-01 2.86246389e-02 -1.42034590e+00
-1.93972126e-01 -5.92369735e-01 -4.21172917e-01 4.05509740e-01
-1.59460828e-01 1.28265321e-01 2.33050123e-01 3.39193761e-01
-3.18091005e-01 1.20622933e+00 7.76640952e-01 -1.72576234e-01
-1.54355764e-01 -2.65638441e-01 -1.22511394e-01 6.74405694e-01
1.22088850e+00 -5.11886418e-01 -2.45437071e-01 -6.63303554e-01
2.84822047e-01 2.26802111e-01 3.29094678e-01 -1.29583538e+00
4.35708493e-01 -3.11894380e-02 3.96909058e-01 -8.38080764e-01
2.26039559e-01 -6.75824702e-01 3.34041804e-01 6.21874511e-01
1.73939958e-01 6.38818860e-01 5.44887364e-01 3.07373613e-01
-4.23079550e-01 5.17260253e-01 8.03766191e-01 1.93753876e-02
-8.56560588e-01 6.94797695e-01 -5.30395269e-01 -3.53044979e-02
8.27171445e-01 2.56533384e-01 -2.03333303e-01 -4.50519562e-01
-5.03962696e-01 5.26635766e-01 1.09483205e-01 5.85862219e-01
3.63250881e-01 -1.20833254e+00 -1.05943894e+00 3.14583093e-01
-1.73406795e-01 -7.48538077e-02 6.81800544e-01 8.90505731e-01
-7.16473639e-01 8.47989619e-01 -2.49025613e-01 -6.78709686e-01
-7.66056895e-01 5.14169157e-01 7.26877570e-01 -3.55323166e-01
-8.55765581e-01 7.74909079e-01 6.63555741e-01 -7.50109434e-01
-1.88969433e-01 -7.13934124e-01 -1.33408189e-01 7.54106790e-02
5.46096921e-01 4.20408696e-01 1.89745963e-01 -6.50375545e-01
-5.84871769e-01 6.53082490e-01 2.47078776e-01 -4.10904922e-02
1.56253338e+00 -1.24622233e-01 -3.02148629e-02 5.29733241e-01
9.62569833e-01 -3.86299402e-01 -1.30185246e+00 -9.93814841e-02
-1.59070864e-01 -1.53459370e-01 5.23124754e-01 -5.81351459e-01
-1.45838177e+00 1.18792796e+00 7.41240799e-01 3.38680446e-01
1.03562164e+00 -4.20754194e-01 9.29561317e-01 3.98414493e-01
4.32955958e-02 -6.79563701e-01 -4.34101731e-01 1.00129867e+00
1.30810928e+00 -1.35137808e+00 8.50244686e-02 4.21541899e-01
-5.48977733e-01 7.70848989e-01 4.93600011e-01 -4.18480396e-01
1.20343685e+00 3.83243859e-01 1.29132092e-01 -3.61478686e-01
-1.02009571e+00 6.56287596e-02 3.72119993e-01 -2.87740007e-02
4.49643999e-01 2.66360551e-01 7.07042068e-02 6.00492716e-01
-2.67437220e-01 -8.47028047e-02 2.03407377e-01 6.29693449e-01
-4.31215554e-01 -3.10435623e-01 -5.88513017e-01 4.05194014e-01
-5.04868567e-01 -3.97761613e-01 1.96833149e-01 6.46718025e-01
-7.00815171e-02 5.55444896e-01 3.80102336e-01 -1.19748056e-01
5.96639328e-02 -2.53685981e-01 -2.11029232e-01 -1.10521719e-01
-7.07419276e-01 -1.22218639e-01 -3.64364445e-01 -6.24647737e-01
-1.04430653e-01 -6.84147835e-01 -1.20795262e+00 -6.43902719e-01
-1.36986390e-01 3.97878498e-01 8.27495515e-01 7.76449025e-01
7.11929560e-01 6.85668230e-01 6.41353965e-01 -1.47705173e+00
-3.50075424e-01 -1.14727664e+00 -4.50676918e-01 -1.31286323e-01
7.21943974e-01 -8.36280465e-01 -3.84040117e-01 -1.04925424e-01] | [6.613004207611084, 2.8383259773254395] |
32433187-d1b8-4c4b-9461-4d7d095dd6fc | dropping-convexity-for-more-efficient-and | 1702.08134 | null | http://arxiv.org/abs/1702.08134v9 | http://arxiv.org/pdf/1702.08134v9.pdf | Dropping Convexity for More Efficient and Scalable Online Multiview Learning | Multiview representation learning is very popular for latent factor analysis.
It naturally arises in many data analysis, machine learning, and information
retrieval applications to model dependent structures among multiple data
sources. For computational convenience, existing approaches usually formulate
the multiview representation learning as convex optimization problems, where
global optima can be obtained by certain algorithms in polynomial time.
However, many pieces of evidence have corroborated that heuristic nonconvex
approaches also have good empirical computational performance and convergence
to the global optima, although there is a lack of theoretical justification.
Such a gap between theory and practice motivates us to study a nonconvex
formulation for multiview representation learning, which can be efficiently
solved by a simple stochastic gradient descent (SGD) algorithm. We first
illustrate the geometry of the nonconvex formulation; Then, we establish
asymptotic global rates of convergence to the global optima by diffusion
approximations. Numerical experiments are provided to support our theory. | ['Chris J. Li', 'Tuo Zhao', 'Lin F. Yang', 'Zhehui Chen'] | 2017-02-27 | null | null | null | null | ['multiview-learning'] | ['computer-vision'] | [-1.32747665e-01 -8.17823783e-02 -7.02908516e-01 -3.11930954e-01
-1.20852077e+00 -5.53437352e-01 3.88316661e-01 6.10589683e-02
1.05363941e-02 6.09464347e-01 3.89821827e-01 -2.43398175e-01
-4.49914485e-01 -4.31047618e-01 -7.04207122e-01 -9.24077272e-01
4.41374071e-02 2.94663697e-01 -5.64382851e-01 1.99240949e-02
2.60487825e-01 7.26220086e-02 -1.16944897e+00 4.21885364e-02
1.02016008e+00 8.21778834e-01 1.11883871e-01 2.19782785e-01
3.53179872e-02 5.15273571e-01 -1.82541057e-01 -2.85733581e-01
4.91606206e-01 -1.93038151e-01 -4.30170417e-01 6.33395195e-01
2.49163479e-01 -1.82744011e-01 -8.41409143e-04 1.29180539e+00
3.38324070e-01 2.62871414e-01 6.38279796e-01 -1.38950384e+00
-9.14187849e-01 3.19941968e-01 -1.20370579e+00 -2.51294337e-02
2.02769145e-01 -3.83564860e-01 1.20847404e+00 -1.31321824e+00
6.13852978e-01 1.28513694e+00 4.83656824e-01 2.83722486e-02
-1.17413962e+00 -2.96753138e-01 4.71236587e-01 1.56950969e-02
-1.23629797e+00 -3.62143606e-01 9.90793169e-01 -4.83931690e-01
2.89010644e-01 2.84752280e-01 4.05373573e-01 7.11157620e-01
1.22803017e-01 1.16847682e+00 1.15049636e+00 -3.83720368e-01
3.05343717e-01 1.96095765e-01 2.71965861e-01 8.43779266e-01
2.65924454e-01 -3.27205628e-01 -5.15034556e-01 -4.88395691e-01
6.48753583e-01 3.34604293e-01 -4.27267343e-01 -8.46889973e-01
-1.10506380e+00 1.35926831e+00 2.64314651e-01 8.49705711e-02
-4.78273988e-01 9.13396254e-02 4.14427996e-01 4.24022108e-01
8.80486906e-01 -1.73543215e-01 -8.00404400e-02 1.79067180e-01
-8.05125773e-01 2.23647982e-01 5.58764517e-01 9.68000889e-01
7.20682919e-01 4.10846353e-01 3.24339271e-01 1.09407365e+00
5.43943405e-01 6.18837595e-01 4.73569989e-01 -1.08764136e+00
6.87631249e-01 3.30793023e-01 2.48359740e-01 -1.48066473e+00
-4.20570187e-02 -3.12469989e-01 -9.92704153e-01 -2.13185549e-02
3.84315044e-01 -1.86990172e-01 -3.63454849e-01 1.53718674e+00
3.77819210e-01 8.22032094e-02 3.68261375e-02 1.11243176e+00
4.02451307e-01 8.30470085e-01 -3.37586939e-01 -9.02390420e-01
9.47482526e-01 -1.06452382e+00 -9.39071596e-01 -3.54367048e-02
5.27779102e-01 -7.21675098e-01 1.02863050e+00 3.64059478e-01
-1.19893646e+00 -2.03263208e-01 -9.16580260e-01 6.71984581e-03
9.55257118e-02 2.04727471e-01 8.34244668e-01 6.31057501e-01
-8.32685769e-01 2.82212764e-01 -8.49690437e-01 -2.98332181e-02
3.56114149e-01 2.39378721e-01 -3.88327420e-01 -3.58200252e-01
-7.10956454e-01 3.32673550e-01 4.68172841e-02 3.12147647e-01
-7.19371378e-01 -7.35230148e-01 -9.07225370e-01 3.40213487e-03
5.35854995e-01 -7.83835351e-01 7.75537550e-01 -1.21803415e+00
-1.35030150e+00 7.35549748e-01 -6.97268844e-01 -1.27662450e-01
4.61534500e-01 -2.77059823e-01 -1.27720967e-01 1.03252806e-01
2.13271946e-01 -1.29802570e-01 1.38101363e+00 -1.56970227e+00
-3.84226769e-01 -5.78772902e-01 4.02293615e-02 3.40563178e-01
-3.85690331e-01 6.87047392e-02 -2.89240897e-01 -8.56674433e-01
5.48769951e-01 -9.26761985e-01 -5.21914721e-01 1.58308377e-03
-3.12005758e-01 -3.38303894e-01 7.98998535e-01 -5.82601190e-01
1.14289415e+00 -2.22808361e+00 6.79380655e-01 3.56759340e-01
3.49437952e-01 -2.51689196e-01 -9.00251977e-03 4.47946608e-01
-1.84410036e-01 2.29915455e-02 -3.26571196e-01 -4.59307134e-01
-3.39538343e-02 1.39133364e-01 -6.76423013e-01 1.14291787e+00
-4.67016667e-01 8.36590350e-01 -9.64955986e-01 -2.23448083e-01
1.58331484e-01 2.64071345e-01 -7.10794747e-01 3.02566718e-02
9.54765901e-02 8.28411728e-02 -7.68623471e-01 7.84495473e-01
6.02057636e-01 -6.38062477e-01 4.93099660e-01 -2.15617105e-01
-1.05604842e-01 -3.37012380e-01 -1.53791356e+00 1.52804649e+00
-3.98606211e-01 5.56882918e-01 5.91591656e-01 -1.61880040e+00
5.47090113e-01 4.32868421e-01 9.10996258e-01 -8.21452290e-02
-9.35487542e-03 3.16426426e-01 -5.85260272e-01 -3.12481165e-01
5.88531137e-01 -5.17739296e-01 4.84226793e-02 5.52068651e-01
-2.69631475e-01 1.29944563e-01 -1.21371143e-01 2.25735188e-01
3.22167724e-01 -2.80383825e-01 3.71134490e-01 -5.63352525e-01
4.05174762e-01 -8.90290216e-02 6.87646329e-01 5.42679012e-01
4.21086438e-02 6.90407753e-01 3.87416750e-01 -5.33978820e-01
-9.62705672e-01 -9.30207789e-01 -3.41967903e-02 1.04456985e+00
2.30953693e-01 -5.47504723e-01 -4.45494354e-01 -5.20065725e-01
-3.64213698e-02 2.56363273e-01 -4.52547073e-01 2.28436574e-01
-4.69389647e-01 -9.71330225e-01 -2.82869101e-01 3.88921171e-01
1.94036841e-01 -3.01777065e-01 -1.61013573e-01 1.02152266e-01
-2.95349002e-01 -9.31096137e-01 -6.12824380e-01 -2.08930507e-01
-1.13107240e+00 -8.56572211e-01 -1.06813669e+00 -6.18532002e-01
9.76223230e-01 8.89725924e-01 7.52835274e-01 2.39901990e-03
1.82728007e-01 8.14754784e-01 -1.82324618e-01 -2.15222344e-01
-3.44916247e-02 -8.27323794e-02 4.82961982e-01 4.20807213e-01
-1.50588721e-01 -5.46500385e-01 -4.37970549e-01 3.94111633e-01
-1.03482354e+00 -8.62667337e-02 1.00075707e-01 8.63674700e-01
9.94033039e-01 -3.20670637e-03 7.88394809e-01 -8.77397120e-01
9.13168848e-01 -7.82678843e-01 -7.62155354e-01 4.82481629e-01
-6.23591244e-01 -1.94602460e-02 5.36696374e-01 -2.59046465e-01
-9.92863417e-01 2.00546514e-02 3.57038409e-01 -9.61857378e-01
4.75468487e-01 1.10087812e+00 -1.75181612e-01 -1.94924921e-02
4.02449399e-01 3.01854789e-01 1.95345312e-01 -3.66203427e-01
7.27936625e-01 4.93096620e-01 8.29209983e-02 -8.61297846e-01
7.49150276e-01 8.24076056e-01 -3.08196321e-02 -1.04080856e+00
-1.18121493e+00 -6.87957287e-01 -3.34600061e-01 -2.78706372e-01
5.83018064e-01 -1.10437250e+00 -5.95202684e-01 5.53261638e-02
-8.10952783e-01 4.57275845e-02 -2.05778241e-01 5.07404447e-01
-9.39593613e-01 6.74943447e-01 -3.71953249e-01 -7.66997516e-01
-1.09398954e-01 -1.21527433e+00 8.13662529e-01 -1.51612118e-01
6.43730387e-02 -1.47244191e+00 -5.57946227e-03 6.31904840e-01
1.60025880e-01 2.46191844e-01 9.68444169e-01 -1.26659364e-01
-5.28802693e-01 -2.50433922e-01 -5.05122729e-02 2.54757911e-01
3.03673327e-01 -9.10616368e-02 -6.50180519e-01 -6.90907657e-01
4.30883855e-01 -3.07117432e-01 7.17369437e-01 8.56933236e-01
1.14464593e+00 -5.43758631e-01 -3.19841385e-01 8.82193744e-01
1.69087958e+00 -3.28235030e-01 2.45244130e-01 1.65184438e-01
8.42773616e-01 5.76786637e-01 5.39781213e-01 7.52377391e-01
4.02053505e-01 4.68848318e-01 3.99474680e-01 8.02976191e-02
3.90318066e-01 -1.27369612e-01 3.27036649e-01 1.33373594e+00
-3.13890874e-01 -1.79634258e-01 -6.12295210e-01 5.94181240e-01
-2.26692057e+00 -1.00964904e+00 -1.78947613e-01 2.41830921e+00
3.37392896e-01 -5.26187181e-01 1.39661908e-01 -9.98216644e-02
8.43870580e-01 5.75524569e-01 -4.30625170e-01 -2.49080539e-01
-3.22773665e-01 -4.43017274e-01 5.59312522e-01 6.50356591e-01
-9.95360017e-01 7.34319031e-01 6.94037247e+00 9.67942178e-01
-1.03458738e+00 1.70649305e-01 7.24718928e-01 -9.78795663e-02
-7.94701934e-01 1.11603015e-03 -5.31767488e-01 2.13756457e-01
4.30851579e-01 -5.85201204e-01 4.99827832e-01 8.74171197e-01
4.27212059e-01 1.17164135e-01 -8.19562554e-01 1.39256597e+00
1.37340114e-01 -1.54130149e+00 1.59490570e-01 2.65041262e-01
1.33237696e+00 -1.30400360e-01 3.89734566e-01 -2.47610901e-02
3.32275212e-01 -8.02957892e-01 4.67113793e-01 3.10114473e-01
4.87725437e-01 -9.50964570e-01 2.13931933e-01 4.76189286e-01
-1.27622330e+00 -3.28272879e-01 -7.26203203e-01 1.83275774e-01
4.50131148e-01 7.73102999e-01 -1.81104124e-01 8.33413124e-01
2.92255998e-01 1.11513662e+00 -1.25690743e-01 8.19607437e-01
-1.14549734e-01 6.83923423e-01 -2.88560033e-01 3.22635412e-01
3.14259440e-01 -9.24589872e-01 8.48003805e-01 6.33137107e-01
3.61797929e-01 1.40828431e-01 4.40001369e-01 6.33534014e-01
-1.96912110e-01 5.45614541e-01 -8.83315146e-01 -1.42045766e-02
1.44493297e-01 1.11741209e+00 -7.54306734e-01 -2.66096532e-01
-8.61381233e-01 7.53563821e-01 4.79869813e-01 8.00236583e-01
-6.34613395e-01 2.60633796e-01 5.98251283e-01 1.77726634e-02
2.56319880e-01 -5.47357917e-01 -3.95430684e-01 -1.74089110e+00
2.90602505e-01 -1.08249736e+00 5.46268225e-01 -3.65049928e-01
-1.41815257e+00 3.53335775e-02 1.24519765e-01 -1.33942354e+00
-2.84258574e-01 -2.76296675e-01 -4.17374939e-01 5.40035844e-01
-1.27588141e+00 -1.06831896e+00 1.85719684e-01 7.14196682e-01
7.67326415e-01 -3.17459911e-01 3.94662231e-01 2.67395556e-01
-5.56237459e-01 5.09396911e-01 8.24800909e-01 -2.80228734e-01
5.00313342e-01 -1.05568457e+00 -3.33360404e-01 7.03863621e-01
1.13476314e-01 6.90358460e-01 7.10304797e-01 -4.14597362e-01
-1.95623541e+00 -8.03167403e-01 6.43688738e-01 -2.01818645e-01
8.26399863e-01 9.22689512e-02 -8.92235398e-01 9.89502311e-01
1.80508032e-01 1.94980219e-01 8.32143068e-01 3.78209442e-01
-2.34160706e-01 5.90277091e-02 -8.60460103e-01 5.53649724e-01
5.86320996e-01 -5.27253091e-01 -2.61869639e-01 6.80884480e-01
3.46361935e-01 -1.49574399e-01 -9.03072834e-01 2.62405425e-01
5.51237285e-01 -6.16015673e-01 1.19750822e+00 -7.90817082e-01
5.34494102e-01 -2.18818352e-01 -6.51414871e-01 -1.49440193e+00
-1.19109422e-01 -7.56251276e-01 -3.75469267e-01 8.81976366e-01
1.55150130e-01 -6.52126551e-01 7.11612880e-01 7.12923467e-01
-6.02568164e-02 -8.78435075e-01 -9.41855252e-01 -8.60255599e-01
2.59445518e-01 -3.56624216e-01 -3.67283681e-03 1.26601171e+00
4.74175923e-02 2.38928869e-01 -8.23248625e-01 2.89863765e-01
1.04032123e+00 7.65070319e-01 7.23416626e-01 -9.62031543e-01
-1.91630572e-01 -3.37616295e-01 -1.43147185e-01 -1.35797119e+00
5.73350489e-01 -1.07164574e+00 -2.95585752e-01 -1.38227308e+00
3.73400420e-01 -4.04952139e-01 -1.62185475e-01 1.13290101e-02
-3.42710078e-01 -1.06772810e-01 3.21833938e-01 5.45785189e-01
-6.51732624e-01 8.46971214e-01 1.38096511e+00 -2.96820223e-01
-2.85213917e-01 1.04412287e-01 -7.87743747e-01 1.01767957e+00
5.86171150e-01 -3.35245490e-01 -6.70805454e-01 -5.88093102e-01
5.55283666e-01 3.88946861e-01 5.55047877e-02 -3.62800926e-01
1.22628197e-01 -4.48012382e-01 -9.44931433e-02 -4.77252156e-01
3.03271532e-01 -8.28310847e-01 2.50570197e-02 2.49929085e-01
-3.27687114e-01 1.50042519e-01 -3.65071028e-01 1.05422604e+00
-4.29380774e-01 -3.75033557e-01 6.81728184e-01 -5.45578524e-02
-3.17607075e-01 5.40154636e-01 -4.68581736e-01 1.25192285e-01
9.01252568e-01 -4.70939912e-02 9.06986371e-02 -8.56299877e-01
-7.67150700e-01 3.15156192e-01 3.46888989e-01 1.76227927e-01
7.76323915e-01 -1.59920943e+00 -7.20991313e-01 9.91991162e-02
-1.23730272e-01 -2.10003316e-01 2.23097131e-01 9.30644393e-01
-1.29608870e-01 2.76885688e-01 3.18729788e-01 -6.58831358e-01
-1.07895219e+00 8.38718534e-01 1.10125683e-01 -4.05560285e-02
-5.20680666e-01 5.15576303e-01 4.66043621e-01 -2.74031520e-01
3.25672962e-02 -9.36619937e-02 -6.45544240e-03 3.38920921e-01
5.17301023e-01 7.96284318e-01 -2.82611132e-01 -7.54575491e-01
-2.72100512e-02 6.43743157e-01 1.39446959e-01 -1.58804402e-01
1.45390713e+00 -3.88690352e-01 -3.18028927e-01 6.24839723e-01
1.53920996e+00 2.60576040e-01 -1.17699802e+00 -4.27399188e-01
-1.63522765e-01 -7.69051611e-01 1.54221684e-01 3.23293149e-01
-1.36047852e+00 6.73202574e-01 2.87268966e-01 3.97953957e-01
9.22236204e-01 -9.02254973e-03 5.38569272e-01 5.64590096e-01
4.58721966e-01 -1.16053152e+00 1.19608648e-01 2.02925682e-01
1.11629379e+00 -1.49589694e+00 4.05941576e-01 -5.83677351e-01
-4.95276392e-01 1.09822607e+00 6.40914589e-02 -4.22526628e-01
9.94489312e-01 -7.99716488e-02 -8.68772119e-02 -2.36833170e-01
-6.10527098e-01 1.93692103e-01 2.90399045e-01 2.77727574e-01
2.96911448e-01 2.35715896e-01 -4.21317011e-01 4.41218317e-01
1.16180889e-01 -2.74931639e-01 4.05367702e-01 7.85304606e-01
-3.91089350e-01 -1.05513847e+00 -4.86755550e-01 3.59084457e-01
-6.68839097e-01 6.99317828e-02 1.30400583e-01 5.61628640e-01
-6.18365586e-01 8.64773750e-01 -2.11331934e-01 2.35995263e-01
-9.19723734e-02 -1.08480871e-01 4.17851806e-01 -3.97876561e-01
-7.09543005e-02 5.13584971e-01 -2.68377215e-01 -4.51991141e-01
-7.63788581e-01 -9.40024197e-01 -9.63696241e-01 -3.31416309e-01
-3.94937485e-01 2.20521629e-01 5.36336541e-01 1.03682625e+00
2.15521410e-01 1.83135327e-02 1.03111207e+00 -7.16255963e-01
-6.67707920e-01 -4.05170709e-01 -8.08441520e-01 4.28510696e-01
4.51532334e-01 -4.99048591e-01 -4.37133372e-01 3.29893768e-01] | [7.179752826690674, 4.476853847503662] |
541e2810-01c3-47a9-bf63-0d37d3f15eb0 | advances-in-collaborative-filtering-and | 2002.12312 | null | https://arxiv.org/abs/2002.12312v1 | https://arxiv.org/pdf/2002.12312v1.pdf | Advances in Collaborative Filtering and Ranking | In this dissertation, we cover some recent advances in collaborative filtering and ranking. In chapter 1, we give a brief introduction of the history and the current landscape of collaborative filtering and ranking; chapter 2 we first talk about pointwise collaborative filtering problem with graph information, and how our proposed new method can encode very deep graph information which helps four existing graph collaborative filtering algorithms; chapter 3 is on the pairwise approach for collaborative ranking and how we speed up the algorithm to near-linear time complexity; chapter 4 is on the new listwise approach for collaborative ranking and how the listwise approach is a better choice of loss for both explicit and implicit feedback over pointwise and pairwise loss; chapter 5 is about the new regularization technique Stochastic Shared Embeddings (SSE) we proposed for embedding layers and how it is both theoretically sound and empirically effectively for 6 different tasks across recommendation and natural language processing; chapter 6 is how we introduce personalization for the state-of-the-art sequential recommendation model with the help of SSE, which plays an important role in preventing our personalized model from overfitting to the training data; chapter 7, we summarize what we have achieved so far and predict what the future directions can be; chapter 8 is the appendix to all the chapters. | ['Liwei Wu'] | 2020-02-27 | null | null | null | null | ['collaborative-ranking'] | ['graphs'] | [-1.91510260e-01 4.07919427e-03 -3.46840546e-02 -4.99324977e-01
-4.31806535e-01 -5.02259195e-01 3.22182745e-01 2.85812169e-01
-2.25651398e-01 3.70979011e-01 9.67404366e-01 -3.12963068e-01
-1.14794672e+00 -1.17999804e+00 -3.69624555e-01 -4.78897870e-01
-6.48873389e-01 5.11087298e-01 8.88474956e-02 -6.85318708e-01
4.55487609e-01 2.52907276e-01 -1.29218745e+00 4.20874894e-01
8.10099304e-01 6.55456424e-01 -8.85183960e-02 8.75902653e-01
-2.15964988e-01 6.06878221e-01 -3.91741216e-01 -1.03297615e+00
5.07695377e-01 -2.98423648e-01 -7.91687846e-01 -3.34642887e-01
7.66881287e-01 -2.27112725e-01 -6.64176226e-01 8.45629036e-01
8.69308650e-01 9.53840494e-01 7.68132389e-01 -8.08106899e-01
-1.50381458e+00 1.31044519e+00 -1.58472940e-01 1.69746596e-02
3.63648295e-01 -4.77672100e-01 1.49736094e+00 -9.09538388e-01
6.46446943e-01 1.14471138e+00 9.06607509e-01 6.64177299e-01
-8.24532986e-01 -4.68947321e-01 7.71147013e-01 5.31186238e-02
-1.07809258e+00 1.67030953e-02 7.85521269e-01 -2.41285861e-01
9.48603034e-01 1.64455712e-01 7.01704144e-01 9.22554374e-01
2.16172844e-01 7.05477417e-01 5.48165739e-01 -2.63251066e-01
-1.03124797e-01 -1.67675987e-01 8.08305502e-01 8.89185131e-01
3.22553426e-01 8.10281709e-02 -5.34660578e-01 -4.28676486e-01
7.14619577e-01 3.45611244e-01 -1.02850102e-01 -3.58402729e-01
-6.96429074e-01 9.18432355e-01 5.31159401e-01 4.35581356e-01
5.37841879e-02 3.13628428e-02 4.10990417e-01 1.05157256e+00
8.72791290e-01 5.17579854e-01 -6.45085275e-01 3.85092378e-01
-7.64106154e-01 2.61842281e-01 1.10847867e+00 6.27915084e-01
3.60757440e-01 -1.08766347e-01 -3.35404992e-01 1.21835911e+00
4.91576076e-01 1.53763175e-01 1.11243203e-01 -8.28291833e-01
3.54767144e-01 4.64847326e-01 9.51681286e-02 -1.40166426e+00
-3.40580225e-01 -9.23136115e-01 -7.36819983e-01 -1.41338274e-01
3.08103651e-01 -3.82050604e-01 -3.41074884e-01 1.44664025e+00
1.40434116e-01 5.59674427e-02 -2.34341890e-01 7.95731366e-01
1.18711388e+00 4.81339037e-01 -1.21329568e-01 -1.61923483e-01
9.13557351e-01 -1.28161907e+00 -6.33045495e-01 6.21072985e-02
6.63974166e-01 -8.66079569e-01 8.73338759e-01 2.02955291e-01
-1.05578208e+00 -6.39059424e-01 -8.18070948e-01 -2.24702194e-01
-6.63720667e-01 5.53674251e-02 1.40064490e+00 8.71088922e-01
-1.29595733e+00 1.23329914e+00 -5.08375823e-01 -4.07194018e-01
3.70539397e-01 6.97186470e-01 -1.79641962e-01 -4.88492459e-01
-1.43855596e+00 6.15221202e-01 -2.72282183e-01 1.15447029e-01
-3.39307606e-01 -8.89158189e-01 -4.46957767e-01 1.79639250e-01
3.66770059e-01 -1.13723803e+00 5.63789904e-01 -5.08004785e-01
-1.78953815e+00 3.51262480e-01 1.18409004e-02 -1.36243150e-01
3.30837309e-01 -6.63691580e-01 -5.64171672e-01 -4.27784234e-01
-3.88654709e-01 5.45227807e-03 5.00813782e-01 -9.30686116e-01
-2.91391104e-01 -4.52055007e-01 3.83279175e-01 4.26745653e-01
-7.68400431e-01 1.67097792e-01 -4.75539267e-01 -8.81704688e-01
3.31529118e-02 -5.20876944e-01 -5.48913717e-01 -3.17722470e-01
3.54711972e-02 -6.74369454e-01 3.45161378e-01 -3.73527408e-01
1.47716081e+00 -1.89005470e+00 3.85192037e-01 2.93940216e-01
3.68192166e-01 1.48741812e-01 -6.41667426e-01 1.24366808e+00
1.24911770e-01 1.33816451e-01 2.11166784e-01 -5.10384202e-01
3.05006832e-01 5.13289534e-02 -2.68861532e-01 4.10928845e-01
-4.86969888e-01 1.05256319e+00 -1.30914891e+00 -4.03048620e-02
2.92014349e-02 7.71497130e-01 -7.71322012e-01 1.50254354e-01
-4.10595024e-03 2.61627018e-01 -3.99402827e-01 1.85262680e-01
6.25463068e-01 -1.98406085e-01 4.46945965e-01 -1.29399374e-01
2.49249831e-01 5.09940863e-01 -1.47964883e+00 1.80404615e+00
-4.47611660e-01 2.26126745e-01 1.36743098e-01 -8.21496665e-01
8.65108430e-01 1.17737271e-01 6.48853719e-01 -2.72243291e-01
3.75136919e-02 -1.18275039e-01 -3.04099739e-01 -1.02247059e-01
7.66805351e-01 4.01005894e-01 2.84124225e-01 9.68207955e-01
3.08768488e-02 5.54685295e-02 2.65764564e-01 8.08246195e-01
1.49156892e+00 -1.37960345e-01 -2.63979822e-01 -2.13807151e-01
4.43266064e-01 -5.07040739e-01 2.62032062e-01 1.21699941e+00
2.07637519e-01 1.99531063e-01 7.31328800e-02 -3.96870494e-01
-5.02066791e-01 -1.02815640e+00 2.80146927e-01 1.89730847e+00
-1.07855551e-01 -9.96246815e-01 -2.46307570e-02 -1.04851568e+00
2.35012665e-01 3.02472800e-01 -7.16472566e-01 3.69021744e-02
-3.62246931e-01 -7.99987912e-01 -4.50673252e-02 6.06822550e-01
2.61339962e-01 -6.50313199e-01 9.21283782e-01 2.16315076e-01
3.32984626e-01 -2.26907343e-01 -8.21527183e-01 3.94273326e-02
-1.47475660e+00 -1.18491161e+00 -4.89517272e-01 -8.89452457e-01
5.65144598e-01 7.06412792e-01 1.37151468e+00 5.45084298e-01
2.82856882e-01 8.48886788e-01 -7.68664539e-01 8.54837298e-02
2.73515463e-01 -5.30757084e-02 7.67136812e-02 -2.09094301e-01
6.06130064e-01 -7.19757378e-01 -8.44799876e-01 2.26547450e-01
-6.26186669e-01 -5.98373413e-01 1.86378732e-01 8.57768118e-01
2.93165773e-01 2.68157244e-01 5.04546225e-01 -1.61585879e+00
1.33105707e+00 -3.96167874e-01 -2.58207232e-01 1.75254777e-01
-9.36390460e-01 -9.28148851e-02 7.36742854e-01 -2.29384258e-01
-8.96550000e-01 -1.67027831e-01 -3.48678112e-01 3.28979902e-02
4.68020111e-01 6.08471334e-01 2.17953727e-01 -2.28573263e-01
6.25997126e-01 -2.99584895e-01 -3.23320597e-01 -9.04732406e-01
1.13074946e+00 5.50227582e-01 5.73308673e-04 -6.19262934e-01
7.29901075e-01 3.74047875e-01 8.59580934e-03 -4.96691108e-01
-1.20766628e+00 -7.46996701e-01 -6.56377733e-01 -1.75712943e-01
4.15297568e-01 -6.95924580e-01 -7.51401246e-01 1.34682342e-01
-8.01333964e-01 -2.89683789e-01 -2.55598813e-01 5.31911492e-01
-1.73244089e-01 7.29901075e-01 -1.30797029e+00 -8.15523803e-01
-6.44514918e-01 -6.74044132e-01 5.63946784e-01 -2.03121584e-02
1.22028954e-01 -1.60901868e+00 3.10506672e-01 4.97627854e-01
7.52686977e-01 -3.12701255e-01 8.81072342e-01 -9.46224034e-01
-4.57097709e-01 -3.93522054e-01 -1.13831177e-01 3.67257327e-01
2.77878586e-02 -4.41570245e-02 -5.24415314e-01 -6.90433741e-01
-5.23017526e-01 -1.19539030e-01 1.23275161e+00 4.10466850e-01
1.00110674e+00 -9.00781974e-02 -3.54636163e-01 5.37596703e-01
1.43479049e+00 -2.76904523e-01 5.63346565e-01 -4.95045520e-02
9.36187685e-01 6.38587236e-01 5.26448965e-01 2.26268500e-01
3.21277589e-01 3.55940789e-01 1.70033917e-01 2.28171483e-01
-4.21374947e-01 -3.76662433e-01 4.69107687e-01 1.61568964e+00
-5.28242588e-01 -5.07451057e-01 1.19730011e-02 1.98803887e-01
-2.19276118e+00 -1.04526079e+00 -4.18214470e-01 2.40346336e+00
5.43675303e-01 -2.47859702e-01 1.66688655e-02 -5.04085235e-02
7.41677642e-01 3.43828917e-01 -2.86416821e-02 -5.65861762e-01
1.45622324e-02 6.14309788e-01 5.89532375e-01 9.01145756e-01
-1.28506255e+00 1.17455161e+00 7.24335766e+00 6.91108823e-01
-3.79661769e-01 2.42461994e-01 -1.84657797e-02 -6.16328381e-02
-4.05767322e-01 -1.99503466e-01 -7.50835657e-01 -3.52990627e-03
8.93910944e-01 5.59048429e-02 7.20792711e-01 7.11545289e-01
8.81828889e-02 4.58831370e-01 -1.12208223e+00 7.04204917e-01
1.82539657e-01 -1.50854492e+00 1.61704525e-01 -3.00235022e-03
1.18051636e+00 2.31305242e-01 1.48382425e-01 3.28330874e-01
1.11547101e+00 -8.54256749e-01 -1.61317736e-01 7.94457197e-01
2.81329691e-01 -7.01530755e-01 9.43822742e-01 -2.04357784e-02
-1.24340081e+00 -3.59190434e-01 -9.74457026e-01 -3.85868371e-01
2.33972073e-03 1.05834007e+00 7.42341578e-02 9.59057629e-01
5.99917591e-01 1.34305370e+00 -4.30069834e-01 1.28078401e+00
-4.61911589e-01 6.33854449e-01 -1.31863743e-01 -4.14076596e-01
-1.45101726e-01 -6.74969971e-01 5.56506932e-01 1.24315500e+00
2.30853464e-02 8.18731636e-02 1.57607883e-01 4.86634038e-02
-3.55606019e-01 4.62938398e-01 -6.11191928e-01 -1.05496123e-01
4.10173088e-01 1.01763153e+00 -5.59297085e-01 -9.17147845e-02
-6.03992462e-01 1.00276399e+00 7.87555158e-01 5.05794764e-01
-3.15713406e-01 -6.97251081e-01 7.35668719e-01 -1.02579631e-01
4.74014521e-01 -5.15748441e-01 -1.07117057e-01 -1.22722018e+00
-4.64610934e-01 -5.22866726e-01 7.66284883e-01 -2.59206533e-01
-2.21127272e+00 4.03652042e-01 -2.93069303e-01 -9.18640733e-01
3.74853998e-01 -6.74016535e-01 -7.10312188e-01 9.36392844e-01
-1.17938745e+00 -8.53474557e-01 1.03343286e-01 6.15869164e-01
3.69709373e-01 -2.70961761e-01 1.04494977e+00 6.62688315e-01
-3.17779183e-01 9.43386376e-01 7.00472295e-01 1.33298114e-01
1.24331486e+00 -1.26404691e+00 3.57510984e-01 4.68540996e-01
5.19805253e-01 1.29096508e+00 3.60916704e-01 -7.30427980e-01
-1.89425194e+00 -1.13568187e+00 1.20682645e+00 -7.33306408e-01
8.31789613e-01 -2.77573436e-01 -6.78954959e-01 4.86435026e-01
1.60211414e-01 -5.71146309e-02 1.11251712e+00 1.06294322e+00
-5.43224275e-01 -2.19437793e-01 -9.97056007e-01 5.61120391e-01
1.67884493e+00 -4.80596751e-01 -5.47016919e-01 9.02838886e-01
7.54660249e-01 -5.28848357e-02 -1.20708179e+00 3.21214758e-02
7.66163349e-01 -9.62053597e-01 1.07862675e+00 -9.43740785e-01
3.12908381e-01 -2.69482762e-01 -1.91927895e-01 -1.48194683e+00
-1.24108517e+00 -7.02014863e-01 -2.23620176e-01 1.05609202e+00
5.05125105e-01 -7.68529832e-01 9.70238507e-01 2.18164578e-01
-1.30434185e-01 -9.23995614e-01 -3.17996621e-01 -5.91256499e-01
3.22959065e-01 -4.36355233e-01 5.34000337e-01 1.05898321e+00
-4.43076864e-02 4.92081016e-01 -4.73572820e-01 5.31905107e-02
4.91006941e-01 1.05413482e-01 7.05662310e-01 -1.38296139e+00
-6.84440732e-01 -3.72881621e-01 -2.09481314e-01 -1.60296440e+00
-1.13608964e-01 -1.27985227e+00 -6.98529124e-01 -2.22716451e+00
1.46989524e-01 -2.89611697e-01 -7.83783317e-01 1.11397117e-01
-1.60350367e-01 4.89226013e-01 8.29425678e-02 6.37174621e-02
-8.92031848e-01 4.60084587e-01 1.51920283e+00 -3.50238740e-01
-4.40126181e-01 3.48505080e-01 -1.07176888e+00 2.29465619e-01
4.30341333e-01 -4.49625522e-01 -7.10549235e-01 -7.64511347e-01
9.48903739e-01 -1.22715682e-01 -3.74723077e-01 -5.09279788e-01
5.99704087e-01 1.09092608e-01 3.62191290e-01 -4.22334343e-01
1.42638847e-01 -7.95305789e-01 -7.84244388e-02 3.26987132e-02
-5.93602598e-01 -2.78300531e-02 -2.65843809e-01 1.17279518e+00
1.59778625e-01 -1.83617175e-01 2.72059351e-01 -1.52085587e-01
-4.40531313e-01 6.44881248e-01 -2.45682940e-01 -3.33045304e-01
3.08599085e-01 8.81956592e-02 -3.18323642e-01 -5.70873678e-01
-1.06629157e+00 4.15070117e-01 1.18639447e-01 5.17329752e-01
5.35887003e-01 -1.21872628e+00 -6.81154013e-01 -1.24181613e-01
-1.21027686e-01 -7.78285861e-01 2.34966144e-01 6.61924303e-01
-1.25535935e-01 1.83968484e-01 3.86573255e-01 3.03051174e-01
-1.39824152e+00 5.15173793e-01 1.01503722e-01 -6.09705269e-01
-6.30070448e-01 1.43333602e+00 -4.42112267e-01 -7.33161509e-01
5.98718464e-01 8.69099945e-02 -7.38500357e-01 -7.25342855e-02
5.46596885e-01 5.84881723e-01 2.69660681e-01 -8.58207569e-02
-3.31093431e-01 8.37355852e-01 -3.97922635e-01 4.01921540e-01
1.63501024e+00 -1.55678570e-01 -5.84954739e-01 2.47300044e-01
1.05309117e+00 6.07500553e-01 -6.08241737e-01 -2.40986466e-01
-3.88008863e-01 -5.97906053e-01 2.72654355e-01 -1.14822531e+00
-1.19647014e+00 5.26526451e-01 4.99001831e-01 2.69867718e-01
8.49577546e-01 -1.06752932e-01 7.18112350e-01 7.52322316e-01
4.73725021e-01 -1.22979283e+00 -1.73080683e-01 9.96587276e-01
7.53584146e-01 -9.93156672e-01 6.68651938e-01 -4.89814967e-01
-3.96485358e-01 1.25075591e+00 7.29412809e-02 -7.57092297e-01
1.45811832e+00 -1.16161123e-01 -2.05014184e-01 -4.53798413e-01
-8.35191548e-01 -1.29969493e-01 6.54969692e-01 7.53752947e-01
1.16215849e+00 1.46159187e-01 -1.03316605e+00 9.02550578e-01
-2.17906103e-01 -2.64215857e-01 2.00847872e-02 5.73370576e-01
-3.77865076e-01 -1.90011990e+00 2.09189042e-01 9.54002678e-01
-4.42250878e-01 -2.81010956e-01 -7.82062292e-01 3.01319420e-01
9.77723375e-02 1.55086601e+00 -3.28003049e-01 -1.05790854e+00
2.11966097e-01 -6.74074739e-02 7.72103667e-01 -1.08206654e+00
-1.09801471e+00 -2.92985290e-01 3.06438208e-01 -5.69485843e-01
-3.43793213e-01 -1.29544884e-01 -8.02274525e-01 -6.84352934e-01
-6.60676599e-01 7.14723110e-01 5.82978964e-01 6.21361971e-01
5.12599289e-01 5.68810463e-01 5.81538916e-01 -6.64867699e-01
-6.04330838e-01 -1.00327885e+00 -9.79027450e-01 2.56776243e-01
-2.34744668e-01 -5.65155149e-01 -6.68488443e-01 -5.82973182e-01] | [10.1134672164917, 5.657388210296631] |
c7dd95a6-5741-44a8-99f1-bdc618564f84 | bert-based-arabic-social-media | 1909.04181 | null | https://arxiv.org/abs/1909.04181v3 | https://arxiv.org/pdf/1909.04181v3.pdf | BERT-Based Arabic Social Media Author Profiling | We report our models for detecting age, language variety, and gender from social media data in the context of the Arabic author profiling and deception detection shared task (APDA). We build simple models based on pre-trained bidirectional encoders from transformers (BERT). We first fine-tune the pre-trained BERT model on each of the three datasets with shared task released data. Then we augment shared task data with in-house data for gender and dialect, showing the utility of augmenting training data. Our best models on the shared task test data are acquired with a majority voting of various BERT models trained under different data conditions. We acquire 54.72% accuracy for age, 93.75% for dialect, 81.67% for gender, and 40.97% joint accuracy across the three tasks. | ['Muhammad Abdul-Mageed', 'Chiyu Zhang'] | 2019-09-09 | null | null | null | null | ['deception-detection'] | ['miscellaneous'] | [-4.27065104e-01 1.42366469e-01 6.94601284e-03 -7.27029145e-01
-9.74857509e-01 -7.23392546e-01 9.62824643e-01 -1.40558124e-01
-8.07635307e-01 6.89352334e-01 2.57914573e-01 -1.53347388e-01
3.80691051e-01 -4.72231120e-01 -4.41110015e-01 -3.57050836e-01
-1.08774588e-01 6.59859598e-01 -3.37054551e-01 -2.77930975e-01
2.96499640e-01 3.58292252e-01 -9.69089448e-01 4.21318203e-01
1.17360008e+00 9.70008433e-01 -7.13019669e-01 1.14798129e+00
2.97233880e-01 7.75504947e-01 -9.71311271e-01 -1.49703312e+00
4.23709629e-03 -6.88667893e-02 -1.00874698e+00 -3.67866248e-01
8.49608660e-01 -9.43524003e-01 -3.21131200e-01 8.27823102e-01
7.47639120e-01 -4.29198712e-01 1.00917304e+00 -1.57651889e+00
-1.00834715e+00 1.08221400e+00 -6.65556669e-01 7.22410753e-02
3.42342615e-01 -9.64066535e-02 1.16277945e+00 -1.11433434e+00
5.21136642e-01 1.41103506e+00 9.67258751e-01 7.58425713e-01
-1.12344038e+00 -1.08947933e+00 -1.77866351e-02 1.95257172e-01
-1.43069363e+00 -8.02100718e-01 4.93546546e-01 -5.61871290e-01
6.82718337e-01 2.36434773e-01 2.71782070e-01 1.82832563e+00
-1.30595490e-01 1.03973329e+00 1.26914680e+00 -1.87161520e-01
-3.59385788e-01 2.86865979e-01 5.50056756e-01 9.66450334e-01
1.79945484e-01 -2.82357782e-01 -1.05152738e+00 -4.02466655e-01
2.02087969e-01 -6.07649088e-01 2.15820540e-02 2.67371535e-01
-9.05864775e-01 8.42286408e-01 1.03903852e-01 -1.43629638e-03
2.68391997e-01 -3.09008896e-01 5.08595467e-01 3.96987081e-01
8.83750141e-01 5.28471470e-01 -5.96671581e-01 -2.53293306e-01
-1.11876762e+00 5.84330380e-01 1.05211949e+00 8.63582790e-01
3.87398601e-01 1.32473812e-01 -3.98009941e-02 1.18575287e+00
3.91515762e-01 7.30011880e-01 5.65959275e-01 -7.34853148e-01
4.95087296e-01 4.05118197e-01 6.54355660e-02 -5.89524150e-01
-5.35288215e-01 -4.69618052e-01 -5.69201589e-01 -2.83841863e-02
1.07362616e+00 -4.81505752e-01 -8.93961370e-01 1.80916715e+00
-1.85123205e-01 -2.83526719e-01 -8.58179480e-02 5.40510297e-01
8.76730204e-01 3.47629488e-01 -2.58287834e-03 2.75489748e-01
1.35306394e+00 -8.03911150e-01 -5.43125927e-01 -3.02752614e-01
9.21898365e-01 -6.00283384e-01 8.15534651e-01 6.39930427e-01
-1.04583347e+00 -3.71911108e-01 -1.13556039e+00 -4.12276775e-01
-5.09140968e-01 4.05786335e-01 3.73402268e-01 1.08878720e+00
-1.04082298e+00 4.93696213e-01 -5.05167663e-01 -2.13478908e-01
5.63297927e-01 3.29292566e-01 -4.68985051e-01 -7.45534077e-02
-1.22546649e+00 8.97961378e-01 -1.88715354e-01 1.12277001e-01
-8.86915743e-01 -6.30705237e-01 -8.19639206e-01 -1.56408429e-01
-3.06492746e-01 -2.50430435e-01 1.19364011e+00 -1.02975249e+00
-1.42116356e+00 1.56991744e+00 -3.05401385e-02 -4.20547843e-01
7.64520407e-01 -6.44833982e-01 -7.12773502e-01 -3.14203441e-01
1.89351931e-01 5.54712832e-01 8.67556095e-01 -9.90100265e-01
-3.85778546e-01 -6.86296701e-01 -3.13894510e-01 -2.64243558e-02
-6.25186741e-01 6.35095298e-01 3.93047929e-02 -4.57609355e-01
-3.04943889e-01 -7.89706826e-01 5.68991125e-01 -4.48174447e-01
-7.27401078e-01 -2.90652722e-01 4.46244121e-01 -1.62382793e+00
1.28160691e+00 -1.99757981e+00 3.47943276e-01 1.85377318e-02
4.78633583e-01 1.11132033e-01 -1.13077335e-01 1.93907589e-01
7.53414556e-02 3.03984970e-01 -1.22603653e-02 -1.06053495e+00
2.72497475e-01 -2.54632503e-01 -1.51451424e-01 4.01766330e-01
2.72720307e-01 7.96656966e-01 -6.40656769e-01 -3.91017348e-01
-5.47203779e-01 1.42334595e-01 -6.16772056e-01 4.41690147e-01
3.78370166e-01 2.90175855e-01 2.48092055e-01 1.09780777e+00
9.52333391e-01 1.39565408e-01 2.49315336e-01 2.34340400e-01
1.36764273e-01 4.01010215e-01 -7.85129964e-01 1.15819669e+00
-2.80197173e-01 7.89442599e-01 5.31550288e-01 -2.15495855e-01
1.25875580e+00 -5.42289093e-02 6.53868541e-02 -5.39862692e-01
3.42860401e-01 6.25372171e-01 3.39968204e-01 -2.00172737e-01
7.28537917e-01 1.25730932e-01 -5.22261798e-01 6.70659602e-01
5.48352540e-01 2.82914013e-01 -7.94011205e-02 4.45289195e-01
8.19936395e-01 -7.17221871e-02 -1.26487091e-01 -2.97766328e-01
6.21901095e-01 -5.69805920e-01 3.16749781e-01 7.13048697e-01
-5.96967220e-01 5.85697830e-01 8.91042769e-01 -4.01093274e-01
-1.10822129e+00 -9.48249876e-01 -2.49424055e-01 1.74704039e+00
-6.10538661e-01 -6.81979179e-01 -6.97513282e-01 -1.02028799e+00
1.46539524e-01 5.83890021e-01 -7.50844240e-01 -1.63384125e-01
-5.36282957e-01 -1.04628670e+00 1.42350519e+00 3.39361131e-01
5.14414728e-01 -5.22863746e-01 1.74454808e-01 -2.30270818e-01
-2.20350772e-01 -1.05573130e+00 -3.34919512e-01 -2.21449621e-02
-1.66410908e-01 -1.14539969e+00 -8.63854051e-01 -7.66323447e-01
1.92862943e-01 -5.12223542e-01 1.27354431e+00 2.36361697e-01
2.22992584e-01 5.70136160e-02 -2.17923790e-01 -5.10630965e-01
-5.46605289e-01 8.40059042e-01 1.43236265e-01 1.73929065e-01
5.19396305e-01 -2.41243333e-01 -1.06477588e-01 2.07777828e-01
-1.03852436e-01 2.68975552e-03 2.08897069e-02 9.92011249e-01
-6.09188497e-01 -8.86434853e-01 7.21841156e-01 -9.08701301e-01
6.77347064e-01 -5.89177191e-01 -2.06528857e-01 5.46280183e-02
-4.54528540e-01 -1.50335923e-01 3.07807058e-01 -2.20227867e-01
-8.22093487e-01 -3.79036099e-01 -4.37686890e-01 1.44063875e-01
-1.13203838e-01 2.31242418e-01 -3.71885777e-01 3.77481878e-01
6.75983012e-01 -1.52232990e-01 1.64896116e-01 -6.75033450e-01
1.39437735e-01 1.46760464e+00 6.58323824e-01 -7.40289152e-01
4.98409897e-01 -7.71775767e-02 -5.11152744e-01 -5.39361179e-01
-8.02200198e-01 9.85173229e-03 -9.44870293e-01 5.14681749e-02
7.01316357e-01 -1.19399023e+00 -8.01062942e-01 1.48269272e+00
-9.45341527e-01 -5.00142336e-01 4.75525737e-01 8.13449472e-02
-4.94999066e-02 2.65861064e-01 -1.11373174e+00 -9.91625547e-01
-6.10576630e-01 -1.00034726e+00 1.14698553e+00 -1.67834684e-01
-6.58280492e-01 -8.89133871e-01 -5.83364852e-02 5.72068572e-01
3.00320029e-01 1.72982663e-01 8.10503364e-01 -1.23730099e+00
1.38682872e-01 -1.67858973e-01 -3.91048312e-01 3.81158829e-01
-1.35659933e-01 4.11539137e-01 -1.48114812e+00 -3.87791693e-01
-7.14675605e-01 -8.78451109e-01 9.76515174e-01 -2.03796685e-01
1.21539438e+00 -3.16938996e-01 -7.61108547e-02 7.69436955e-01
6.83366776e-01 -3.67426664e-01 3.80503356e-01 3.01855862e-01
9.70287204e-01 5.23877084e-01 4.01117168e-02 5.46281099e-01
8.50321889e-01 5.55331945e-01 2.18341991e-01 1.61270365e-01
1.42795607e-01 -1.85921431e-01 6.88336194e-01 6.19737864e-01
-1.11431427e-01 -1.79925367e-01 -1.30520976e+00 5.47331035e-01
-1.52298319e+00 -7.30736732e-01 -3.28799039e-01 1.88120019e+00
1.09627450e+00 4.30455394e-02 7.62908340e-01 2.34868288e-01
5.77665687e-01 1.82340220e-01 -3.02129686e-01 -1.11722636e+00
-3.93816471e-01 2.33483598e-01 5.53210974e-01 8.76428664e-01
-1.22587502e+00 9.31236327e-01 6.89849615e+00 4.24068153e-01
-9.08403039e-01 1.41098335e-01 1.00466156e+00 -3.31244320e-01
-1.10614844e-01 -5.36322773e-01 -1.14165270e+00 7.21139610e-01
1.36128938e+00 -1.13992728e-01 6.20112300e-01 7.03604043e-01
-5.94560862e-01 1.19050041e-01 -1.39519739e+00 1.13265276e+00
4.92607653e-01 -6.57368898e-01 -3.03112060e-01 2.84124374e-01
5.44027209e-01 1.58168282e-02 3.51503849e-01 6.68779492e-01
6.63218856e-01 -1.43923032e+00 1.05901587e+00 2.55911350e-01
1.09650230e+00 -9.64085579e-01 9.10658717e-01 3.53690356e-01
-1.53036758e-01 -3.62430573e-01 1.07433230e-01 -3.01023960e-01
-2.31723487e-01 4.22611982e-01 -9.56476748e-01 3.84920448e-01
7.06831932e-01 8.87732565e-01 -1.16864491e+00 3.63495618e-01
-3.34924102e-01 8.28431904e-01 -2.62637764e-01 1.68451428e-01
-7.81587586e-02 1.16750754e-01 2.44886711e-01 1.22869742e+00
2.03037471e-01 -6.38367891e-01 -2.40841478e-01 4.32229966e-01
-5.27718842e-01 -1.10291086e-01 -2.72901028e-01 -4.32570232e-03
5.88963926e-01 1.40400875e+00 1.00309923e-01 -4.63818729e-01
-2.53605068e-01 1.26085138e+00 8.99089158e-01 8.16126987e-02
-8.12857091e-01 -3.07629794e-01 8.01615655e-01 -1.11804776e-01
-2.32664168e-01 -4.11300659e-01 -5.97920656e-01 -1.27964187e+00
-4.00705218e-01 -1.15321565e+00 6.46402597e-01 -6.67040467e-01
-1.65967178e+00 3.42353135e-01 -3.53149533e-01 -4.21871722e-01
-7.02246368e-01 -1.09444809e+00 -4.32438016e-01 1.25715113e+00
-1.06763589e+00 -1.72113442e+00 -1.84806466e-01 4.72375065e-01
1.67755142e-01 -6.17595255e-01 9.69373584e-01 3.97644669e-01
-9.97852743e-01 1.39375377e+00 1.23744361e-01 8.76199305e-01
1.62501824e+00 -1.65345919e+00 5.77223003e-01 7.98604250e-01
-2.32157677e-01 5.96757650e-01 3.83728266e-01 -5.12399912e-01
-8.50080907e-01 -7.22409189e-01 1.42543054e+00 -1.19390357e+00
7.73034394e-01 -1.14599860e+00 -8.05504441e-01 1.16430724e+00
3.25260401e-01 -3.99124593e-01 9.21973765e-01 1.02735507e+00
-1.04557443e+00 1.81965847e-02 -1.22363150e+00 4.82985526e-01
1.09478879e+00 -8.24052930e-01 -2.65462786e-01 1.07915156e-01
1.72499612e-01 -5.45811534e-01 -9.98308361e-01 -9.67886765e-03
1.14472198e+00 -9.83440399e-01 6.49448872e-01 -8.11393678e-01
9.20533240e-01 5.68769693e-01 -1.83901817e-01 -1.64224768e+00
-3.53488088e-01 -3.16748053e-01 -2.99259394e-01 1.62966132e+00
5.56688845e-01 -7.13313997e-01 5.46691000e-01 7.37607002e-01
3.03127002e-02 -5.64299464e-01 -7.83781707e-01 -4.39585418e-01
8.32188785e-01 -1.25286773e-01 8.84192109e-01 1.31141746e+00
1.01594225e-01 5.50637901e-01 -5.72092235e-01 9.74507332e-02
5.55140436e-01 -3.04715633e-01 1.03136790e+00 -1.33168077e+00
-1.33663282e-01 -3.75025272e-01 -1.31135136e-01 -4.02401447e-01
5.62780976e-01 -9.98413622e-01 -4.60134178e-01 -1.01336908e+00
3.57735306e-01 -3.82416010e-01 -8.79701376e-02 6.73435211e-01
-3.56179774e-01 5.63543737e-01 1.14009447e-01 1.05309926e-01
-4.51667100e-01 1.83426276e-01 6.94395363e-01 -2.09298387e-01
1.08617783e-01 -2.80351967e-01 -1.13099754e+00 4.73640591e-01
1.11951840e+00 -1.73826694e-01 2.14185610e-01 -6.39652550e-01
5.14909565e-01 -4.22155231e-01 4.45687562e-01 -7.59913683e-01
-2.62110472e-01 3.57255489e-01 1.01054752e+00 -4.14674848e-01
4.67964917e-01 -1.36753604e-01 -6.45814896e-01 3.84652406e-01
-3.63043159e-01 2.52697617e-01 6.22800514e-02 -1.74188495e-01
2.10423619e-01 -6.10814476e-03 7.31307685e-01 1.81805626e-01
-3.27607661e-01 1.10871069e-01 -3.60390842e-01 1.66960850e-01
3.32111508e-01 3.71083021e-02 -7.12106764e-01 -4.59648728e-01
-8.54776919e-01 3.78902048e-01 5.36377311e-01 7.16918826e-01
5.74940853e-02 -1.19808722e+00 -1.26923525e+00 5.21644473e-01
1.29499525e-01 -5.50622106e-01 -9.07565802e-02 5.37805200e-01
-2.15582013e-01 1.20932326e-01 -6.30536258e-01 -3.50076258e-01
-1.56438625e+00 1.44765019e-01 5.28544724e-01 -2.90296882e-01
2.41121396e-01 1.02520132e+00 -3.20332676e-01 -9.09212947e-01
-6.84675649e-02 2.35309348e-01 -1.10159568e-01 4.70007896e-01
5.79747319e-01 8.06235969e-01 2.78448462e-01 -8.22487473e-01
-5.64113438e-01 -3.52380797e-02 -5.55970967e-01 -4.84373003e-01
1.11354744e+00 -1.13146044e-01 -2.53383100e-01 4.89719242e-01
1.16999900e+00 5.66074669e-01 -8.97007406e-01 -1.15940526e-01
-1.10355265e-01 -3.18845004e-01 -2.45739788e-01 -1.30785251e+00
-5.74383438e-01 9.10834014e-01 4.10940856e-01 9.22759771e-02
4.12106544e-01 -2.05649123e-01 4.56794471e-01 -7.13694021e-02
1.96636856e-01 -1.30699074e+00 7.57924020e-02 1.04629183e+00
8.75059664e-01 -1.33001184e+00 6.01650290e-02 -2.46948794e-01
-5.82602561e-01 8.41707230e-01 9.57792938e-01 2.69569270e-02
4.37568814e-01 9.71111804e-02 2.14321852e-01 5.20456806e-02
-5.25495768e-01 9.90727097e-02 2.60413617e-01 6.43844485e-01
7.39156842e-01 3.29917461e-01 -2.76751101e-01 1.30375147e+00
-1.02106011e+00 -4.02779222e-01 6.52805746e-01 5.06802797e-01
4.82334988e-03 -1.06906915e+00 -4.10695702e-01 6.35581791e-01
-7.82884717e-01 -1.38743833e-01 -1.07363200e+00 7.37932265e-01
2.60288626e-01 8.31011534e-01 3.62301528e-01 -7.52023518e-01
-1.81230739e-01 7.96535075e-01 6.91405773e-01 -3.57179582e-01
-9.42709923e-01 -6.38524592e-01 8.53963315e-01 -1.47775874e-01
3.67520899e-01 -9.04098809e-01 -5.36142468e-01 -8.87746274e-01
8.56589992e-03 -1.81432962e-01 6.09061718e-01 7.34180570e-01
4.42435682e-01 1.38611302e-01 6.23622119e-01 -6.74480438e-01
-6.19718015e-01 -1.54350889e+00 -7.71857321e-01 4.80418772e-01
4.44424421e-01 -4.36170489e-01 -4.01464731e-01 -3.01554769e-01] | [9.318061828613281, 10.499059677124023] |
e4e4d1ac-d38e-4350-9c43-ec5026b3b327 | legal-case-document-summarization-extractive | 2210.07544 | null | https://arxiv.org/abs/2210.07544v1 | https://arxiv.org/pdf/2210.07544v1.pdf | Legal Case Document Summarization: Extractive and Abstractive Methods and their Evaluation | Summarization of legal case judgement documents is a challenging problem in Legal NLP. However, not much analyses exist on how different families of summarization models (e.g., extractive vs. abstractive) perform when applied to legal case documents. This question is particularly important since many recent transformer-based abstractive summarization models have restrictions on the number of input tokens, and legal documents are known to be very long. Also, it is an open question on how best to evaluate legal case document summarization systems. In this paper, we carry out extensive experiments with several extractive and abstractive summarization methods (both supervised and unsupervised) over three legal summarization datasets that we have developed. Our analyses, that includes evaluation by law practitioners, lead to several interesting insights on legal summarization in specific and long document summarization in general. | ['Saptarshi Ghosh', 'Pawan Goyal', 'Kripabandhu Ghosh', 'Rajdeep Mukherjee', 'Soham Poddar', 'Paheli Bhattacharya', 'Abhay Shukla'] | 2022-10-14 | null | null | null | null | ['abstractive-text-summarization', 'document-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 5.26549339e-01 2.75066704e-01 -6.61226749e-01 -1.95765078e-01
-1.18748939e+00 -9.70586479e-01 6.91743970e-01 7.13345647e-01
-3.02204162e-01 1.24819291e+00 1.02816474e+00 -5.47842741e-01
-5.88065922e-01 -4.96049762e-01 -2.07598671e-01 -2.64016896e-01
1.82444349e-01 7.34251678e-01 2.11521521e-01 -3.67605835e-01
9.38290238e-01 6.38385952e-01 -1.02647257e+00 5.18069446e-01
1.70787096e+00 8.02925602e-02 -3.77313823e-01 7.98356712e-01
-3.73632431e-01 1.13004553e+00 -1.17144144e+00 -8.54407012e-01
7.67679838e-03 -5.20120203e-01 -1.27852166e+00 -5.89130856e-02
7.52512097e-01 -4.73937243e-01 -1.88126653e-01 9.61103022e-01
6.33520603e-01 -9.69640166e-02 1.07677341e+00 -9.26648796e-01
-7.10499763e-01 1.04015338e+00 -6.57684982e-01 8.65965188e-01
7.97082067e-01 -1.91985905e-01 1.32386792e+00 8.82392451e-02
1.06487429e+00 1.21948540e+00 3.79324168e-01 4.38492924e-01
-1.03237450e+00 -2.16618910e-01 2.81230593e-03 4.26497042e-01
-7.48356104e-01 -7.11537480e-01 7.04039574e-01 -4.47345763e-01
1.14714003e+00 4.72762972e-01 5.96159458e-01 7.96579123e-01
5.93454659e-01 1.14582396e+00 8.19168270e-01 -5.02171636e-01
2.57538110e-01 -6.24698810e-02 7.11725891e-01 7.62089044e-02
1.14857984e+00 -5.14068186e-01 -1.88875556e-01 -8.14218640e-01
4.11730930e-02 -4.09399420e-01 -2.42791280e-01 2.38007963e-01
-5.56659698e-01 1.10277903e+00 -4.65356618e-01 6.36558354e-01
-4.71413314e-01 6.94371909e-02 8.76530290e-01 3.28703672e-01
4.92765814e-01 6.54404938e-01 -6.54816180e-02 -4.84702319e-01
-1.33141077e+00 8.97515833e-01 1.18410349e+00 8.43469501e-01
2.05546066e-01 -2.06267595e-01 -6.18381321e-01 8.87644410e-01
-4.37898338e-02 6.42584085e-01 2.19503641e-01 -1.11811936e+00
1.12521374e+00 5.26412547e-01 -8.12573135e-02 -8.36745083e-01
-1.17273718e-01 -7.16510937e-02 -6.35591269e-01 -2.18432486e-01
1.20730758e-01 -3.11143875e-01 -4.86480981e-01 1.11015475e+00
-2.63700008e-01 -5.53868175e-01 5.45410156e-01 1.53225765e-01
1.04140186e+00 6.33798003e-01 7.10535896e-05 -6.69347823e-01
1.29568362e+00 -5.02574980e-01 -9.02997732e-01 -2.57912606e-01
8.28262508e-01 -8.94788682e-01 4.38451797e-01 1.80693626e-01
-1.35686147e+00 6.45357594e-02 -7.26107657e-01 -2.85175353e-01
-5.03593162e-02 -4.11849506e-02 6.15670741e-01 7.86049247e-01
-7.23956466e-01 5.41935325e-01 -5.15638471e-01 -7.30749905e-01
7.12711394e-01 -4.03979747e-03 -2.13434041e-01 -1.68520629e-01
-1.19297791e+00 1.13291180e+00 5.89250505e-01 -2.84455270e-01
1.21271554e-02 -5.27999640e-01 -8.06448221e-01 2.24200845e-01
6.54197633e-01 -7.31625617e-01 1.57807887e+00 -1.54940099e-01
-8.83978844e-01 6.76526546e-01 -4.49992239e-01 -8.03741038e-01
4.45837736e-01 -1.90591589e-01 -3.02489489e-01 5.09756744e-01
7.06320941e-01 2.36684717e-02 2.28458539e-01 -1.05220902e+00
-8.46602678e-01 -3.12861502e-01 3.65080804e-01 1.24400668e-01
-1.52436584e-01 5.62295675e-01 -1.95860993e-02 -5.68993151e-01
-4.37997431e-01 -5.77247202e-01 -3.48609626e-01 -7.98915446e-01
-7.78110564e-01 -5.69824696e-01 4.70165700e-01 -8.70349467e-01
1.76362133e+00 -1.64264429e+00 -8.32439139e-02 1.46584839e-01
7.53158107e-02 4.40626860e-01 1.10493042e-01 1.28661847e+00
6.18229844e-02 6.26326025e-01 -4.61161375e-01 7.00967759e-02
2.17247412e-01 3.87628168e-01 -9.08369303e-01 1.81331038e-01
-7.18340799e-02 9.29791927e-01 -9.01010394e-01 -1.16204464e+00
-1.11415774e-01 -2.20507562e-01 -6.71569228e-01 -4.17109996e-01
-1.08351886e-01 1.76060542e-01 -9.87192929e-01 6.33649528e-01
4.51897025e-01 2.56871253e-01 1.41805544e-01 9.20633823e-02
-2.90479273e-01 6.28234923e-01 -6.00914061e-01 1.49771333e+00
-1.20172184e-03 7.62017727e-01 -4.32643354e-01 -9.67650652e-01
3.49773437e-01 4.79847342e-01 4.90905225e-01 -5.73645532e-01
7.35582188e-02 2.82926291e-01 6.49629682e-02 -9.40426767e-01
1.08548021e+00 -2.16521278e-01 -5.47591627e-01 8.44387114e-01
-3.10873657e-01 -3.91320437e-01 1.56636584e+00 7.84805119e-01
1.28207111e+00 -3.69853228e-01 8.44700038e-01 -1.16921432e-01
4.58944112e-01 4.86839384e-01 5.74810922e-01 1.04542220e+00
7.71558136e-02 5.82371235e-01 1.37765968e+00 4.97975796e-02
-7.91356802e-01 -7.35087931e-01 -2.67144322e-01 4.35674310e-01
-1.98181421e-01 -7.92630374e-01 -8.98288190e-01 -7.80127585e-01
1.43315047e-01 1.15962589e+00 -3.74974042e-01 -1.21003307e-01
-9.58584428e-01 -7.29730725e-01 8.63016188e-01 5.54164946e-01
1.93727091e-01 -1.16329968e+00 -7.47024953e-01 4.41918582e-01
-3.16642672e-01 -1.12163031e+00 -3.59387368e-01 -2.20949396e-01
-9.28823411e-01 -1.36157548e+00 -4.80874270e-01 -2.90310651e-01
4.20577019e-01 -1.93731207e-02 8.08408439e-01 6.04311377e-02
7.26325577e-03 4.44547683e-01 -7.46785343e-01 -7.93497205e-01
-9.21151638e-01 6.47520840e-01 -2.55065680e-01 -5.71752489e-01
3.77400160e-01 -5.97213149e-01 -5.05160615e-02 -5.07344246e-01
-1.30533338e+00 -4.00456548e-01 7.63536751e-01 2.42410615e-01
3.72382365e-02 7.10655451e-02 9.12151754e-01 -1.47564745e+00
1.74256861e+00 -3.72530818e-01 -2.65926927e-01 7.86188185e-01
-4.93122905e-01 2.16976807e-01 7.12148726e-01 1.60465196e-01
-1.20189166e+00 -6.87374055e-01 -4.10189927e-01 4.06332999e-01
-2.15477750e-01 8.53472173e-01 1.27517417e-01 6.03955626e-01
7.32597411e-01 1.53321087e-01 -1.95288718e-01 -3.74953151e-01
5.30186236e-01 9.05720174e-01 4.50569332e-01 -6.62909508e-01
8.19646239e-01 4.33088273e-01 -2.81917214e-01 -1.13042438e+00
-9.93328154e-01 -5.91434062e-01 -6.79288685e-01 -1.21314235e-01
8.85513484e-01 -1.84926271e-01 -2.17361003e-01 -7.64424447e-03
-1.48792803e+00 1.71279266e-01 -7.01579690e-01 2.49310210e-01
-4.92848068e-01 1.16901004e+00 -5.21469891e-01 -7.61111319e-01
-8.53105485e-01 -8.52122903e-01 8.95989001e-01 3.55269045e-01
-8.73122752e-01 -9.16382909e-01 4.59796995e-01 7.75436044e-01
1.64168194e-01 3.54278058e-01 1.21183276e+00 -1.06724691e+00
-6.56814352e-02 -6.56777322e-01 -1.57817882e-02 3.08484107e-01
2.29362205e-01 3.72218996e-01 -4.09103841e-01 -3.63408066e-02
-1.18940040e-01 -1.32585347e-01 1.12047386e+00 6.86033130e-01
3.60122263e-01 -6.83802724e-01 -4.02016580e-01 -2.09933236e-01
9.65867877e-01 3.46257687e-01 7.93151140e-01 2.33296409e-01
2.19558820e-01 6.54789746e-01 7.24427402e-01 3.72715265e-01
2.49359980e-01 3.71362031e-01 -1.90416351e-01 4.05662656e-01
1.04897534e-02 -4.59251739e-02 2.65943706e-01 7.74033844e-01
-4.29922134e-01 -5.82979739e-01 -7.92647243e-01 6.62068367e-01
-1.95957947e+00 -1.77933836e+00 -3.11008185e-01 1.63585579e+00
7.92482436e-01 3.05849344e-01 1.87060311e-01 3.57324034e-01
5.52123070e-01 5.49277782e-01 -3.78420621e-01 -9.93040562e-01
-2.25928456e-01 2.72472680e-01 3.48781377e-01 3.55587751e-01
-8.65120888e-01 8.78646314e-01 5.96551323e+00 1.04258800e+00
-5.47908545e-01 -8.41725916e-02 1.80842400e-01 -1.52790010e-01
-5.42975247e-01 3.39335203e-01 -8.85533333e-01 3.42220277e-01
7.46736765e-01 -9.75299418e-01 -2.64939010e-01 6.82692707e-01
3.57889652e-01 -4.36741173e-01 -9.45425868e-01 5.58419228e-01
4.65209782e-01 -1.55552351e+00 4.79236513e-01 3.62742901e-01
9.39240098e-01 -2.62690336e-01 -4.12579745e-01 3.90893400e-01
2.73489207e-01 -7.17923880e-01 5.72310865e-01 4.08945888e-01
4.30278420e-01 -7.23678589e-01 1.03327298e+00 5.11803627e-01
-7.35895991e-01 -2.07883030e-01 -5.56997180e-01 1.62089899e-01
9.31819201e-01 6.37499571e-01 -5.44953883e-01 1.09889519e+00
1.52954057e-01 8.18243682e-01 -5.28229475e-01 1.28762567e+00
-4.55100834e-01 7.59163260e-01 3.55005846e-03 -1.42968550e-01
3.62591982e-01 -2.66156346e-01 9.23766732e-01 1.22879350e+00
1.17944717e-01 3.66066635e-01 -6.17232993e-02 4.23490793e-01
-7.64699727e-02 3.49694788e-01 -8.01415443e-01 -1.51719496e-01
2.57076383e-01 8.97839904e-01 -6.45756423e-01 -6.19355381e-01
-1.92343608e-01 4.48006392e-01 1.10035082e-02 3.05776715e-01
-5.25454700e-01 -6.46427214e-01 2.15401709e-01 3.74897718e-01
1.94322318e-01 3.33234556e-02 -2.34881532e-03 -1.23651278e+00
2.49185741e-01 -8.51490259e-01 7.99809813e-01 -5.25439620e-01
-1.13560736e+00 4.15298373e-01 6.98548257e-01 -1.06261194e+00
-6.43594861e-01 -9.85924453e-02 -1.08938766e+00 2.43800566e-01
-1.56377852e+00 -6.81427002e-01 4.92999583e-01 2.13697236e-02
7.09442675e-01 -1.12617694e-01 3.48094910e-01 2.24575192e-01
-5.41623771e-01 2.40915030e-01 1.59586966e-01 9.83577520e-02
5.25753081e-01 -1.20858753e+00 1.45483330e-01 8.89162719e-01
7.67319947e-02 8.72302294e-01 9.27689731e-01 -9.70886171e-01
-9.28314745e-01 -8.32848668e-01 1.45620918e+00 -6.14671767e-01
5.45715749e-01 2.35365227e-01 -8.88539672e-01 7.68960536e-01
7.15165794e-01 -9.44321752e-01 8.70251656e-01 5.69417849e-02
1.77695509e-02 -4.13851999e-02 -1.03067076e+00 6.86688423e-01
1.01526105e+00 -1.84120983e-01 -1.48929787e+00 6.41630888e-01
3.64088267e-01 -1.17477946e-01 -8.20017636e-01 2.47031897e-01
4.74924624e-01 -9.29805815e-01 6.24424100e-01 -8.16695631e-01
7.20430732e-01 2.39447486e-02 3.17681372e-01 -1.06173718e+00
6.89425878e-03 -5.38778126e-01 1.77913859e-01 1.59696162e+00
6.65737391e-01 -9.19631958e-01 4.98625338e-01 5.87975860e-01
-2.78281093e-01 -8.45025897e-01 -9.24796999e-01 -8.91967356e-01
4.94148433e-01 -3.46590847e-01 4.08404857e-01 7.14431047e-01
5.60228288e-01 6.56297565e-01 -1.31245613e-01 -3.86769801e-01
5.20080388e-01 5.24739504e-01 7.42097616e-01 -1.48567808e+00
2.02848330e-01 -6.88804567e-01 -3.68877441e-01 -7.88552582e-01
5.48524022e-01 -1.06859469e+00 -4.58592713e-01 -2.40884662e+00
6.16085410e-01 2.36402959e-01 3.82997245e-01 5.20662725e-01
-1.42190173e-01 -4.09786493e-01 9.06197876e-02 5.20871699e-01
-7.48542964e-01 3.73566866e-01 1.03497899e+00 -4.37607348e-01
-3.20960820e-01 2.99661398e-01 -1.21816754e+00 7.84761846e-01
1.09254146e+00 -7.81644404e-01 -4.34811175e-01 -3.27827841e-01
4.35627609e-01 2.97568202e-01 -2.30602115e-01 -7.37788081e-01
6.01179898e-01 -3.26174468e-01 -3.79354209e-01 -9.64154184e-01
-1.82757437e-01 -1.63282290e-01 -2.62226462e-01 4.23806757e-01
-5.48835516e-01 -2.21142378e-02 8.49796981e-02 4.04250622e-01
-4.83558357e-01 -1.03947628e+00 3.96091908e-01 -1.42818823e-01
-1.77096784e-01 -3.95075381e-02 -9.35624242e-01 6.68773353e-01
8.94933939e-01 -5.07209003e-01 -7.66986907e-01 -4.56062853e-01
-3.68404657e-01 5.21194875e-01 3.35525215e-01 8.18840712e-02
3.97820503e-01 -8.61028612e-01 -1.35036099e+00 -7.05687582e-01
5.08958474e-02 -1.25343278e-01 3.14854324e-01 9.54123557e-01
-7.55866945e-01 9.79935825e-01 1.07083455e-01 -9.27461609e-02
-1.51747060e+00 2.87720025e-01 -2.31776655e-01 -9.03022230e-01
-7.89884865e-01 2.32095718e-02 -4.28819954e-01 -7.03562573e-02
-2.42767915e-01 -4.60804999e-01 -7.64356554e-01 6.47479057e-01
4.54494685e-01 6.59535348e-01 -1.47345075e-02 -7.55216837e-01
-2.74164081e-01 8.05985630e-01 -5.85022688e-01 -1.53667852e-01
1.49371624e+00 3.11375588e-01 -6.56995356e-01 9.20329094e-02
9.30373967e-01 5.21260977e-01 -9.00669694e-02 7.10713118e-02
4.69692707e-01 -9.23032835e-02 -5.29563427e-01 -4.92415756e-01
-6.92775667e-01 4.51005816e-01 -6.13384008e-01 7.29101479e-01
8.39288592e-01 3.11989903e-01 7.82736659e-01 8.78261507e-01
1.29656019e-02 -1.33246934e+00 -4.95638192e-01 5.10122895e-01
1.27785993e+00 -6.33216739e-01 7.00873315e-01 -3.78009260e-01
-7.48576880e-01 1.06312931e+00 -3.79171409e-02 -9.58019774e-03
1.70588955e-01 -3.07817850e-02 -6.12627901e-02 -6.60359740e-01
-7.36243486e-01 -1.65607736e-01 1.29868045e-01 2.21411794e-01
2.88848609e-01 -4.20237869e-01 -1.26276731e+00 2.78668880e-01
-6.43461645e-01 4.15861085e-02 1.18137455e+00 1.26467395e+00
-5.44092417e-01 -1.21082628e+00 -4.71658677e-01 1.13123286e+00
-9.10100639e-01 -1.68600380e-01 -8.96947563e-01 1.01787531e+00
-2.12889522e-01 1.37665546e+00 -3.52102280e-01 4.52319652e-01
7.19874263e-01 2.78124344e-02 4.46669251e-01 -1.09597635e+00
-8.33116114e-01 -1.83365345e-01 7.69470930e-01 -1.38149951e-02
-6.30167782e-01 -1.21483505e+00 -1.38506794e+00 -2.73703277e-01
-5.81333339e-01 9.20117557e-01 1.89465284e-01 1.16272855e+00
5.12714237e-02 3.72410834e-01 6.39463887e-02 -2.44741797e-01
-6.87551022e-01 -1.08260489e+00 -7.34045863e-01 2.44701713e-01
1.87438548e-01 -7.66730979e-02 -4.16761965e-01 -2.55891550e-02] | [12.125465393066406, 9.586036682128906] |
f028722f-f60f-441b-b673-9f893910c223 | using-known-information-to-accelerate | 1811.03322 | null | https://arxiv.org/abs/1811.03322v2 | https://arxiv.org/pdf/1811.03322v2.pdf | Using Known Information to Accelerate HyperParameters Optimization Based on SMBO | Automl is the key technology for machine learning problem. Current state of art hyperparameter optimization methods are based on traditional black-box optimization methods like SMBO (SMAC, TPE). The objective function of black-box optimization is non-smooth, or time-consuming to evaluate, or in some way noisy. Recent years, many researchers offered the work about the properties of hyperparameters. However, traditional hyperparameter optimization methods do not take those information into consideration. In this paper, we use gradient information and machine learning model analysis information to accelerate traditional hyperparameter optimization methods SMBO. In our L2 norm experiments, our method yielded state-of-the-art performance, and in many cases outperformed the previous best configuration approach. | ['Zhang Yunquan', 'Xia Fen', 'Li Shigang', 'Cheng Daning', 'Zhang Hanping'] | 2018-11-08 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-5.98744571e-01 -3.77725601e-01 -4.47422326e-01 -7.91586936e-02
-6.92454278e-01 -3.45096022e-01 2.01641902e-01 1.47606153e-02
-5.58636844e-01 1.01592982e+00 -1.39762670e-01 -2.59840310e-01
-4.02382791e-01 -5.21186650e-01 -6.66359365e-01 -1.03807223e+00
1.87880158e-01 7.33161986e-01 6.91105872e-02 -4.53796834e-01
4.55731481e-01 3.74448419e-01 -1.22227585e+00 -2.19967365e-01
9.72806036e-01 1.00630164e+00 -1.91780716e-01 5.62185585e-01
-1.94088310e-01 3.14643519e-04 -5.31535625e-01 -3.11874986e-01
3.89200509e-01 -2.87491888e-01 -7.55500674e-01 -2.50364870e-01
6.09137192e-02 2.80880779e-01 -2.85636317e-02 1.09389472e+00
7.67691255e-01 4.21142012e-01 4.98455405e-01 -1.24979377e+00
-2.77774453e-01 6.07345700e-01 -5.68104506e-01 2.17246309e-01
1.63413271e-01 4.93187338e-01 8.91704082e-01 -8.12454522e-01
3.29166561e-01 1.34650064e+00 7.99428165e-01 3.89852285e-01
-1.14062345e+00 -5.20877421e-01 2.14162901e-01 4.84107763e-01
-1.45373309e+00 -8.69708657e-02 6.96898639e-01 -3.02337170e-01
8.42492580e-01 3.96020025e-01 9.58106697e-01 7.38049150e-01
1.71386912e-01 8.57988656e-01 1.14496708e+00 -4.80145097e-01
3.26951444e-01 4.02881294e-01 3.36834967e-01 7.88890362e-01
3.32006842e-01 1.72958151e-01 -3.51301134e-01 -4.93812799e-01
6.11639500e-01 -4.52642351e-01 -2.07892343e-01 -3.24572951e-01
-1.20861149e+00 1.08297169e+00 1.65477499e-01 1.83470845e-01
-2.17863888e-01 1.98157609e-01 2.84980685e-01 1.22317187e-01
4.49468404e-01 8.92882109e-01 -6.37425125e-01 -3.89861971e-01
-6.17680669e-01 4.44969088e-01 9.71510053e-01 7.52036273e-01
7.36278713e-01 7.32765358e-04 -1.80098578e-01 9.04626966e-01
4.72663730e-01 4.33799297e-01 6.92967236e-01 -7.09847808e-01
4.34567600e-01 5.97492337e-01 2.86991119e-01 -7.97799647e-01
-8.64929020e-01 -5.07625997e-01 -7.78742313e-01 8.64522681e-02
5.55653811e-01 -3.47378701e-01 -7.87162006e-01 1.23298192e+00
7.21243560e-01 1.27032161e-01 -2.21160665e-01 1.14346468e+00
6.66718304e-01 7.63400912e-01 -1.01951122e-01 -3.25319648e-01
1.02728546e+00 -1.36235189e+00 -8.17231178e-01 3.80873382e-02
6.73924208e-01 -9.65415537e-01 1.38634336e+00 5.71542740e-01
-1.36229277e+00 -1.97160274e-01 -1.15083671e+00 3.12041551e-01
-3.62174928e-01 -1.36681661e-01 8.11591625e-01 8.23170483e-01
-6.48050606e-01 7.41510093e-01 -8.14169407e-01 -1.31399650e-02
1.77584305e-01 5.25976777e-01 3.45794596e-02 3.68707865e-01
-1.15605843e+00 1.00566888e+00 6.43282115e-01 2.02341035e-01
-5.90743721e-01 -1.03986609e+00 -6.03395462e-01 5.70258982e-02
7.63339102e-01 -8.71580124e-01 1.28459477e+00 -6.48536742e-01
-2.28215909e+00 3.09493095e-01 -8.84186998e-02 -2.26448029e-01
6.87431753e-01 -4.40827310e-01 -2.60070860e-01 -2.68345386e-01
-6.89273655e-01 3.50296319e-01 8.18971872e-01 -1.02113783e+00
-3.33816439e-01 -1.16040081e-01 -3.40619475e-01 2.61367887e-01
-2.29362085e-01 1.05128802e-01 -6.52844250e-01 -6.03519976e-01
6.00504130e-02 -1.07393420e+00 -5.00580728e-01 -3.90439391e-01
-5.94014764e-01 -3.45973760e-01 6.25323057e-01 -3.18224996e-01
1.67680919e+00 -1.71333921e+00 1.81230471e-01 5.56168616e-01
-1.40463650e-01 5.87848902e-01 9.63734090e-02 3.67826313e-01
-2.43269056e-02 5.00649512e-01 -1.49615243e-01 -3.58991958e-02
1.47605240e-01 -2.01817632e-01 -9.74244326e-02 7.25973845e-01
-2.25633264e-01 9.06091392e-01 -7.07804143e-01 -6.71684146e-01
2.54078925e-01 3.71383369e-01 -7.94583440e-01 1.15804419e-01
-4.08111364e-01 4.87271696e-01 -5.96341908e-01 6.81029022e-01
5.19871891e-01 -3.50014180e-01 -1.98094323e-01 -2.94027239e-01
2.79794484e-02 -4.64036986e-02 -1.64555156e+00 1.38990343e+00
-4.48196054e-01 5.97481072e-01 -1.60222709e-01 -9.20623302e-01
7.96260417e-01 2.83170849e-01 6.79060102e-01 -2.95332283e-01
2.83254892e-01 2.96099097e-01 -1.08395673e-01 -6.63192391e-01
1.81619763e-01 1.01551503e-01 2.20374584e-01 1.59514770e-01
-2.43411511e-01 -3.84891897e-01 3.10912490e-01 -4.30532843e-01
5.51838160e-01 4.87733223e-02 5.78749299e-01 -3.93844515e-01
7.63338208e-01 2.57644951e-01 5.98544002e-01 7.20142245e-01
-3.31324898e-02 5.58463573e-01 4.04083073e-01 -5.93486905e-01
-1.14842379e+00 -4.66897964e-01 -4.36699480e-01 8.89613092e-01
3.09087545e-01 -5.11798084e-01 -9.52636898e-01 -4.66988415e-01
1.39762834e-01 5.79443216e-01 -3.39117795e-01 -1.11304946e-01
-9.53862667e-01 -1.38864803e+00 3.85565966e-01 1.12452410e-01
5.23805201e-01 -8.54150474e-01 -2.86698431e-01 4.35859889e-01
7.72921965e-02 -8.82775486e-01 -6.70693040e-01 -2.49501273e-01
-1.28130019e+00 -9.45047498e-01 -7.80549228e-01 -2.65173554e-01
4.93173957e-01 -2.48220444e-01 1.04811716e+00 2.61991590e-01
-4.91228044e-01 -8.30132812e-02 -1.64592430e-01 -4.42913860e-01
-8.26530084e-02 3.16273630e-01 1.85184330e-01 -1.65362004e-02
1.51572451e-01 -1.87031895e-01 -7.60021329e-01 9.24196422e-01
-5.97799659e-01 -1.46913618e-01 5.52166700e-01 1.08668602e+00
8.00638378e-01 1.39586786e-02 4.83444124e-01 -7.84314513e-01
8.54478717e-01 -3.74623388e-01 -1.13349926e+00 4.07434613e-01
-1.09813213e+00 4.60660845e-01 5.61437249e-01 -7.81042874e-01
-9.64510977e-01 -1.12769090e-01 -4.49226834e-02 -3.55814308e-01
2.64286608e-01 5.25920212e-01 -2.14158237e-01 -3.94552916e-01
6.54037297e-01 -2.54859447e-01 -7.56858364e-02 -6.35172606e-01
9.08107013e-02 5.55528820e-01 5.39135821e-02 -6.86140954e-01
5.89743137e-01 2.52096742e-01 2.99825102e-01 -7.06909597e-01
-8.43304396e-01 -6.44540370e-01 -3.22762400e-01 -1.04048982e-01
6.38154328e-01 -2.52037197e-01 -1.14768136e+00 4.69828665e-01
-7.78793275e-01 -3.36472601e-01 2.20713057e-02 7.59470463e-01
-6.07225895e-01 2.55100042e-01 -1.38813540e-01 -7.22834766e-01
-4.88343686e-01 -1.42724979e+00 7.47701049e-01 5.42275906e-01
4.65867389e-03 -1.22142553e+00 2.97335505e-01 5.47060013e-01
6.26443267e-01 6.97659701e-02 8.58276188e-01 -5.41685283e-01
-3.68040353e-01 -3.26141417e-01 7.23029077e-02 1.80618297e-02
-3.77726108e-01 2.66384840e-01 -6.83331430e-01 -3.67811948e-01
-2.49944106e-02 -4.63073291e-02 6.47311389e-01 7.11762667e-01
1.57134867e+00 -3.67534071e-01 -4.45296645e-01 1.23539507e+00
1.54421389e+00 1.05386533e-01 5.06408036e-01 8.18055987e-01
4.73811001e-01 2.39349216e-01 7.83446670e-01 6.53719723e-01
-1.51562812e-02 9.86746430e-01 2.20526218e-01 -1.73541456e-01
3.44205230e-01 -3.28953043e-02 7.22131208e-02 5.99323153e-01
-1.48738742e-01 -2.52056926e-01 -1.10264587e+00 2.14159459e-01
-2.26911497e+00 -5.10096014e-01 -1.89879671e-01 2.21382189e+00
1.01547492e+00 1.32734492e-01 2.42909834e-01 -7.75891319e-02
8.04480314e-01 -1.75140664e-01 -7.17698336e-01 -4.68867540e-01
-2.06150010e-01 -1.65099069e-01 9.13530648e-01 5.19803882e-01
-1.14697409e+00 1.14423966e+00 7.25493431e+00 1.11026609e+00
-9.52337563e-01 3.14549543e-02 4.83589470e-01 -5.51493466e-01
-1.63589463e-01 8.21418911e-02 -1.20028245e+00 5.79482555e-01
7.12075114e-01 -2.70518601e-01 7.69091010e-01 9.37275708e-01
3.41480404e-01 -3.15169543e-01 -8.00255120e-01 1.27584898e+00
-1.79420769e-01 -1.33125567e+00 -4.43617672e-01 -4.70849909e-02
1.00937045e+00 -1.49285316e-01 -4.55599912e-02 3.55299830e-01
2.61249930e-01 -1.36889088e+00 2.18362436e-01 6.00772262e-01
3.00062537e-01 -8.69435132e-01 9.42506790e-01 3.27339739e-01
-8.27801228e-01 -1.39327198e-01 -5.39366066e-01 3.59078854e-01
3.51225823e-01 8.05871367e-01 -6.58549607e-01 2.61862040e-01
8.26985121e-01 1.81717843e-01 -4.47859377e-01 1.84690011e+00
-2.29234040e-01 1.02254605e+00 -6.30920947e-01 -6.08087659e-01
4.37377185e-01 -5.64935148e-01 9.96337593e-01 9.90364552e-01
6.10557720e-02 -9.13464427e-02 1.66932032e-01 6.78795874e-01
1.74930096e-01 4.46573317e-01 -4.41539846e-02 5.63251078e-02
4.42619532e-01 1.11945105e+00 -4.14777279e-01 -1.94279224e-01
-1.02578789e-01 2.67936349e-01 1.98331639e-01 4.79191631e-01
-9.66803730e-01 -5.64634740e-01 6.25831008e-01 -1.82627201e-01
9.97869223e-02 -1.79105476e-01 -6.44677162e-01 -1.02014565e+00
1.20604243e-02 -9.58316624e-01 5.47257900e-01 -4.14216459e-01
-1.10117674e+00 2.43389904e-01 3.56490850e-01 -9.77893054e-01
-1.00175843e-01 -7.18826175e-01 -7.63828039e-01 7.19997585e-01
-1.39534879e+00 -5.36593616e-01 -2.35914633e-01 3.88020396e-01
4.85360503e-01 -4.56226856e-01 4.17160749e-01 1.63823813e-01
-1.03526175e+00 8.89530361e-01 6.54924691e-01 -4.07762140e-01
7.67933846e-01 -1.08503902e+00 1.08818382e-01 3.73977274e-01
-2.70092875e-01 5.61668813e-01 9.16555285e-01 -4.91632849e-01
-1.69156528e+00 -5.16471744e-01 3.33085507e-01 -1.23526074e-01
7.17683792e-01 -3.68863009e-02 -9.88791645e-01 2.09756359e-01
1.50323994e-02 5.77228004e-03 5.02900243e-01 3.02371114e-01
1.59115627e-01 -1.72257468e-01 -1.08331203e+00 7.39583850e-01
6.09876513e-01 2.18093544e-01 -3.67875576e-01 5.17536104e-01
6.52047336e-01 -5.48063338e-01 -1.02089775e+00 5.16725063e-01
5.29721498e-01 -5.15382349e-01 1.18787324e+00 -9.69487548e-01
-1.56150714e-01 -1.31300047e-01 3.18691462e-01 -1.49955976e+00
-4.74142544e-02 -1.25170255e+00 -3.63143951e-01 8.56908500e-01
6.40560567e-01 -8.57604563e-01 9.13691461e-01 7.42187381e-01
-9.15451571e-02 -1.20066166e+00 -9.01147425e-01 -1.09067857e+00
1.78871825e-01 -2.57323027e-01 9.25110996e-01 6.69348657e-01
1.10494412e-01 6.30967617e-02 -3.38223070e-01 -5.21902144e-02
6.71657026e-01 -1.45989479e-02 8.89482975e-01 -1.03349280e+00
-4.27286655e-01 -1.05453336e+00 -1.66724652e-01 -7.75739789e-01
1.24106698e-01 -5.23265958e-01 -1.02399714e-01 -1.20117331e+00
8.97839293e-02 -6.37247860e-01 -2.72328824e-01 2.29912400e-01
-6.90677822e-01 -3.33347209e-02 -6.15422018e-02 2.76442885e-01
-4.36754882e-01 6.35120511e-01 1.44508851e+00 5.41225867e-03
-6.72776222e-01 2.51037717e-01 -3.24500114e-01 7.92471528e-01
1.17442393e+00 -5.83538055e-01 -1.79281667e-01 -1.73529983e-01
6.09640479e-01 -1.04452752e-01 -4.90665510e-02 -7.08997905e-01
3.12970608e-01 -6.95286751e-01 4.54938598e-02 -5.35799503e-01
3.26003373e-01 -6.62877023e-01 8.73718876e-03 3.62552196e-01
-1.18456028e-01 5.32391705e-02 2.44186997e-01 3.65802407e-01
-2.25575089e-01 -6.49127781e-01 1.08648825e+00 2.06283610e-02
-3.32002252e-01 3.40931326e-01 -1.40005484e-01 3.75015110e-01
1.04692340e+00 -1.21417619e-01 -3.97681624e-01 -1.43416449e-01
-3.52105826e-01 5.14179528e-01 3.28969300e-01 1.01107456e-01
3.34458530e-01 -1.23567009e+00 -7.11689949e-01 -4.39965688e-02
-1.48412347e-01 1.59732997e-01 -3.75685245e-02 1.20322835e+00
-8.38530660e-01 6.20324612e-01 1.64095357e-01 -5.95282197e-01
-1.15893674e+00 5.53453565e-01 6.18256271e-01 -4.23370510e-01
-5.14660656e-01 9.21994150e-01 -2.32601672e-01 -4.51274008e-01
4.75529075e-01 -4.71428856e-02 -1.47828326e-01 -1.21050484e-01
2.88064867e-01 9.12307918e-01 1.39893621e-01 -6.27371669e-02
-1.97587594e-01 1.02896869e+00 1.19600005e-01 -3.35024983e-01
1.17354536e+00 1.90516472e-01 -2.99252998e-02 3.12632859e-01
1.28218925e+00 -2.23104775e-01 -9.36773777e-01 -2.24794313e-01
1.45278513e-01 -5.73524117e-01 4.95940149e-01 -7.35771179e-01
-1.00730777e+00 6.11305475e-01 5.12549877e-01 6.64415881e-02
9.24094141e-01 -2.79444128e-01 8.01445067e-01 8.56752813e-01
1.07305951e-01 -1.66885018e+00 -8.27160478e-02 6.23799145e-01
8.50841105e-01 -1.56489801e+00 4.29410428e-01 -2.87235171e-01
-6.37240469e-01 1.25802064e+00 7.35711873e-01 -4.39482555e-02
8.72300625e-01 4.45715822e-02 1.98361743e-02 -5.68065643e-02
-7.40831673e-01 7.61719048e-02 7.39811778e-01 6.81634769e-02
2.42067292e-01 -1.54693797e-01 -7.06754565e-01 4.30965722e-01
-4.08724695e-01 -2.84861445e-01 1.05180927e-01 7.53737271e-01
-5.41630268e-01 -1.08640194e+00 -7.92052865e-01 2.40745038e-01
-5.18353999e-01 -6.87644184e-02 -1.43632710e-01 1.04226089e+00
-4.58442748e-01 8.10392320e-01 -1.84511960e-01 -1.05713882e-01
2.95156211e-01 6.31253347e-02 3.19170892e-01 -1.34533569e-01
-9.02707934e-01 1.50217220e-01 3.80653888e-02 -6.71986759e-01
1.02114774e-01 -5.75659096e-01 -1.20402217e+00 -3.74256939e-01
-7.90334523e-01 3.16812724e-01 1.31629646e+00 8.96608770e-01
1.70602649e-01 2.60230511e-01 7.02863276e-01 -6.73036873e-01
-9.17014182e-01 -9.67071533e-01 -1.60599336e-01 2.02969760e-01
1.74339592e-01 -8.67485046e-01 -6.19792759e-01 -4.37548399e-01] | [6.5842390060424805, 4.104918956756592] |
87706498-f3db-4d22-a6ac-b0d0ecf4e64f | boltzman-tuning-of-generative-models | null | null | https://openreview.net/forum?id=hP-pn8bKffe | https://openreview.net/pdf?id=hP-pn8bKffe | Boltzman Tuning of Generative Models | The paper focuses on the a posteriori tuning of a generative model in order to favor the generation of good instances in the sense of some external differentiable criterion. The proposed approach, called Boltzmann Tuning of Generative Models (BTGM) applies to a wide range of applications. It covers conditional generative modelling as a particular case, and offers an affordable alternative to rejection sampling. The contribution of the paper is twofold. Firstly, the objective is formalized and tackled as a well-posed optimization problem; a practical methodology is proposed to choose among the candidate criteria representing the same goal, the one best suited to efficiently learn a tuned generative model. Secondly, the merits of the approach are demonstrated on a real-world application, in the context of robust design for energy policies, showing the ability of BTGM to sample the extreme regions of the considered criteria. | ['Michele Sebag', 'Victor Berger'] | 2021-01-01 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [ 3.45886916e-01 4.18663710e-01 -4.64450344e-02 -5.92732653e-02
-9.40853834e-01 -8.50513577e-02 8.51408362e-01 2.33376492e-02
-5.95810235e-01 9.46034014e-01 -1.33646697e-01 1.68635882e-02
-7.09031105e-01 -9.25598443e-01 -4.86061186e-01 -1.24764347e+00
1.92753285e-01 8.57729256e-01 -2.33369231e-01 -2.11482689e-01
3.24121207e-01 6.31704867e-01 -1.70931184e+00 -3.03500235e-01
1.15544069e+00 7.21469104e-01 2.48884723e-01 2.62307882e-01
2.56574094e-01 7.63176149e-03 -6.26418889e-01 -2.01379269e-01
-2.75033806e-02 -5.86559117e-01 -5.07091641e-01 1.52109444e-01
-1.70447513e-01 4.24250782e-01 5.93926907e-01 9.84436393e-01
6.89484715e-01 5.57314813e-01 1.34861624e+00 -8.47656608e-01
1.44111468e-02 4.42968905e-01 1.00165002e-01 3.33060585e-02
3.62708792e-02 7.58165419e-02 9.27096963e-01 -7.85676301e-01
2.70578295e-01 1.06672227e+00 1.81029886e-02 5.24567962e-01
-1.75664139e+00 -2.06769988e-01 -3.38165127e-02 8.68281722e-02
-1.49358571e+00 -2.94354439e-01 8.07101488e-01 -6.24285102e-01
5.93957484e-01 3.12843800e-01 6.75923586e-01 9.78442967e-01
2.73864478e-01 4.78346318e-01 1.46893442e+00 -5.55010676e-01
1.05059803e+00 4.00237501e-01 -1.77848816e-01 1.40173838e-01
3.24459702e-01 3.83443505e-01 -3.30539703e-01 -3.96937609e-01
2.67364353e-01 -4.97045249e-01 -1.39383957e-01 -7.88400292e-01
-6.52861178e-01 1.04267156e+00 2.28036270e-01 6.14991009e-01
-5.30416310e-01 1.45445317e-01 1.13309063e-01 -5.80817237e-02
7.03242779e-01 4.83480364e-01 1.09077170e-01 -3.57848331e-02
-1.16212678e+00 4.87982810e-01 7.88390756e-01 4.46791410e-01
5.71101427e-01 4.27287608e-01 -2.53394127e-01 6.62553251e-01
5.73277712e-01 5.69755256e-01 3.46648633e-01 -1.87877998e-01
2.19921038e-01 1.39374048e-01 2.33314127e-01 -6.75169110e-01
-1.91484004e-01 -9.57893848e-01 -6.35571837e-01 5.69334865e-01
2.29640603e-01 -2.54631847e-01 -4.69766527e-01 1.74669135e+00
5.87040067e-01 -1.98273823e-01 -3.00819594e-02 8.12519252e-01
2.95928538e-01 6.97713614e-01 2.47324035e-01 -5.49305320e-01
9.30368304e-01 -6.42338172e-02 -4.83427942e-01 -4.79824319e-02
-4.12187316e-02 -5.64260602e-01 8.26660275e-01 6.21645808e-01
-1.13179660e+00 -6.05588377e-01 -9.71002400e-01 7.36747384e-01
-2.33385772e-01 2.60869443e-01 1.70874298e-01 8.78682017e-01
-8.22824478e-01 5.91184199e-01 -5.61517715e-01 -2.87972331e-01
-3.96212041e-02 3.77631813e-01 2.09714159e-01 3.17984641e-01
-1.10920656e+00 1.07485557e+00 7.18927324e-01 4.45646971e-01
-9.22315300e-01 -3.48006606e-01 -4.61474597e-01 3.11890841e-01
3.93024623e-01 -7.33298361e-01 8.28269482e-01 -1.12024391e+00
-1.93174720e+00 6.84796572e-01 1.52336910e-01 -4.01741475e-01
9.51998770e-01 -7.32410327e-02 -2.71060139e-01 -1.88325971e-01
-2.51678377e-01 2.91006237e-01 1.29810047e+00 -1.12269545e+00
-2.84274161e-01 -2.68099219e-01 -3.45290542e-01 1.24741055e-01
-1.64604798e-01 -2.24578351e-01 1.24674521e-01 -5.62946439e-01
-2.69857764e-01 -7.93903172e-01 -3.55857044e-01 -7.79493988e-01
-4.24775064e-01 -3.75350356e-01 1.78597003e-01 -2.48359844e-01
1.27807522e+00 -1.83459926e+00 8.62114072e-01 9.65840280e-01
-3.79306316e-01 2.40284801e-01 3.40807438e-01 8.00498426e-01
-4.30795178e-02 -1.49189115e-01 -5.05241513e-01 -1.43065661e-01
4.47711319e-01 3.99745926e-02 -3.94839108e-01 7.78033197e-01
2.40037605e-01 4.64786947e-01 -5.84707320e-01 -3.57838064e-01
5.20852387e-01 6.19285166e-01 -4.40076053e-01 2.85666436e-01
-5.79011202e-01 5.35759211e-01 -7.70145595e-01 1.38514666e-02
2.94994503e-01 2.14040577e-01 2.89603084e-01 1.33446425e-01
-2.97048628e-01 -2.17349321e-01 -1.46838832e+00 8.31097364e-01
-5.10999143e-01 4.54161316e-02 -2.39057541e-01 -1.27900088e+00
1.43404400e+00 2.14606076e-01 3.82597983e-01 -4.11414742e-01
4.97243077e-01 5.16589344e-01 -5.83876260e-02 -2.29778200e-01
2.23211840e-01 -3.86023194e-01 -1.00637980e-01 1.60367653e-01
1.53972507e-01 -4.53775465e-01 2.43980497e-01 -4.72605437e-01
2.36494288e-01 4.47456956e-01 7.33011425e-01 -7.02636838e-01
7.12034345e-01 -3.66956264e-01 2.27514297e-01 5.68821251e-01
2.73887306e-01 8.95238742e-02 5.28510928e-01 -1.10467546e-01
-9.08780575e-01 -9.51639652e-01 -3.39624286e-01 6.18616879e-01
-1.24349177e-01 1.38569400e-01 -8.86247516e-01 -3.33762109e-01
-1.44431025e-01 1.20967960e+00 -6.86112106e-01 -1.83318406e-01
-5.10973155e-01 -1.26178074e+00 -7.62133673e-02 -7.68699870e-02
1.52930379e-01 -1.16394043e+00 -1.20809805e+00 3.43764901e-01
1.25812754e-01 -3.94173682e-01 1.55605718e-01 4.59221244e-01
-9.64289904e-01 -8.26678753e-01 -1.03195524e+00 -2.65986353e-01
4.55071956e-01 -6.43877149e-01 1.23853469e+00 -4.38233435e-01
-2.40158603e-01 4.71131921e-01 -9.09489170e-02 -3.62747431e-01
-7.93699145e-01 2.69217819e-01 -1.70091346e-01 4.22656208e-01
1.83820501e-01 -4.95507747e-01 -2.58153677e-01 3.42949450e-01
-8.72200131e-01 -2.96881050e-01 5.72239399e-01 8.96349490e-01
7.23523915e-01 1.26383960e-01 8.52712989e-01 -7.82231629e-01
7.52508998e-01 -3.96065116e-01 -1.08068120e+00 2.47376919e-01
-8.74957383e-01 1.80802748e-01 5.85076630e-01 -3.08610201e-01
-1.07283843e+00 -6.54402375e-02 -3.09545189e-01 2.48919334e-02
-3.50091338e-01 1.70312613e-01 -5.77215016e-01 -1.01139292e-01
5.75417757e-01 3.61435950e-01 -2.29626685e-01 -5.50819457e-01
3.71029884e-01 1.29985943e-01 1.54567406e-01 -9.05076146e-01
7.83181667e-01 7.67523870e-02 4.69211251e-01 -1.07693923e+00
-2.85036713e-01 -1.89100340e-01 -3.68593395e-01 -5.95869958e-01
7.42192388e-01 -2.58399040e-01 -7.36909389e-01 4.06219997e-02
-5.83293796e-01 -2.76347220e-01 -6.71251178e-01 4.91527379e-01
-1.29780996e+00 -8.26509763e-03 7.60757104e-02 -1.40926719e+00
-2.77261406e-01 -1.16901481e+00 8.20203662e-01 3.22699249e-01
-1.81177855e-01 -1.04701006e+00 3.99447292e-01 -1.13688640e-01
2.88413107e-01 3.84198695e-01 1.23099077e+00 -6.18576705e-01
-5.10052025e-01 -8.58240724e-02 6.07985377e-01 3.66255313e-01
-1.84596464e-01 2.68654615e-01 -1.01439047e+00 -4.98357624e-01
1.66266665e-01 -9.70397368e-02 7.69163966e-01 6.47491574e-01
6.35625660e-01 -8.22423920e-02 -1.92837983e-01 2.57447898e-01
1.69297540e+00 2.92115629e-01 6.81376040e-01 3.67930949e-01
-7.93071762e-02 5.97120464e-01 9.28307831e-01 5.99318385e-01
-4.40451086e-01 1.12086141e+00 4.95559245e-01 -7.92089924e-02
3.39579225e-01 -1.42407557e-02 1.55287713e-01 3.40119451e-01
-1.18283987e-01 -3.56554300e-01 -6.04798436e-01 3.81687164e-01
-1.80131090e+00 -8.11120450e-01 2.12394759e-01 2.56598186e+00
3.84472519e-01 2.80477703e-01 5.21817923e-01 2.43526906e-01
8.52897644e-01 2.87286900e-02 -1.98214144e-01 -6.43712580e-01
1.26725212e-02 4.80726421e-01 2.38110289e-01 5.97285509e-01
-9.11362231e-01 2.91420430e-01 5.94864416e+00 1.16911423e+00
-9.57038343e-01 -8.00663531e-02 6.32866025e-01 1.02529414e-01
-3.69447410e-01 -3.24186683e-02 -8.83148909e-01 6.08833432e-01
9.22868073e-01 -1.48869917e-01 3.88250798e-01 7.34141469e-01
4.33039159e-01 -2.39715934e-01 -7.40642190e-01 6.42440438e-01
-7.65472054e-02 -8.35802794e-01 -1.92040093e-02 2.74029791e-01
7.22484231e-01 -5.14564514e-01 4.72935528e-01 3.84655185e-02
-1.74591243e-01 -1.04281855e+00 9.78998840e-01 9.10219908e-01
3.71605724e-01 -1.28455520e+00 6.95748925e-01 5.89029610e-01
-6.37868881e-01 -9.21875834e-02 -2.58931577e-01 4.09107387e-01
2.83606887e-01 8.44730198e-01 -8.11409533e-01 7.88473845e-01
2.05104098e-01 3.71301100e-02 -2.34158263e-01 1.33408952e+00
-2.23136246e-01 7.99096167e-01 -3.04946750e-01 -5.90840220e-01
3.12575191e-01 -6.17003322e-01 1.02959096e+00 1.26064277e+00
4.86254513e-01 -4.87993926e-01 -8.58698785e-02 1.27169001e+00
5.90933084e-01 5.84074855e-01 -4.41215694e-01 4.18191254e-02
7.27526471e-02 9.43573356e-01 -7.72300899e-01 -8.84517580e-02
3.80078256e-01 2.71247417e-01 1.12819135e-01 4.02628839e-01
-7.45508194e-01 -8.48984495e-02 1.69900537e-01 -3.49810123e-02
5.00949383e-01 5.32211810e-02 -1.16581462e-01 -6.93808436e-01
-2.64395952e-01 -8.91799927e-01 4.93978411e-01 -2.17436105e-01
-9.39897835e-01 6.02353811e-01 4.64679360e-01 -1.10915256e+00
-8.67387474e-01 -4.28585231e-01 -6.32740140e-01 1.23796749e+00
-1.16044343e+00 -6.11497223e-01 -4.61792536e-02 2.14661151e-01
4.63735938e-01 -1.86233073e-01 8.79094243e-01 1.21153697e-01
-4.67621475e-01 8.00153166e-02 5.47878742e-01 -7.51958132e-01
1.88335896e-01 -1.42865551e+00 -6.85259253e-02 8.33464563e-01
6.58760294e-02 5.08304834e-01 1.36153495e+00 -4.58440423e-01
-9.85226154e-01 -8.14593732e-01 7.50106812e-01 1.31057218e-01
2.43238330e-01 -2.41134912e-01 -6.82918191e-01 1.24788880e-02
6.41103089e-02 -4.64298189e-01 4.44120854e-01 -2.74689347e-02
3.61514211e-01 -1.09160103e-01 -1.13692272e+00 4.63236123e-01
2.78805315e-01 -1.07436724e-01 -3.63753319e-01 2.12316200e-01
-1.58325359e-01 -1.70976743e-01 -9.18769538e-01 3.91504198e-01
3.14194769e-01 -9.82206047e-01 9.68861163e-01 -4.27317530e-01
-3.57840508e-02 -2.63487965e-01 1.72665976e-02 -1.62329459e+00
-1.58619136e-01 -8.89848590e-01 -1.09065279e-01 1.23720336e+00
1.94210693e-01 -8.54760468e-01 6.57977939e-01 2.12868094e-01
3.43743056e-01 -1.05312490e+00 -1.37371159e+00 -6.37411416e-01
6.43744841e-02 -2.07153276e-01 4.85062957e-01 2.52205074e-01
-4.45638567e-01 1.89597309e-01 -3.23879778e-01 -7.45055676e-02
7.82716036e-01 3.54114681e-01 6.63390815e-01 -1.18325448e+00
-7.63316214e-01 -6.45996988e-01 -4.40381110e-01 -5.53625762e-01
5.71048670e-02 -5.81237793e-01 2.57659256e-01 -1.14937282e+00
-1.75363138e-01 -4.60758716e-01 -3.20293635e-01 -4.07083452e-01
-9.61534083e-02 -1.33037925e-01 1.74152926e-01 -1.41289368e-01
-6.21251650e-02 9.01913583e-01 8.37768912e-01 -6.03172109e-02
-1.66183025e-01 8.73636425e-01 -4.53869313e-01 4.95939612e-01
7.44240701e-01 -4.36021954e-01 -6.70227170e-01 4.27793264e-01
3.63656729e-01 9.47475061e-02 5.85882902e-01 -9.25051928e-01
-3.33823115e-01 -2.12879270e-01 2.68296242e-01 -5.13767421e-01
3.43731940e-01 -9.43600595e-01 6.67178273e-01 5.46704173e-01
-4.73158926e-01 -2.03165218e-01 -1.17636703e-01 5.84641337e-01
-5.22982515e-02 -8.08464706e-01 1.10876977e+00 6.53580576e-02
-5.20242989e-01 -1.42958239e-01 -3.92997593e-01 1.34582788e-01
1.00662172e+00 -2.03748450e-01 2.95846373e-01 -3.20488870e-01
-9.85716879e-01 -1.20424271e-01 1.91998348e-01 -3.42455506e-02
3.46214116e-01 -9.82057035e-01 -6.64394498e-01 2.01846170e-03
6.84886277e-02 -4.51313734e-01 3.05105060e-01 7.14233518e-01
-8.04575831e-02 5.43185890e-01 5.49704954e-03 -6.38547719e-01
-9.44213390e-01 5.68991601e-01 7.15970099e-01 -2.84737617e-01
-5.50189376e-01 3.85902494e-01 9.86758769e-02 1.04533471e-01
1.48942828e-01 -1.66101888e-01 -5.45388281e-01 7.52418414e-02
7.13355169e-02 4.80133861e-01 4.13724095e-01 -5.76573253e-01
-2.53135562e-01 5.65713406e-01 7.75846004e-01 -3.36162239e-01
1.25302458e+00 9.17334929e-02 1.34386897e-01 4.76760149e-01
8.33032489e-01 2.03367211e-02 -1.28770351e+00 8.34752992e-02
1.08741418e-01 -2.87333429e-01 2.64460016e-02 -7.35169172e-01
-6.10469460e-01 6.56655014e-01 7.15500593e-01 4.17708486e-01
1.09259105e+00 -2.98752099e-01 -6.82591274e-02 1.41562074e-01
5.33573806e-01 -1.27892792e+00 -3.02845329e-01 5.94633892e-02
9.54068124e-01 -5.85308552e-01 -3.93652022e-02 -1.71862379e-01
-4.26875293e-01 1.22536922e+00 1.28055543e-01 -3.05420071e-01
3.76342624e-01 -3.82612715e-03 -3.50947291e-01 -1.93525646e-02
-5.25178194e-01 -3.46982747e-01 6.97100580e-01 5.03269672e-01
1.70461833e-01 2.61565894e-01 -8.69461536e-01 3.17512006e-02
8.26585144e-02 -1.01035073e-01 -1.17863761e-02 5.55602312e-01
-5.22565067e-01 -1.00075817e+00 -5.59997857e-01 2.03556612e-01
-1.99696720e-01 1.42957374e-01 -1.23541899e-01 9.61549461e-01
9.03226361e-02 8.50872397e-01 -4.07771468e-01 2.37353027e-01
4.28524107e-01 2.19923750e-01 5.47563255e-01 -4.40737188e-01
-6.52255476e-01 4.66876477e-01 8.52733999e-02 -3.70938241e-01
-4.80169356e-01 -7.56806254e-01 -6.29938662e-01 2.77672172e-01
-4.46664900e-01 6.50766253e-01 7.15027392e-01 1.03514826e+00
-2.23263919e-01 7.94488609e-01 6.16998136e-01 -9.34435606e-01
-9.17894065e-01 -6.33873820e-01 -7.29164600e-01 3.01380977e-02
6.81704357e-02 -9.32357490e-01 -4.07557249e-01 -4.79497284e-01] | [6.091376304626465, 3.6492083072662354] |
b6495759-bd9c-45c1-a1c9-4cbc24d28494 | equal-confusion-fairness-measuring-group | 2307.00472 | null | https://arxiv.org/abs/2307.00472v1 | https://arxiv.org/pdf/2307.00472v1.pdf | Equal Confusion Fairness: Measuring Group-Based Disparities in Automated Decision Systems | As artificial intelligence plays an increasingly substantial role in decisions affecting humans and society, the accountability of automated decision systems has been receiving increasing attention from researchers and practitioners. Fairness, which is concerned with eliminating unjust treatment and discrimination against individuals or sensitive groups, is a critical aspect of accountability. Yet, for evaluating fairness, there is a plethora of fairness metrics in the literature that employ different perspectives and assumptions that are often incompatible. This work focuses on group fairness. Most group fairness metrics desire a parity between selected statistics computed from confusion matrices belonging to different sensitive groups. Generalizing this intuition, this paper proposes a new equal confusion fairness test to check an automated decision system for fairness and a new confusion parity error to quantify the extent of any unfairness. To further analyze the source of potential unfairness, an appropriate post hoc analysis methodology is also presented. The usefulness of the test, metric, and post hoc analysis is demonstrated via a case study on the controversial case of COMPAS, an automated decision system employed in the US to assist judges with assessing recidivism risks. Overall, the methods and metrics provided here may assess automated decision systems' fairness as part of a more extensive accountability assessment, such as those based on the system accountability benchmark. | ['Ioannis A. Kakadiaris', 'Furkan Gursoy'] | 2023-07-02 | null | null | null | null | ['fairness', 'fairness'] | ['computer-vision', 'miscellaneous'] | [ 1.38441294e-01 1.21496230e-01 -1.39429763e-01 -6.62694812e-01
-1.61566690e-01 -6.02015853e-01 5.52345037e-01 7.77203083e-01
-6.90858603e-01 7.59796619e-01 2.47639969e-01 -6.54435992e-01
-5.67281544e-01 -5.38717866e-01 3.36760193e-01 -3.73265475e-01
4.46698606e-01 3.49441111e-01 -4.93197322e-01 -1.09564245e-01
9.77839231e-01 6.05080307e-01 -1.49493408e+00 3.91516760e-02
1.33085215e+00 8.69770885e-01 -7.74217665e-01 3.94621760e-01
1.11461729e-01 1.22078264e+00 -6.97750926e-01 -1.11662889e+00
4.91043687e-01 -5.73771179e-01 -9.80902970e-01 -3.00986350e-01
6.35145426e-01 -5.44813156e-01 2.83463567e-01 1.32133818e+00
6.34341896e-01 -6.01525865e-02 7.80982375e-01 -1.59813380e+00
-6.74877942e-01 6.66839838e-01 -2.50065327e-01 2.51510412e-01
2.64254689e-01 3.81992221e-01 1.17030454e+00 -4.34246987e-01
4.53837961e-02 1.20421112e+00 7.19816744e-01 3.67104471e-01
-1.08347297e+00 -8.48178625e-01 -2.02907845e-01 3.01037014e-01
-1.22175848e+00 -6.14589751e-01 4.31394070e-01 -8.82033408e-01
5.10379255e-01 6.98738992e-01 5.25080085e-01 1.29079863e-01
2.86935270e-01 -3.25962715e-02 1.73431897e+00 -5.41249812e-01
4.00325090e-01 3.51961106e-01 6.37682974e-01 2.42456406e-01
1.00537574e+00 2.39193335e-01 -1.42558180e-02 -4.92373466e-01
2.48838961e-01 -2.63308465e-01 1.24575503e-01 -1.12830810e-01
-6.75400794e-01 1.06327701e+00 2.85926372e-01 2.82980978e-01
-3.88288498e-01 -3.00902277e-01 5.75563252e-01 3.63241345e-01
4.81898874e-01 8.99949372e-01 8.64032283e-02 -2.75104165e-01
-8.22973013e-01 4.07337576e-01 8.96979570e-01 6.73152655e-02
4.81153339e-01 -7.11582601e-02 -7.94832408e-01 5.75113416e-01
1.97259456e-01 4.07564938e-01 1.92289591e-01 -1.47866929e+00
2.87255347e-01 9.97826993e-01 4.27719414e-01 -1.47144794e+00
-3.09871823e-01 -3.64699453e-01 -8.06161106e-01 7.55716741e-01
6.44046128e-01 1.84177309e-02 -2.50975609e-01 1.63857675e+00
1.57283157e-01 -7.34328926e-01 -1.30397215e-01 1.08886838e+00
3.46474200e-01 -6.75370246e-02 6.12918615e-01 -3.17170173e-01
1.09840155e+00 -4.20608610e-01 -8.49857032e-01 7.89226405e-03
4.92800087e-01 -6.06662333e-01 8.97376001e-01 2.70880640e-01
-1.08715641e+00 -3.84828150e-01 -7.65932798e-01 -9.48385075e-02
-1.40547708e-01 -3.67412537e-01 7.30250239e-01 1.30770826e+00
-7.26144433e-01 6.57234490e-01 -4.18170914e-02 -3.25525969e-01
5.82822025e-01 1.53965771e-01 -2.34790057e-01 1.44243181e-01
-1.29923034e+00 1.53944230e+00 2.68140882e-02 8.95540044e-02
-1.27419755e-01 -4.28109974e-01 -5.46811163e-01 4.04158980e-01
-7.57754548e-03 -6.48638368e-01 1.02055180e+00 -1.33040440e+00
-9.52997923e-01 1.14601171e+00 2.27544010e-01 -5.66578388e-01
9.90061820e-01 1.36954248e-01 -3.96786600e-01 -1.42446265e-01
5.51385045e-01 -2.76681446e-02 3.21575612e-01 -1.04358792e+00
-7.09260166e-01 -6.94846213e-01 1.20283119e-01 4.44856256e-01
-2.08179444e-01 7.24421918e-01 7.47688711e-01 -4.09290582e-01
-2.58782953e-01 -7.53285348e-01 -2.25399762e-01 -1.40115649e-01
-2.58766085e-01 -1.41206875e-01 1.63186699e-01 -7.19750166e-01
1.88208544e+00 -1.99551964e+00 -5.94535530e-01 5.68263292e-01
4.85288948e-01 4.56123561e-01 3.94668758e-01 2.05278844e-01
-2.82396555e-01 5.25586367e-01 -4.03004795e-01 1.27592474e-01
4.25424576e-01 -2.19654813e-01 4.91252460e-04 6.18980646e-01
-7.22730756e-02 4.42580670e-01 -6.41727149e-01 -4.65674967e-01
2.63140857e-01 -1.17644154e-01 -4.56157804e-01 8.02000836e-02
7.74041176e-01 9.48381051e-02 -7.74030611e-02 5.00098109e-01
7.01144874e-01 9.21309590e-02 2.18157157e-01 4.97290976e-02
-3.86437684e-01 1.30031452e-01 -9.82544065e-01 6.02576375e-01
2.11480245e-01 4.81983453e-01 -1.22092277e-01 -9.27783787e-01
9.90333974e-01 1.49257764e-01 2.16118231e-01 -9.12235081e-01
5.26900947e-01 3.14929157e-01 5.70363700e-01 -3.93319428e-01
7.49075711e-01 -5.34353197e-01 -3.55096579e-01 8.32013369e-01
-5.89795053e-01 -1.88018247e-01 2.87081957e-01 1.07792728e-01
8.66622865e-01 -5.76927781e-01 1.00921178e+00 -5.72780848e-01
6.86556518e-01 1.75653584e-02 6.64752901e-01 8.83342981e-01
-1.21460915e+00 3.50046486e-01 6.27549231e-01 -6.18634343e-01
-1.09523189e+00 -8.06358874e-01 -4.01194125e-01 7.23995984e-01
1.15114965e-01 -4.18353789e-02 -9.41362977e-01 -5.18226504e-01
4.98902231e-01 1.01064670e+00 -7.75726199e-01 -2.73356408e-01
9.10911784e-02 -7.11007535e-01 6.95019305e-01 1.27846077e-01
6.73475802e-01 -7.68136919e-01 -1.10592544e+00 -6.21424578e-02
-1.43215373e-01 -6.58507943e-01 -1.46785229e-01 -3.53495091e-01
-5.97535551e-01 -1.43494892e+00 -2.02122539e-01 5.03974780e-02
5.76092243e-01 8.63870159e-02 1.04762173e+00 3.80615056e-01
-2.05790088e-01 2.94505000e-01 -5.73122092e-02 -7.55299866e-01
-7.63876021e-01 -5.02228856e-01 2.32969388e-01 -1.19616993e-01
8.77936721e-01 -7.17766732e-02 -5.45565665e-01 3.75871867e-01
-6.61170483e-01 -1.84936643e-01 3.23271096e-01 6.02681816e-01
-1.73006982e-01 -1.56831443e-02 7.95955241e-01 -9.74046767e-01
1.33845258e+00 -4.63263094e-01 -4.53367710e-01 2.16858491e-01
-1.46738780e+00 -3.39559793e-01 3.93966258e-01 2.64314592e-01
-1.05827868e+00 -6.60596848e-01 4.31987882e-01 1.61325306e-01
-1.46525919e-01 5.53799093e-01 1.10592442e-02 -1.60978526e-01
9.55596209e-01 -7.19086170e-01 5.29514432e-01 1.36429304e-02
1.23951435e-01 9.65149820e-01 2.98974723e-01 -4.78228211e-01
4.82094646e-01 8.57070982e-02 -1.35784773e-02 -1.48383588e-01
-8.64776552e-01 -3.15769494e-01 -3.44854116e-01 -6.69702172e-01
6.45100653e-01 -5.31963944e-01 -1.03191793e+00 5.16125500e-01
-9.14929271e-01 1.74575865e-01 -3.03800166e-01 5.93177140e-01
-2.58198649e-01 5.20862579e-01 -3.57643962e-01 -1.34225655e+00
-5.41199327e-01 -9.76068795e-01 8.91671777e-02 3.43564570e-01
-9.15295899e-01 -7.12036371e-01 1.23484984e-01 8.14344168e-01
7.24155128e-01 3.77248198e-01 8.97358954e-01 -9.80545461e-01
1.41266093e-01 -5.24255157e-01 -5.42683363e-01 6.05102479e-01
2.69890666e-01 1.95474520e-01 -9.02918875e-01 -1.16694130e-01
1.06804885e-01 -9.44332406e-02 3.10927212e-01 2.45460570e-01
8.19281340e-01 -7.27742255e-01 2.94329107e-01 -1.95091248e-01
1.27377963e+00 3.91623944e-01 6.84407890e-01 5.44108987e-01
3.18332464e-01 1.17889428e+00 7.72662640e-01 7.59860218e-01
5.51647127e-01 3.80122215e-01 4.37323093e-01 1.36488182e-02
3.30223590e-01 2.52141416e-01 -3.33925672e-02 4.79170829e-01
-5.27078152e-01 2.76980013e-01 -1.19806445e+00 1.56970412e-01
-1.70959747e+00 -1.44301748e+00 -2.58281440e-01 2.64661336e+00
5.72712481e-01 7.92088434e-02 2.90158898e-01 6.37464166e-01
1.28758812e+00 -2.34306872e-01 -3.41175526e-01 -1.19038880e+00
1.47517296e-02 -3.31152156e-02 3.96460295e-01 6.00646555e-01
-9.23805058e-01 3.91153693e-01 6.90236712e+00 2.94243425e-01
-6.64349437e-01 -1.85520768e-01 1.20612919e+00 1.26043916e-01
-3.26743931e-01 2.00746998e-01 -9.95855480e-02 6.56240404e-01
8.78023803e-01 -9.54429269e-01 2.11658895e-01 6.68006003e-01
5.82476139e-01 -4.44229633e-01 -8.19366097e-01 8.13872099e-01
5.46839423e-02 -8.29839647e-01 -1.80989634e-02 1.88425139e-01
6.15835786e-01 -7.37757325e-01 1.36002123e-01 1.86832085e-01
4.89786327e-01 -1.23172092e+00 8.51236165e-01 6.96836591e-01
6.18964732e-01 -9.34878647e-01 1.21974027e+00 -7.10753798e-02
-3.15010816e-01 -4.14076924e-01 -1.65597081e-01 -1.01352823e+00
-1.31169274e-01 8.51540387e-01 -4.16130394e-01 3.73077154e-01
5.47605813e-01 1.93674102e-01 -6.50961578e-01 1.14448988e+00
-5.66060580e-02 3.78945351e-01 3.15873682e-01 3.18432529e-03
-3.66359428e-02 -3.99986386e-01 3.19308400e-01 9.64041173e-01
8.67314562e-02 2.18744412e-01 -4.36036170e-01 9.99669909e-01
1.61228180e-01 3.75970721e-01 -5.92706442e-01 -5.59082404e-02
7.23482192e-01 1.19215763e+00 -5.55112183e-01 -3.77518564e-01
-1.48478985e-01 4.53446656e-01 1.81371465e-01 7.22301528e-02
-7.40387201e-01 -3.58681560e-01 8.72893333e-01 4.49308418e-02
-8.23802531e-01 2.98222333e-01 -1.13567531e+00 -6.77836418e-01
-1.85648471e-01 -1.15799892e+00 4.03404444e-01 -3.65171880e-01
-1.52893209e+00 1.94677651e-01 -3.44475284e-02 -1.30596387e+00
-4.93471650e-03 -3.30852628e-01 -5.06764650e-01 1.29373097e+00
-1.08337784e+00 -5.94841480e-01 -5.02957582e-01 1.77523613e-01
-3.97599697e-01 -3.21177036e-01 7.36923635e-01 3.70085806e-01
-5.60233414e-01 6.48554325e-01 -2.96127975e-01 6.89114816e-03
9.99033988e-01 -1.20676279e+00 -4.34318930e-02 1.00602353e+00
-8.14574003e-01 7.88831770e-01 8.38504851e-01 -5.93628168e-01
-4.06821251e-01 -6.76855505e-01 1.28411806e+00 -6.17532313e-01
4.52004492e-01 4.52789873e-01 -6.81647360e-01 2.34583616e-01
1.48313910e-01 -4.31333601e-01 1.25397062e+00 8.40889588e-02
-2.66870767e-01 -1.61764011e-01 -1.75257373e+00 5.47164083e-01
7.60869324e-01 -5.49262702e-01 -7.02551484e-01 -2.85885572e-01
2.72934213e-02 2.75925323e-02 -8.40149581e-01 3.64154041e-01
8.31836939e-01 -1.56857145e+00 5.47585070e-01 -6.94850743e-01
6.26351655e-01 -2.41173059e-01 -9.84348431e-02 -1.12026894e+00
-8.15913975e-01 -3.32384318e-01 5.30329227e-01 1.29879725e+00
2.35388130e-01 -1.04829335e+00 3.83169711e-01 1.70951617e+00
2.00272948e-01 -2.16986179e-01 -8.46242666e-01 -4.36836511e-01
2.05105767e-01 -2.97998279e-01 8.66744339e-01 1.55181265e+00
2.94944227e-01 8.02977681e-02 -2.21566498e-01 -1.64226905e-01
8.39049339e-01 -1.45431891e-01 7.15705395e-01 -1.70910573e+00
3.07818830e-01 -1.02409554e+00 -7.92974472e-01 4.24287826e-01
7.68023059e-02 -6.81329250e-01 -3.09230775e-01 -1.39903557e+00
5.74156344e-01 -5.21156609e-01 -5.53661764e-01 4.46451366e-01
-5.71138203e-01 8.78906548e-02 5.87109387e-01 4.31458563e-01
-2.44069383e-01 2.02841148e-01 8.84233594e-01 -4.85143550e-02
2.26647973e-01 -9.82332453e-02 -1.55343735e+00 7.22736478e-01
8.80099237e-01 -4.25207466e-01 -1.49242535e-01 -3.04212980e-02
3.31941009e-01 -1.65300831e-01 4.27898824e-01 -1.02365947e+00
2.27509245e-01 -5.88771284e-01 1.15396276e-01 2.51714230e-01
-3.46314579e-01 -9.47955549e-01 3.25578153e-01 7.05010414e-01
-5.77649057e-01 1.63397402e-01 -1.83189027e-02 -8.86340719e-03
-3.15306455e-01 -2.93814957e-01 9.62349296e-01 2.47037224e-02
-3.54727030e-01 1.67404842e-02 -3.60006303e-01 -2.76992507e-02
1.21364737e+00 -3.97832751e-01 -6.66719139e-01 -4.15536702e-01
-4.39089239e-01 1.72178626e-01 7.14514613e-01 1.24092214e-01
1.93992943e-01 -1.38387680e+00 -1.03066695e+00 -2.51746148e-01
3.15053195e-01 -7.81144679e-01 2.01789647e-01 9.14955556e-01
-6.09431684e-01 3.96098763e-01 -7.41424322e-01 -7.17712892e-03
-1.30740631e+00 3.22844058e-01 4.20073539e-01 -1.78917214e-01
1.27199605e-01 3.32349509e-01 6.47698343e-02 -2.29649618e-01
-8.67306739e-02 1.01188146e-01 -5.35593748e-01 2.96681166e-01
6.49663568e-01 1.03074431e+00 -4.51106206e-03 -9.93779659e-01
-3.84555876e-01 1.69823498e-01 2.80587405e-01 -2.04181656e-01
7.37103939e-01 -3.42090309e-01 -7.02581525e-01 3.10206920e-01
4.74040121e-01 1.21669374e-01 -6.40736222e-01 2.31948912e-01
1.87477991e-01 -1.07973039e+00 -1.61611184e-01 -1.13551605e+00
-6.37626112e-01 6.37048244e-01 6.56363189e-01 7.04953730e-01
1.05465531e+00 -8.20928156e-01 -2.97991812e-01 -2.15366781e-02
4.11218703e-01 -1.44632971e+00 -4.27616924e-01 2.42839947e-01
9.51369882e-01 -1.32950866e+00 2.16247380e-01 -1.32447451e-01
-1.01947188e+00 9.52327609e-01 6.00057304e-01 1.36688307e-01
2.89839089e-01 -1.55277178e-01 2.79221684e-01 -5.97474501e-02
-2.11704671e-01 1.80803448e-01 1.47358105e-01 3.96932155e-01
8.18123639e-01 6.69639826e-01 -1.36332941e+00 8.81513536e-01
-3.70578229e-01 1.78083554e-01 7.83280075e-01 6.36342347e-01
-4.23076689e-01 -8.20892334e-01 -5.96328855e-01 8.78496826e-01
-7.24470139e-01 -3.49377021e-02 -7.58122742e-01 6.12053454e-01
1.23783492e-01 1.24564862e+00 -8.71457346e-03 -4.32983041e-01
3.72886658e-01 -7.61085600e-02 9.35117006e-02 -3.70272815e-01
-1.04479373e+00 -7.63833463e-01 4.66492802e-01 -5.66411614e-01
-6.70711815e-01 -7.13203132e-01 -8.47592294e-01 -1.07533669e+00
-1.96288973e-01 5.15322685e-01 4.56593007e-01 9.51930106e-01
2.72277117e-01 1.71965525e-01 7.00559199e-01 -2.81218767e-01
-8.14140916e-01 -9.17545140e-01 -7.96021402e-01 7.84236431e-01
2.59000119e-02 -6.57351077e-01 -7.08037734e-01 -4.85957831e-01] | [8.880069732666016, 5.475654602050781] |
0f0bef1f-f827-4ee0-b886-09c71fb027a2 | acm-net-action-context-modeling-network-for | 2104.02967 | null | https://arxiv.org/abs/2104.02967v1 | https://arxiv.org/pdf/2104.02967v1.pdf | ACM-Net: Action Context Modeling Network for Weakly-Supervised Temporal Action Localization | Weakly-supervised temporal action localization aims to localize action instances temporal boundary and identify the corresponding action category with only video-level labels. Traditional methods mainly focus on foreground and background frames separation with only a single attention branch and class activation sequence. However, we argue that apart from the distinctive foreground and background frames there are plenty of semantically ambiguous action context frames. It does not make sense to group those context frames to the same background class since they are semantically related to a specific action category. Consequently, it is challenging to suppress action context frames with only a single class activation sequence. To address this issue, in this paper, we propose an action-context modeling network termed ACM-Net, which integrates a three-branch attention module to measure the likelihood of each temporal point being action instance, context, or non-action background, simultaneously. Then based on the obtained three-branch attention values, we construct three-branch class activation sequences to represent the action instances, contexts, and non-action backgrounds, individually. To evaluate the effectiveness of our ACM-Net, we conduct extensive experiments on two benchmark datasets, THUMOS-14 and ActivityNet-1.3. The experiments show that our method can outperform current state-of-the-art methods, and even achieve comparable performance with fully-supervised methods. Code can be found at https://github.com/ispc-lab/ACM-Net | ['Alois Knoll', 'Fan Lu', 'Lijun Zhang', 'Zhijun Li', 'Guang Chen', 'Sanqing Qu'] | 2021-04-07 | null | null | null | null | ['weakly-supervised-action-localization', 'weakly-supervised-temporal-action'] | ['computer-vision', 'computer-vision'] | [ 5.64853370e-01 -1.70821637e-01 -5.57871521e-01 -2.13046893e-01
-6.50825143e-01 -4.03370857e-01 5.59467137e-01 -3.99530381e-02
-3.29867780e-01 6.02111161e-01 2.87661403e-01 -9.77954939e-02
2.04382613e-01 -3.51661682e-01 -5.46774209e-01 -8.82719696e-01
8.32888186e-02 -1.13340542e-01 8.39429259e-01 3.55761379e-01
1.18713669e-01 2.28525653e-01 -1.27433479e+00 7.67517805e-01
5.09127021e-01 9.80834246e-01 1.28668249e-01 2.72465527e-01
-1.39949232e-01 1.27008355e+00 -5.89069784e-01 9.67884138e-02
4.85703610e-02 -8.77411544e-01 -7.98011303e-01 2.91420639e-01
4.40417707e-01 -4.37861919e-01 -3.78126919e-01 1.06423557e+00
2.48159021e-01 3.26621503e-01 3.33715618e-01 -1.37532949e+00
-2.94389814e-01 4.31071997e-01 -7.47157276e-01 7.51529336e-01
3.43354493e-01 3.79340261e-01 9.67364013e-01 -8.26139748e-01
5.48361659e-01 1.09162736e+00 2.38347322e-01 5.79562485e-01
-1.04005969e+00 -6.06692612e-01 9.33411896e-01 4.45536375e-01
-1.27158785e+00 -2.45978907e-01 8.42147589e-01 -3.44277024e-01
6.80234313e-01 1.23516493e-01 8.09066355e-01 1.34344757e+00
1.04121584e-03 1.25238109e+00 9.24895525e-01 -9.84355733e-02
2.80541927e-01 -4.17178780e-01 6.90860450e-02 5.75525403e-01
-1.29529715e-01 -1.43694341e-01 -5.30425549e-01 3.34650837e-02
7.62774527e-01 2.09749475e-01 -3.82073194e-01 -2.63065368e-01
-1.47414279e+00 4.59710896e-01 3.06039393e-01 6.53588295e-01
-4.22768682e-01 4.00340855e-01 3.97337586e-01 -3.27518940e-01
4.13066953e-01 -2.87217289e-01 -4.80099887e-01 -3.19683164e-01
-7.16575682e-01 3.07949092e-02 3.19718063e-01 9.13505137e-01
6.81437552e-01 -7.99603462e-02 -6.64857864e-01 5.67144394e-01
1.57139838e-01 2.00558349e-01 3.70114237e-01 -9.82227445e-01
4.00393516e-01 8.14779937e-01 9.02436748e-02 -9.00883734e-01
-2.49835953e-01 -2.60613412e-01 -5.01057923e-01 -7.58811533e-02
4.84934509e-01 -1.42430827e-01 -9.08681571e-01 1.78289473e+00
4.91924673e-01 9.34162736e-01 -1.21302105e-01 1.08435774e+00
7.39162445e-01 6.72736645e-01 4.26633596e-01 -3.64355206e-01
1.32984555e+00 -1.31805646e+00 -8.27319801e-01 -3.93793076e-01
6.61162615e-01 -5.67982614e-01 1.19825852e+00 1.25381783e-01
-8.36478591e-01 -6.11132145e-01 -6.66579664e-01 1.11765243e-01
-7.40863308e-02 2.72833526e-01 5.41216195e-01 6.73932731e-02
-5.14989436e-01 4.12569344e-01 -1.07108080e+00 -2.41026536e-01
6.38807118e-01 2.91838590e-02 -1.78710416e-01 1.42929954e-02
-1.01740301e+00 4.10449147e-01 5.88735282e-01 1.62322685e-01
-1.19026375e+00 -3.08855802e-01 -7.87878454e-01 -4.05960833e-04
8.37206662e-01 -2.76998281e-01 1.24152124e+00 -1.54629505e+00
-1.13715947e+00 6.33646429e-01 -4.40378010e-01 -2.95641094e-01
4.10342842e-01 -2.38749623e-01 -5.19322455e-01 4.32948351e-01
2.61998534e-01 7.26327598e-01 6.35685802e-01 -1.05414033e+00
-1.06288791e+00 -1.29373387e-01 2.97455698e-01 2.48483837e-01
-1.82065636e-01 3.62235248e-01 -9.01561916e-01 -7.71135032e-01
1.44060552e-01 -8.39735568e-01 -1.35451317e-01 -9.45060253e-02
-3.39240670e-01 -4.16114330e-01 1.00602531e+00 -4.97913241e-01
1.49006665e+00 -2.38383722e+00 -4.68950756e-02 -2.51003265e-01
-1.83900744e-02 3.21479827e-01 -1.30628780e-01 -6.18169531e-02
-1.37195766e-01 5.51034212e-02 -3.78206670e-01 -4.38634828e-02
-2.22848654e-01 2.72048146e-01 -8.93526077e-02 4.25955683e-01
3.50745976e-01 7.78144062e-01 -1.27948999e+00 -6.25478685e-01
2.77465791e-01 3.88779134e-01 -3.45348448e-01 -4.68493812e-02
-4.20715779e-01 7.86441684e-01 -7.31964588e-01 9.46668267e-01
3.10056925e-01 -3.07194948e-01 2.80172646e-01 -2.74226695e-01
-3.02541698e-03 3.06554794e-01 -1.25164986e+00 1.67635417e+00
-4.39640805e-02 6.40988052e-01 -2.22496584e-01 -1.07167614e+00
4.39556301e-01 3.32112670e-01 7.62545168e-01 -5.72125614e-01
1.56593695e-01 7.10385591e-02 1.15447037e-01 -6.22569442e-01
8.27266425e-02 1.22662317e-02 1.46698635e-02 2.14452833e-01
5.46118543e-02 7.75902987e-01 4.54943478e-01 1.17874622e-01
1.21219754e+00 5.76673806e-01 1.86379030e-01 2.11324561e-02
8.53744864e-01 -3.17871153e-01 1.15193403e+00 5.15680194e-01
-7.45155990e-01 6.54787719e-01 8.41509521e-01 -4.33930516e-01
-4.08671677e-01 -8.61422122e-01 1.87303379e-01 1.28580034e+00
5.77563643e-01 -5.11427820e-01 -7.90319502e-01 -1.13500011e+00
-3.48671019e-01 5.93182504e-01 -6.15026474e-01 -1.65771842e-01
-8.31960618e-01 -5.28119564e-01 3.48568887e-01 7.64971793e-01
5.97282946e-01 -1.44451714e+00 -6.96096241e-01 2.17537522e-01
-4.90333110e-01 -1.41346526e+00 -8.29191506e-01 -3.94689292e-02
-7.74667799e-01 -1.26403451e+00 -6.48573697e-01 -7.73193002e-01
5.84922194e-01 3.57182711e-01 9.59522486e-01 2.24610642e-01
-4.53179423e-03 2.20923007e-01 -4.64702457e-01 -5.01511246e-02
-6.01667762e-02 -1.95375144e-01 -1.87204659e-01 4.80149537e-01
5.64762831e-01 -5.09127617e-01 -8.97680461e-01 6.33619249e-01
-8.36119890e-01 2.01161176e-01 4.80301261e-01 4.46543276e-01
9.47270691e-01 -2.84465533e-02 4.42120552e-01 -5.93308687e-01
-1.72960293e-02 -4.32651550e-01 -4.23229516e-01 3.18778098e-01
-3.90638188e-02 -4.13770407e-01 3.33806425e-01 -6.67837501e-01
-1.04860759e+00 3.63029599e-01 1.36443809e-01 -6.80734873e-01
-5.40915549e-01 3.62523109e-01 -4.64248091e-01 3.49678606e-01
2.75195211e-01 2.37435892e-01 -5.00716031e-01 -3.36892664e-01
2.64622778e-01 3.35156143e-01 5.45364678e-01 -4.80456710e-01
3.48522544e-01 6.84667706e-01 -2.25733936e-01 -5.71755469e-01
-1.15641034e+00 -6.43266380e-01 -6.20272815e-01 -5.61534107e-01
1.29935312e+00 -9.09944832e-01 -3.58475238e-01 4.68062669e-01
-1.01115334e+00 -5.84576905e-01 -1.92509443e-01 5.49383223e-01
-5.03838718e-01 5.36162972e-01 -5.27278185e-01 -7.08202720e-01
4.68300842e-02 -1.17102683e+00 1.07983220e+00 3.30890268e-01
-1.09256439e-01 -7.62850642e-01 -2.37684891e-01 3.81967247e-01
-2.07959458e-01 3.71934950e-01 4.19460207e-01 -5.61681092e-01
-8.21167588e-01 -2.78068297e-02 -1.56884223e-01 2.63589114e-01
4.02831912e-01 2.41534766e-02 -7.88249314e-01 -3.21754999e-02
-2.17779279e-01 -1.11544549e-01 1.01897752e+00 5.23934126e-01
1.47370613e+00 -1.35480836e-01 -4.32123870e-01 5.31878293e-01
1.03869152e+00 4.89988953e-01 6.35219574e-01 8.57765004e-02
8.79697978e-01 3.54400098e-01 9.33115184e-01 2.74164528e-01
3.52922231e-02 8.44997704e-01 4.13102627e-01 -1.03027292e-01
-1.47253022e-01 -2.43502080e-01 6.44594193e-01 2.89075553e-01
-2.58039236e-01 -3.00206721e-01 -7.61452973e-01 5.58802068e-01
-2.09433174e+00 -1.25536311e+00 -2.05371559e-01 2.01862526e+00
7.49598503e-01 4.50005919e-01 3.48160625e-01 -8.67353827e-02
9.51076627e-01 3.79508406e-01 -6.44426167e-01 2.83452123e-01
-1.67133063e-01 -1.19182616e-01 2.69148827e-01 2.69651711e-01
-1.47813928e+00 1.03206825e+00 4.78183174e+00 1.00126326e+00
-9.45048273e-01 3.29355091e-01 8.20545673e-01 -4.29003268e-01
6.05236255e-02 2.52800822e-01 -8.19997728e-01 7.95278788e-01
5.42353094e-01 1.26947746e-01 1.64668307e-01 7.71535039e-01
4.88530129e-01 -3.30116749e-01 -1.20874619e+00 7.47901797e-01
1.26536461e-02 -1.06781244e+00 -1.44918421e-02 -1.49100319e-01
6.84240103e-01 -9.85126346e-02 -3.73369843e-01 3.88356507e-01
-1.93339884e-01 -6.00088775e-01 8.98130715e-01 3.39255691e-01
4.53563333e-01 -5.66357017e-01 5.25911629e-01 2.59465784e-01
-1.59472299e+00 -2.38072112e-01 -3.05472706e-02 -2.52335500e-02
3.65386069e-01 3.30024362e-01 -3.13241594e-02 4.42215592e-01
8.05251479e-01 1.08435023e+00 -4.36595201e-01 9.70419049e-01
-5.79111576e-01 8.50998640e-01 -6.48756623e-02 1.68243006e-01
4.58031416e-01 -1.39647812e-01 4.90421653e-01 1.24557352e+00
5.17269149e-02 4.58421856e-01 5.80355465e-01 7.66414464e-01
9.04492587e-02 -1.28702760e-01 -2.32913271e-01 -7.93377459e-02
2.11488441e-01 1.19994152e+00 -1.18327165e+00 -5.73663652e-01
-7.58377254e-01 1.03904283e+00 -5.74023277e-03 5.91979027e-01
-1.35320461e+00 -8.84822458e-02 6.11769259e-01 -2.52855159e-02
3.96020770e-01 -7.99491405e-02 -2.89203748e-02 -1.37325847e+00
2.06437618e-01 -8.54936838e-01 6.54182911e-01 -7.72572339e-01
-9.14960623e-01 3.77288938e-01 1.25983745e-01 -1.44040215e+00
1.61225975e-01 -2.81481445e-01 -7.74011672e-01 4.89919215e-01
-1.35385215e+00 -1.03800511e+00 -4.66349930e-01 6.35351479e-01
8.85810971e-01 1.37820452e-01 1.95757195e-01 4.53451306e-01
-9.68284965e-01 3.69944125e-01 -3.84306073e-01 4.28399593e-01
5.54450095e-01 -9.88900244e-01 7.15939328e-02 1.14258230e+00
2.52671480e-01 3.15817326e-01 2.85410911e-01 -7.95137703e-01
-8.26886117e-01 -1.22581267e+00 6.67494118e-01 -4.34970856e-01
6.88560247e-01 -2.02431127e-01 -1.01987600e+00 7.63535202e-01
1.37342915e-01 3.80853027e-01 4.93806213e-01 -2.23180220e-01
7.63213187e-02 -3.40296254e-02 -6.49695277e-01 7.15879738e-01
1.43933904e+00 -4.78913784e-01 -5.98378837e-01 4.40160930e-01
7.65846252e-01 -3.12535614e-01 -4.01130438e-01 5.94337702e-01
3.37752014e-01 -1.04341471e+00 9.23605323e-01 -5.49579382e-01
6.15158677e-01 -7.58344233e-01 -2.01839190e-02 -5.99367261e-01
-3.29768598e-01 -4.17320490e-01 -4.04314190e-01 1.30767870e+00
1.62405267e-01 -2.59113222e-01 8.23081791e-01 3.59084487e-01
-2.65962034e-01 -9.40954089e-01 -9.20630336e-01 -8.42186451e-01
-2.64209270e-01 -4.17111725e-01 3.38860065e-01 9.77467239e-01
-2.30205476e-01 3.23949754e-01 -2.65905440e-01 1.94642022e-01
2.83476740e-01 2.94022769e-01 5.27349055e-01 -7.91729271e-01
-3.15297514e-01 -6.11635089e-01 -3.62980157e-01 -1.11790764e+00
3.54331613e-01 -6.40089393e-01 2.43195638e-01 -1.44782245e+00
3.32147032e-01 -1.25520706e-01 -8.65435064e-01 9.10989404e-01
-5.64536273e-01 3.26520413e-01 1.91526204e-01 1.60961941e-01
-1.04613733e+00 5.51447034e-01 1.24170160e+00 -1.80264547e-01
-2.89889365e-01 -8.66797660e-03 -3.97240609e-01 1.03257561e+00
7.51159668e-01 -4.44784105e-01 -5.74765027e-01 -2.78235614e-01
-2.75724858e-01 8.66698921e-02 5.98395824e-01 -1.10200751e+00
5.06469682e-02 -5.36201239e-01 3.67469966e-01 -6.36840463e-01
1.43376350e-01 -7.28269577e-01 6.16342500e-02 4.24831718e-01
-4.24828649e-01 -3.65640551e-01 1.30962625e-01 7.96163142e-01
-2.48074487e-01 -1.47003278e-01 6.90198720e-01 -1.67171091e-01
-1.19963920e+00 4.77031946e-01 -5.14314115e-01 1.81342468e-01
1.28055787e+00 -3.00964564e-01 -1.80509627e-01 -8.29704180e-02
-9.08964038e-01 2.41979450e-01 2.95701802e-01 4.98956710e-01
4.71260875e-01 -1.41783166e+00 -5.10399401e-01 -1.78165305e-02
2.46802971e-01 3.18456702e-02 4.84580755e-01 1.15444636e+00
-2.04046085e-01 2.03492343e-01 5.30408733e-02 -7.15810776e-01
-1.36156714e+00 6.10197008e-01 4.99955326e-01 -4.92783226e-02
-7.62781918e-01 7.49439836e-01 7.54195988e-01 1.37526587e-01
3.97367805e-01 -5.73738039e-01 -2.53316969e-01 -8.32970068e-02
6.06746376e-01 2.19391242e-01 -3.49528015e-01 -8.31265152e-01
-6.14135265e-01 5.20294905e-01 2.90554851e-01 1.12167783e-01
8.39251399e-01 1.74586549e-02 -1.54261077e-02 4.18686360e-01
9.80538428e-01 -1.57928184e-01 -1.78388214e+00 -2.65199393e-01
1.12438291e-01 -5.63168883e-01 -1.04778990e-01 -6.80030346e-01
-1.33355224e+00 7.70937622e-01 6.49560869e-01 -5.90449087e-02
1.48186100e+00 2.31751099e-01 6.89334154e-01 7.58924931e-02
8.97186249e-02 -1.08511496e+00 4.67702478e-01 4.15316850e-01
7.18685806e-01 -1.07732832e+00 -8.29345919e-03 -4.86760646e-01
-7.46502519e-01 7.40237355e-01 1.07611549e+00 6.77272603e-02
4.39575315e-01 1.85277425e-02 3.98528110e-03 -1.03933699e-01
-6.62793279e-01 -3.88809204e-01 3.30483884e-01 3.53952169e-01
4.73245382e-01 -1.51456475e-01 -4.76305008e-01 6.27283752e-01
6.90615118e-01 2.61392415e-01 3.24953467e-01 1.17615533e+00
-4.23347116e-01 -9.61890638e-01 -1.05794609e-01 2.65474796e-01
-7.17146099e-01 -4.15739082e-02 -3.96762013e-01 7.02651262e-01
4.59193677e-01 9.16860282e-01 9.26799104e-02 -3.07011366e-01
2.29792982e-01 1.33749649e-01 3.09112608e-01 -6.13265872e-01
-3.83101314e-01 5.77965021e-01 9.12929997e-02 -8.51065993e-01
-8.96708548e-01 -8.77808869e-01 -1.61817443e+00 1.77355871e-01
-2.05249161e-01 -1.33691400e-01 -9.35054570e-02 9.93247688e-01
2.81408995e-01 7.19798923e-01 3.62519532e-01 -7.30633497e-01
3.08277570e-02 -8.77350807e-01 -3.95972610e-01 5.58569491e-01
1.33217096e-01 -8.19618523e-01 -3.48537982e-01 4.11608100e-01] | [8.543376922607422, 0.6376709342002869] |
b88f12a5-27c5-4f21-b691-83d5135aaa73 | deep-hr-fast-heart-rate-estimation-from-face | 2002.04821 | null | https://arxiv.org/abs/2002.04821v1 | https://arxiv.org/pdf/2002.04821v1.pdf | Deep-HR: Fast Heart Rate Estimation from Face Video Under Realistic Conditions | This paper presents a novel method for remote heart rate (HR) estimation. Recent studies have proved that blood pumping by the heart is highly correlated to the intense color of face pixels, and surprisingly can be utilized for remote HR estimation. Researchers successfully proposed several methods for this task, but making it work in realistic situations is still a challenging problem in computer vision community. Furthermore, learning to solve such a complex task on a dataset with very limited annotated samples is not reasonable. Consequently, researchers do not prefer to use the deep learning approaches for this problem. In this paper, we propose a simple yet efficient approach to benefit the advantages of the Deep Neural Network (DNN) by simplifying HR estimation from a complex task to learning from very correlated representation to HR. Inspired by previous work, we learn a component called Front-End (FE) to provide a discriminative representation of face videos, afterward a light deep regression auto-encoder as Back-End (BE) is learned to map the FE representation to HR. Regression task on the informative representation is simple and could be learned efficiently on limited training samples. Beside of this, to be more accurate and work well on low-quality videos, two deep encoder-decoder networks are trained to refine the output of FE. We also introduce a challenging dataset (HR-D) to show that our method can efficiently work in realistic conditions. Experimental results on HR-D and MAHNOB datasets confirm that our method could run as a real-time method and estimate the average HR better than state-of-the-art ones. | ['Xiaobai Li', 'Mahmood Fathy', 'Mohammad Sabokrou', 'Masoud Pourreza', 'Guoying Zhao'] | 2020-02-12 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [ 3.31166759e-02 5.05583212e-02 -4.89514433e-02 -4.75810409e-01
-4.00383979e-01 2.67863005e-01 1.32890761e-01 -5.68687260e-01
-2.71106362e-01 7.88594186e-01 -1.45942152e-01 1.51961550e-01
3.18992108e-01 -7.19253123e-01 -5.32458603e-01 -8.26885402e-01
4.70832810e-02 -1.59304515e-01 -8.22922811e-02 -6.88498169e-02
1.24143921e-02 6.31101131e-01 -1.65017831e+00 -4.05186377e-02
6.95542753e-01 1.14968669e+00 1.16373822e-01 5.76006293e-01
1.90490857e-01 1.11701453e+00 -5.53052366e-01 -2.83000201e-01
4.51802850e-01 -1.03395700e+00 -4.41559702e-01 -8.96137394e-03
3.88565689e-01 -7.10089028e-01 -4.56532985e-01 6.46984935e-01
1.07720292e+00 1.27070382e-01 4.12609190e-01 -9.26179826e-01
-3.91754836e-01 1.08151548e-01 -7.50908256e-01 6.58715218e-02
1.28525972e-01 1.43566236e-01 4.61458057e-01 -9.13928807e-01
3.19657773e-01 1.04298663e+00 5.98177671e-01 8.14932108e-01
-9.20355499e-01 -8.70930016e-01 -1.96502298e-01 3.51962566e-01
-1.56598103e+00 -7.50193954e-01 1.13988364e+00 -1.57327443e-01
2.68138230e-01 2.23334461e-01 7.99204469e-01 1.12143683e+00
-7.04823211e-02 5.13119936e-01 1.25045168e+00 -2.77119517e-01
4.92788330e-02 3.73414218e-01 -4.40814734e-01 9.28162575e-01
6.08602120e-03 6.17862940e-02 -4.99615878e-01 4.74383265e-01
1.09426165e+00 4.51390892e-02 -6.69577479e-01 -8.61668214e-02
-7.75277853e-01 7.38162577e-01 5.59938550e-01 2.84995973e-01
-3.09784144e-01 6.36038110e-02 1.98287353e-01 1.79544508e-01
4.36105877e-01 1.14762418e-01 -2.20493838e-01 -2.08074637e-02
-1.17548525e+00 -1.07350834e-01 8.38375986e-01 5.53774416e-01
7.25607336e-01 2.99790978e-01 -3.85276586e-01 1.06550908e+00
2.31279448e-01 6.09548151e-01 3.57024223e-01 -9.84339952e-01
6.95275068e-02 1.91946015e-01 -7.83141479e-02 -1.03259456e+00
-5.20991027e-01 -6.42450571e-01 -1.25229526e+00 3.52780521e-01
5.24043381e-01 -1.82245210e-01 -5.93706071e-01 1.61491752e+00
3.69202644e-01 4.51513201e-01 -1.05336130e-01 1.32873929e+00
9.63478684e-01 6.84940159e-01 -1.25934526e-01 -7.24591553e-01
1.16591680e+00 -9.80886877e-01 -7.49410212e-01 -5.39751463e-02
1.52963236e-01 -6.17038548e-01 1.03023982e+00 6.00461245e-01
-1.02031362e+00 -1.00860262e+00 -9.74950016e-01 -2.15853319e-01
9.14819613e-02 4.72460121e-01 6.07695639e-01 6.31724596e-01
-9.25711274e-01 7.27445662e-01 -5.74148118e-01 -3.98399323e-01
5.44784606e-01 5.22858873e-02 -2.49195889e-01 -1.82484612e-01
-1.16327524e+00 8.93307805e-01 -2.02384824e-03 7.15060890e-01
-1.09318018e+00 -4.42032397e-01 -6.74984574e-01 -2.35323049e-02
2.21622691e-01 -6.71526253e-01 8.01085174e-01 -1.16245723e+00
-2.08663440e+00 7.68955410e-01 -5.94895147e-02 -1.97623745e-01
5.39457381e-01 -3.59138787e-01 -2.77965426e-01 4.81252998e-01
-4.67183292e-01 5.84598601e-01 1.10863900e+00 -1.23734987e+00
-9.62660983e-02 -2.53227293e-01 -4.26869616e-02 1.84531763e-01
-5.06425500e-01 -2.76716594e-02 -6.30902290e-01 -2.82267749e-01
-1.43344954e-01 -7.09185064e-01 7.88167305e-03 4.27045166e-01
-1.82822406e-01 2.80760862e-02 5.03883779e-01 -9.90207076e-01
1.13654613e+00 -2.01985168e+00 5.11338860e-02 -1.48874700e-01
6.48693368e-02 4.63756979e-01 -7.52078444e-02 3.44089009e-02
-5.89323267e-02 -1.49163380e-01 -2.23026186e-01 -3.15562546e-01
-3.01835179e-01 1.07633993e-01 -7.35535845e-02 8.36514592e-01
2.33831957e-01 6.58406079e-01 -6.48561716e-01 -7.79118896e-01
4.60459292e-01 1.06850600e+00 -3.53167504e-01 8.39225829e-01
2.93377042e-01 8.39225173e-01 -1.42827407e-01 6.55338943e-01
7.64799953e-01 -5.18232211e-02 1.40248194e-01 -5.24513781e-01
2.95844972e-02 -3.74596864e-01 -9.51388419e-01 1.55140018e+00
-8.30463886e-01 6.43793881e-01 -4.61146943e-02 -1.26297724e+00
1.37983906e+00 3.24477077e-01 5.82817197e-01 -1.02480662e+00
1.38936028e-01 2.91185100e-02 -7.10295066e-02 -9.68565702e-01
-1.79838881e-01 -4.61545914e-01 5.59249580e-01 -1.97300385e-03
1.12713441e-01 -3.67470942e-02 -1.32534251e-01 -4.06422704e-01
7.22413182e-01 5.93422174e-01 3.23178768e-01 1.29751638e-01
8.04917097e-01 -7.49228656e-01 8.36125135e-01 3.64285707e-01
-3.47056836e-01 8.72505307e-01 4.21082318e-01 -5.85083485e-01
-7.83410192e-01 -6.99040890e-01 -3.22373658e-01 7.71436274e-01
1.61545113e-01 -1.96371615e-01 -7.75424242e-01 -5.94335139e-01
-4.65115637e-01 3.26281399e-01 -5.94150245e-01 -7.45485127e-02
-6.29478931e-01 -8.82711947e-01 3.55862826e-01 3.95187616e-01
8.86904776e-01 -1.16767466e+00 -9.02490914e-01 2.45140512e-02
-3.12302738e-01 -1.15303147e+00 -5.21786511e-02 -5.63425012e-03
-8.01732898e-01 -1.09525478e+00 -1.07455683e+00 -5.03560185e-01
4.34776187e-01 1.58010229e-01 1.13409662e+00 2.61786044e-01
-7.10428715e-01 -4.70753349e-02 -3.65808368e-01 -3.20252150e-01
1.82907172e-02 -3.51230770e-01 -3.21579427e-01 3.15138012e-01
1.83351785e-01 -5.38234472e-01 -1.08028913e+00 3.26328158e-01
-6.38108492e-01 1.72005594e-01 6.63474798e-01 7.96018481e-01
4.79885578e-01 -1.62125882e-02 7.34282076e-01 -6.96548998e-01
-1.08182378e-01 -1.62206694e-01 -5.18173873e-01 1.92035124e-01
-4.99238521e-01 -1.62014976e-01 8.16222191e-01 -1.57042757e-01
-1.19820917e+00 3.50536376e-01 -3.70070428e-01 -7.15092957e-01
-1.35685712e-01 1.60031855e-01 -1.81506947e-01 -1.03202313e-01
5.36129594e-01 2.95475066e-01 2.25000873e-01 -4.24490839e-01
1.84770584e-01 7.17732072e-01 4.26469266e-01 -2.74328381e-01
8.42064381e-01 4.90443528e-01 4.80854034e-01 -1.01477230e+00
-1.07253826e+00 -2.61602044e-01 -5.84348619e-01 -6.19636297e-01
1.00248611e+00 -1.27776039e+00 -8.52288127e-01 3.47265631e-01
-9.03487027e-01 -5.78946233e-01 -8.09753612e-02 6.93949103e-01
-6.50666237e-01 4.92499471e-01 -6.67791843e-01 -1.10330606e+00
-4.33806598e-01 -8.26016486e-01 9.15901601e-01 6.49210989e-01
3.48304421e-01 -1.01342642e+00 8.53844807e-02 4.99520719e-01
6.88858271e-01 5.21282315e-01 3.32303703e-01 3.35880704e-02
-4.58374947e-01 1.40208423e-01 -3.65631700e-01 6.89141870e-01
-1.42939156e-02 1.75658822e-01 -1.55205750e+00 -2.76856929e-01
3.32476586e-01 -5.43726325e-01 1.02129412e+00 4.25632596e-01
1.77924025e+00 -3.02506648e-02 5.32908924e-02 9.52009499e-01
1.56416917e+00 -1.28994510e-01 1.17189062e+00 -8.10693800e-02
7.27663934e-01 7.50072956e-01 8.06053698e-01 5.64489722e-01
2.82556057e-01 8.64417672e-01 5.18304348e-01 -7.09957719e-01
-4.84610915e-01 1.24062449e-02 7.01245546e-01 7.42317319e-01
-7.22733200e-01 1.37591027e-02 -2.71139175e-01 1.46667694e-03
-1.52843285e+00 -1.19410944e+00 -7.68641159e-02 2.36060476e+00
9.76492167e-01 -3.00222486e-01 2.64882773e-01 2.29455560e-01
5.33544123e-01 8.80988464e-02 -3.60158384e-01 -3.08429152e-01
6.18709065e-02 4.66286063e-01 1.52258635e-01 1.86252236e-01
-1.00551426e+00 5.65653741e-01 5.39033699e+00 6.93817377e-01
-1.54199314e+00 1.85069472e-01 1.00927913e+00 -4.74992692e-02
1.07304156e-01 -2.72466868e-01 -7.13793874e-01 4.34105963e-01
1.03419173e+00 3.86307657e-01 4.88892078e-01 8.04156959e-01
4.71432447e-01 -1.75791636e-01 -1.08901227e+00 1.48746085e+00
5.01217127e-01 -6.66327000e-01 -4.45359766e-01 -1.31364152e-01
3.79304349e-01 -4.35473144e-01 -9.36318636e-02 5.11498809e-01
-5.22575438e-01 -1.28960943e+00 2.34354541e-01 9.30464506e-01
1.06899738e+00 -7.28659749e-01 9.10493672e-01 1.48489460e-01
-1.10320950e+00 -1.10786945e-01 -8.67928863e-01 1.11901894e-01
-3.92206237e-02 9.46512401e-01 -5.57866693e-01 5.58563113e-01
7.75808930e-01 9.56976056e-01 -7.00699210e-01 1.00237513e+00
-4.04119909e-01 5.17014205e-01 2.86707119e-03 3.58442456e-01
-3.99670810e-01 -3.24055612e-01 -1.73912086e-02 1.14587653e+00
4.66512471e-01 6.95232153e-02 -2.25615241e-02 8.39206636e-01
-6.67400062e-02 2.20671296e-01 -5.74600160e-01 2.22966298e-01
-1.20827593e-01 1.92702484e+00 -6.05122685e-01 -1.52114913e-01
-5.16219437e-01 1.07341957e+00 3.05112064e-01 3.33547443e-01
-1.28319848e+00 -4.78032202e-01 2.02492744e-01 1.28997311e-01
2.16231138e-01 1.00604601e-01 1.04851596e-01 -1.22805929e+00
-3.87715846e-02 -6.86845183e-01 3.02595109e-01 -9.71878290e-01
-1.09263921e+00 8.82270753e-01 -4.23790544e-01 -1.42214298e+00
-6.70770779e-02 -4.37328845e-01 -5.19872487e-01 9.68202651e-01
-2.19308257e+00 -1.01388812e+00 -1.01213396e+00 8.09150517e-01
5.53618789e-01 7.86869004e-02 8.07253420e-01 6.43932700e-01
-1.04318988e+00 6.45213366e-01 -4.12303329e-01 2.13348687e-01
1.09853792e+00 -1.04982293e+00 -5.39601088e-01 8.74304473e-01
7.72000179e-02 3.01403046e-01 4.60909605e-01 -1.58453077e-01
-1.49546611e+00 -1.11385596e+00 5.66666842e-01 -1.17874168e-01
6.79784501e-03 -3.38181227e-01 -8.17474186e-01 2.20073923e-01
2.62641430e-01 6.52260065e-01 6.31603122e-01 -9.03040692e-02
-1.12133272e-01 -7.64266729e-01 -1.01354408e+00 2.60730088e-01
8.98961544e-01 -4.24890906e-01 -1.16751730e-01 2.32196718e-01
2.33717218e-01 -2.96872467e-01 -1.02519667e+00 4.58926886e-01
5.92382908e-01 -1.41911161e+00 9.62294102e-01 -3.44209606e-03
7.18896210e-01 -4.51308757e-01 1.96763743e-02 -1.13357759e+00
2.68582180e-02 -5.87596118e-01 -4.12058562e-01 1.18368542e+00
-9.44727138e-02 -2.49727964e-01 6.23138607e-01 3.94701630e-01
-4.73012514e-02 -8.84088218e-01 -7.00040162e-01 -5.72140157e-01
-2.82195121e-01 -2.51213945e-02 2.00400233e-01 7.91976631e-01
-2.92408735e-01 4.37686682e-01 -1.00216615e+00 3.98223028e-02
6.95582092e-01 3.38506073e-01 9.86277938e-01 -1.17606628e+00
-3.90318811e-01 -1.02582015e-01 -2.24910185e-01 -9.54915345e-01
1.25805348e-01 -5.79275489e-01 1.95028827e-01 -1.38160443e+00
3.78934145e-01 -3.39323312e-01 -3.69663805e-01 4.25993085e-01
-3.14280927e-01 7.12291777e-01 2.20140949e-01 4.27746996e-02
-4.30503190e-01 7.22393274e-01 1.57850516e+00 1.47006556e-01
-1.52170435e-01 -1.41878314e-02 -4.90948349e-01 4.52735722e-01
6.93259239e-01 -2.15750799e-01 -4.17258322e-01 -2.24463418e-02
7.92531222e-02 5.26093781e-01 5.08995891e-01 -1.28104961e+00
-2.16809232e-02 -1.29652575e-01 8.92812192e-01 -2.49030799e-01
4.95441258e-01 -9.23602223e-01 9.79105756e-02 5.06343603e-01
-2.57850379e-01 -6.08162344e-01 -1.93356827e-01 2.61537403e-01
-2.78351873e-01 -8.21079239e-02 1.22630548e+00 -1.59067288e-01
-5.40601194e-01 5.42609513e-01 2.09254578e-01 -6.37537912e-02
1.00871122e+00 -2.51703084e-01 -2.24426821e-01 -4.84974116e-01
-5.74188173e-01 -5.62558249e-02 2.30169609e-01 1.10175505e-01
7.59098887e-01 -1.14059484e+00 -8.52340162e-01 2.39977717e-01
-3.84442993e-02 -4.20202427e-02 4.85321581e-01 1.12927604e+00
-6.55634880e-01 2.69259680e-02 -5.13015449e-01 -5.85434020e-01
-1.05400586e+00 7.94975519e-01 6.31086826e-01 6.73320219e-02
-7.77912974e-01 6.73719764e-01 1.93399131e-01 3.86978388e-02
3.49193424e-01 -3.18369865e-02 -6.20102942e-01 1.24005780e-01
7.95219958e-01 2.83456326e-01 -7.66993687e-02 -5.26425004e-01
-2.48215452e-01 9.49908376e-01 5.22065341e-01 3.78733993e-01
1.53733122e+00 -3.80155087e-01 -7.80677050e-02 4.34279948e-01
1.36619473e+00 -1.92986764e-02 -1.54891026e+00 4.19326313e-02
-5.97022653e-01 -5.87082028e-01 1.54071122e-01 -6.61962092e-01
-1.85660172e+00 1.40388811e+00 1.00847042e+00 -1.58266693e-01
1.66261864e+00 -3.36727262e-01 6.44022644e-01 2.17222571e-01
8.50588009e-02 -1.10553777e+00 3.73298973e-01 -1.18473589e-01
9.27926958e-01 -1.53293824e+00 1.42642736e-01 -4.64046180e-01
-6.97373629e-01 1.69443130e+00 8.65171134e-01 7.41137043e-02
5.00863254e-01 1.07289935e-02 1.98071629e-01 -3.72309907e-04
-5.16357720e-01 -3.60641152e-01 1.84228942e-01 5.83964825e-01
6.85718060e-01 -3.06728870e-01 -1.98234171e-01 3.48367602e-01
1.86537310e-01 4.59627002e-01 6.55216932e-01 3.07963729e-01
-3.54042977e-01 -7.59211957e-01 -1.23702466e-01 1.06075406e-01
-6.37526333e-01 1.63759515e-01 2.69515932e-01 7.43512034e-01
1.84162825e-01 9.42463398e-01 -6.55935705e-02 -2.70265937e-01
1.47359461e-01 -7.13522285e-02 6.16841316e-01 -3.53614211e-01
-3.49396586e-01 2.55703807e-01 -9.81072336e-02 -9.38985944e-01
-8.08265626e-01 -2.11345077e-01 -9.83804584e-01 -2.31977195e-01
-1.31172448e-01 -1.23448081e-01 6.82401836e-01 6.69395566e-01
-2.95531135e-02 6.24905765e-01 1.21861732e+00 -8.27009141e-01
-3.63353610e-01 -9.09294367e-01 -8.99069548e-01 5.03595293e-01
2.64608711e-01 -6.63297236e-01 -5.03768802e-01 1.28363803e-01] | [13.890766143798828, 2.694209098815918] |
4a31e987-34d4-4c95-84ee-6c6be365bba3 | topic-modeling-based-sentiment-analysis-on | null | null | https://aclanthology.org/P15-1131 | https://aclanthology.org/P15-1131.pdf | Topic Modeling based Sentiment Analysis on Social Media for Stock Market Prediction | null | ['Thien Hai Nguyen', 'Kiyoaki Shirai'] | 2015-07-01 | topic-modeling-based-sentiment-analysis-on-1 | https://aclanthology.org/P15-1131 | https://aclanthology.org/P15-1131.pdf | ijcnlp-2015-7 | ['stock-market-prediction', 'stock-prediction'] | ['time-series', 'time-series'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.359321117401123, 3.7348501682281494] |
687c4bee-82e4-47ce-a165-f8ba42b222bd | g2netpl-generic-game-theoretic-network-for | 2210.11469 | null | https://arxiv.org/abs/2210.11469v1 | https://arxiv.org/pdf/2210.11469v1.pdf | G2NetPL: Generic Game-Theoretic Network for Partial-Label Image Classification | Multi-label image classification aims to predict all possible labels in an image. It is usually formulated as a partial-label learning problem, since it could be expensive in practice to annotate all the labels in every training image. Existing works on partial-label learning focus on the case where each training image is labeled with only a subset of its positive/negative labels. To effectively address partial-label classification, this paper proposes an end-to-end Generic Game-theoretic Network (G2NetPL) for partial-label learning, which can be applied to most partial-label settings, including a very challenging, but annotation-efficient case where only a subset of the training images are labeled, each with only one positive label, while the rest of the training images remain unlabeled. In G2NetPL, each unobserved label is associated with a soft pseudo label, which, together with the network, formulates a two-player non-zero-sum non-cooperative game. The objective of the network is to minimize the loss function with given pseudo labels, while the pseudo labels will seek convergence to 1 (positive) or 0 (negative) with a penalty of deviating from the predicted labels determined by the network. In addition, we introduce a confidence-aware scheduler into the loss of the network to adaptively perform easy-to-hard learning for different labels. Extensive experiments demonstrate that our proposed G2NetPL outperforms many state-of-the-art multi-label classification methods under various partial-label settings on three different datasets. | ['Song Wang', 'XiaoFeng Wang', 'Mostafa M. Fouda', 'Xin Zhang', 'Rabab Abdelfattah'] | 2022-10-20 | null | null | null | null | ['multi-label-image-classification', 'partial-label-learning'] | ['computer-vision', 'methodology'] | [ 5.02097726e-01 4.50523227e-01 -6.40600801e-01 -4.78209764e-01
-8.92503023e-01 -5.69570422e-01 7.49219134e-02 5.87929003e-02
-4.38730955e-01 7.47537553e-01 -7.56189167e-01 -1.31829187e-01
-1.44689083e-01 -5.40883601e-01 -4.75257427e-01 -1.12236977e+00
4.64209393e-02 7.64991879e-01 -1.55937998e-02 2.48626113e-01
-1.91513285e-01 -1.68335378e-01 -1.47493613e+00 1.91286370e-01
5.95695376e-01 1.40815306e+00 2.56873071e-01 5.09224057e-01
1.79968774e-01 1.31700277e+00 -3.89959514e-01 -5.32069743e-01
4.07132149e-01 -4.88802552e-01 -9.97723281e-01 4.81205404e-01
1.53050601e-01 3.59172635e-02 2.55147517e-01 1.46343029e+00
4.41948771e-01 -9.00802575e-03 7.75160730e-01 -1.70739961e+00
-3.77965450e-01 5.88161349e-01 -9.53604639e-01 -4.38091069e-01
-1.33659855e-01 -1.82449654e-01 1.44683993e+00 -5.52679241e-01
4.71584052e-01 1.06366670e+00 5.86448789e-01 8.78540933e-01
-1.13638735e+00 -8.95556033e-01 3.73125702e-01 3.06377541e-02
-1.55659688e+00 6.99742138e-02 7.93361545e-01 -4.07125860e-01
1.38162136e-01 8.08181316e-02 2.41676211e-01 6.22663796e-01
-6.70610070e-02 1.02253902e+00 1.27378368e+00 -5.37606120e-01
3.44417810e-01 2.86327064e-01 2.86524370e-02 9.07695770e-01
-2.83681154e-01 -3.68748009e-02 -1.85409516e-01 -2.34684601e-01
3.29354465e-01 -9.62502733e-02 -2.00031430e-01 -6.67269647e-01
-8.42682660e-01 1.00871289e+00 1.84539467e-01 -9.33149904e-02
-3.07221651e-01 2.90961444e-01 4.66251582e-01 4.65078592e-01
9.03372467e-01 9.09771919e-02 -7.31334686e-01 2.58426249e-01
-6.76524818e-01 -6.27857819e-02 6.56889141e-01 9.54688966e-01
1.08534527e+00 -4.82318401e-01 -2.03069031e-01 1.12152457e+00
4.00101155e-01 3.60079110e-02 2.18291983e-01 -1.21025586e+00
2.46361271e-01 6.26078129e-01 9.98766795e-02 -6.44037306e-01
-6.74424887e-01 -6.78236842e-01 -9.12748277e-01 3.48258883e-01
4.08977449e-01 -4.29085314e-01 -7.21742332e-01 2.10102439e+00
5.18209279e-01 4.73958373e-01 -4.30389568e-02 9.52402592e-01
5.77470422e-01 4.50622022e-01 3.88637215e-01 -6.67664587e-01
1.24426103e+00 -1.30416393e+00 -3.92740875e-01 -3.64086628e-01
1.05094743e+00 -4.97126818e-01 8.49951386e-01 4.31502312e-01
-8.31105769e-01 -3.61227036e-01 -8.47007096e-01 4.15862978e-01
6.39990866e-02 3.91759366e-01 3.65559220e-01 7.90368736e-01
-1.04662037e+00 3.21905673e-01 -1.49347886e-01 1.10900970e-02
5.47513008e-01 6.46220028e-01 -2.49783516e-01 -4.32072163e-01
-1.04805267e+00 4.47642207e-01 7.10885346e-01 -1.91042945e-01
-9.51429725e-01 -4.08963561e-01 -7.60254145e-01 -6.60972074e-02
1.08030319e+00 -4.09785569e-01 1.44980681e+00 -1.45731974e+00
-1.10745156e+00 1.41627586e+00 2.38881215e-01 -3.38800885e-02
6.10829294e-01 6.12967730e-01 -9.06122401e-02 1.03221841e-01
5.11493266e-01 1.17917705e+00 7.18747854e-01 -1.55035877e+00
-1.27333224e+00 -1.24575883e-01 3.74377728e-01 5.51640093e-01
-2.24222988e-01 -2.83437856e-02 -4.77078944e-01 -4.70282733e-01
2.19219476e-02 -1.30745268e+00 -3.67817909e-01 1.13774456e-01
-6.03988111e-01 -6.30529463e-01 7.19783127e-01 -7.34226331e-02
7.90662706e-01 -2.09253979e+00 6.88257208e-03 5.06543778e-02
3.31370294e-01 1.54502904e-02 -3.89735550e-01 -1.19373284e-01
-1.56455532e-01 2.80919939e-01 -8.30596909e-02 -8.40458632e-01
-6.73913285e-02 3.77774924e-01 2.70168126e-01 6.46650374e-01
3.16161439e-02 7.51636505e-01 -1.07610881e+00 -7.81618297e-01
5.03393412e-02 -6.32827282e-02 -3.65736783e-01 1.69409871e-01
-3.93086970e-01 6.28429711e-01 -3.99215072e-01 8.23508382e-01
7.84810245e-01 -7.62500763e-01 4.00199205e-01 2.11080492e-01
3.26254278e-01 -3.75514179e-01 -1.31792700e+00 1.31993234e+00
-4.11693096e-01 -1.33905811e-02 2.64343947e-01 -1.38835204e+00
5.78454614e-01 5.48499703e-01 8.06437254e-01 -4.16278362e-01
3.93869162e-01 2.08176449e-01 -3.00283343e-01 -1.54543072e-01
4.98938300e-02 -4.55102146e-01 -2.94208467e-01 7.49621034e-01
2.49347389e-01 1.97107300e-01 2.28984296e-01 2.51423214e-02
8.60267580e-01 -1.73958853e-01 4.10697699e-01 2.30988935e-02
5.17493784e-01 -2.77823925e-01 8.83843780e-01 9.63486195e-01
-4.81042296e-01 4.40111309e-01 8.08621228e-01 -1.94772601e-01
-6.71350420e-01 -4.07169074e-01 -2.18150355e-02 1.57676959e+00
4.54733163e-01 -9.45209432e-03 -6.51582420e-01 -1.28360057e+00
-2.84080148e-01 2.80558378e-01 -6.36336982e-01 -1.45781711e-01
-4.47540283e-02 -8.53817642e-01 2.86964595e-01 -3.27600427e-02
5.73589921e-01 -1.14659250e+00 -2.77515620e-01 2.13166878e-01
-4.52353686e-01 -9.93579865e-01 -6.68282211e-01 8.33433867e-01
-3.65662575e-01 -1.34958065e+00 -6.02419734e-01 -1.29448891e+00
8.31454277e-01 3.57394367e-01 9.43311512e-01 1.49546593e-01
6.47829399e-02 1.75936371e-01 -5.07874370e-01 -9.48790982e-02
-5.25053620e-01 2.26806462e-01 -2.70836323e-01 4.33943123e-01
6.33763596e-02 -1.36412993e-01 -2.65441895e-01 7.74432719e-01
-7.25554228e-01 1.76027820e-01 2.75282472e-01 1.20911443e+00
1.02200925e+00 6.03668630e-01 8.73370171e-01 -1.28958213e+00
4.49893743e-01 -7.20237732e-01 -4.93830681e-01 5.55682838e-01
-6.84588253e-01 -3.24662387e-01 7.29202390e-01 -8.97620916e-01
-6.51023030e-01 5.70274711e-01 2.03957126e-01 -5.56693375e-01
-5.81271574e-02 5.69802046e-01 -2.79678762e-01 -4.54991490e-01
3.06652427e-01 -1.53011069e-01 -3.90516706e-02 -4.83708568e-02
4.16615993e-01 7.77008414e-01 2.17543676e-01 -4.64834839e-01
3.71895730e-01 1.72306016e-01 2.98005551e-01 -3.37769799e-02
-1.58103168e+00 -7.85781622e-01 -5.30999482e-01 -6.97828650e-01
7.35323906e-01 -1.05146039e+00 -1.04404581e+00 7.09120035e-01
-7.32380331e-01 -6.39414847e-01 -4.34300780e-01 2.13452175e-01
-6.87876761e-01 2.61816055e-01 -6.83837116e-01 -8.94971490e-01
-4.10111696e-02 -1.30666006e+00 1.08087468e+00 3.18650514e-01
2.69912601e-01 -1.19119346e+00 -2.27361128e-01 6.47445977e-01
-1.46892026e-01 -2.37103906e-02 9.82076526e-01 -7.64566898e-01
-3.15459132e-01 -2.29193240e-01 -3.07838798e-01 6.54409289e-01
-1.01272864e-02 -6.19556129e-01 -9.30174887e-01 -5.22295058e-01
3.33698690e-02 -1.25868773e+00 7.86005318e-01 4.30909395e-01
1.27663064e+00 -2.73085862e-01 -3.13891560e-01 2.86788046e-01
1.48176658e+00 2.39216313e-01 -4.75737676e-02 1.41905993e-01
6.63427413e-01 6.39455378e-01 1.14016521e+00 6.20933294e-01
6.09532773e-01 6.74584985e-01 1.01748633e+00 -2.32812420e-01
1.36008054e-01 6.32542670e-02 5.31129539e-02 4.28138971e-01
4.54661995e-01 -7.93057740e-01 -5.99079967e-01 2.14476526e-01
-2.18482590e+00 -6.02729201e-01 -5.13293371e-02 2.04093647e+00
9.15927887e-01 2.17563696e-02 1.09399118e-01 1.20173894e-01
1.18289840e+00 2.27509633e-01 -9.42497194e-01 1.18459344e-01
-2.33860135e-01 -1.82155427e-02 6.08702838e-01 2.57230848e-01
-1.74353540e+00 9.59300280e-01 5.49416590e+00 1.24841499e+00
-1.08564639e+00 5.60513079e-01 1.27991426e+00 2.06982787e-03
1.32322282e-01 -2.37829946e-02 -6.96114719e-01 4.03238624e-01
6.02795482e-01 2.24710703e-01 4.76488829e-01 1.04703188e+00
-1.28405586e-01 -3.40573549e-01 -9.85323131e-01 9.46604669e-01
1.40851155e-01 -8.36642683e-01 -4.73701388e-01 1.41060889e-01
1.04070628e+00 -4.29836586e-02 1.65776014e-01 4.91927594e-01
6.74130917e-01 -7.11380303e-01 9.53876734e-01 5.52545972e-02
1.23027563e+00 -9.35653090e-01 8.57570112e-01 8.68844330e-01
-1.17202508e+00 -5.49101532e-01 -3.44391912e-01 -5.33895195e-02
9.95068401e-02 5.55866480e-01 -5.94887197e-01 3.46548706e-01
3.14883679e-01 6.91407621e-01 -3.94950360e-01 8.64458263e-01
-3.61508667e-01 5.93048573e-01 -5.57362325e-02 4.60899621e-02
4.21549559e-01 -9.35365334e-02 -9.79882479e-02 7.30989635e-01
1.31517470e-01 1.60632834e-01 1.19960225e+00 3.30923498e-01
-6.38500750e-01 3.66098613e-01 -1.45831227e-01 3.04613739e-01
4.15494204e-01 1.51373708e+00 -1.05102849e+00 -2.16653422e-01
-4.51155037e-01 9.02406275e-01 5.75475395e-01 1.39871970e-01
-8.02425921e-01 4.09451425e-02 1.82962105e-01 -3.15068305e-01
2.73724888e-02 6.00850582e-01 -1.41871065e-01 -9.24989164e-01
-1.32502764e-01 -5.12615740e-01 7.87270486e-01 -8.17008317e-01
-1.43536687e+00 2.64432073e-01 -2.84561247e-01 -1.35563529e+00
-1.15950212e-01 -3.36362004e-01 -1.47735819e-01 6.98969007e-01
-1.73711824e+00 -1.40513921e+00 -1.86547637e-01 5.11092365e-01
4.05394375e-01 -1.40747726e-01 7.38045096e-01 6.07371211e-01
-5.41377664e-01 8.21740746e-01 9.34733376e-02 3.14522237e-02
7.58476675e-01 -1.30775332e+00 -4.06315833e-01 3.59476507e-01
-6.44327477e-02 -4.27692056e-01 1.65198669e-01 -4.06468868e-01
-3.60621542e-01 -1.63711798e+00 8.40103328e-01 2.46350542e-01
3.81359160e-01 -2.70486236e-01 -5.72276890e-01 6.00409031e-01
-2.23353595e-01 5.89099109e-01 8.36993635e-01 -9.60053727e-02
-1.64131984e-01 2.48243511e-02 -1.33871233e+00 3.53542268e-01
8.43991160e-01 -3.59271646e-01 3.88394713e-01 8.33087444e-01
6.71739757e-01 -3.23084205e-01 -5.96524596e-01 3.94155502e-01
2.02828690e-01 -5.34221828e-01 7.03527629e-01 -2.55293965e-01
2.66977549e-01 -3.19274694e-01 4.11010124e-02 -1.40687895e+00
-5.08312404e-01 -1.47158414e-01 2.68156439e-01 1.15330052e+00
4.60206419e-01 -4.62257087e-01 1.09346831e+00 3.77060324e-01
3.91112901e-02 -1.03132057e+00 -1.19735157e+00 -6.09070539e-01
-2.22972371e-02 -4.11853552e-01 1.45088866e-01 1.22153008e+00
-1.77619755e-01 4.64534879e-01 -9.16736841e-01 1.51691526e-01
7.47915566e-01 4.43413891e-02 1.98845223e-01 -1.49953628e+00
-3.42210144e-01 -3.29735577e-01 -2.95312643e-01 -1.08550751e+00
7.92028368e-01 -1.20600820e+00 3.67381781e-01 -1.09011352e+00
7.55311966e-01 -1.15334690e+00 -5.57500005e-01 1.03966355e+00
-2.14369446e-01 6.21114433e-01 1.99397206e-01 2.84207642e-01
-1.36536896e+00 2.74772137e-01 1.35399270e+00 -4.61076647e-01
-3.74642923e-03 4.04434979e-01 -8.01529586e-01 6.36052191e-01
7.12746501e-01 -1.02265692e+00 -6.65476918e-01 1.29298419e-01
3.58623952e-01 4.59344000e-01 1.41741544e-01 -7.01903582e-01
3.20445955e-01 -4.30710346e-01 -2.46019468e-01 -3.66015196e-01
2.59087831e-01 -8.87335777e-01 9.05192122e-02 3.19833338e-01
-7.22127795e-01 -6.49731338e-01 -4.96384442e-01 7.45598733e-01
-1.09343052e-01 -7.92776585e-01 1.19674969e+00 -2.68008411e-01
-5.65129936e-01 5.49440622e-01 -1.62579909e-01 2.76212782e-01
1.63409376e+00 -4.74690273e-02 -2.05744714e-01 -4.57411200e-01
-9.65839088e-01 7.45750546e-01 3.84523958e-01 1.28421128e-01
1.58409461e-01 -1.36362338e+00 -5.77426672e-01 -1.67285815e-01
3.10164750e-01 1.54331878e-01 5.09891450e-01 5.85555077e-01
-1.98434480e-02 8.37096423e-02 5.26315086e-02 -5.99923968e-01
-1.44944382e+00 8.30993056e-01 5.55978775e-01 -8.36485028e-01
-2.37245820e-02 1.12104785e+00 4.43949461e-01 -8.60300958e-01
5.19491911e-01 1.79025173e-01 -3.81389946e-01 3.34990025e-01
2.09554210e-01 1.77004874e-01 -1.35765404e-01 -7.98213422e-01
-6.20835945e-02 7.06935346e-01 -1.18871279e-01 1.18903644e-01
9.74359989e-01 -3.08354616e-01 -1.96293712e-01 5.40184736e-01
1.32399213e+00 -5.90641320e-01 -1.42028046e+00 -5.79278350e-01
-1.94624647e-01 -1.05116099e-01 3.88825648e-02 -9.55535293e-01
-1.55311835e+00 5.69366932e-01 6.25463188e-01 2.07039714e-01
1.39491165e+00 2.78076291e-01 4.33405489e-01 2.57050902e-01
7.53798544e-01 -1.21726501e+00 3.66049349e-01 4.53307033e-01
1.93285808e-01 -1.69129598e+00 -2.87072778e-01 -6.52276635e-01
-9.15057778e-01 7.62559831e-01 7.43909895e-01 3.37122351e-01
9.39388633e-01 7.16931885e-03 1.48440078e-01 -1.19704917e-01
-8.33194733e-01 -2.00020134e-01 -2.29336604e-01 4.35655415e-01
-2.58674193e-02 4.14935797e-01 -2.57688701e-01 5.58066607e-01
4.22078043e-01 -4.04488519e-02 5.85771024e-01 7.53053248e-01
-3.96450937e-01 -1.32486105e+00 -1.43694669e-01 5.92170358e-01
-6.49227202e-01 1.78829238e-01 -1.52244076e-01 3.71218443e-01
7.49277711e-01 1.26852536e+00 -1.39521658e-01 -4.83685404e-01
-3.06641459e-02 1.77006815e-02 1.29150063e-01 -9.01476860e-01
-5.03820479e-01 2.68878967e-01 -1.58422645e-02 -1.53508410e-01
-8.02599609e-01 -5.62284470e-01 -1.19551277e+00 -4.67362292e-02
-9.44394290e-01 9.17496458e-02 4.18149590e-01 9.26515162e-01
-3.83984856e-02 4.87302721e-01 1.03729415e+00 -6.77495480e-01
-5.69835246e-01 -8.54080915e-01 -1.17976880e+00 3.90706599e-01
1.71734735e-01 -8.08327973e-01 -3.91744971e-01 -3.36138308e-02] | [9.50788688659668, 4.115539073944092] |
3a9a6285-4d93-4b2a-aa5d-f02baa509da0 | diachronic-sense-modeling-with-deep | null | null | https://aclanthology.org/P19-1379 | https://aclanthology.org/P19-1379.pdf | Diachronic Sense Modeling with Deep Contextualized Word Embeddings: An Ecological View | Diachronic word embeddings have been widely used in detecting temporal changes. However, existing methods face the meaning conflation deficiency by representing a word as a single vector at each time period. To address this issue, this paper proposes a sense representation and tracking framework based on deep contextualized embeddings, aiming at answering not only what and when, but also how the word meaning changes. The experiments show that our framework is effective in representing fine-grained word senses, and it brings a significant improvement in word change detection task. Furthermore, we model the word change from an ecological viewpoint, and sketch two interesting sense behaviors in the process of language evolution, i.e. sense competition and sense cooperation. | ['Renfen Hu', 'Shichen Liang', 'Shen Li'] | 2019-07-01 | null | null | null | acl-2019-7 | ['diachronic-word-embeddings'] | ['natural-language-processing'] | [ 8.27973112e-02 -7.26642609e-01 -1.86337471e-01 -9.72817391e-02
2.13143677e-01 -7.01076388e-01 7.03430593e-01 7.01264858e-01
-6.00112021e-01 3.98753226e-01 6.65647924e-01 -1.32622331e-01
1.90342851e-02 -8.90436769e-01 -1.11386076e-01 -6.92299902e-01
2.23589957e-01 -7.83051029e-02 2.63335168e-01 -8.07942867e-01
3.17481816e-01 4.97882515e-02 -1.46285403e+00 -2.42706493e-01
8.47025037e-01 5.79334557e-01 4.10866439e-01 1.70471698e-01
-8.00625443e-01 4.54385281e-02 -7.61726797e-01 -2.41014138e-01
-2.32747823e-01 -3.35280865e-01 -5.56488156e-01 -1.19736344e-01
-2.06505954e-01 2.85181135e-01 -2.17457250e-01 1.35133624e+00
4.20048952e-01 1.66796684e-01 3.98148984e-01 -1.04409969e+00
-1.45160234e+00 7.33113825e-01 -5.84198356e-01 5.91964245e-01
5.62570035e-01 1.02439404e-01 1.30300677e+00 -7.60152519e-01
5.86465299e-01 1.42663395e+00 5.77863574e-01 4.75497633e-01
-9.42257166e-01 -4.99987274e-01 7.81623483e-01 2.49236405e-01
-1.29085350e+00 3.70727032e-02 8.66597414e-01 -5.95072329e-01
9.53232765e-01 2.31817424e-01 1.28401625e+00 1.43763876e+00
4.37778175e-01 7.18583822e-01 8.64465833e-01 -1.68535501e-01
2.43045345e-01 -3.12236339e-01 6.32940471e-01 1.47059426e-01
5.69793284e-01 -1.12755373e-01 -4.95722979e-01 -3.08363289e-01
4.86343265e-01 4.48749483e-01 -2.46828422e-01 2.54325867e-02
-1.28253746e+00 8.48519504e-01 4.00581568e-01 1.07591820e+00
-4.70457375e-01 2.09944502e-01 5.17886400e-01 3.00024092e-01
8.02531481e-01 4.49059874e-01 -6.91452980e-01 -4.11428958e-01
-3.90116811e-01 1.37812376e-01 1.56501323e-01 3.91502023e-01
5.96087992e-01 7.69957006e-02 -1.54299125e-01 9.63642299e-01
4.72011149e-01 4.60174233e-01 1.20985663e+00 -1.42197549e-01
-2.59179682e-01 9.34556007e-01 1.21857367e-01 -1.33048344e+00
-3.46893102e-01 -3.93821716e-01 -7.30747819e-01 -3.03707987e-01
-1.87320113e-01 2.42965482e-02 -8.25784504e-01 2.29625988e+00
5.18346250e-01 3.84571016e-01 2.81751249e-02 6.93677604e-01
4.83151913e-01 7.05394983e-01 3.01378399e-01 -4.76222634e-01
1.72238541e+00 -3.16017598e-01 -1.13318694e+00 -4.32495445e-01
5.09759009e-01 -6.09729052e-01 1.41184473e+00 5.49892820e-02
-5.07851601e-01 -5.02531946e-01 -9.30756032e-01 8.65429565e-02
-9.33136106e-01 -5.77855349e-01 7.21385241e-01 2.85291642e-01
-9.54236627e-01 1.30818903e-01 -5.92882812e-01 -8.46272409e-01
-4.09324542e-02 -1.82079509e-01 -1.44327343e-01 2.79064089e-01
-1.76974940e+00 7.53956795e-01 5.12141049e-01 3.13158594e-02
-2.77593881e-01 -6.43674314e-01 -8.56567621e-01 7.47871697e-02
2.42925122e-01 -6.53037012e-01 1.01209950e+00 -8.88572335e-01
-9.98579979e-01 8.10646415e-01 -3.35251451e-01 -7.81680048e-02
-5.57528436e-02 -1.71547174e-01 -1.02570963e+00 -5.28809130e-01
3.09535891e-01 2.44242892e-01 4.07401443e-01 -1.08669341e+00
-6.07881069e-01 -5.71328223e-01 1.34354923e-02 5.19527569e-02
-9.32554543e-01 3.65713313e-02 -2.00976789e-01 -1.16652298e+00
-3.49263996e-02 -6.33005261e-01 -3.22847456e-01 -2.19566002e-02
-9.86186936e-02 -8.20241928e-01 7.59541154e-01 -3.04575443e-01
1.81471086e+00 -2.33023190e+00 2.18993381e-01 -2.25116506e-01
3.44924122e-01 4.39513147e-01 -3.53526831e-01 8.40078056e-01
-8.76265764e-03 5.48109531e-01 -5.04818499e-01 -5.92481121e-02
-3.89073491e-02 5.96906662e-01 -4.20731843e-01 2.06373155e-01
8.63263011e-02 8.48403454e-01 -1.41685390e+00 -1.27861634e-01
1.72803868e-02 4.62083429e-01 -2.78348327e-01 5.06268367e-02
-1.98461547e-01 -2.13104282e-02 -6.88100278e-01 5.11971414e-01
5.25710166e-01 -1.18721768e-01 4.44955856e-01 3.38632278e-02
-4.33052152e-01 4.16862406e-02 -1.04397678e+00 1.77808893e+00
-4.06812370e-01 6.65836513e-01 -2.26538092e-01 -8.67242277e-01
9.55998182e-01 2.83748582e-02 4.31018084e-01 -6.20277345e-01
1.61479227e-02 -6.16596974e-02 5.00959270e-02 -5.69534123e-01
7.29588389e-01 -3.28259021e-01 -4.39043552e-01 5.70736587e-01
-2.56877720e-01 1.09183699e-01 2.98929885e-02 1.31659761e-01
1.07853150e+00 -4.01401341e-01 7.99853086e-01 -3.95302534e-01
3.48610967e-01 -9.59138796e-02 1.01609766e+00 2.85485029e-01
-4.67042208e-01 5.37107326e-02 3.81128520e-01 -5.72493970e-01
-5.79324007e-01 -1.15325999e+00 -9.71568376e-02 1.25934422e+00
6.89363360e-01 -6.94738269e-01 -2.59833902e-01 -3.46791923e-01
2.44108915e-01 8.07411134e-01 -1.00803924e+00 -6.51060164e-01
-3.17269385e-01 -8.51649165e-01 2.96308935e-01 5.68386495e-01
2.79177189e-01 -1.18072891e+00 -5.22177935e-01 4.43695426e-01
-3.29373665e-02 -5.04348636e-01 -7.86638021e-01 -7.71388337e-02
-5.49443483e-01 -9.10919487e-01 -3.13639820e-01 -7.68864870e-01
2.73647994e-01 6.59163654e-01 1.00554454e+00 1.15271047e-01
-4.00518954e-01 2.85875529e-01 -7.84169495e-01 -5.30188620e-01
-6.65518343e-02 -6.55398220e-02 5.35405695e-01 9.77644697e-02
7.08166659e-01 -5.79722404e-01 -6.58866525e-01 -4.29132255e-03
-1.42664659e+00 -4.54382002e-01 2.37168357e-01 8.45729232e-01
4.58162308e-01 3.08204368e-02 6.37763917e-01 -5.65265715e-01
1.48000574e+00 -8.23589087e-01 -9.17211920e-02 4.18857545e-01
-1.07135832e+00 4.69613746e-02 3.10103297e-01 -6.61042750e-01
-1.02302253e+00 -4.50398266e-01 -3.61416757e-01 -6.37692288e-02
2.15952881e-02 7.43405282e-01 1.80751346e-02 5.33976376e-01
3.55826139e-01 5.54508507e-01 -1.85336769e-01 -5.51934421e-01
6.77739263e-01 7.96459675e-01 1.58764228e-01 -3.07429612e-01
6.35018170e-01 3.37761521e-01 -6.45833611e-01 -8.86471152e-01
-6.50690794e-01 -6.60359502e-01 -4.89315063e-01 -5.02503589e-02
8.88562560e-01 -6.94227397e-01 -3.65625143e-01 4.97266829e-01
-1.32192242e+00 1.59568414e-01 -3.70123774e-01 2.25870356e-01
8.02704021e-02 4.51614112e-01 -3.32840174e-01 -7.83298135e-01
-3.85554165e-01 -7.39287138e-01 9.50176060e-01 3.06317657e-01
-4.75956589e-01 -1.40546167e+00 7.88902640e-01 -3.54494512e-01
4.77057964e-01 1.84390351e-01 9.60301518e-01 -6.11429572e-01
1.77955225e-01 7.85502717e-02 6.01934753e-02 -1.33672372e-01
7.62149215e-01 1.31873280e-01 -6.94954991e-01 -2.27049679e-01
-7.61682987e-02 2.01940119e-01 1.02970636e+00 6.21889383e-02
8.46850991e-01 -1.62775293e-01 -5.05663395e-01 3.13468456e-01
1.29677737e+00 4.87181872e-01 2.57322967e-01 3.25788826e-01
4.78389114e-01 3.40823710e-01 6.10277534e-01 6.00040317e-01
4.32210028e-01 5.83917260e-01 4.29666877e-01 3.27200070e-02
6.24830425e-02 -3.58855993e-01 3.48634750e-01 1.23949862e+00
4.80397016e-01 -5.64975500e-01 -1.06355274e+00 9.71155584e-01
-2.03054190e+00 -9.30057824e-01 -5.41235581e-02 1.59142923e+00
1.00617218e+00 1.54294055e-02 -2.76734144e-03 -1.21874139e-01
6.82612300e-01 7.05791473e-01 -8.94687414e-01 -5.70301354e-01
-3.12232763e-01 7.33108539e-03 3.16915661e-02 3.69385242e-01
-7.21402764e-01 1.36044025e+00 6.70960665e+00 6.33376002e-01
-1.26511502e+00 4.15487945e-01 3.59861390e-03 2.28416041e-01
-1.05089629e+00 -1.52302623e-01 -1.66689605e-01 7.41501033e-01
3.95742357e-01 -5.51442742e-01 1.33656815e-01 4.57018793e-01
3.02779496e-01 3.59758288e-01 -4.72293496e-01 1.19121361e+00
1.05386250e-01 -1.14832425e+00 4.91385758e-01 -6.29236400e-02
4.19232517e-01 -1.19365398e-02 -3.91339771e-02 2.38939926e-01
5.11059463e-01 -6.83962166e-01 5.17251313e-01 6.83417916e-01
5.50768673e-01 -5.24446189e-01 6.33629322e-01 1.14832506e-01
-1.65984452e+00 -2.90808473e-02 -4.65952545e-01 -5.02350986e-01
2.20630139e-01 7.87943542e-01 -2.97073483e-01 4.91053790e-01
5.75645149e-01 1.17560399e+00 -5.67804337e-01 5.96246183e-01
-2.75749922e-01 6.43632531e-01 6.51805699e-02 -4.04667288e-01
1.95197597e-01 -2.56128341e-01 5.55371523e-01 1.18637168e+00
4.96178597e-01 6.93275630e-02 2.49014586e-01 8.81878376e-01
8.07241872e-02 2.67491758e-01 -8.77961636e-01 -5.47353327e-01
8.31621230e-01 9.89210069e-01 -6.65788710e-01 -2.11752430e-01
-2.77238488e-01 1.20770800e+00 2.51984507e-01 3.62473637e-01
-5.35812378e-01 -3.63980353e-01 1.79420960e+00 -3.12619537e-01
7.98522010e-02 -4.73411947e-01 -2.44129285e-01 -1.40842950e+00
5.04840165e-03 -2.76931971e-01 3.66667539e-01 -6.01259530e-01
-1.55204213e+00 3.09117675e-01 -2.29241669e-01 -1.08463621e+00
1.05910763e-01 -6.51589155e-01 -1.06549454e+00 5.43152750e-01
-1.36816037e+00 -9.10917222e-01 -1.07551821e-01 3.95321280e-01
7.64329433e-01 -3.88221852e-02 8.62782419e-01 5.14665693e-02
-8.11691642e-01 6.00087345e-01 2.28167728e-01 -1.61546394e-01
5.45447886e-01 -1.20943308e+00 6.93206012e-01 7.96970963e-01
3.20937127e-01 1.12831402e+00 8.30137074e-01 -7.46051788e-01
-1.39288056e+00 -9.36930537e-01 1.06749284e+00 -2.45298758e-01
1.12902331e+00 -4.57639337e-01 -1.03444874e+00 3.81995678e-01
3.73475432e-01 -4.18066263e-01 1.01021206e+00 4.14176404e-01
-4.66942310e-01 -8.05220753e-02 -1.00059330e+00 9.25544679e-01
1.47694218e+00 -5.65441906e-01 -1.08797288e+00 1.50152877e-01
1.48402786e+00 4.71894026e-01 -7.64064610e-01 3.84976119e-02
5.41005373e-01 -4.96065557e-01 8.99230421e-01 -8.28115880e-01
3.95382196e-02 -4.25368756e-01 -3.15614700e-01 -1.82325780e+00
-7.02646255e-01 -3.55760545e-01 6.10861331e-02 1.52867770e+00
-2.03973521e-02 -1.07852268e+00 1.32123278e-02 -3.23940873e-01
3.43186073e-02 -6.18329525e-01 -1.09027350e+00 -6.63647473e-01
2.82768995e-01 -4.15520519e-01 1.09888887e+00 1.59261560e+00
1.24729887e-01 4.85312700e-01 -1.87275067e-01 3.35685462e-02
5.82913458e-02 2.07916632e-01 1.24685422e-01 -1.53169143e+00
1.71937466e-01 -9.25075233e-01 -6.20285928e-01 -9.00270522e-01
3.53157312e-01 -7.84546435e-01 -5.42599037e-02 -1.61139476e+00
2.00356558e-01 3.14183608e-02 -9.84330416e-01 4.19230908e-01
-8.01482856e-01 -1.83078781e-01 5.68088628e-02 1.71983302e-01
-4.53609914e-01 8.17479908e-01 9.40094233e-01 -3.51701826e-01
-6.38488755e-02 -6.31367564e-01 -1.10490108e+00 5.91414988e-01
9.75868046e-01 -2.31913328e-01 -4.78368074e-01 -6.46802187e-01
6.02156281e-01 -5.31690359e-01 1.00261591e-01 -4.36190665e-01
-1.73572488e-02 -7.09685326e-01 2.50806864e-02 -2.17199966e-01
8.12022835e-02 -7.38151610e-01 8.38151202e-03 8.07423472e-01
-3.28751057e-01 6.13174140e-01 1.66587427e-01 9.49348807e-01
-2.71857083e-01 1.24222964e-01 4.17304546e-01 -1.56768411e-02
-1.27628434e+00 3.11434299e-01 -5.70373416e-01 7.97607377e-02
1.00532711e+00 -1.44630566e-01 -1.95243344e-01 -1.07894681e-01
-4.60055739e-01 3.99085909e-01 6.05236650e-01 1.18898213e+00
7.09445953e-01 -1.71061456e+00 -5.33486664e-01 2.17095822e-01
6.84117854e-01 -6.04056835e-01 2.90206671e-01 1.94066793e-01
5.42758405e-02 -2.11763382e-02 -2.30044909e-02 -3.29941183e-01
-1.13777435e+00 8.14151704e-01 2.13568449e-01 -9.65322088e-03
-3.61464113e-01 8.40325892e-01 4.00845259e-01 -5.32944024e-01
-2.44627550e-01 -4.67290133e-01 -5.78706086e-01 5.89050353e-01
6.12909973e-01 1.75424200e-02 -4.26191032e-01 -6.70171738e-01
-7.21796155e-01 9.40839827e-01 1.55712724e-01 1.20407373e-01
1.14741921e+00 -3.74899298e-01 -3.29583794e-01 1.14798903e+00
1.09650159e+00 -4.47064675e-02 -5.77799141e-01 -4.07035291e-01
2.66482294e-01 -5.07968545e-01 -6.22046590e-02 -7.28746712e-01
-8.23739231e-01 7.20711052e-01 9.06996846e-01 7.18872428e-01
9.90367889e-01 1.95235059e-01 1.14478993e+00 2.67154336e-01
2.39202484e-01 -9.89972711e-01 8.87576342e-02 7.13704765e-01
8.72686088e-01 -1.14838028e+00 -4.00247902e-01 1.38659477e-01
-4.69353467e-01 8.91624272e-01 5.93556106e-01 1.24242149e-01
7.86001742e-01 -1.09166697e-01 2.73387104e-01 -4.15460855e-01
-9.07978237e-01 -6.21614993e-01 -4.65176478e-02 5.33117175e-01
6.10442340e-01 5.26041985e-01 -1.03306305e+00 6.11338615e-01
-1.47811189e-01 -3.42497051e-01 3.21382225e-01 8.68778288e-01
-6.71858847e-01 -1.36112201e+00 -6.39786273e-02 3.08434606e-01
-1.16025291e-01 -2.28152171e-01 -9.38587725e-01 5.23770154e-01
4.38364387e-01 1.08466506e+00 2.85133213e-01 -8.43889177e-01
3.87587190e-01 1.42083511e-01 -2.58362025e-01 -7.07495153e-01
-4.74983156e-01 -1.71593830e-01 -3.58691901e-01 -3.54099691e-01
-3.04029554e-01 -6.56753361e-01 -1.23067391e+00 -1.42748892e-01
-1.77199990e-01 2.67673522e-01 4.33646232e-01 8.85409296e-01
5.33590853e-01 7.03614712e-01 7.45130956e-01 -6.80915713e-02
-5.03227115e-01 -1.11725867e+00 -6.61005676e-01 8.64667952e-01
4.26318496e-01 -8.54147673e-01 -2.22876117e-01 -3.86463016e-01] | [10.319664001464844, 8.902417182922363] |
78ea34b8-1faa-4d80-958b-4c4d527b214e | the-ntt-dcase2020-challenge-task-6-system | 2007.00225 | null | https://arxiv.org/abs/2007.00225v1 | https://arxiv.org/pdf/2007.00225v1.pdf | The NTT DCASE2020 Challenge Task 6 system: Automated Audio Captioning with Keywords and Sentence Length Estimation | This technical report describes the system participating to the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 Challenge, Task 6: automated audio captioning. Our submission focuses on solving two indeterminacy problems in automated audio captioning: word selection indeterminacy and sentence length indeterminacy. We simultaneously solve the main caption generation and sub indeterminacy problems by estimating keywords and sentence length through multi-task learning. We tested a simplified model of our submission using the development-testing dataset. Our model achieved 20.7 SPIDEr score where that of the baseline system was 5.4. | ['Yasunori Ohishi', 'Kunio Kashino', 'Yuma Koizumi', 'Daiki Takeuchi', 'Noboru Harada'] | 2020-07-01 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.18552816e-01 -3.36024836e-02 3.42225879e-01 -2.09961399e-01
-2.01330757e+00 -8.35917950e-01 9.81234014e-02 -5.31754754e-02
-2.63166100e-01 8.56754065e-01 4.94464010e-01 -2.08552480e-01
6.12675920e-02 9.53002796e-02 -7.25703001e-01 -2.25984350e-01
-1.22302569e-01 6.91985250e-01 5.11709303e-02 -1.90836206e-01
1.64001226e-01 -1.39461398e-01 -1.73930931e+00 9.36174273e-01
3.99084479e-01 1.16476989e+00 2.09719792e-01 1.57764292e+00
-1.27217779e-02 6.22566402e-01 -1.35123324e+00 -2.06862196e-01
-1.33329928e-01 -2.80629784e-01 -1.04762316e+00 -2.39005104e-01
1.03870034e+00 -1.13056809e-01 7.44432360e-02 6.94289804e-01
1.17625237e+00 -2.03558028e-01 3.70432526e-01 -1.55634522e+00
-2.96780914e-01 1.12326956e+00 -2.98353344e-01 6.15623534e-01
8.58700454e-01 -9.63198394e-02 1.46912479e+00 -1.00371611e+00
2.09320895e-02 1.24268305e+00 7.37286568e-01 6.24271333e-01
-1.07805157e+00 -9.33853507e-01 -8.73550996e-02 3.18036675e-01
-1.56966114e+00 -8.00359249e-01 5.65760791e-01 -5.57696402e-01
1.24623084e+00 7.50779808e-01 1.27084211e-01 1.61811340e+00
-1.37778059e-01 8.16476583e-01 5.05541801e-01 -3.73130739e-01
3.32512483e-02 -3.62652652e-02 4.05973047e-02 5.63948900e-02
-3.09318334e-01 -1.53517053e-01 -9.06271458e-01 -4.63505685e-01
1.09168552e-01 -1.26710689e+00 2.90421713e-02 6.17142677e-01
-1.34909785e+00 4.85502422e-01 -6.81767881e-01 1.73398517e-02
-2.04902157e-01 3.83583099e-01 7.17184842e-01 5.03496885e-01
2.71040350e-01 1.06618214e+00 -7.03159809e-01 -5.84685087e-01
-1.00406528e+00 4.30260837e-01 7.67659843e-01 1.13155830e+00
-2.38148998e-02 5.17397761e-01 -9.03738618e-01 1.17753553e+00
1.37071013e-01 7.89821684e-01 6.27733052e-01 -1.04583228e+00
9.40410793e-01 -4.88179982e-01 3.73540223e-01 -5.87903440e-01
-4.49270636e-01 -5.11937141e-01 -3.15277904e-01 -5.06545246e-01
3.12059343e-01 -7.81010866e-01 -7.41187692e-01 1.82343447e+00
-3.33410829e-01 4.20526773e-01 1.34076504e-02 4.93828565e-01
1.13978493e+00 1.21491730e+00 2.63525486e-01 -3.71661246e-01
1.50890803e+00 -8.76773059e-01 -9.63780880e-01 -4.17340219e-01
3.24711114e-01 -1.33873391e+00 1.07818747e+00 8.00126791e-01
-1.30295944e+00 -7.00399399e-01 -1.18361235e+00 2.17626676e-01
2.08033741e-01 3.94049466e-01 2.23961085e-01 7.98763037e-01
-1.23493385e+00 -1.52290002e-01 -7.15612620e-02 1.64086670e-01
-2.56755501e-01 3.80404234e-01 8.61792266e-02 6.36771858e-01
-1.66143870e+00 5.35212457e-01 5.39140344e-01 -2.64288127e-01
-1.35544491e+00 -8.07941139e-01 -5.22603035e-01 1.29366100e-01
2.81680882e-01 -3.69337618e-01 2.27027130e+00 -5.46424627e-01
-1.40806353e+00 8.12276244e-01 3.39318477e-02 -6.68480158e-01
1.64922938e-01 -5.85368335e-01 -1.10260618e+00 1.67954579e-01
1.55767694e-01 7.29830742e-01 9.17872906e-01 -8.61925066e-01
-7.81280041e-01 2.44078875e-01 -2.71534562e-01 4.11683291e-01
-1.92206323e-01 6.13524497e-01 -2.86212772e-01 -8.87351930e-01
-4.71046627e-01 -7.89328396e-01 2.17688784e-01 -6.99243307e-01
-6.19131029e-01 -3.92394990e-01 6.41121447e-01 -9.91021752e-01
1.71195781e+00 -2.13506293e+00 -2.63230383e-01 -2.63228536e-01
-2.25870073e-01 9.06243399e-02 -6.90716505e-01 1.73229516e-01
-3.62969667e-01 3.92978877e-01 1.83642089e-01 -4.59318846e-01
-6.04378432e-02 -4.16952878e-01 -6.44042611e-01 -1.34174332e-01
4.59129035e-01 4.96182263e-01 -1.01586413e+00 -7.87714601e-01
-1.71672300e-01 3.86920422e-02 -3.88109684e-01 4.08306897e-01
-5.08039415e-01 -1.17405869e-01 -7.82084987e-02 8.87295902e-01
3.17512006e-01 1.86312005e-01 -2.08714753e-01 -2.25579254e-02
-9.95761603e-02 6.52379692e-01 -1.06346512e+00 1.78440475e+00
-6.91053450e-01 8.63793254e-01 3.30238402e-01 -5.43659925e-01
8.96520019e-01 1.06011581e+00 6.57565415e-01 -4.72246319e-01
-7.10070878e-02 3.58961821e-01 -1.72782913e-01 -8.78430367e-01
7.50030816e-01 9.97241959e-02 -6.70787334e-01 4.26396281e-01
3.32529426e-01 -6.40792787e-01 1.28058523e-01 1.87458724e-01
1.08770800e+00 -3.80749404e-01 -4.72512422e-03 -5.35818711e-02
4.05575931e-01 -3.18763524e-01 2.19107911e-01 1.03170788e+00
-4.37363088e-01 1.36615789e+00 3.98712397e-01 -2.15141669e-01
-1.18426239e+00 -1.34118807e+00 1.77878998e-02 1.45389080e+00
-6.06619418e-01 -7.22534776e-01 -6.65175319e-01 -3.24389517e-01
-4.10541832e-01 1.01321256e+00 -2.53896236e-01 -2.30609085e-02
-6.46300912e-01 -6.79491103e-01 1.41949332e+00 1.63873389e-01
-1.00063965e-01 -1.27516890e+00 -3.51018250e-01 5.36992013e-01
-9.09905910e-01 -1.36757278e+00 -8.28969240e-01 1.05489351e-01
-1.60578713e-01 -4.87159252e-01 -6.35971308e-01 -1.00777578e+00
-4.79302764e-01 -1.04340285e-01 1.44438303e+00 -6.30527437e-01
-3.37836474e-01 5.69397509e-01 -3.97061974e-01 -9.62096512e-01
-7.58460760e-01 4.37170267e-01 1.28877565e-01 -9.75742340e-02
3.13126206e-01 -6.01958394e-01 -1.18108518e-01 2.68380255e-01
-5.17158985e-01 -1.60901800e-01 6.18200779e-01 6.10037088e-01
4.31328833e-01 -4.78052497e-01 1.26876879e+00 -3.16172183e-01
1.18148053e+00 -5.11114120e-01 -3.83686960e-01 4.87509936e-01
-3.76307040e-01 -1.31327868e-01 3.09945524e-01 -6.31924033e-01
-7.34324574e-01 4.36770350e-01 -5.88185430e-01 -1.94064811e-01
-2.20166415e-01 2.12208584e-01 -1.02016985e-01 5.01955092e-01
7.89013088e-01 2.94576585e-01 -3.12625110e-01 -4.44974720e-01
8.46221149e-02 1.43682444e+00 1.10372639e+00 -7.46931791e-01
3.75176609e-01 -3.37664664e-01 -5.80122173e-01 -8.77307951e-01
-9.96987224e-01 -7.48724997e-01 5.27729169e-02 -4.65976208e-01
7.06129909e-01 -1.35956550e+00 -5.35651207e-01 3.45604688e-01
-1.90417552e+00 2.56298900e-01 -2.45568842e-01 2.83914745e-01
-8.03776920e-01 1.91057414e-01 -3.55472505e-01 -9.84890997e-01
-7.82634258e-01 -1.00876081e+00 1.64431465e+00 -3.49406332e-01
-8.36338520e-01 -2.21047059e-01 4.02478755e-01 5.64288080e-01
3.80313963e-01 1.54205421e-02 5.36055028e-01 -9.12847996e-01
1.46791190e-01 -2.56342024e-01 1.26059964e-01 3.45193297e-01
-2.29607850e-01 2.20845751e-02 -1.62261117e+00 -1.67752635e-02
-1.65445104e-01 -6.78062558e-01 6.66431904e-01 3.95534188e-01
1.55800414e+00 -4.82974976e-01 3.21836889e-01 1.43467471e-01
8.96214843e-01 3.93202752e-01 3.92443299e-01 6.18423931e-02
2.61584103e-01 4.98796940e-01 5.58126509e-01 7.59671867e-01
1.52160391e-01 9.01005030e-01 4.30279285e-01 3.95686805e-01
-4.10511464e-01 -4.22623783e-01 6.98940039e-01 1.13898647e+00
4.84845698e-01 -5.85958540e-01 -1.21147466e+00 1.01758873e+00
-1.36561596e+00 -1.08427036e+00 -1.92467302e-01 1.89997554e+00
1.25307322e+00 2.81527936e-01 4.15352046e-01 4.72587794e-01
1.00611544e+00 6.26719072e-02 -7.44301900e-02 -8.42489600e-01
-4.15724516e-01 1.14805892e-01 1.46210760e-01 6.08304739e-01
-1.37009633e+00 6.23351336e-01 7.90720320e+00 1.07967746e+00
-8.01340818e-01 3.92196059e-01 5.01716554e-01 -6.40729189e-01
-2.99046338e-01 -5.89824319e-01 -1.09565616e+00 5.61729968e-01
1.85338676e+00 -4.65844780e-01 2.28137493e-01 5.40241778e-01
3.30874741e-01 2.43224889e-01 -1.19973457e+00 1.31172025e+00
3.30256164e-01 -1.11911750e+00 2.05982611e-01 -6.46519423e-01
5.84114254e-01 2.45103896e-01 3.47101808e-01 5.84871829e-01
-1.98354125e-01 -9.74586904e-01 1.24425089e+00 4.15991507e-02
1.20917904e+00 -3.83005381e-01 6.70217514e-01 -6.76289946e-02
-1.07428777e+00 -1.83642849e-01 8.78822953e-02 4.54950854e-02
6.13154531e-01 5.22264719e-01 -1.49368215e+00 1.23367205e-01
6.15756214e-01 4.67051938e-02 -7.24148750e-01 1.41524291e+00
5.43987863e-02 1.02490330e+00 -3.75952572e-01 -1.85970023e-01
5.59561327e-02 6.87591434e-01 1.04282749e+00 1.66535938e+00
6.62613988e-01 -3.04710329e-01 -1.17267460e-01 3.86157840e-01
-2.14707792e-01 1.05082653e-01 -3.74710917e-01 -8.69003907e-02
8.55807722e-01 8.17109525e-01 -1.16934597e-01 -1.52730167e-01
-8.64404142e-02 7.47664273e-01 -2.14926571e-01 2.65019834e-01
-9.37587798e-01 -7.29490519e-01 4.06731904e-01 -2.02183679e-01
8.97886232e-02 1.06147029e-01 -3.39609981e-01 -6.91514909e-01
7.74786994e-02 -1.13026834e+00 6.56755745e-01 -9.84708965e-01
-1.07785976e+00 9.11047697e-01 1.57819122e-01 -1.34847176e+00
-6.20506108e-01 -2.26012498e-01 -5.49015522e-01 6.37209475e-01
-1.03748024e+00 -8.42678308e-01 1.25805125e-01 1.42437354e-01
1.16337919e+00 -3.53379190e-01 1.17333794e+00 8.43690813e-01
-3.43290508e-01 1.16585815e+00 -2.95779616e-01 -1.51034400e-01
9.77377176e-01 -1.33960104e+00 6.30622447e-01 6.31651640e-01
3.11847121e-01 -7.11114705e-02 1.19076979e+00 -2.98150092e-01
-9.60663915e-01 -1.27144837e+00 1.49948728e+00 -7.14655757e-01
9.09926713e-01 -6.57200694e-01 -4.67810720e-01 2.79093742e-01
3.11351866e-01 -6.06105447e-01 7.74081469e-01 2.15646669e-01
-3.58949065e-01 -2.37058729e-01 -6.06163800e-01 9.19235125e-02
5.56728899e-01 -8.77139151e-01 -6.84329569e-01 6.74018264e-01
1.41057277e+00 -5.05917966e-01 -5.67362785e-01 2.44296104e-01
7.88517714e-01 -2.12374181e-01 9.60397124e-01 -6.66509271e-01
2.67042667e-01 -6.42874911e-02 -2.31278718e-01 -1.08665287e+00
7.99582824e-02 -1.28847265e+00 -3.56655449e-01 1.43690407e+00
1.22120404e+00 2.90398985e-01 4.93104219e-01 8.36962312e-02
-5.67456543e-01 -1.25196904e-01 -1.35208833e+00 -1.07414186e+00
-8.45336169e-02 -1.03291500e+00 5.95608711e-01 4.91262257e-01
5.09347022e-02 9.63622928e-01 -8.06719840e-01 2.74097651e-01
3.31900150e-01 -2.48973116e-01 2.59534210e-01 -1.08965790e+00
-4.05345112e-01 -3.48587900e-01 -6.84285611e-02 -6.81391120e-01
6.96656993e-03 -4.56988811e-01 4.89122003e-01 -1.08202076e+00
-1.28351316e-01 6.64121583e-02 -4.15603846e-01 4.93625969e-01
4.33485173e-02 1.49450555e-01 1.98196381e-01 -1.06662966e-01
-1.09819055e+00 2.95881063e-01 5.61371446e-01 -6.27989948e-01
-3.95040393e-01 6.06027246e-01 -6.87867045e-01 7.63412639e-02
1.09821343e+00 -6.39152348e-01 -3.62351686e-01 -6.80930018e-01
7.05765665e-01 4.46009696e-01 1.34989262e-01 -1.23590529e+00
2.11690336e-01 1.63729101e-01 -4.50855583e-01 -8.10632467e-01
5.10236800e-01 -3.89688790e-01 5.71595840e-02 1.09879926e-01
-9.24031258e-01 3.81553471e-01 6.28376424e-01 3.78239512e-01
-5.97927153e-01 -4.50282604e-01 2.54486978e-01 7.94397071e-02
-4.31643009e-01 -1.58952419e-02 -8.29895616e-01 3.89644951e-01
7.58235037e-01 2.98746347e-01 -2.31021583e-01 -7.80510068e-01
-9.45424676e-01 3.90892416e-01 -6.60619140e-01 9.79595125e-01
6.96508169e-01 -1.29875982e+00 -1.52066040e+00 -1.22201219e-01
3.82988662e-01 -4.05101717e-01 4.32345048e-02 3.39780688e-01
-3.07582021e-01 7.97856331e-01 -8.28818139e-03 -5.41280329e-01
-1.47240210e+00 3.98599580e-02 2.31946960e-01 -1.10609598e-01
2.45074242e-01 1.20491469e+00 -4.44266170e-01 -1.02982251e-02
8.35451901e-01 -1.66814223e-01 -4.51388091e-01 3.18984598e-01
6.80173814e-01 4.60425615e-01 4.45648432e-01 -4.25040156e-01
-2.68200666e-01 4.72996831e-02 -8.45762938e-02 -6.49727762e-01
9.93338525e-01 -1.78568423e-01 2.58896589e-01 7.42387772e-01
1.12659526e+00 1.65674016e-02 -5.30810237e-01 2.19995558e-01
3.30883950e-01 1.81077600e-01 1.74513847e-01 -1.20451713e+00
-1.74854070e-01 9.64540184e-01 7.48001575e-01 5.36794186e-01
1.15979147e+00 1.03777878e-01 9.16180193e-01 4.27447706e-01
7.32560009e-02 -1.31383085e+00 2.29467422e-01 7.56258070e-01
1.30407989e+00 -1.07265687e+00 -2.71667570e-01 -2.60899037e-01
-6.19945943e-01 9.14592862e-01 7.64774203e-01 6.10066652e-01
2.60166883e-01 3.65540951e-01 8.27143416e-02 -7.68927559e-02
-1.21930444e+00 3.11156929e-01 4.56019402e-01 6.26345336e-01
4.76820707e-01 2.53256112e-01 -1.53883651e-01 1.02989221e+00
-6.82360351e-01 -3.02325219e-01 7.66548514e-01 3.99897158e-01
-3.97424310e-01 -8.98933947e-01 -5.32067657e-01 4.30970669e-01
-1.02984571e+00 -4.05907035e-01 -7.12351978e-01 1.00748101e-02
-2.07363889e-02 1.32353771e+00 7.23856362e-03 -7.12818265e-01
3.47818553e-01 2.30843782e-01 1.16681613e-01 -8.96609604e-01
-7.57750273e-01 2.01011255e-01 7.55659163e-01 -2.46281132e-01
-1.37687281e-01 -7.91342080e-01 -6.86044216e-01 5.06902397e-01
-4.52564746e-01 7.22122014e-01 8.05712044e-01 4.84346360e-01
4.16743189e-01 6.66986465e-01 7.24345148e-01 -4.08681095e-01
-9.87385154e-01 -1.27331519e+00 -3.67397189e-01 -7.76032507e-02
7.68708885e-01 3.49521041e-02 -5.85212409e-01 2.86634952e-01] | [15.276212692260742, 4.893064498901367] |
28ecbecc-05ec-4672-8aae-c8c56e3e296f | perturbation-based-two-stage-multi-domain | 2306.10700 | null | https://arxiv.org/abs/2306.10700v1 | https://arxiv.org/pdf/2306.10700v1.pdf | Perturbation-Based Two-Stage Multi-Domain Active Learning | In multi-domain learning (MDL) scenarios, high labeling effort is required due to the complexity of collecting data from various domains. Active Learning (AL) presents an encouraging solution to this issue by annotating a smaller number of highly informative instances, thereby reducing the labeling effort. Previous research has relied on conventional AL strategies for MDL scenarios, which underutilize the domain-shared information of each instance during the selection procedure. To mitigate this issue, we propose a novel perturbation-based two-stage multi-domain active learning (P2S-MDAL) method incorporated into the well-regarded ASP-MTL model. Specifically, P2S-MDAL involves allocating budgets for domains and establishing regions for diversity selection, which are further used to select the most cross-domain influential samples in each region. A perturbation metric has been introduced to evaluate the robustness of the shared feature extractor of the model, facilitating the identification of potentially cross-domain influential samples. Experiments are conducted on three real-world datasets, encompassing both texts and images. The superior performance over conventional AL strategies shows the effectiveness of the proposed strategy. Additionally, an ablation study has been carried out to demonstrate the validity of each component. Finally, we outline several intriguing potential directions for future MDAL research, thus catalyzing the field's advancement. | ['Ke Tang', 'Shan He', 'Zeyu Dai', 'Rui He'] | 2023-06-19 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 3.91190916e-01 2.92814404e-01 -5.46838582e-01 -2.77569592e-01
-1.33296168e+00 -4.07488734e-01 5.67750752e-01 3.75745922e-01
-3.28218102e-01 9.34743226e-01 1.86526462e-01 1.01399511e-01
-4.62877810e-01 -4.15975541e-01 -4.16040689e-01 -1.00924158e+00
-6.71398714e-02 6.42824590e-01 3.10208440e-01 1.71635225e-01
3.93482208e-01 4.23397124e-01 -1.38381135e+00 1.82100981e-01
1.24508774e+00 8.68598938e-01 9.99078825e-02 6.30080095e-03
-2.47398227e-01 6.29615843e-01 -8.14229548e-01 -3.60705018e-01
3.82070631e-01 -3.24524075e-01 -6.89681590e-01 4.72367853e-01
1.33733705e-01 -6.02001920e-02 2.11956352e-01 9.15544331e-01
7.27500677e-01 3.51978242e-01 6.55004680e-01 -1.37622118e+00
-2.56380886e-01 6.20710075e-01 -8.01175714e-01 1.26108170e-01
3.18671286e-01 1.95345297e-01 1.20980024e+00 -1.06293154e+00
6.39352918e-01 1.26149869e+00 4.70088363e-01 3.43879342e-01
-1.28857768e+00 -7.54675329e-01 4.92323011e-01 3.80015463e-01
-1.37160981e+00 -6.34748161e-01 1.17389774e+00 -3.84283721e-01
6.33536100e-01 7.05265105e-02 3.65108639e-01 1.22069693e+00
-1.85534135e-01 1.36071277e+00 1.03831017e+00 -7.40665734e-01
6.87960923e-01 5.35381317e-01 1.98427528e-01 3.11791778e-01
4.33020055e-01 -1.31505445e-01 -9.44357336e-01 -6.26451373e-01
2.67554522e-01 -4.46702123e-01 -7.93237090e-02 -8.35639477e-01
-7.16460586e-01 9.55600679e-01 -1.50830001e-01 1.25045806e-01
-4.84376997e-01 -7.44044542e-01 3.87944907e-01 2.03455508e-01
1.09436584e+00 6.68898582e-01 -4.75074530e-01 -9.82482433e-02
-1.08260834e+00 2.26149276e-01 5.94744742e-01 8.64757061e-01
6.06443524e-01 -3.48099858e-01 -9.47201103e-02 1.05333364e+00
5.43994784e-01 -2.58613154e-02 1.90753683e-01 -8.61223876e-01
6.50264263e-01 1.26833940e+00 9.30973664e-02 -7.53478229e-01
-2.21373066e-01 -4.84642625e-01 -4.96435016e-01 1.44895464e-01
3.06894243e-01 -1.62885085e-01 -4.26948160e-01 1.70044243e+00
7.05242872e-01 1.87937152e-02 1.55880585e-01 7.15583682e-01
4.59137648e-01 4.21158850e-01 2.17672944e-01 -5.51369905e-01
8.74443412e-01 -6.96155965e-01 -4.65964288e-01 -3.09471697e-01
5.87157845e-01 -5.67883253e-01 1.04586923e+00 6.22452378e-01
-9.76979077e-01 -3.13841730e-01 -1.06924868e+00 4.04224396e-01
-1.88465565e-01 9.31826085e-02 3.43981087e-01 6.33819342e-01
-5.15446723e-01 2.11070463e-01 -5.28444350e-01 -2.03113332e-01
8.63277555e-01 2.90238798e-01 -9.06815454e-02 -1.39639437e-01
-1.41902590e+00 7.44257808e-01 5.15690684e-01 -8.89382064e-02
-7.86104143e-01 -8.41198027e-01 -5.63174129e-01 -1.56392187e-01
6.08470678e-01 -2.20827878e-01 8.75866830e-01 -1.28804636e+00
-1.22131777e+00 8.35659444e-01 -8.62769261e-02 -4.86862570e-01
8.15414429e-01 -3.22083831e-01 -5.22325158e-01 1.74030125e-01
2.83848733e-01 5.58016598e-01 8.83721590e-01 -1.43544185e+00
-7.45870411e-01 -3.87766093e-01 2.76265629e-02 4.95384574e-01
-6.93303466e-01 -7.20797526e-03 -4.04543787e-01 -7.02391386e-01
4.57220078e-02 -8.43607783e-01 -3.24346662e-01 -7.00414032e-02
-3.00746053e-01 -5.06636500e-01 9.27315056e-01 -3.10640633e-01
1.49079084e+00 -2.32304478e+00 1.60864621e-01 2.93983459e-01
2.49912366e-01 4.76023793e-01 -1.00828335e-01 3.91915202e-01
4.23448049e-02 -1.04203805e-01 -4.04398233e-01 -4.20948744e-01
-6.72316626e-02 -2.74604764e-02 -2.91943222e-01 5.64499080e-01
5.30795157e-01 3.86598736e-01 -8.12934041e-01 -7.70716786e-01
1.06913112e-01 5.37923723e-03 -4.94218856e-01 1.28810585e-01
-3.28485966e-01 5.24141610e-01 -4.86449808e-01 7.49773383e-01
7.67519057e-01 -3.23126405e-01 4.11640197e-01 -5.60025051e-02
-3.09081804e-02 2.56153673e-01 -1.44452262e+00 1.42760718e+00
-2.24169374e-01 4.94167864e-01 -6.87530637e-02 -1.28525293e+00
1.11634970e+00 1.74689770e-01 9.11352456e-01 -7.02890813e-01
-1.21146195e-01 2.59052873e-01 -4.52506058e-02 -4.30302441e-01
3.61697406e-01 1.82392895e-01 -2.69817770e-01 4.98583704e-01
2.64940076e-02 3.31239492e-01 3.13963264e-01 3.00903618e-01
9.75677848e-01 1.79609239e-01 6.26395166e-01 -3.36651385e-01
7.19338298e-01 2.47271001e-01 7.63593018e-01 7.71803856e-01
-7.47566640e-01 2.81681269e-01 6.18387222e-01 -3.59224640e-02
-1.01771474e+00 -8.92235518e-01 -2.32596621e-01 1.04763520e+00
3.08690012e-01 -8.31909999e-02 -4.49152499e-01 -1.02644670e+00
-5.34381755e-02 9.37301338e-01 -4.41393673e-01 -3.21441978e-01
-4.88557905e-01 -1.09346640e+00 4.81599838e-01 9.37660038e-02
5.77866673e-01 -8.29769254e-01 -5.62080443e-01 2.06142277e-01
-2.23155141e-01 -9.23624456e-01 5.12119420e-02 3.28743577e-01
-8.42475832e-01 -1.16988194e+00 -6.84278607e-01 -5.47677517e-01
6.73138499e-01 5.98393055e-03 1.01195645e+00 -5.72621882e-01
-6.44842759e-02 2.73210227e-01 -6.55613124e-01 -4.40611750e-01
-4.76269126e-01 3.16856295e-01 3.84341627e-02 1.70698777e-01
7.97630906e-01 -2.63276786e-01 -3.13739717e-01 4.16434765e-01
-5.75326860e-01 -1.21998839e-01 7.63962626e-01 9.71711218e-01
5.84112287e-01 4.01952237e-01 1.07219601e+00 -1.20761371e+00
8.02988172e-01 -7.71519423e-01 -5.49270034e-01 5.35845339e-01
-7.88434803e-01 -7.43588284e-02 6.14865243e-01 -7.21733451e-01
-1.38305962e+00 1.03001453e-01 2.72109002e-01 -3.25334221e-01
-6.36158362e-02 4.30086553e-01 -5.68593264e-01 6.75680488e-02
7.74585724e-01 1.39361784e-01 -5.83078898e-02 -3.35229993e-01
1.07205980e-01 8.02638352e-01 -1.50089720e-02 -6.33363605e-01
6.19999468e-01 2.76141375e-01 -3.38015586e-01 -8.16151798e-01
-9.96519744e-01 -7.11378455e-01 -7.23682404e-01 -4.13236409e-01
5.05071171e-02 -9.72287536e-01 -1.61723554e-01 3.76011759e-01
-8.34455073e-01 -4.50949967e-02 -4.70427096e-01 4.69610214e-01
-3.10110211e-01 5.31175196e-01 -2.84204520e-02 -1.10186744e+00
-1.87007412e-01 -1.13090730e+00 8.37305129e-01 1.69390768e-01
-6.08630717e-01 -1.06475115e+00 1.71979517e-01 6.63846612e-01
1.39045445e-02 1.26109108e-01 9.71903384e-01 -1.27034247e+00
-4.11274582e-01 -5.45395277e-02 -2.51729437e-03 3.33610803e-01
7.99124911e-02 -2.22973600e-01 -1.03369856e+00 -4.48728830e-01
-2.78034955e-02 -4.89795685e-01 5.63760579e-01 3.55193347e-01
9.95235145e-01 -9.60403755e-02 -4.38738614e-01 5.16134724e-02
1.20199144e+00 4.72618192e-01 3.66106063e-01 6.23747885e-01
3.15652788e-01 8.11700284e-01 1.32356763e+00 8.83477926e-01
6.37400299e-02 5.79744637e-01 1.72095180e-01 -3.75273973e-01
-1.65124256e-02 1.12844527e-01 4.49868649e-01 5.27761340e-01
4.06347841e-01 -3.99840355e-01 -9.98417497e-01 5.76138854e-01
-1.89246035e+00 -7.32697427e-01 9.60357562e-02 2.26315403e+00
9.94558096e-01 4.97294366e-01 2.89752334e-01 4.04884636e-01
7.91316569e-01 1.69203326e-01 -1.07802880e+00 -6.95558116e-02
-3.55673194e-01 4.11278084e-02 2.11553305e-01 3.21132571e-01
-1.32792664e+00 7.03102291e-01 5.99322510e+00 1.09333205e+00
-9.14529204e-01 7.64523074e-02 6.60965562e-01 -2.64105320e-01
-1.51479706e-01 -3.34650800e-02 -9.31712687e-01 5.76027930e-01
6.92326665e-01 -3.90266150e-01 -7.80050680e-02 1.05738008e+00
4.08301413e-01 -4.35466528e-01 -9.79446828e-01 6.97425306e-01
8.17564726e-02 -1.03500223e+00 9.20454338e-02 8.38309899e-03
7.19544828e-01 -2.99915761e-01 2.17087548e-02 2.88907707e-01
3.44660401e-01 -3.55309337e-01 6.09749973e-01 1.75096184e-01
6.00197196e-01 -9.45427597e-01 5.25124133e-01 5.45889616e-01
-8.89165282e-01 -3.41627210e-01 -2.81128317e-01 2.62312442e-01
-5.48886061e-02 9.59396422e-01 -1.29768729e+00 6.77558541e-01
4.50783253e-01 6.78300023e-01 -6.56824708e-01 9.21566546e-01
7.46021196e-02 8.49462807e-01 -1.75004557e-01 1.99099407e-02
1.55392304e-01 -4.10642847e-02 8.54346037e-01 1.13367355e+00
-4.19532880e-02 -2.36185163e-01 3.25675547e-01 8.63468647e-01
6.22856803e-02 1.90717831e-01 -4.76582170e-01 -8.77134651e-02
1.04031777e+00 9.18684900e-01 -6.40159309e-01 -2.46120781e-01
-4.31017965e-01 6.97344601e-01 2.38371700e-01 1.57809287e-01
-7.99079180e-01 -2.47089371e-01 4.98803794e-01 3.34406942e-02
-5.34526911e-03 7.90083483e-02 -5.59765756e-01 -9.29884195e-01
4.78654110e-05 -1.12619412e+00 5.84159434e-01 -2.27018744e-01
-1.50888240e+00 2.75146067e-01 2.23062277e-01 -1.61181569e+00
-1.30573764e-01 -2.07917392e-01 -1.70339972e-01 8.14664245e-01
-1.44793952e+00 -9.25666273e-01 -3.15028466e-02 6.05322540e-01
9.71085668e-01 -6.07760191e-01 6.26067519e-01 5.05589843e-01
-9.86721933e-01 8.81111860e-01 3.07803601e-01 -1.70893833e-01
8.46998155e-01 -9.91805434e-01 7.29686916e-02 1.07113743e+00
-1.68400258e-01 6.11116886e-01 5.77054203e-01 -7.23644376e-01
-1.03619516e+00 -1.08826816e+00 7.74054468e-01 -2.69220114e-01
4.64655668e-01 -4.21826720e-01 -9.10966575e-01 2.70032108e-01
-1.26457155e-01 -4.30202216e-01 1.08228493e+00 1.21557750e-01
-4.99974638e-02 -1.43369138e-01 -1.33354199e+00 4.94181305e-01
7.82090366e-01 -2.41320238e-01 -6.51738882e-01 1.62299171e-01
4.71993893e-01 3.58545296e-02 -7.32044339e-01 4.62912589e-01
3.12598139e-01 -8.69045615e-01 9.07570183e-01 -3.06706399e-01
3.25804353e-01 -4.59136404e-02 -1.47831753e-01 -1.19118333e+00
-3.10533404e-01 -4.14722949e-01 -2.80172497e-01 1.59045672e+00
5.32373607e-01 -6.56155407e-01 1.00545990e+00 4.50670689e-01
6.32034913e-02 -8.24996114e-01 -9.37641561e-01 -7.01358438e-01
-9.25370678e-02 -2.74977177e-01 3.56152385e-01 1.22729015e+00
-8.64685327e-02 4.23039943e-01 -2.87004411e-01 1.82882592e-01
9.07641351e-01 -3.21785770e-02 7.08404541e-01 -1.55660129e+00
-1.24786898e-01 -3.00764859e-01 4.55394723e-02 -9.11618590e-01
3.09535205e-01 -6.43928111e-01 -1.99375469e-02 -1.10678911e+00
2.46346205e-01 -8.02445829e-01 -4.04031694e-01 3.00188214e-01
-3.45206380e-01 -1.25388101e-01 -5.99863194e-03 5.30542433e-01
-8.55114102e-01 4.96163964e-01 1.02986300e+00 -1.02734253e-01
-4.08937246e-01 1.31359085e-01 -8.19446683e-01 6.28928602e-01
6.80611312e-01 -5.97137094e-01 -8.82610083e-01 5.29698282e-02
-4.01740931e-02 -2.35396475e-01 -2.06081364e-02 -8.52284491e-01
2.38900647e-01 -3.83553892e-01 4.40790087e-01 -7.55903423e-01
2.71635950e-01 -9.33070481e-01 5.26636392e-02 2.96679348e-01
-6.69545710e-01 -4.33394313e-01 1.17204711e-01 6.27403557e-01
-3.02705914e-01 -2.50320464e-01 1.11547339e+00 2.52002701e-02
-1.01621723e+00 5.36933467e-02 -2.25679651e-01 3.06194603e-01
1.35201025e+00 -4.55266207e-01 -1.24352269e-01 1.09196611e-01
-5.46773851e-01 2.68728048e-01 3.45602125e-01 2.82926977e-01
4.50522840e-01 -1.04763508e+00 -6.88659072e-01 2.23780617e-01
3.57019991e-01 1.53071433e-01 2.09463835e-01 7.84534752e-01
-9.80294123e-02 1.92772910e-01 -2.59623360e-02 -7.51085222e-01
-1.35782826e+00 3.11922818e-01 9.12021101e-02 -5.93677044e-01
-4.38903064e-01 8.36855650e-01 6.23562671e-02 -3.97291034e-01
4.12612170e-01 4.29602772e-01 -3.70099604e-01 6.22122049e-01
1.76278159e-01 6.88695550e-01 2.10067123e-01 -3.45985502e-01
-6.47556007e-01 5.27635962e-02 -4.04598206e-01 -8.54773149e-02
1.35204804e+00 -3.38990837e-01 2.16297954e-01 5.06331146e-01
8.31035435e-01 1.65892150e-02 -1.44721317e+00 -5.00916243e-01
5.51855803e-01 -4.00111020e-01 6.11302480e-02 -9.02155995e-01
-6.87191188e-01 5.27618051e-01 5.85493445e-01 8.59879702e-02
1.38739586e+00 -1.05790392e-01 3.61716688e-01 9.62673649e-02
3.76742512e-01 -1.55229700e+00 3.60886931e-01 2.17693090e-01
6.12978458e-01 -1.29930103e+00 3.64154547e-01 -3.35931093e-01
-9.62781489e-01 8.46870065e-01 7.89894104e-01 1.18160076e-01
4.11841571e-01 2.49934852e-01 -7.01863021e-02 -8.30920413e-02
-9.32473004e-01 1.01117976e-01 6.80407435e-02 4.39362377e-01
2.40050927e-01 2.34209467e-02 -4.34484214e-01 5.64491987e-01
3.38225424e-01 -1.44367889e-01 2.17987403e-01 1.05554450e+00
-4.43986684e-01 -1.28472364e+00 -4.28955853e-01 3.81500810e-01
-2.41079047e-01 2.25434318e-01 -7.55109012e-01 8.19660902e-01
3.18248481e-01 9.46076214e-01 -1.10511385e-01 -2.47997299e-01
3.28318268e-01 2.02622280e-01 2.91495621e-01 -5.97235024e-01
-5.18822908e-01 2.62603074e-01 2.01540157e-01 -2.43600726e-01
-7.32926607e-01 -8.83540869e-01 -8.88684511e-01 -1.65186405e-01
-5.56237519e-01 2.04905897e-01 2.69425243e-01 9.16106045e-01
5.12583196e-01 2.98151165e-01 9.33853626e-01 -4.87230271e-01
-6.71674788e-01 -7.99258292e-01 -5.73474705e-01 5.03025830e-01
8.38407651e-02 -1.03698111e+00 -4.39148962e-01 -1.75237998e-01] | [10.145723342895508, 3.469902992248535] |
20628905-cad7-4d57-bf38-8256523f1344 | on-using-context-for-automatic-correction-of | null | null | https://aclanthology.org/W12-2012 | https://aclanthology.org/W12-2012.pdf | On using context for automatic correction of non-word misspellings in student essays | null | ['Yoko Futagi', 'Michael Flor'] | 2012-06-01 | null | null | null | ws-2012-6 | ['lexical-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.318655490875244, 3.686575174331665] |
f0e86f2d-11cf-4431-8316-8ed735b4109d | mass-producing-failures-of-multimodal-systems | 2306.12105 | null | https://arxiv.org/abs/2306.12105v1 | https://arxiv.org/pdf/2306.12105v1.pdf | Mass-Producing Failures of Multimodal Systems with Language Models | Deployed multimodal systems can fail in ways that evaluators did not anticipate. In order to find these failures before deployment, we introduce MultiMon, a system that automatically identifies systematic failures -- generalizable, natural-language descriptions of patterns of model failures. To uncover systematic failures, MultiMon scrapes a corpus for examples of erroneous agreement: inputs that produce the same output, but should not. It then prompts a language model (e.g., GPT-4) to find systematic patterns of failure and describe them in natural language. We use MultiMon to find 14 systematic failures (e.g., "ignores quantifiers") of the CLIP text-encoder, each comprising hundreds of distinct inputs (e.g., "a shelf with a few/many books"). Because CLIP is the backbone for most state-of-the-art multimodal systems, these inputs produce failures in Midjourney 5.1, DALL-E, VideoFusion, and others. MultiMon can also steer towards failures relevant to specific use cases, such as self-driving cars. We see MultiMon as a step towards evaluation that autonomously explores the long tail of potential system failures. Code for MULTIMON is available at https://github.com/tsb0601/MultiMon. | ['Jacob Steinhardt', 'Erik Jones', 'Shengbang Tong'] | 2023-06-21 | null | null | null | null | ['self-driving-cars'] | ['computer-vision'] | [-3.66429448e-01 5.63921146e-02 -1.14631809e-01 -5.97857893e-01
-1.08372927e+00 -9.04521286e-01 5.07423341e-01 6.26450032e-02
1.25807256e-01 4.79309410e-01 2.96949595e-01 -7.07314372e-01
2.25442559e-01 -2.56275177e-01 -9.38358486e-01 -8.14370587e-02
8.04462060e-02 4.33424503e-01 1.03915662e-01 -5.28065026e-01
1.36803210e-01 -1.27158046e-01 -1.50880146e+00 8.40682983e-01
4.82559592e-01 6.35444343e-01 -1.03898495e-01 8.72356117e-01
1.15969833e-02 1.20154285e+00 -7.58506954e-01 -5.92085481e-01
-4.32593301e-02 -2.28057608e-01 -9.72951829e-01 -7.90817067e-02
5.59154689e-01 -5.70024908e-01 -3.53845984e-01 8.44203293e-01
3.31858069e-01 -2.54488140e-01 2.84602106e-01 -2.10163140e+00
-3.59838843e-01 8.31598282e-01 -2.44565964e-01 8.03534016e-02
8.93868864e-01 9.88220215e-01 1.08566797e+00 -9.00358558e-01
5.78532577e-01 1.39544606e+00 7.89137065e-01 5.77084422e-01
-1.29265606e+00 -7.10144281e-01 -2.07936913e-01 -1.72466412e-02
-1.51235187e+00 -8.78391743e-01 2.41695076e-01 -6.09417856e-01
1.41384399e+00 5.17779052e-01 5.29952012e-02 1.45220828e+00
4.72097099e-01 8.17568004e-01 6.69499099e-01 -2.82466173e-01
1.71341792e-01 1.11936464e-03 2.77266920e-01 6.82355583e-01
-4.82201017e-02 -5.42838536e-02 -9.02823150e-01 -3.94668669e-01
-2.34296620e-02 -3.42786998e-01 -1.15027577e-01 3.18711996e-01
-1.08906007e+00 3.72518003e-01 -4.86828201e-02 1.37998685e-01
-1.12579279e-01 5.97417176e-01 4.94866461e-01 4.73863333e-01
-2.07497720e-02 5.43341935e-01 -5.70207298e-01 -8.35593224e-01
-9.50168490e-01 3.86134446e-01 1.02995420e+00 1.60673904e+00
6.66967571e-01 -2.07733989e-01 -1.04595117e-01 5.94702065e-01
4.37930167e-01 4.06541109e-01 1.48519576e-01 -1.24583137e+00
6.94244385e-01 6.49262667e-01 4.82744515e-01 -6.81552470e-01
-6.64737165e-01 8.92873630e-02 -2.53035456e-01 7.91075602e-02
2.67660052e-01 -2.94597268e-01 -6.73835576e-01 1.61746633e+00
-3.55187386e-01 -2.10137233e-01 7.07918555e-02 1.02762735e+00
8.56167734e-01 5.34548759e-01 2.51266748e-01 1.54584318e-01
1.26519072e+00 -6.67633116e-01 -5.85451961e-01 -3.80065948e-01
9.79361653e-01 -9.35554147e-01 1.39501321e+00 5.67198217e-01
-1.21693039e+00 -4.40617830e-01 -9.21403289e-01 1.24349661e-01
-1.98105112e-01 1.02820523e-01 2.34406382e-01 2.64630914e-01
-1.33320987e+00 3.71027887e-01 -8.54731798e-01 -8.29152226e-01
-2.82961549e-03 3.18169236e-01 -3.49064410e-01 -1.40658066e-01
-1.17347741e+00 9.84940410e-01 2.63835698e-01 -5.42735495e-02
-9.93954003e-01 -3.30333710e-01 -9.31072354e-01 -5.64097650e-02
5.78249454e-01 -5.86598635e-01 1.98094308e+00 -9.31384683e-01
-9.90297556e-01 5.27901411e-01 -3.41665119e-01 -1.90101981e-01
3.27402949e-01 -2.94563383e-01 -8.72211993e-01 -7.55543858e-02
2.31874123e-01 8.74920487e-01 3.98346215e-01 -1.34697843e+00
-7.38076031e-01 1.02249660e-01 2.94980794e-01 2.30164547e-02
-2.37599928e-02 4.12664771e-01 -7.76194751e-01 -2.28925254e-02
-2.27081463e-01 -8.73249114e-01 1.60242274e-01 -4.68133479e-01
-9.05815780e-01 -3.78213257e-01 4.65089560e-01 -4.76907760e-01
1.66980922e+00 -2.41657257e+00 -7.63038695e-02 2.09472179e-01
2.59127706e-01 -3.54979157e-01 -4.17140275e-01 9.90730286e-01
-3.89329679e-02 6.92640662e-01 8.75066146e-02 -6.52591228e-01
3.71960133e-01 2.21091807e-01 -3.95839036e-01 3.48315656e-01
4.52195108e-01 7.66428232e-01 -8.88377964e-01 -4.55429941e-01
2.14693144e-01 3.49177197e-02 -4.07635540e-01 2.57686496e-01
-5.88214695e-01 -7.43171647e-02 -6.43345863e-02 1.09782112e+00
4.45580989e-01 -3.17819685e-01 8.21163878e-02 -3.39625001e-01
-4.82279032e-01 2.76182055e-01 -8.98062289e-01 1.54655373e+00
-2.18628407e-01 7.55823314e-01 -1.09401442e-01 -2.94602066e-02
5.42280018e-01 3.53763223e-01 1.16768837e-01 -6.45431697e-01
1.03254721e-01 3.25119644e-01 -1.08777851e-01 -1.06276798e+00
7.80043483e-01 9.83028188e-02 -4.87860352e-01 2.90182471e-01
1.65128350e-01 -1.91811711e-01 4.00495142e-01 6.80034995e-01
1.74071145e+00 -7.63754621e-02 -1.27541907e-02 8.71393830e-02
-1.56178221e-01 4.16963518e-01 3.54057550e-01 1.04148686e+00
-1.68662027e-01 8.91874075e-01 7.16615319e-01 -1.23507142e-01
-1.19903386e+00 -7.68305600e-01 7.37160519e-02 1.10849559e+00
2.50407159e-01 -1.26427805e+00 -4.44196463e-01 -5.86318910e-01
2.78522283e-01 1.18070710e+00 -2.88320750e-01 -2.62949139e-01
-7.57889822e-02 -8.92211124e-02 1.13698506e+00 5.11355579e-01
8.44711661e-02 -9.64697778e-01 -6.60178363e-01 1.50535107e-01
-3.93040359e-01 -1.11158621e+00 -4.19536173e-01 3.12595487e-01
-3.81103188e-01 -1.18366218e+00 1.37821361e-01 -1.92708224e-01
3.74511987e-01 -3.58719341e-02 1.50366676e+00 6.55006647e-01
1.34329408e-01 6.06276512e-01 -5.16095042e-01 -1.33576378e-01
-7.92363405e-01 -1.09025806e-01 1.19819582e-01 -4.31421936e-01
7.31844068e-01 1.37508094e-01 -1.91591322e-01 7.22294569e-01
-8.89593601e-01 1.10390976e-01 2.16380835e-01 7.59984612e-01
-6.33723056e-03 1.45060480e-01 2.86434740e-01 -6.98700786e-01
9.82141256e-01 -9.77346480e-01 -4.56008762e-01 3.07021707e-01
-6.18313253e-01 -1.87242001e-01 6.56978428e-01 -1.81915149e-01
-4.81958985e-01 -4.01452221e-02 -1.49750486e-01 -5.12105107e-01
-5.21075428e-01 9.23877120e-01 5.68438545e-02 4.89481360e-01
8.98371637e-01 3.01011112e-02 -5.36192767e-02 -2.13859349e-01
3.04400951e-01 1.01899230e+00 7.42470503e-01 -7.23459601e-01
4.13329810e-01 -1.34498850e-01 -6.66496933e-01 -3.94824594e-01
-1.65639296e-01 -4.86488193e-01 6.11877032e-02 -5.24090052e-01
4.09787834e-01 -9.62762713e-01 -1.14205503e+00 3.51601064e-01
-1.35859632e+00 -7.52623856e-01 8.25190321e-02 1.34415478e-01
-5.13704658e-01 4.65685315e-03 -7.95345724e-01 -9.63270366e-01
2.46895570e-02 -1.45921111e+00 1.19989169e+00 3.10611397e-01
-1.10968602e+00 -5.15738487e-01 -1.62657976e-01 1.75417915e-01
5.22431970e-01 -2.49964163e-01 8.39884520e-01 -6.31900907e-01
-2.85587817e-01 -1.03958361e-01 4.26617190e-02 1.86069071e-01
-2.53744721e-01 7.29320586e-01 -8.50412130e-01 -3.19961667e-01
-5.69286108e-01 -6.90330029e-01 2.14472279e-01 -6.06053062e-02
7.94532239e-01 -4.09019202e-01 -4.08192664e-01 8.03268030e-02
1.21152854e+00 3.75461876e-01 6.30017459e-01 3.76641154e-01
2.70983100e-01 6.11072302e-01 6.13486528e-01 7.95401275e-01
7.46159315e-01 7.51528561e-01 5.20569623e-01 2.28014618e-01
1.46372378e-01 -3.29316765e-01 8.03407848e-01 5.00911772e-01
5.62654614e-01 -7.32586443e-01 -1.51289189e+00 6.05588078e-01
-1.97239590e+00 -8.04664314e-01 -4.60467190e-01 1.82845962e+00
7.42603064e-01 4.35398757e-01 2.01348558e-01 -2.95346707e-01
4.50218469e-01 -1.57534882e-01 -4.77463961e-01 -4.60809916e-01
-1.23481326e-01 -2.51392275e-01 3.54208320e-01 5.87141752e-01
-6.99883044e-01 1.00427806e+00 6.77379942e+00 4.46608365e-01
-9.10323143e-01 9.38458294e-02 4.50022846e-01 -2.75200665e-01
-5.88753104e-01 3.74974459e-01 -7.85086095e-01 9.26509082e-01
1.26615381e+00 4.10363898e-02 4.78764147e-01 5.43330669e-01
3.70400399e-01 -5.50476968e-01 -1.57754505e+00 9.17593420e-01
1.24985408e-02 -1.14571762e+00 -3.35112691e-01 -3.48978579e-01
2.16900513e-01 2.92913765e-01 -1.13352694e-01 5.81827521e-01
6.02108121e-01 -1.25010884e+00 1.31401217e+00 7.26673901e-01
9.09010708e-01 -4.54709113e-01 8.32975149e-01 4.71789360e-01
-7.33943224e-01 -1.36125997e-01 -5.80200031e-02 1.21206539e-02
3.45468849e-01 3.99873525e-01 -8.11559081e-01 4.61213201e-01
9.98542488e-01 5.15896976e-01 -8.24040174e-01 8.72094035e-01
-9.04456452e-02 8.82194340e-01 -5.84630370e-01 -5.61853759e-02
-9.71529186e-02 4.67324853e-01 5.69155514e-01 1.39225554e+00
3.14706028e-01 1.80560306e-01 -4.09557223e-02 9.87579703e-01
-3.94224077e-02 -3.35033327e-01 -8.28997195e-01 -1.62400588e-01
9.27762628e-01 9.83500361e-01 -8.21247548e-02 -3.58583570e-01
-5.39954603e-01 8.54174733e-01 7.41215423e-02 5.65311253e-01
-1.00565755e+00 -1.49409890e-01 4.26369429e-01 2.33899251e-01
-2.93378025e-01 -2.68866897e-01 -1.31414637e-01 -1.02366436e+00
1.31930947e-01 -1.45351577e+00 3.02317768e-01 -1.56580424e+00
-1.36397874e+00 7.61351645e-01 1.86481073e-01 -1.22630465e+00
-3.27167004e-01 -2.80408949e-01 -6.18005753e-01 8.08906376e-01
-6.84758663e-01 -8.78662586e-01 -4.84886706e-01 3.86983126e-01
6.74200416e-01 -2.28618219e-01 6.09828413e-01 2.84053147e-01
-6.45454228e-01 6.88451588e-01 -5.14708996e-01 -2.50589550e-01
1.14572942e+00 -1.17034328e+00 5.89280784e-01 7.57921040e-01
-2.17312634e-01 9.41499233e-01 1.14128399e+00 -8.35998774e-01
-1.67233133e+00 -7.56623864e-01 1.06008649e+00 -8.63927126e-01
9.49243665e-01 -3.15553546e-01 -6.93685174e-01 9.54885781e-01
5.81272125e-01 -3.28785062e-01 5.29375136e-01 2.33742505e-01
-3.32849771e-01 3.30736041e-02 -7.43996918e-01 7.26258337e-01
7.52242327e-01 -7.31542051e-01 -3.52714449e-01 3.49660993e-01
6.52804911e-01 -7.65146852e-01 -7.15257585e-01 2.50626147e-01
6.20181978e-01 -9.67697859e-01 2.09030166e-01 -7.30377555e-01
8.46735418e-01 -5.67044318e-01 -5.62547505e-01 -1.28712809e+00
-2.32436582e-02 -9.59850669e-01 -1.60763040e-01 1.33807302e+00
8.26054513e-01 -3.80387098e-01 2.72981524e-01 1.37500131e+00
-5.36019385e-01 -6.90851390e-01 -6.69566274e-01 -7.02139437e-01
-2.91061908e-01 -9.39729810e-01 5.80063045e-01 8.12004089e-01
6.42726421e-01 2.63354123e-01 -4.18445170e-01 2.55789280e-01
1.61193773e-01 -5.15553534e-01 1.15569806e+00 -4.92678583e-01
-4.13502216e-01 -3.48531514e-01 -3.50690156e-01 -1.14883661e+00
-4.44814973e-02 -6.13882661e-01 3.34798247e-01 -1.27253723e+00
2.14268312e-01 -4.44641799e-01 -3.99444113e-03 1.20861697e+00
1.39990464e-01 1.58191547e-02 4.38140392e-01 4.91572201e-01
-1.12854862e+00 2.04057828e-01 3.42580825e-01 -2.34789193e-01
-5.44283316e-02 -2.89948642e-01 -9.26050663e-01 4.68836248e-01
7.72940874e-01 -4.99046624e-01 -1.00055605e-01 -6.19026661e-01
7.56909132e-01 3.52804542e-01 6.60895586e-01 -1.13895154e+00
6.04162097e-01 -1.23726256e-01 -1.51783377e-01 -5.03901839e-01
1.84387099e-02 -8.89358282e-01 6.05680943e-01 5.13034463e-02
-4.76434410e-01 7.12137043e-01 5.92503726e-01 -1.40507475e-01
-4.35131758e-01 -4.05228108e-01 2.01248556e-01 2.39213437e-01
-8.36474657e-01 -2.71880597e-01 -6.74477577e-01 -7.74269179e-02
9.27060843e-01 1.52481169e-01 -1.21436667e+00 -5.84186852e-01
-5.75819612e-01 9.29885626e-01 8.01447213e-01 6.60904884e-01
6.59137487e-01 -1.09828103e+00 -5.30973196e-01 -1.88499270e-03
6.21167064e-01 -2.43815720e-01 2.27414548e-01 1.08617330e+00
-4.74927038e-01 1.74085975e-01 7.57440329e-02 -7.71825969e-01
-1.20455062e+00 1.09517813e-01 6.37615144e-01 2.61787206e-01
-3.64042491e-01 7.35346079e-01 -2.52977103e-01 -3.18793893e-01
3.08156937e-01 -4.52143580e-01 3.61232162e-01 -3.26759875e-01
3.24848264e-01 -3.96179333e-02 1.99659809e-01 -4.73990351e-01
-8.17410290e-01 -6.52821213e-02 -7.50342682e-02 -4.23187882e-01
9.30323303e-01 -2.57569432e-01 -2.24204194e-02 7.91077912e-01
1.07606530e+00 -1.65639296e-01 -1.12805939e+00 1.45940855e-01
1.47346465e-03 -3.23799551e-01 -3.20515960e-01 -1.35864508e+00
-6.24628425e-01 4.90640044e-01 4.21742082e-01 4.66060400e-01
9.67787743e-01 4.63162929e-01 5.18301904e-01 2.85140932e-01
5.87015629e-01 -1.03418350e+00 -1.23255938e-01 5.64902902e-01
1.15992951e+00 -1.43583989e+00 -2.77901858e-01 7.45256543e-02
-8.09225976e-01 1.12390947e+00 1.04112554e+00 4.03216839e-01
1.11754164e-01 7.36754835e-01 1.87270805e-01 -6.23998225e-01
-1.53269255e+00 1.19730905e-01 -1.36179209e-01 1.02703065e-01
5.97291827e-01 3.16391230e-01 -2.70841699e-02 8.77813697e-01
-2.15012282e-01 2.50246823e-01 7.54763544e-01 1.05034983e+00
-9.21054259e-02 -9.11173344e-01 -4.65215057e-01 5.47407448e-01
-2.11365625e-01 -2.30361298e-01 -5.43993235e-01 8.24669898e-01
1.11880086e-01 1.47915173e+00 1.17408037e-02 -1.18116820e+00
8.46946001e-01 2.16813266e-01 1.28080383e-01 -6.01394057e-01
-8.85022640e-01 -1.27460966e-02 6.26149297e-01 -9.18683708e-01
-4.77758795e-02 -7.76740551e-01 -1.31498361e+00 -7.39366710e-01
-2.04933509e-01 -5.42422757e-02 5.32014966e-01 9.34322715e-01
5.82940936e-01 2.88746238e-01 3.43014508e-01 -3.52701604e-01
-5.33991635e-01 -1.05852103e+00 -2.00932607e-01 6.29598200e-01
3.41186643e-01 -4.79705364e-01 -4.07041162e-01 2.47488096e-01] | [8.314888954162598, 7.793886184692383] |
d70044a5-ffaa-4f48-b03c-032d5ff826e7 | deep-learning-based-super-resolution-for | 2210.08080 | null | https://arxiv.org/abs/2210.08080v1 | https://arxiv.org/pdf/2210.08080v1.pdf | Deep Learning based Super-Resolution for Medical Volume Visualization with Direct Volume Rendering | Modern-day display systems demand high-quality rendering. However, rendering at higher resolution requires a large number of data samples and is computationally expensive. Recent advances in deep learning-based image and video super-resolution techniques motivate us to investigate such networks for high-fidelity upscaling of frames rendered at a lower resolution to a higher resolution. While our work focuses on super-resolution of medical volume visualization performed with direct volume rendering, it is also applicable for volume visualization with other rendering techniques. We propose a learning-based technique where our proposed system uses color information along with other supplementary features gathered from our volume renderer to learn efficient upscaling of a low-resolution rendering to a higher-resolution space. Furthermore, to improve temporal stability, we also implement the temporal reprojection technique for accumulating history samples in volumetric rendering. | ['Sumanta Pattanaik', 'Sudarshan Devkota'] | 2022-10-14 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 5.08256018e-01 -3.64335686e-01 3.34482074e-01 -1.87086508e-01
-8.81929398e-01 -5.11897653e-02 4.23405081e-01 -3.09104808e-02
-1.99119657e-01 1.00539279e+00 -7.93040916e-02 -2.51154184e-01
5.98554313e-02 -1.18503225e+00 -4.62458491e-01 -6.67124391e-01
-4.09355402e-01 3.39885384e-01 3.59912992e-01 -1.62594169e-01
2.70512681e-02 8.39069545e-01 -1.68531227e+00 6.36727512e-01
7.72238374e-01 9.29718018e-01 1.81006953e-01 1.06374896e+00
-2.06509292e-01 5.98205447e-01 -5.98012090e-01 1.60602450e-01
3.03540707e-01 -4.92799968e-01 -5.63505590e-01 -3.38973440e-02
8.69886577e-01 -9.34897959e-01 -1.72680378e-01 6.70478940e-01
5.61150372e-01 2.90251970e-01 2.69976258e-01 -7.31420577e-01
-4.22868818e-01 4.15582582e-02 -8.21865499e-01 6.46462977e-01
3.32809567e-01 9.08831954e-02 3.38872999e-01 -6.83446169e-01
8.68982017e-01 1.25116181e+00 4.44453031e-01 6.66747153e-01
-1.57215035e+00 -6.82348967e-01 -1.19715683e-01 -9.32069346e-02
-1.14247012e+00 -5.21528274e-02 8.25390935e-01 -4.27769005e-01
6.32162213e-01 6.12691939e-01 1.14823627e+00 7.15926290e-01
3.79972041e-01 2.54165288e-02 1.53825712e+00 2.59228013e-02
3.34790111e-01 -7.19055310e-02 -3.55726272e-01 9.03355420e-01
1.32272802e-02 3.67450774e-01 -7.21601129e-01 -2.92311579e-01
1.96779060e+00 1.01902865e-01 -3.84207070e-01 -1.93395883e-01
-1.18577075e+00 8.16179395e-01 6.05872452e-01 2.82247335e-01
-4.55066442e-01 3.32372457e-01 3.69737685e-01 1.61439836e-01
1.01938796e+00 3.58990252e-01 -2.78469980e-01 -2.21080780e-01
-1.37621891e+00 1.67223483e-01 3.95366460e-01 3.14084291e-01
7.27770865e-01 5.16109049e-01 -3.82614881e-01 3.86696905e-01
-1.33501858e-01 9.88973752e-02 1.23207837e-01 -1.39577055e+00
2.35488981e-01 2.37160578e-01 3.62487212e-02 -9.44070756e-01
-4.82342660e-01 -3.54747832e-01 -1.23599923e+00 1.04641211e+00
2.20645174e-01 2.82122847e-02 -9.12872851e-01 1.21527278e+00
7.03899205e-01 7.49086738e-01 -1.84684247e-01 1.15226209e+00
8.83151829e-01 8.15144777e-01 -5.67613877e-02 -5.82830727e-01
1.31895638e+00 -5.61918020e-01 -7.71436572e-01 5.80507755e-01
3.08941543e-01 -4.56042916e-01 1.31880569e+00 5.32794237e-01
-1.24185491e+00 -6.31714523e-01 -1.15873075e+00 -4.57914293e-01
-7.22268037e-03 -2.71532804e-01 8.25986564e-01 4.03363973e-01
-1.26294088e+00 1.08304644e+00 -1.11621058e+00 1.34385303e-01
6.50055826e-01 1.45010114e-01 -1.76536173e-01 8.41941983e-02
-8.46544743e-01 5.31823874e-01 -1.15904711e-01 -1.92222506e-01
-7.29307055e-01 -1.45182204e+00 -6.71641707e-01 -1.12958893e-01
-7.98608735e-02 -1.02408528e+00 8.37414563e-01 -6.65488303e-01
-1.70929229e+00 6.41980767e-01 -1.79943398e-01 -2.77323812e-01
8.71614039e-01 1.02543430e-02 -2.25660816e-01 4.92512196e-01
-3.15202713e-01 2.55908400e-01 1.10371280e+00 -1.43424344e+00
-5.63404500e-01 -3.90120834e-01 9.57471654e-02 3.78586084e-01
-1.22040123e-01 -2.93110430e-01 -1.69960499e-01 -5.86392522e-01
1.63208529e-01 -4.48852718e-01 -2.99750775e-01 6.81942463e-01
-7.13921413e-02 2.88470507e-01 1.09253597e+00 -9.03690338e-01
1.04340243e+00 -1.81871021e+00 1.18848354e-01 5.23775630e-03
8.96345139e-01 7.06365854e-02 1.07682921e-01 -5.94822653e-02
-1.50464430e-01 -5.95994145e-02 -2.75368869e-01 -4.72961575e-01
-8.13498855e-01 1.17594622e-01 -1.20719135e-01 2.72162795e-01
3.65714096e-02 6.59652710e-01 -1.04210222e+00 -7.58131921e-01
7.06217170e-01 1.45133352e+00 -6.82976723e-01 1.67797610e-01
5.03845625e-02 1.26750517e+00 -2.79749393e-01 2.86889464e-01
9.59738374e-01 -4.79248375e-01 -5.68919908e-03 -5.23666978e-01
-3.54992777e-01 8.61418769e-02 -1.02666676e+00 1.70643878e+00
-8.72254670e-01 6.82155788e-01 3.22505146e-01 -5.41486621e-01
6.74764872e-01 2.42552385e-01 9.22597170e-01 -7.95364022e-01
-5.45797162e-02 -2.71536320e-01 -3.71181011e-01 -2.89395541e-01
7.78967142e-01 -4.48098183e-01 5.48843384e-01 5.06295562e-01
-4.29717749e-01 -3.51325363e-01 -6.39596730e-02 2.55505055e-01
8.68319213e-01 2.58899957e-01 3.53000760e-02 -7.27493241e-02
2.93364108e-01 -2.96008915e-01 2.50339180e-01 3.86652321e-01
1.58248663e-01 7.58870840e-01 3.53539824e-01 -7.21915901e-01
-1.49544120e+00 -1.26155841e+00 -5.40314198e-01 1.00993562e+00
-3.17605957e-02 -3.26871395e-01 -5.80633402e-01 -2.96934787e-02
-9.00233164e-02 4.22631264e-01 -8.22352588e-01 2.38127083e-01
-9.50409234e-01 -8.96863282e-01 2.36015357e-02 5.63707352e-01
4.34397250e-01 -7.61465847e-01 -1.09911156e+00 2.81488240e-01
6.09679632e-02 -8.45585048e-01 -1.52478382e-01 -1.89194962e-01
-1.47466075e+00 -9.03896332e-01 -9.73405778e-01 -3.11285466e-01
5.87458253e-01 8.07420537e-03 1.28334224e+00 3.95071149e-01
-6.97876871e-01 2.77077734e-01 4.93978225e-02 4.18826863e-02
-4.70294625e-01 -2.96462864e-01 -1.43556923e-01 -1.43713087e-01
-5.29002011e-01 -1.00515318e+00 -7.86148369e-01 -2.05614895e-01
-9.16841447e-01 7.44048655e-01 2.68913284e-02 6.74999237e-01
1.00694466e+00 2.10319296e-03 1.33309901e-01 -9.22226787e-01
6.83145583e-01 5.06453856e-04 -8.59300256e-01 -1.28800750e-01
-4.63352680e-01 3.34202982e-02 7.16589808e-01 -5.34033358e-01
-1.33326399e+00 -2.39318773e-01 -1.61814332e-01 -7.43203044e-01
1.46202490e-01 1.28151298e-01 4.40348268e-01 -3.38907570e-01
6.93084598e-01 1.39017954e-01 -8.65122676e-02 -4.92897451e-01
3.59799176e-01 -9.80017558e-02 4.98898178e-01 -3.74293357e-01
9.33047712e-01 1.05971122e+00 5.06377876e-01 -7.61476815e-01
-5.19192517e-01 8.51743072e-02 -8.17573965e-01 -3.22833538e-01
9.63195145e-01 -6.69715524e-01 -1.08515179e+00 1.85706944e-03
-9.87831891e-01 -5.80877364e-01 -5.31129479e-01 3.79531562e-01
-5.76282799e-01 2.89589345e-01 -8.98453832e-01 -7.71724880e-01
-5.54017782e-01 -1.04250920e+00 1.26775658e+00 1.16964616e-01
7.40222260e-02 -1.06200659e+00 1.69926956e-01 4.16313894e-02
6.94126844e-01 8.60164523e-01 9.05934393e-01 5.17165601e-01
-9.79965687e-01 2.11897120e-01 -5.63377976e-01 8.22554063e-03
3.53109151e-01 1.63242251e-01 -9.36427116e-01 -3.75616014e-01
8.10574833e-03 4.09900062e-02 6.82471633e-01 7.01162338e-01
1.59177089e+00 -2.83055186e-01 -3.17784399e-01 1.26868153e+00
1.65245557e+00 5.76408170e-02 4.94652748e-01 1.37314186e-01
1.11888099e+00 3.36911380e-01 3.79699767e-01 6.74463212e-01
2.37766013e-01 7.02766418e-01 3.13588738e-01 -5.97082019e-01
-4.20135558e-01 1.79522540e-02 -3.10227275e-01 6.74121261e-01
-6.90063179e-01 3.29009056e-01 -5.64708889e-01 -2.68611107e-02
-1.43143857e+00 -9.70019042e-01 -1.13926508e-01 2.40446830e+00
1.00272310e+00 -4.34480831e-02 1.12470612e-01 4.52519879e-02
4.58940506e-01 4.10630107e-01 -7.24906564e-01 -3.50399166e-01
1.52508542e-01 5.75831473e-01 2.08176464e-01 7.53893256e-01
-9.30368781e-01 5.43081701e-01 6.85096884e+00 6.09120548e-01
-1.49308836e+00 3.20486158e-01 8.48570526e-01 -6.35007381e-01
-5.48613191e-01 -5.21940649e-01 -2.94113070e-01 2.55374998e-01
9.33600128e-01 -3.45942616e-01 6.69035852e-01 5.63854635e-01
3.78698081e-01 -1.36202827e-01 -1.15888166e+00 1.41607726e+00
-1.29429430e-01 -2.01287723e+00 3.05872709e-01 7.41584301e-02
7.53016293e-01 -1.97153926e-01 3.62538487e-01 -1.02919219e-02
-2.15640068e-02 -1.08681345e+00 3.81826907e-02 5.69061100e-01
1.43987584e+00 -9.66448307e-01 5.57363071e-02 2.99661364e-02
-1.43227875e+00 3.19273710e-01 -3.28162521e-01 -6.63688481e-02
4.89349872e-01 7.27235079e-01 -7.08179891e-01 4.36800241e-01
7.51626015e-01 5.02384782e-01 -4.19235587e-01 6.89500332e-01
2.50219196e-01 6.82789311e-02 -1.82767972e-01 4.52409387e-01
-2.38655001e-01 -3.61063331e-01 4.86634433e-01 9.40677643e-01
3.71593803e-01 3.45713139e-01 1.17879391e-01 9.67614114e-01
7.06482604e-02 -4.54437584e-02 -4.61682767e-01 4.02216852e-01
2.15020552e-01 1.42685819e+00 -8.01502705e-01 -7.93641388e-01
-2.05413565e-01 1.05631042e+00 2.81831503e-01 4.01251435e-01
-7.88661003e-01 8.79809931e-02 7.42215455e-01 4.78189558e-01
-7.31517524e-02 -4.80982244e-01 -5.46041369e-01 -1.21028090e+00
-9.23334360e-02 -3.38713825e-01 1.79176152e-01 -9.74785745e-01
-9.70706582e-01 1.00045609e+00 -6.19166903e-03 -1.38545728e+00
-3.62975419e-01 -1.62911609e-01 -3.78121138e-01 1.15106142e+00
-1.71733272e+00 -8.72041464e-01 -7.28839874e-01 8.41513276e-01
4.66857731e-01 3.34069550e-01 8.78267586e-01 2.94554651e-01
5.65981446e-03 2.61724174e-01 4.64734249e-02 -1.03821889e-01
3.92113060e-01 -1.15123284e+00 3.39632511e-01 5.48489273e-01
-1.45457029e-01 1.99953914e-01 7.00584292e-01 -7.98566520e-01
-1.25557446e+00 -1.02889287e+00 3.26730013e-02 -2.92167902e-01
2.95625806e-01 -3.12821358e-01 -1.35219884e+00 4.22643751e-01
-4.33857553e-02 4.94913161e-01 4.73224312e-01 -8.09558928e-02
-9.00870189e-02 -1.70030311e-01 -1.47660422e+00 5.54159462e-01
8.69577944e-01 -6.30410731e-01 -1.56996306e-02 3.68458658e-01
8.12445283e-01 -9.86560941e-01 -1.16877019e+00 2.32165918e-01
5.00325024e-01 -1.07916653e+00 1.21330476e+00 -2.94230402e-01
7.38943756e-01 -3.26048970e-01 2.23386690e-01 -1.17072952e+00
-3.75658959e-01 -4.96922463e-01 -4.80197370e-01 5.74636936e-01
-2.39834264e-01 -5.47822773e-01 8.94891143e-01 6.53116167e-01
1.28542483e-01 -8.25625837e-01 -1.27772737e+00 -4.20570970e-01
-2.21994277e-02 -3.65072668e-01 5.78027487e-01 1.09674764e+00
-3.29489291e-01 -1.45958528e-01 -5.14960587e-01 2.28428364e-01
1.14697373e+00 3.70908231e-01 4.40186143e-01 -1.30395293e+00
-2.47168943e-01 -2.14824215e-01 -2.53540784e-01 -6.71509504e-01
-2.98808813e-01 -6.52325988e-01 -3.07846397e-01 -1.73737919e+00
8.70204493e-02 -4.78462756e-01 -8.50085318e-02 -1.15805618e-01
-6.54169843e-02 7.84696281e-01 1.60269931e-01 1.14957169e-01
-7.01701194e-02 3.26191545e-01 1.96015167e+00 2.41659239e-01
-4.89953518e-01 -1.99874461e-01 -1.69648364e-01 6.63387477e-01
6.80623472e-01 -8.25165212e-02 -4.46062088e-01 -5.25636792e-01
5.21998256e-02 7.61008322e-01 5.55959404e-01 -9.14144099e-01
-1.25537008e-01 -9.43320319e-02 9.07804191e-01 -7.90320456e-01
8.98786902e-01 -6.16676033e-01 2.08205312e-01 3.42826843e-01
-3.04787546e-01 1.86603472e-01 2.97132581e-01 3.97274464e-01
1.65562794e-01 4.51253116e-01 1.08594048e+00 -3.22110355e-01
-3.93147737e-01 6.78848565e-01 -1.18278660e-01 -1.52007744e-01
7.54520059e-01 -3.10659289e-01 -9.03623849e-02 -3.78787279e-01
-9.85706151e-01 -2.75560558e-01 5.87080717e-01 1.26400098e-01
9.77929235e-01 -1.44675529e+00 -7.68746197e-01 4.15336668e-01
-4.19397742e-01 1.25834346e-01 7.66166866e-01 5.93540967e-01
-9.43154871e-01 -3.59364040e-02 -4.81052965e-01 -8.33540142e-01
-1.40693450e+00 5.99475980e-01 3.77083242e-01 -3.87172639e-01
-1.49402857e+00 7.48633683e-01 4.30782109e-01 5.90846911e-02
-1.83296818e-02 -3.09154391e-01 -2.17267364e-01 -1.31934986e-01
1.06615639e+00 5.76937973e-01 2.55261585e-02 -4.56393123e-01
-1.36777952e-01 6.99537516e-01 1.95992962e-01 -2.96451002e-01
1.41625011e+00 -1.25793755e-01 -5.29785603e-02 7.46225476e-01
1.25930107e+00 -2.40318865e-01 -1.68278718e+00 -1.57927662e-01
-6.37993634e-01 -8.62011731e-01 4.54706341e-01 -4.42725450e-01
-1.56318927e+00 1.04758108e+00 8.65214288e-01 1.42534405e-01
1.30744302e+00 -2.44175777e-01 7.17844367e-01 -1.15968034e-01
5.16481221e-01 -8.57065797e-01 1.26337022e-01 9.35924128e-02
9.56211209e-01 -1.19673991e+00 4.64256883e-01 -4.76463318e-01
-1.61371633e-01 1.24690807e+00 6.43200696e-01 -2.84816891e-01
6.85663760e-01 7.81148493e-01 3.99137139e-02 -1.89910084e-01
-8.27713311e-01 2.30585888e-01 3.15168083e-01 8.58345270e-01
6.91848218e-01 -4.73144092e-02 -1.16110124e-01 -3.54798317e-01
-2.59721130e-01 2.44371921e-01 7.36130238e-01 6.64558351e-01
-2.10417006e-02 -6.77926362e-01 -4.39931542e-01 7.57538140e-01
-3.09139669e-01 -2.16152027e-01 6.03133798e-01 5.60927629e-01
-5.58799431e-02 2.77259052e-01 6.77350521e-01 4.56812717e-02
5.02495728e-02 -3.27212155e-01 7.69699574e-01 -5.51481903e-01
-4.32971507e-01 3.37179482e-01 -2.00670600e-01 -1.07858300e+00
-6.13422096e-01 -3.62333000e-01 -1.30360639e+00 -4.86004770e-01
3.02259207e-01 -1.58610225e-01 6.88864887e-01 4.75669354e-01
2.47389197e-01 1.29270124e+00 4.27929938e-01 -1.56852901e+00
1.72745705e-01 -5.37320435e-01 -8.28155696e-01 2.86907911e-01
8.37224245e-01 -4.74386245e-01 -2.98437148e-01 1.07051872e-01] | [9.446389198303223, -3.1085453033447266] |
c2b13a3d-1e85-4e5b-8f6f-b23c6304a4f7 | robust-face-recognition-via-adaptive-sparse | 1404.4780 | null | https://arxiv.org/abs/1404.4780v1 | https://arxiv.org/pdf/1404.4780v1.pdf | Robust Face Recognition via Adaptive Sparse Representation | Sparse Representation (or coding) based Classification (SRC) has gained great success in face recognition in recent years. However, SRC emphasizes the sparsity too much and overlooks the correlation information which has been demonstrated to be critical in real-world face recognition problems. Besides, some work considers the correlation but overlooks the discriminative ability of sparsity. Different from these existing techniques, in this paper, we propose a framework called Adaptive Sparse Representation based Classification (ASRC) in which sparsity and correlation are jointly considered. Specifically, when the samples are of low correlation, ASRC selects the most discriminative samples for representation, like SRC; when the training samples are highly correlated, ASRC selects most of the correlated and discriminative samples for representation, rather than choosing some related samples randomly. In general, the representation model is adaptive to the correlation structure, which benefits from both $\ell_1$-norm and $\ell_2$-norm. Extensive experiments conducted on publicly available data sets verify the effectiveness and robustness of the proposed algorithm by comparing it with state-of-the-art methods. | ['Can-Yi Lu', 'Pei-Pei Li', 'Jing Wang', 'Meng Wang', 'Xuegang Hu', 'Shuicheng Yan'] | 2014-04-18 | null | null | null | null | ['robust-face-recognition', 'sparse-representation-based-classification'] | ['computer-vision', 'computer-vision'] | [ 2.49438137e-01 -4.07623917e-01 -3.29196334e-01 -3.82032067e-01
-4.45543051e-01 2.49152303e-01 2.30540484e-01 -3.01602423e-01
-8.46355036e-03 5.05821347e-01 3.27803552e-01 2.40329459e-01
-4.36826885e-01 -6.95571423e-01 -1.54436588e-01 -1.04118228e+00
-8.08166116e-02 -9.08488333e-02 -2.37207770e-01 -7.39175975e-02
3.43510121e-01 3.92897278e-01 -1.59253907e+00 1.60920665e-01
7.54411399e-01 1.24745440e+00 1.07431665e-01 -2.17898324e-01
-1.55511588e-01 6.61977708e-01 -5.71926355e-01 -2.74712313e-02
3.05872649e-01 -5.86613774e-01 -1.31569117e-01 2.21745566e-01
5.12286663e-01 1.99860692e-01 -4.36632454e-01 1.29471564e+00
4.42833751e-01 1.40250787e-01 4.15636510e-01 -1.16153705e+00
-6.76704168e-01 3.27148408e-01 -9.90671992e-01 3.67295831e-01
2.65664369e-01 -2.36028731e-01 7.49042392e-01 -1.14698672e+00
2.82030314e-01 1.25570691e+00 5.91842294e-01 1.95453525e-01
-1.03147614e+00 -1.04242206e+00 3.03789347e-01 3.38827670e-01
-1.82113230e+00 -5.63919961e-01 1.16432822e+00 -3.66960704e-01
3.03679168e-01 1.83844119e-01 5.49908638e-01 6.77602351e-01
-1.34108812e-01 8.02426159e-01 1.14262128e+00 -3.30823511e-01
1.69930220e-01 -3.18523161e-02 1.31660596e-01 6.03224337e-01
4.48892742e-01 8.61709341e-02 -6.58737421e-01 -3.21125686e-01
7.46908128e-01 5.47928512e-01 -6.09019935e-01 -2.86443412e-01
-1.06834459e+00 8.95429313e-01 2.52155185e-01 5.68073630e-01
-4.89463240e-01 -1.81102812e-01 3.10613930e-01 1.77607283e-01
3.13243359e-01 -6.44756332e-02 5.24419069e-04 6.06463626e-02
-1.15876365e+00 -1.90650344e-01 2.73241371e-01 8.96729648e-01
8.50484073e-01 4.30419385e-01 -1.08697928e-01 1.32047105e+00
3.17162752e-01 4.44214493e-01 6.51468635e-01 -6.10813916e-01
4.90753800e-01 7.03517258e-01 -3.14036608e-01 -1.60182726e+00
-5.37485182e-02 -8.47069979e-01 -1.35066259e+00 -2.60209352e-01
-1.20255947e-01 9.79771093e-02 -5.83517551e-01 1.57284677e+00
4.45327498e-02 6.86276734e-01 8.79494324e-02 9.74412620e-01
8.82386506e-01 5.96480131e-01 -1.40119299e-01 -7.18088269e-01
1.13513243e+00 -6.49520636e-01 -8.03328156e-01 -1.28136054e-01
1.75553977e-01 -9.09366131e-01 6.20570362e-01 3.84294689e-01
-7.43419230e-01 -7.72422314e-01 -1.01402950e+00 5.05879700e-01
1.18010409e-01 4.33342725e-01 7.90086031e-01 6.57799065e-01
-6.73964977e-01 2.26901457e-01 -4.67058927e-01 -1.18014917e-01
5.25194824e-01 3.18357974e-01 -4.88986850e-01 -6.38502955e-01
-9.37770605e-01 2.66739070e-01 1.72298506e-01 3.79584491e-01
-5.34841537e-01 -4.88359272e-01 -7.89847434e-01 1.87190920e-01
4.05505538e-01 5.15188277e-02 4.92153436e-01 -1.07148147e+00
-1.23761201e+00 2.60124683e-01 -5.63514352e-01 -2.14546025e-02
5.98329306e-02 -4.46545072e-02 -7.15144873e-01 2.68536896e-01
3.41530479e-02 1.50467828e-01 9.77087438e-01 -1.14901841e+00
-4.00048524e-01 -5.14761090e-01 -3.01232576e-01 -2.49456372e-02
-6.95675492e-01 2.91427672e-02 -5.03951967e-01 -1.02957749e+00
7.07456708e-01 -7.69872665e-01 -2.14842573e-01 -2.39042612e-03
-1.67552769e-01 -1.91584751e-01 1.28136230e+00 -3.65059465e-01
1.35527253e+00 -2.57428122e+00 3.45429368e-02 6.25901103e-01
5.18128611e-02 3.85679454e-01 -2.87708133e-01 4.26273823e-01
-3.82317007e-01 -1.50686771e-01 -3.70376527e-01 -1.02629744e-01
-5.26147842e-01 9.12571773e-02 -1.98347226e-01 6.57509983e-01
6.57665282e-02 2.62030452e-01 -8.21567595e-01 -4.87962216e-01
1.52798861e-01 7.95750916e-01 -4.95540380e-01 9.40267295e-02
3.54523689e-01 3.10506225e-01 -7.96180248e-01 9.37198162e-01
1.03276896e+00 -2.31753081e-01 3.31968695e-01 -4.84170824e-01
4.37313616e-02 -1.97107852e-01 -1.58315361e+00 1.22447729e+00
-3.05305034e-01 3.86934787e-01 4.29871008e-02 -1.36986244e+00
1.39721632e+00 3.84394407e-01 8.01652193e-01 -5.93987286e-01
-5.18523641e-02 1.77341476e-01 -8.95547401e-03 -2.43553132e-01
1.83943406e-01 -1.10701330e-01 3.38114411e-01 2.39970058e-01
-8.70802179e-02 3.82511914e-01 9.44579840e-02 9.71936807e-02
7.95471609e-01 -2.56496519e-01 5.38021564e-01 -1.93522692e-01
9.17085826e-01 -3.43168974e-01 1.05984819e+00 1.39889449e-01
-1.68255895e-01 7.42412090e-01 2.45131999e-01 -2.19145760e-01
-4.26630139e-01 -7.30227649e-01 -2.63499916e-01 5.36138415e-01
2.83782393e-01 -5.90783060e-01 -2.28139311e-01 -7.28624940e-01
-5.95790781e-02 2.82770097e-01 -5.93003929e-01 -2.84840703e-01
-6.44348681e-01 -8.49750102e-01 1.11108564e-01 4.08323973e-01
7.07376838e-01 -8.19333732e-01 -4.52900976e-02 1.36479750e-01
-3.93397436e-02 -8.93426061e-01 -5.39555013e-01 -2.20479667e-01
-8.98396313e-01 -1.18317485e+00 -8.02448630e-01 -8.32329035e-01
1.11278260e+00 1.05510974e+00 5.45335233e-01 3.94684523e-01
-2.61517316e-01 1.34481415e-01 -7.56605566e-01 3.68458480e-02
3.66514504e-01 -3.33999604e-01 4.02728245e-02 5.54343879e-01
5.23540080e-01 -5.62320590e-01 -5.79127848e-01 5.51679134e-01
-7.76237607e-01 -3.14319223e-01 7.20518768e-01 1.14562607e+00
7.65646458e-01 4.97902155e-01 7.65646100e-01 -8.80286634e-01
4.17738140e-01 -7.51036644e-01 -3.41472477e-01 6.32475317e-02
-5.27629435e-01 -3.25179160e-01 1.01467896e+00 -4.07017142e-01
-8.37808549e-01 8.86265635e-02 7.52468081e-03 -7.31771231e-01
1.34213910e-01 7.16544569e-01 -3.16070080e-01 -2.42846712e-01
1.14532977e-01 7.84482658e-01 1.04697615e-01 -5.63694775e-01
-1.45735800e-01 7.90926695e-01 7.17684478e-02 -5.48218966e-01
8.43747914e-01 4.92711097e-01 8.16565976e-02 -1.04953182e+00
-6.32452130e-01 -6.44545019e-01 -3.43060434e-01 -1.28908277e-01
2.52666533e-01 -1.07124007e+00 -5.33731580e-01 1.71378389e-01
-5.07885933e-01 6.18537962e-01 -1.24017164e-01 7.84913659e-01
-8.37630853e-02 5.96745431e-01 -2.40520179e-01 -9.94549811e-01
-9.31827426e-02 -1.31049716e+00 7.66225040e-01 3.67794126e-01
9.20623019e-02 -5.66147685e-01 -3.04108679e-01 3.38807315e-01
4.74260360e-01 4.58595939e-02 8.18138242e-01 -5.10343552e-01
-4.12463814e-01 -2.99540728e-01 -3.36066097e-01 4.32011038e-01
5.51405907e-01 -2.36887671e-02 -6.33569002e-01 -5.77399313e-01
1.61606595e-01 -1.61096036e-01 7.74824321e-01 1.70734495e-01
1.54259038e+00 -4.06515837e-01 -3.35219353e-01 5.44024944e-01
1.51961386e+00 4.87694323e-01 8.60696375e-01 -5.37547618e-02
5.33215523e-01 5.81552029e-01 6.63423598e-01 5.98367929e-01
6.97014555e-02 7.95897603e-01 2.04204261e-01 -7.79943913e-02
-7.84563497e-02 -1.91131920e-01 3.23891550e-01 1.05692923e+00
-1.56516451e-02 2.31994390e-01 -4.15180951e-01 3.21146101e-01
-1.65597963e+00 -1.23237598e+00 2.85836775e-02 2.25006366e+00
6.24450266e-01 -2.98248142e-01 -4.86157313e-02 5.23825109e-01
9.49963272e-01 5.21144807e-01 -3.55993807e-01 7.68378153e-02
-2.15753391e-01 3.32156688e-01 7.23240376e-02 9.12200734e-02
-7.88366854e-01 6.93058014e-01 5.74950075e+00 1.22771537e+00
-1.38329411e+00 -5.39875738e-02 7.62729108e-01 4.10631523e-02
-1.11808635e-01 4.22239304e-02 -8.08251619e-01 8.22979629e-01
2.39250049e-01 4.85132635e-02 3.44599515e-01 1.06759036e+00
1.91533491e-01 -1.42918462e-02 -7.88446665e-01 1.32484138e+00
4.57643300e-01 -1.01994503e+00 1.35490790e-01 5.44012338e-02
8.02941024e-01 -5.40835202e-01 1.65156111e-01 1.88948527e-01
-1.30156636e-01 -1.15228558e+00 3.43937427e-01 5.20179033e-01
7.22492695e-01 -8.19444716e-01 9.28476214e-01 1.56149849e-01
-1.61648166e+00 -1.04462437e-01 -6.63154781e-01 2.65687406e-01
-1.74584299e-01 1.05006790e+00 -3.22692037e-01 7.31063724e-01
5.24395823e-01 1.29964864e+00 -5.17153442e-01 1.01094460e+00
-1.75016537e-01 7.94720888e-01 -1.62252977e-01 2.00727373e-01
2.91258320e-02 -6.07862473e-01 4.38656747e-01 1.21551919e+00
5.88192761e-01 3.80980760e-01 4.31645364e-01 4.19820696e-01
-4.59358236e-03 4.67578769e-01 -4.86202925e-01 -1.80360768e-02
8.34776580e-01 1.10438669e+00 -4.42306966e-01 -2.06207678e-01
-5.80248475e-01 5.98928332e-01 7.64923468e-02 3.86894882e-01
-3.61976594e-01 -3.16567212e-01 6.37216389e-01 -7.54247680e-02
6.56303525e-01 -3.56500983e-01 -4.68494780e-02 -1.22336245e+00
1.20857395e-01 -1.00304031e+00 4.24426168e-01 -2.53483146e-01
-1.37345243e+00 5.95189631e-01 -5.39137796e-02 -1.79828572e+00
1.83151215e-01 -2.29962245e-01 -6.41911566e-01 8.28909814e-01
-1.64195311e+00 -8.63764942e-01 -4.05059814e-01 6.92970991e-01
5.92053950e-01 -6.34499550e-01 5.51085770e-01 4.83250439e-01
-8.50087047e-01 6.28713131e-01 2.21771270e-01 1.84147060e-01
4.80285615e-01 -4.25892651e-01 -5.10900438e-01 7.61894107e-01
1.25682548e-01 1.13933015e+00 3.20797563e-01 -3.86564225e-01
-1.63245487e+00 -1.05273020e+00 7.02483654e-01 5.25713146e-01
1.47459835e-01 4.50632162e-02 -1.01960123e+00 2.22423315e-01
-1.81656420e-01 4.52106953e-01 1.04252720e+00 1.08366735e-01
-6.50101840e-01 -6.48852885e-01 -1.26237893e+00 4.13799047e-01
9.68265951e-01 -3.87233078e-01 -3.68930340e-01 1.75122827e-01
1.10543650e-02 1.77134812e-01 -8.69478524e-01 5.76563358e-01
5.64351022e-01 -1.09262323e+00 9.87845600e-01 -7.20308647e-02
2.23396614e-01 -7.92910576e-01 -5.77020943e-01 -1.11320019e+00
-5.58240652e-01 -1.47540435e-01 -6.87577501e-02 1.26871800e+00
-7.40377372e-03 -7.52635479e-01 8.31073940e-01 -6.81620762e-02
8.36967900e-02 -9.56381023e-01 -1.02911437e+00 -8.56221437e-01
-5.20301998e-01 -9.69035178e-02 5.66932619e-01 1.14930296e+00
-9.93204340e-02 4.36536632e-02 -5.49177885e-01 1.63567036e-01
7.44135380e-01 4.91464287e-01 5.38208783e-01 -1.23595393e+00
-1.79146096e-01 -3.10485661e-01 -7.42215633e-01 -1.06808090e+00
1.62324771e-01 -7.39377379e-01 -2.13667169e-01 -1.32943726e+00
3.63382965e-01 -7.99810886e-01 -5.84054112e-01 5.36095858e-01
-2.52178043e-01 3.47007513e-01 2.63963252e-01 6.14586532e-01
-3.14694822e-01 8.50628316e-01 1.10492587e+00 -1.98569089e-01
-3.07590216e-02 -6.10899366e-02 -9.87104595e-01 5.18099427e-01
5.72410762e-01 -4.06246006e-01 -4.08905476e-01 -1.65904209e-01
-3.08873206e-01 9.53983143e-02 -1.32621145e-02 -1.15424669e+00
1.58308864e-01 -4.12682652e-01 5.71124971e-01 -4.11684364e-01
5.49654067e-01 -1.02792799e+00 2.92474151e-01 4.78068978e-01
-1.54294387e-01 -2.11403534e-01 -2.16088489e-01 6.02121651e-01
-6.52460814e-01 -2.01438740e-01 1.00587142e+00 -8.61912072e-02
-5.99611342e-01 5.84815085e-01 3.01803742e-02 -1.86314985e-01
1.03210890e+00 -5.36201537e-01 -6.12151437e-02 -3.27948034e-01
-2.78945118e-01 -3.11662294e-02 2.27187246e-01 3.15835953e-01
1.03790092e+00 -1.47820306e+00 -8.38052273e-01 5.85056782e-01
2.61040390e-01 -2.61801571e-01 3.46162587e-01 8.67572188e-01
1.58832576e-02 3.28495741e-01 -7.18982890e-02 -4.82649654e-01
-1.36392379e+00 4.28606242e-01 -6.56964108e-02 8.89898315e-02
-5.64816177e-01 7.12782979e-01 1.22972138e-01 1.69433564e-01
1.28753990e-01 2.21820563e-01 -6.21587634e-01 1.45961314e-01
7.23622799e-01 3.34323972e-01 -2.30571162e-02 -9.58101690e-01
-5.37077785e-01 9.72327054e-01 -2.07782939e-01 2.88692951e-01
1.37257564e+00 2.21629933e-01 -3.29398096e-01 1.00152314e-01
1.38469517e+00 3.86724442e-01 -9.87038791e-01 -4.72740769e-01
-2.77172863e-01 -1.15437186e+00 1.60495684e-01 -3.61574292e-01
-1.72925842e+00 7.55021513e-01 5.78918695e-01 -9.02695879e-02
1.32481670e+00 -4.28820550e-01 7.02068806e-01 1.76574633e-01
7.59855270e-01 -7.44324446e-01 2.28510439e-01 3.36671293e-01
9.18985844e-01 -1.13462937e+00 3.44881386e-01 -8.42611492e-01
-5.81839979e-01 1.03802311e+00 7.00931430e-01 -3.50015670e-01
9.00150180e-01 1.09657403e-02 -1.48583502e-01 3.65199298e-02
-5.51656067e-01 4.81470488e-03 1.92682832e-01 6.36049092e-01
5.70473433e-01 -5.25120348e-02 -6.07969522e-01 4.62018788e-01
-5.79419248e-02 -7.12500652e-03 1.46964192e-01 1.01204038e+00
-3.96495104e-01 -1.25471342e+00 -5.73822498e-01 6.45484984e-01
-2.82660633e-01 -1.17868908e-01 -5.77202141e-02 4.41186219e-01
2.83385843e-01 1.08552861e+00 -1.13708563e-01 -5.13407886e-01
4.14658561e-02 -3.10919523e-01 1.90674245e-01 -6.61499858e-01
-3.40607375e-01 1.85294300e-01 -2.93895900e-01 -3.88591200e-01
-5.84495485e-01 -7.59054661e-01 -1.02822268e+00 -8.16380307e-02
-3.73824656e-01 6.04638755e-01 4.70426023e-01 7.44097710e-01
4.86419916e-01 2.72453904e-01 1.26535773e+00 -4.98323768e-01
-4.12741959e-01 -9.22303081e-01 -8.37704957e-01 2.38264754e-01
1.54408753e-01 -9.39366043e-01 -4.47475284e-01 -4.55202870e-02] | [12.478398323059082, 0.426063597202301] |
53810929-3e14-455a-9dd7-7c21f6f7acc7 | bigearthnet-mm-a-large-scale-multi-modal | 2105.07921 | null | https://arxiv.org/abs/2105.07921v2 | https://arxiv.org/pdf/2105.07921v2.pdf | BigEarthNet-MM: A Large Scale Multi-Modal Multi-Label Benchmark Archive for Remote Sensing Image Classification and Retrieval | This paper presents the multi-modal BigEarthNet (BigEarthNet-MM) benchmark archive made up of 590,326 pairs of Sentinel-1 and Sentinel-2 image patches to support the deep learning (DL) studies in multi-modal multi-label remote sensing (RS) image retrieval and classification. Each pair of patches in BigEarthNet-MM is annotated with multi-labels provided by the CORINE Land Cover (CLC) map of 2018 based on its thematically most detailed Level-3 class nomenclature. Our initial research demonstrates that some CLC classes are challenging to be accurately described by only considering (single-date) BigEarthNet-MM images. In this paper, we also introduce an alternative class-nomenclature as an evolution of the original CLC labels to address this problem. This is achieved by interpreting and arranging the CLC Level-3 nomenclature based on the properties of BigEarthNet-MM images in a new nomenclature of 19 classes. In our experiments, we show the potential of BigEarthNet-MM for multi-modal multi-label image retrieval and classification problems by considering several state-of-the-art DL models. We also demonstrate that the DL models trained from scratch on BigEarthNet-MM outperform those pre-trained on ImageNet, especially in relation to some complex classes, including agriculture and other vegetated and natural environments. We make all the data and the DL models publicly available at https://bigearth.net, offering an important resource to support studies on multi-modal image scene classification and retrieval problems in RS. | ['Volker Markl', 'Begüm Demir', 'Mário Caetano', 'Pedro Benevides', 'Hugo Costa', 'Filipe Marcelino', 'Tristan Kreuziger', 'Arne de Wall', 'Gencer Sumbul'] | 2021-05-17 | null | null | null | null | ['multi-label-image-retrieval', 'remote-sensing-image-classification'] | ['computer-vision', 'miscellaneous'] | [ 1.66468874e-01 -3.75791699e-01 -2.42858931e-01 -4.74905521e-01
-9.96993482e-01 -6.91527128e-01 8.06696832e-01 4.73472983e-01
-5.44560730e-01 4.90650594e-01 -1.74297363e-01 -4.29185480e-01
-2.31969312e-01 -1.22089684e+00 -5.85053444e-01 -8.13802242e-01
-2.48795152e-01 4.88402963e-01 -1.57029852e-01 -4.44934636e-01
-4.04131711e-01 7.62435198e-01 -1.97090364e+00 6.58742428e-01
4.38423038e-01 1.35145617e+00 3.35889906e-01 4.72058415e-01
-7.86400512e-02 6.81516707e-01 -4.74775471e-02 -5.75027689e-02
6.84802830e-01 7.96985999e-02 -8.09495270e-01 -3.99750769e-01
8.69932413e-01 -3.95583689e-01 1.47679776e-01 1.15363574e+00
5.56084871e-01 -1.03967369e-01 8.42906833e-01 -1.25165284e+00
-6.35228693e-01 2.69720674e-01 -5.90705752e-01 -1.10191181e-01
-1.53665841e-01 -2.60073513e-01 1.29544318e+00 -7.64900684e-01
6.67699814e-01 1.35079312e+00 1.06185257e+00 2.05865696e-01
-8.01259160e-01 -8.94784391e-01 8.39660317e-02 2.17624158e-02
-1.83998477e+00 3.03264782e-02 6.37569418e-03 -4.81044233e-01
8.59480381e-01 4.84291285e-01 4.93911415e-01 6.26344264e-01
2.74335802e-01 7.15136409e-01 1.32202375e+00 -3.05334598e-01
1.84662770e-02 -2.19944969e-01 5.67710437e-02 6.37783408e-01
6.22774549e-02 2.65334755e-01 2.60524787e-02 -3.82819742e-01
2.69275814e-01 4.59731966e-01 2.89632138e-02 -2.32507393e-01
-1.07125640e+00 1.13699532e+00 9.67108905e-01 4.17490393e-01
-6.18547916e-01 2.47708723e-01 4.50849921e-01 4.26726192e-01
1.01211059e+00 1.17159568e-01 -6.80145085e-01 7.15497911e-01
-1.06594670e+00 2.56610036e-01 5.31883717e-01 9.15135920e-01
1.47031116e+00 -1.79757386e-01 9.76530463e-02 1.18571138e+00
5.87461829e-01 1.36798239e+00 -7.14948699e-02 -7.00444996e-01
-8.18075426e-03 5.54098189e-01 -9.81826615e-03 -1.07428384e+00
-7.19329715e-01 -4.97898370e-01 -9.70963061e-01 -1.52790561e-01
-1.38390124e-01 3.96452807e-02 -1.25913095e+00 1.45704901e+00
2.52913415e-01 -1.39452904e-01 9.93966535e-02 8.21147799e-01
1.29859626e+00 8.66270304e-01 5.86091399e-01 3.96660388e-01
1.52152050e+00 -7.70274878e-01 -3.78413022e-01 -2.86457419e-01
9.49750066e-01 -6.26584530e-01 7.10431993e-01 -1.36578515e-01
2.48471880e-03 -4.72701520e-01 -7.14849770e-01 2.52993494e-01
-1.30338240e+00 2.18014568e-01 8.48958015e-01 2.96885520e-01
-1.27250826e+00 1.26638547e-01 -4.16459024e-01 -9.72049892e-01
4.77803528e-01 -1.08241059e-01 -5.81888139e-01 -4.84043658e-01
-1.51943254e+00 9.22462046e-01 5.03968775e-01 3.83406907e-01
-1.26815975e+00 -8.66432905e-01 -9.98145580e-01 -1.68033034e-01
4.80980203e-02 -2.33371958e-01 9.03036833e-01 -8.51680040e-01
-5.20360768e-01 1.60870564e+00 4.14597541e-01 -3.05735737e-01
-6.77742864e-05 6.93423748e-02 -7.59937406e-01 3.46803993e-01
6.36568546e-01 1.21612024e+00 4.19212371e-01 -1.64499843e+00
-1.03927946e+00 -4.53324139e-01 3.32639843e-01 8.49175230e-02
-1.76668927e-01 2.91717909e-02 1.16493553e-01 -5.52768767e-01
1.39272466e-01 -1.03803360e+00 -2.88821369e-01 1.04059622e-01
-1.59123927e-01 -2.02140629e-01 7.36806691e-01 -1.84818104e-01
7.14095175e-01 -2.21128058e+00 -4.52001125e-01 1.52978107e-01
7.52142891e-02 3.55534106e-01 -9.03563321e-01 6.94720209e-01
-8.67532566e-02 3.61140847e-01 -2.75521070e-01 -1.10096440e-01
7.26554543e-02 4.50458944e-01 -3.88522208e-01 8.48738909e-01
-5.72451726e-02 9.63104367e-01 -1.05183709e+00 -3.11997980e-01
3.34468126e-01 5.33507228e-01 7.28691444e-02 -2.54928879e-02
-4.87760425e-01 1.78362310e-01 -4.29380566e-01 1.33410203e+00
1.16328919e+00 -1.64667338e-01 -6.23325780e-02 -4.20495778e-01
-2.93931067e-01 -5.08106649e-01 -6.54660761e-01 1.58670521e+00
-6.16393805e-01 4.71841604e-01 1.38602704e-01 -9.81384993e-01
9.72334564e-01 6.53680936e-02 7.20005512e-01 -9.71796751e-01
2.84196250e-03 6.08829737e-01 -4.79416668e-01 -2.56210804e-01
5.77536523e-01 -3.58662277e-01 -2.70171344e-01 5.40132582e-01
1.34000465e-01 -2.05700725e-01 1.01520665e-01 -4.83872481e-02
6.25837147e-01 -9.83825028e-02 2.00826958e-01 -6.54206038e-01
4.06550169e-01 4.53663141e-01 2.45434836e-01 1.18068480e+00
-2.62328655e-01 7.30044246e-01 -1.14119627e-01 -1.05697274e+00
-9.26094472e-01 -8.20854008e-01 -7.70613611e-01 1.49881542e+00
-3.52420174e-02 -2.21664339e-01 -1.89011425e-01 -4.18117195e-01
3.94043803e-01 3.20050687e-01 -9.29330945e-01 3.25364888e-01
5.11802323e-02 -1.27796471e+00 1.11426783e+00 -4.97331880e-02
8.63790929e-01 -1.14692223e+00 -5.27999580e-01 4.77380008e-02
-2.06881613e-01 -1.07973027e+00 2.42334872e-01 2.51294434e-01
-2.65248686e-01 -1.26812851e+00 -8.52021396e-01 -6.77559078e-01
3.55489254e-01 6.16227567e-01 1.40216625e+00 2.21422613e-01
-4.22874600e-01 5.78319490e-01 -8.63870621e-01 -4.45031762e-01
-3.30946207e-01 3.25564444e-01 -4.04974669e-01 1.42193496e-01
6.97084367e-01 -2.28734344e-01 -6.30353093e-01 2.54659891e-01
-1.44857979e+00 -1.61604404e-01 5.71712971e-01 5.74274242e-01
9.12078857e-01 -1.13412254e-02 3.52942556e-01 -9.55691993e-01
-8.07953533e-03 -9.43305016e-01 -6.61302567e-01 6.04276299e-01
-5.15341818e-01 -4.38833028e-01 1.66771069e-01 9.14220586e-02
-5.86294770e-01 1.16159432e-01 -3.01069617e-01 -3.18740368e-01
-4.97056246e-01 1.04208326e+00 2.75400639e-01 -4.28167135e-01
6.68034375e-01 1.75553665e-01 -1.84872687e-01 -4.95594054e-01
5.51115990e-01 1.13560295e+00 1.78776965e-01 -5.27132809e-01
5.68818986e-01 7.93838084e-01 1.00637659e-01 -9.53010738e-01
-1.46508956e+00 -1.01337981e+00 -5.63325942e-01 -3.68578136e-01
1.00115871e+00 -1.59331048e+00 -4.28487331e-01 8.88513267e-01
-8.85396957e-01 -5.35945475e-01 -5.41594028e-02 1.34221762e-01
-1.56994611e-01 1.30624799e-02 -4.37019974e-01 -6.49726033e-01
-5.89954197e-01 -1.02251017e+00 1.86577940e+00 -1.13718338e-01
5.30222893e-01 -1.12972975e+00 2.60523081e-01 8.81583020e-02
8.13649178e-01 6.53338373e-01 8.34534407e-01 -6.37187362e-01
-3.04281324e-01 -2.17612877e-01 -6.94746137e-01 4.39873427e-01
1.46183863e-01 -1.95925757e-01 -1.17147744e+00 -6.49603963e-01
-4.62654173e-01 -7.00307012e-01 1.43870962e+00 3.55046839e-01
9.31101322e-01 -7.58607686e-03 -1.59119874e-01 8.51685941e-01
2.13043666e+00 -1.64092109e-01 5.21484852e-01 6.99830115e-01
7.64778554e-01 6.70440376e-01 8.81932855e-01 4.07641321e-01
7.07183540e-01 3.50575775e-01 1.08994412e+00 -6.33310914e-01
-6.94711581e-02 1.07695632e-01 -8.47590938e-02 3.93479407e-01
1.28453791e-01 -4.12507296e-01 -1.38861990e+00 7.74053574e-01
-1.76419020e+00 -8.85998785e-01 -3.81274223e-01 1.78128028e+00
5.31631589e-01 -7.75504708e-01 -3.84275854e-01 -4.70950365e-01
5.85562825e-01 7.97833860e-01 -4.11890149e-01 9.48783085e-02
-7.81275511e-01 3.66786003e-01 1.22133291e+00 4.70772654e-01
-1.80278409e+00 1.22641504e+00 5.91657019e+00 1.08724236e+00
-1.29894054e+00 4.44539309e-01 4.70117718e-01 3.92364860e-01
-3.19087833e-01 5.67263598e-03 -9.47896957e-01 -1.28444396e-02
8.78141165e-01 4.80488598e-01 2.08403453e-01 6.87518835e-01
-2.34458551e-01 -3.10676217e-01 -5.82935393e-01 9.20331478e-01
1.96941709e-03 -1.31784570e+00 4.91795003e-01 1.35893315e-01
1.01800680e+00 1.15794361e+00 -1.14137540e-02 3.45136672e-01
5.27115762e-01 -9.53500092e-01 6.86710835e-01 3.19922447e-01
1.24376476e+00 -2.66608417e-01 1.11661756e+00 5.08353226e-02
-1.48642325e+00 -1.55566990e-01 -7.24176407e-01 2.54736394e-01
-2.74929374e-01 6.90770447e-01 -2.94700235e-01 1.05721557e+00
1.12730598e+00 1.20195794e+00 -8.06150496e-01 8.99129450e-01
3.24186757e-02 3.34545285e-01 -4.82296735e-01 2.74679571e-01
8.89347434e-01 -1.32894844e-01 9.14299935e-02 1.36057317e+00
4.70204175e-01 1.03056571e-02 5.84354579e-01 5.42489767e-01
7.01952144e-04 1.77500978e-01 -9.24325705e-01 5.91937080e-03
4.12820637e-01 1.62960565e+00 -5.84573507e-01 -2.86150813e-01
-4.07767534e-01 4.44823295e-01 -7.90811405e-02 5.35984576e-01
-5.46599984e-01 -1.03701897e-01 6.62410855e-01 -4.35251147e-02
7.81769678e-02 5.20497076e-02 3.13229918e-01 -1.02166915e+00
-5.82332134e-01 -7.88461685e-01 7.45006680e-01 -1.13399088e+00
-1.58147359e+00 9.43476081e-01 2.98031807e-01 -1.35324907e+00
6.77447245e-02 -7.57672369e-01 1.03659980e-01 9.23470438e-01
-2.49373126e+00 -2.08395886e+00 -6.59094512e-01 6.14501953e-01
3.60716134e-02 -2.20034584e-01 1.42478704e+00 6.17962897e-01
-7.96027407e-02 4.16076854e-02 5.76753020e-01 2.44847294e-02
7.61564791e-01 -1.03737116e+00 1.05286483e-02 3.61091584e-01
-4.60308678e-02 1.72439754e-01 2.02795058e-01 -3.78926009e-01
-1.04592514e+00 -1.82330513e+00 7.81572521e-01 -4.81830575e-02
9.52847838e-01 -2.92132527e-01 -4.83402193e-01 7.43609905e-01
9.41084176e-02 3.82428437e-01 9.14095521e-01 -3.61381806e-02
-6.76614106e-01 -6.10007524e-01 -1.17429006e+00 -1.06776267e-01
7.46643543e-01 -9.27418232e-01 1.16746444e-02 1.08124292e+00
5.04960179e-01 -7.35941082e-02 -1.09052980e+00 8.35206568e-01
5.88598490e-01 -7.00129449e-01 1.01586401e+00 -2.79069602e-01
3.61284852e-01 -3.11970413e-01 -9.46997166e-01 -1.26254380e+00
-3.39460194e-01 3.30109894e-01 8.83958876e-01 7.48289764e-01
2.54097968e-01 -8.04385841e-01 1.69742808e-01 -2.03459993e-01
-2.12844059e-01 -2.59185731e-01 -8.18664908e-01 -6.68296278e-01
4.74774539e-01 -4.30788815e-01 8.30326319e-01 1.14753795e+00
-1.00454998e+00 -1.24741852e-01 -2.72868872e-01 5.11948705e-01
6.95039213e-01 5.29920936e-01 5.43280661e-01 -1.56360292e+00
5.02769947e-01 -2.04623073e-01 -3.25029969e-01 -6.23081446e-01
3.96759540e-01 -1.29220533e+00 9.35299844e-02 -1.59542191e+00
3.24882478e-01 -1.06428254e+00 -6.05503201e-01 1.17308474e+00
3.00770551e-01 8.85239244e-01 2.15527937e-01 7.30166912e-01
-8.43178988e-01 4.43681955e-01 9.75902796e-01 -7.51523912e-01
5.08587003e-01 -3.37936580e-01 -4.25618470e-01 2.71030545e-01
7.33211696e-01 -7.17109442e-01 1.81750823e-02 -7.33051717e-01
5.44972062e-01 -3.25253218e-01 6.45748436e-01 -9.03560758e-01
-3.82567421e-02 -1.83732554e-01 -1.17697112e-01 -1.07350814e+00
1.27649739e-01 -1.06154501e+00 5.58494508e-01 3.85212213e-01
-1.80241570e-01 -1.40140191e-01 4.77866739e-01 2.56524444e-01
-5.31277597e-01 -5.70471697e-02 7.73183703e-01 -4.38010275e-01
-1.22556353e+00 7.63348639e-01 -4.01864469e-01 -3.11576873e-01
7.92494655e-01 2.96633512e-01 -6.56529665e-01 -1.40265644e-01
-6.54756725e-01 4.85638142e-01 4.17026699e-01 3.70859087e-01
2.25947514e-01 -1.45630920e+00 -1.09226191e+00 3.06207985e-01
8.79282296e-01 7.57245943e-02 3.90207022e-01 5.64444661e-01
-1.04294944e+00 8.05400074e-01 -3.21088910e-01 -9.48287666e-01
-8.54715347e-01 1.82179525e-01 6.86953008e-01 -3.24145257e-01
-2.29123220e-01 6.15039468e-01 3.67476642e-01 -1.29671848e+00
-7.63684511e-02 -7.31486529e-02 -3.01716805e-01 4.48753387e-01
3.70056957e-01 -2.40304291e-01 3.05428863e-01 -1.05141962e+00
-5.43970942e-01 8.04924786e-01 2.54001558e-01 2.81206012e-01
1.80514419e+00 -5.08042350e-02 -8.27606499e-01 4.69622612e-01
1.50129950e+00 -6.28330886e-01 -6.72620773e-01 -3.43264699e-01
-7.08159357e-02 -2.63046622e-01 3.58502805e-01 -1.06867146e+00
-1.38388944e+00 7.49130845e-01 1.06907225e+00 3.29220116e-01
1.13506806e+00 7.68694207e-02 5.24765015e-01 6.45747900e-01
6.36033714e-01 -7.31422961e-01 -4.96990085e-01 7.50838459e-01
8.70654702e-01 -1.60232353e+00 -7.91173987e-03 -6.20281883e-02
4.45379950e-02 8.41894984e-01 1.47138119e-01 1.98625829e-02
1.28500187e+00 -1.14310637e-01 6.38460219e-01 -8.64384174e-01
-3.28453988e-01 -8.15514743e-01 8.37214217e-02 5.53752482e-01
2.46102557e-01 5.71909010e-01 -1.10240296e-01 -6.02028035e-02
2.86511302e-01 -1.71869934e-01 1.08503021e-01 1.03468692e+00
-5.35564363e-01 -1.10466337e+00 -4.45444077e-01 5.07760823e-01
-3.45302999e-01 -6.59571886e-01 -1.62568033e-01 8.38842690e-01
3.59171987e-01 9.33887959e-01 3.22328284e-02 -2.84915805e-01
1.35408029e-01 -3.36073697e-01 -8.42305308e-04 -6.84947193e-01
-5.20421267e-01 -1.72625259e-01 -7.93006793e-02 -5.39751351e-01
-1.26366663e+00 -4.53268677e-01 -6.72295570e-01 -4.44054753e-01
-1.95744053e-01 1.10331401e-01 8.54579568e-01 7.43027985e-01
3.20666820e-01 3.22780833e-02 6.18048847e-01 -9.33417439e-01
-3.45188975e-01 -1.16052866e+00 -1.17000830e+00 3.91203821e-01
4.65040028e-01 -6.77112877e-01 -4.52916652e-01 -3.85341763e-01] | [9.599770545959473, -1.46067214012146] |
a730938d-b2ac-4d03-a81c-bdf01776d429 | region-refinement-network-for-salient-object | 1906.11443 | null | https://arxiv.org/abs/1906.11443v2 | https://arxiv.org/pdf/1906.11443v2.pdf | Region Refinement Network for Salient Object Detection | Albeit intensively studied, false prediction and unclear boundaries are still major issues of salient object detection. In this paper, we propose a Region Refinement Network (RRN), which recurrently filters redundant information and explicitly models boundary information for saliency detection. Different from existing refinement methods, we propose a Region Refinement Module (RRM) that optimizes salient region prediction by incorporating supervised attention masks in the intermediate refinement stages. The module only brings a minor increase in model size and yet significantly reduces false predictions from the background. To further refine boundary areas, we propose a Boundary Refinement Loss (BRL) that adds extra supervision for better distinguishing foreground from background. BRL is parameter free and easy to train. We further observe that BRL helps retain the integrity in prediction by refining the boundary. Extensive experiments on saliency detection datasets show that our refinement module and loss bring significant improvement to the baseline and can be easily applied to different frameworks. We also demonstrate that our proposed model generalizes well to portrait segmentation and shadow detection tasks. | ['Ruiyu Li', 'Hengshuang Zhao', 'Xiaoyong Shen', 'Jiaya Jia', 'Zhuotao Tian', 'Jiaze Wang', 'Michelle Shu'] | 2019-06-27 | null | null | null | null | ['shadow-detection'] | ['computer-vision'] | [ 5.77161551e-01 4.88085836e-01 -3.96428078e-01 -4.22253728e-01
-6.15551054e-01 -1.47594109e-01 3.96653771e-01 1.40440375e-01
-1.51672199e-01 6.30185187e-01 2.29317829e-01 -9.37249139e-02
3.88782263e-01 -5.61826408e-01 -8.46006095e-01 -4.07196492e-01
1.95269659e-01 4.51904014e-02 1.08575392e+00 -2.51349688e-01
2.47652680e-01 4.54421878e-01 -1.40278900e+00 2.70716071e-01
1.26044834e+00 1.05643368e+00 5.77299476e-01 3.40394258e-01
1.70527890e-01 1.11179864e+00 -3.63719553e-01 -2.03583583e-01
3.11275959e-01 -1.11194447e-01 -9.82109249e-01 3.56012195e-01
6.32437229e-01 -4.89604414e-01 -5.71556874e-02 9.12286699e-01
2.40248412e-01 7.27566257e-02 3.38158667e-01 -1.20938957e+00
-5.86359262e-01 4.64289516e-01 -1.17654085e+00 3.68002832e-01
6.61879405e-02 -6.81378320e-02 1.18653321e+00 -1.16986549e+00
5.73111176e-01 1.20150352e+00 7.50121057e-01 5.56116343e-01
-1.33105397e+00 -5.89044273e-01 8.21372569e-01 2.22144768e-01
-1.33592534e+00 -3.30856621e-01 8.62064540e-01 -6.24803528e-02
3.99885595e-01 4.35883224e-01 5.81428528e-01 8.73434365e-01
-4.86937957e-03 1.53570795e+00 8.67525578e-01 -2.10327968e-01
-3.13533880e-02 2.11649522e-01 2.32897729e-01 7.80179501e-01
6.93798587e-02 -1.46583870e-01 -5.10245144e-01 7.63519481e-02
1.00178885e+00 1.84982434e-01 -4.88461554e-01 -6.78534210e-01
-1.08866382e+00 7.40019560e-01 8.46998572e-01 -1.35082573e-01
-1.41778335e-01 1.12263039e-01 -4.00747620e-02 -2.22852975e-01
6.59019113e-01 3.94341111e-01 -2.80070841e-01 4.37077790e-01
-1.42990601e+00 1.85037345e-01 3.42806399e-01 1.12641335e+00
9.16035771e-01 -8.34460184e-02 -6.88518226e-01 8.59148622e-01
2.29609236e-01 2.90492892e-01 8.42992887e-02 -1.06283593e+00
1.83270246e-01 7.80821145e-01 4.26333308e-01 -9.99655843e-01
-4.24986631e-01 -7.90168345e-01 -6.94993019e-01 1.49555638e-01
1.30477145e-01 -4.30058734e-03 -1.17833221e+00 1.64602184e+00
5.42926013e-01 5.36564469e-01 -3.08372676e-01 1.19417882e+00
8.98872256e-01 3.96556139e-01 1.28327712e-01 1.56202585e-01
1.06690979e+00 -1.61458230e+00 -4.78009045e-01 -6.02078199e-01
2.37654939e-01 -6.39506102e-01 1.05928981e+00 3.42037603e-02
-1.27887201e+00 -4.81261849e-01 -8.97132516e-01 -5.05761445e-01
9.18989629e-02 2.27064744e-01 7.43627131e-01 8.85096565e-02
-1.27001286e+00 4.13209260e-01 -8.24544311e-01 -1.27335772e-01
8.14337611e-01 4.12748784e-01 6.05235212e-02 1.70141041e-01
-1.06849349e+00 8.97588909e-01 1.77440479e-01 1.75078019e-01
-9.77614462e-01 -9.54078615e-01 -1.22117102e+00 2.02861354e-01
6.61603391e-01 -7.03994393e-01 1.34392345e+00 -1.20201218e+00
-1.18811738e+00 7.12119222e-01 -6.42558932e-01 -7.29815960e-01
6.30533636e-01 -5.10438919e-01 -3.15957852e-02 2.24398673e-01
3.98098171e-01 1.31060469e+00 1.12732756e+00 -1.46544039e+00
-1.00614238e+00 1.35059521e-01 1.13273554e-01 1.81817621e-01
-2.56058257e-02 -6.97702542e-02 -7.27641344e-01 -8.96744311e-01
2.51252145e-01 -7.58456171e-01 -6.68956757e-01 2.84914106e-01
-5.25136471e-01 2.84527689e-02 1.02475488e+00 -7.19764173e-01
1.25732708e+00 -2.01573658e+00 2.18228456e-02 8.62432867e-02
4.86863703e-01 1.70171678e-01 -1.98463663e-01 -3.05694073e-01
2.05874637e-01 3.22557092e-02 -5.99702060e-01 -5.69992602e-01
-1.51373088e-01 1.19244039e-01 -5.74140668e-01 1.69953763e-01
7.73891568e-01 1.30060840e+00 -8.97747934e-01 -7.51280129e-01
1.10327519e-01 5.06167948e-01 -6.79719210e-01 1.04238054e-04
-2.66332269e-01 3.79593670e-01 -3.02211642e-01 8.26246321e-01
9.33111548e-01 -7.78421104e-01 -3.60931396e-01 -6.23300001e-02
-8.77134055e-02 3.31217915e-01 -9.56697404e-01 1.50720108e+00
-1.45407051e-01 6.97648704e-01 2.11122200e-01 -6.91812277e-01
8.17003787e-01 -2.13354275e-01 1.47057250e-01 -6.63116872e-01
-9.74739566e-02 -1.76067144e-01 -2.31020480e-01 4.47909459e-02
9.60201263e-01 6.97442144e-02 2.77027022e-02 1.91282332e-01
-1.64754331e-01 1.52619064e-01 -1.30178943e-01 3.53470743e-01
9.27705705e-01 3.50661725e-01 1.46928072e-01 -3.54265153e-01
5.39423823e-01 -7.18029365e-02 1.17605841e+00 8.47166836e-01
-5.16839623e-01 1.02536607e+00 3.59671772e-01 -2.67288685e-01
-7.63555527e-01 -1.03521597e+00 -3.13448124e-02 1.27716744e+00
9.79848504e-01 -2.37529173e-01 -7.11091101e-01 -9.16935027e-01
-2.57527381e-02 5.83355069e-01 -6.89214468e-01 -1.21856116e-01
-7.28190958e-01 -5.91374218e-01 3.86959182e-05 8.48847628e-01
7.17927337e-01 -1.10390007e+00 -6.97641790e-01 2.37440895e-02
-3.84541839e-01 -1.10442674e+00 -8.16340029e-01 4.14337069e-02
-8.89888108e-01 -7.87961662e-01 -9.01299477e-01 -8.95382941e-01
7.92923927e-01 8.12967300e-01 1.25341117e+00 2.78112054e-01
-2.41177067e-01 9.06851068e-02 -2.70052195e-01 -5.87078869e-01
-4.02182229e-02 1.68095082e-01 -3.96451473e-01 3.85462083e-02
-5.40484488e-02 -2.57106841e-01 -8.83307457e-01 5.89650333e-01
-6.44764364e-01 6.04299128e-01 6.25511646e-01 7.61548758e-01
6.86376810e-01 -3.09238404e-01 5.79869092e-01 -1.01092553e+00
-4.66360040e-02 -2.81363040e-01 -6.54625297e-01 2.70184934e-01
-4.89661455e-01 6.39443174e-02 1.42092854e-01 -1.61157414e-01
-1.29444063e+00 8.36547688e-02 1.12006282e-02 -5.11155009e-01
-7.63687864e-02 -1.44409806e-01 -8.29334259e-02 -3.16704512e-01
1.53929800e-01 1.43758118e-01 -2.37260461e-01 -5.00255525e-01
2.40742490e-01 1.43070951e-01 5.93405008e-01 -6.72314242e-02
8.77089918e-01 8.36478472e-01 -2.53119349e-01 -5.03532290e-01
-1.38474071e+00 -6.47886634e-01 -6.00467682e-01 -9.25333127e-02
5.05292475e-01 -1.19260442e+00 -3.75836104e-01 2.01441288e-01
-1.04291403e+00 -6.16871893e-01 -3.51088375e-01 -1.46161228e-01
-3.62823844e-01 6.20153904e-01 -6.80298150e-01 -7.78877795e-01
-5.22008955e-01 -9.50121343e-01 1.55285847e+00 5.40651798e-01
-1.63823232e-01 -8.55315506e-01 -3.67742002e-01 3.20042372e-01
5.25952399e-01 1.37261197e-01 3.86465669e-01 -1.53943405e-01
-9.46252167e-01 2.68216699e-01 -6.76391482e-01 1.31068006e-01
1.08297721e-01 -2.00321093e-01 -1.04683864e+00 -2.40088657e-01
-2.24627666e-02 -6.79007545e-02 1.50565720e+00 5.12515604e-01
1.24610555e+00 -2.88270801e-01 -5.88403165e-01 6.78351402e-01
1.11223722e+00 -2.06110954e-01 6.57042265e-01 4.42609340e-01
9.83554661e-01 4.95962590e-01 1.07838154e+00 3.13622594e-01
4.80170548e-01 6.38764739e-01 5.02912521e-01 -9.25606906e-01
-3.09582621e-01 -2.48901442e-01 4.07859445e-01 1.38907522e-01
1.36388645e-01 4.79401089e-02 -6.36978507e-01 7.58815348e-01
-2.03430343e+00 -8.94685090e-01 -2.81236097e-02 1.88459623e+00
9.98049855e-01 3.90677541e-01 3.38738948e-01 -1.07121110e-01
9.77778077e-01 3.27375531e-01 -7.16436148e-01 -4.68090810e-02
-2.54747152e-01 1.76674992e-01 5.54751337e-01 8.16525936e-01
-1.48054230e+00 1.49504125e+00 6.53463221e+00 8.40500593e-01
-1.07830608e+00 8.11561048e-02 8.69174480e-01 -8.04036483e-02
-3.96170110e-01 1.09459758e-01 -1.11328673e+00 4.15476054e-01
2.33729661e-01 2.45991871e-01 1.56981088e-02 9.63739038e-01
3.77769887e-01 -3.00177902e-01 -8.20864320e-01 6.22453511e-01
5.77047728e-02 -1.47189140e+00 1.45926625e-01 -2.58862197e-01
9.08966899e-01 6.79855049e-02 2.00449988e-01 2.60618359e-01
2.65629679e-01 -7.95365036e-01 1.03737223e+00 4.38011497e-01
2.92880028e-01 -7.25632668e-01 5.98555386e-01 2.38029242e-01
-1.25837517e+00 -1.28954619e-01 -4.74323034e-01 -4.40405309e-02
2.50052989e-01 6.11397743e-01 -1.03730369e+00 3.07898909e-01
7.00158477e-01 1.07679701e+00 -1.07726848e+00 1.28614140e+00
-4.29136783e-01 5.98236799e-01 -1.73471123e-01 2.42863402e-01
2.74831384e-01 3.76004702e-03 7.02668726e-01 1.31139910e+00
-2.43324980e-01 -2.00495690e-01 2.75352508e-01 1.21994853e+00
-3.56416814e-02 -1.37628913e-01 -1.52715474e-01 6.61871433e-01
2.98337340e-01 1.38921261e+00 -9.93758798e-01 -3.57648045e-01
-3.37141693e-01 1.20771182e+00 5.50100625e-01 4.30411726e-01
-1.19401228e+00 -1.46667764e-01 5.67533910e-01 2.81161547e-01
8.10633540e-01 2.23183602e-01 -4.96871293e-01 -1.16613543e+00
4.74644266e-02 -4.38390225e-01 1.08815290e-01 -7.13642180e-01
-1.00901949e+00 3.36906910e-01 -1.84823081e-01 -1.07776141e+00
3.04815561e-01 -2.17531413e-01 -7.13673115e-01 6.32000387e-01
-2.17077994e+00 -1.31191707e+00 -4.02840972e-01 2.47339576e-01
8.57893944e-01 4.68733639e-01 1.81577697e-01 7.03534856e-02
-7.83359230e-01 6.63627505e-01 -4.08807158e-01 1.65100127e-01
7.39603460e-01 -1.41813457e+00 6.91442251e-01 1.16507661e+00
-9.15261060e-02 4.91779596e-01 6.51984215e-01 -6.73892975e-01
-5.53587377e-01 -1.58911502e+00 7.62469113e-01 -3.70564044e-01
4.32981908e-01 -5.04223466e-01 -1.13687813e+00 6.50741160e-01
1.70715809e-01 -2.42057294e-02 4.25657392e-01 1.23433219e-02
-2.07286060e-01 -1.14121905e-03 -9.46125031e-01 8.48379135e-01
1.09438038e+00 -2.07398385e-01 -5.04888773e-01 1.06596269e-01
1.11777079e+00 -4.43919897e-01 -4.78176504e-01 8.51633966e-01
2.58762032e-01 -1.02797544e+00 1.04285407e+00 -1.99060902e-01
3.94103229e-01 -5.29679298e-01 2.90305972e-01 -7.46252894e-01
-6.42075717e-01 -7.26002991e-01 -4.37022179e-01 1.31009436e+00
5.50505698e-01 -2.62371689e-01 9.96537149e-01 7.66472876e-01
-3.06290060e-01 -9.88349140e-01 -7.39053428e-01 -5.18776715e-01
-3.34077984e-01 -1.57364264e-01 3.77353549e-01 6.44295096e-01
-2.74314195e-01 4.16231781e-01 -5.39508045e-01 3.81096989e-01
6.73639238e-01 6.56672955e-01 6.26839101e-01 -1.28934836e+00
-2.08734766e-01 -4.83270943e-01 -8.29600617e-02 -1.49428463e+00
1.00327805e-01 -6.41641319e-01 3.87297720e-01 -1.56154668e+00
5.43827713e-01 -5.33358693e-01 -5.42658925e-01 7.82568514e-01
-8.24700356e-01 4.99706507e-01 3.14380288e-01 1.45423025e-01
-1.18850851e+00 7.49831319e-01 1.53149390e+00 -1.63933843e-01
-3.25244963e-01 1.25562951e-01 -9.89197433e-01 9.85115886e-01
7.01427758e-01 -3.47200483e-01 -2.89651066e-01 -2.90997773e-01
-1.16813675e-01 -3.08093071e-01 7.43237734e-01 -9.39728796e-01
1.22220330e-01 -1.07755959e-01 5.61628699e-01 -8.87541831e-01
2.76021302e-01 -5.96072733e-01 -4.50437069e-01 2.35959753e-01
-3.30175400e-01 -2.68525481e-01 3.75653297e-01 8.04721892e-01
-6.98399171e-02 -5.06189233e-03 9.69937980e-01 7.13069886e-02
-9.82452333e-01 3.21990013e-01 -1.14280440e-01 1.24208838e-01
9.60354984e-01 -3.57877105e-01 -1.67893633e-01 -2.81218439e-01
-6.67139709e-01 7.48368263e-01 8.67625892e-01 4.56215084e-01
7.79392719e-01 -1.17579091e+00 -5.39764822e-01 1.91486806e-01
-6.11161776e-02 4.25957918e-01 1.30503401e-01 1.12786603e+00
-3.21737349e-01 2.18543127e-01 9.96223465e-02 -7.96032548e-01
-1.33593512e+00 5.54079592e-01 1.34716362e-01 -2.33988523e-01
-8.30780089e-01 1.07551861e+00 9.21657741e-01 -5.47442921e-02
4.61278826e-01 -6.66763067e-01 -3.46781701e-01 -2.17971295e-01
5.35131991e-01 2.62689024e-01 -3.56504500e-01 -7.11248457e-01
-4.45050210e-01 4.26277220e-01 -5.15016973e-01 1.85671702e-01
1.22618628e+00 -3.94836992e-01 2.88737789e-02 2.40066931e-01
6.38416648e-01 5.42752328e-04 -1.89301276e+00 -2.18439564e-01
1.69653505e-01 -3.74577194e-01 2.03895599e-01 -6.20166659e-01
-1.09251177e+00 7.28184998e-01 3.42960477e-01 -1.47275254e-01
1.29782176e+00 7.24487752e-02 9.72030282e-01 2.19482034e-02
2.83228964e-01 -1.07736874e+00 1.22149453e-01 5.08162618e-01
8.74532878e-01 -1.54326129e+00 7.50379413e-02 -9.03767943e-01
-9.37406361e-01 5.82778692e-01 8.65750134e-01 -2.66326129e-01
5.11139393e-01 2.66859412e-01 -1.27495468e-01 -4.04209942e-02
-6.72980845e-01 -4.83698338e-01 5.76493144e-01 4.30544704e-01
1.49957746e-01 -2.92355984e-01 -1.79801032e-01 6.20949924e-01
1.34419441e-01 -2.00011835e-01 4.35260981e-01 8.97161245e-01
-8.25195372e-01 -8.26934278e-01 -3.39237005e-01 3.39771926e-01
-4.35856402e-01 -3.57153386e-01 -5.93154788e-01 7.52948582e-01
1.29234776e-01 8.99317920e-01 1.20833382e-01 -1.27718821e-01
8.20288509e-02 -3.50937665e-01 1.36226490e-01 -7.74325788e-01
-5.38874865e-01 2.73807168e-01 -1.90532818e-01 -8.45813632e-01
-4.79536146e-01 -5.98316252e-01 -1.29992974e+00 2.19173729e-01
-7.90632904e-01 1.89589545e-01 -7.15645403e-02 7.82288492e-01
5.90662897e-01 6.76729679e-01 5.17448902e-01 -1.11144376e+00
-6.13750629e-02 -7.75244534e-01 -4.28541183e-01 3.11215013e-01
7.67201543e-01 -8.88236463e-01 -2.92728513e-01 -3.33805457e-02] | [9.765791893005371, -0.2849918007850647] |
3c86efb3-f955-4148-ba24-11e1cbe49f63 | what-would-harry-say-building-dialogue-agents | 2211.06869 | null | https://arxiv.org/abs/2211.06869v3 | https://arxiv.org/pdf/2211.06869v3.pdf | What would Harry say? Building Dialogue Agents for Characters in a Story | We have a Christmas gift for Harry Potter fans all over the world. In this paper, we present Harry Potter Dialogue (HPD), a dataset that helps train Harry Potter-like dialogue agents. Such a task is typically viewed as a variant of personalized dialogue agents, but they differ significantly in three respects: 1) Harry lived in a virtual world of wizards, thus, real-world commonsense may not apply to Harry's conversations; 2) Harry's behavior is strongly linked to background information in conversations: the scene, its attributes and its relationship to other speakers; and 3) Such backgrounds are dynamically altered as the storyline goes on. The HPD dataset, as the first dataset to facilitate the study of dialogue agent construction for characters within a story, provides rich contextual information about each dialogue session such as scenes, character attributes, and relations. More importantly, all the background information will change over the course of the story. In addition, HPD could support both dialogue generation and retrieval tasks. We evaluate baselines such as Dialog-GPT and BOB to determine the extent to which they can generate Harry Potter-like responses. The experimental results disappoint us in that although the generated responses are fluent, they still seem out of character for Harry. Besides, we validate the current most robust dialogue agent, ChatGPT, which also can't generate plausible Harry-Potter-like responses in some cases, either. Our results suggest that there is much scope for future research. | ['Longyue Wang', 'Jia Li', 'Ziyang Chen', 'Deng Cai', 'Haiyun Jiang', 'Yan Wang', 'Nuo Chen'] | 2022-11-13 | null | null | null | null | ['pesona-dialogue-in-story'] | ['natural-language-processing'] | [-2.64521211e-01 4.15536761e-01 2.82676965e-01 -4.71813440e-01
-7.59811282e-01 -8.36152852e-01 1.06856561e+00 -3.42620797e-02
-1.14386939e-01 1.02989292e+00 8.80991220e-01 -4.16081445e-03
3.60024810e-01 -8.01948249e-01 -1.99304163e-01 -3.67819697e-01
2.01983348e-01 1.04617643e+00 2.69143224e-01 -1.08640587e+00
9.43617672e-02 2.15656430e-01 -1.00915396e+00 6.56159043e-01
3.70763153e-01 1.15181707e-01 2.25081205e-01 9.98750567e-01
-7.65951425e-02 1.40415859e+00 -1.20455813e+00 -6.52534842e-01
-8.83855298e-02 -8.22073281e-01 -1.40575254e+00 7.67815635e-02
1.11597247e-01 -7.35053897e-01 -5.15809417e-01 4.75221246e-01
5.11152685e-01 5.61839521e-01 6.35139227e-01 -1.22825444e+00
-4.48584050e-01 1.08104277e+00 -1.01367466e-01 -6.55032024e-02
8.68523896e-01 6.04851961e-01 1.12829137e+00 -6.44191921e-01
1.02446175e+00 1.76566112e+00 5.32079577e-01 9.24942374e-01
-1.04945004e+00 -3.97785306e-01 5.36222719e-02 -1.23211965e-01
-9.64426398e-01 -5.70006490e-01 6.28853381e-01 -3.38711828e-01
9.25338149e-01 6.47733927e-01 6.56209409e-01 1.72573864e+00
-2.50218391e-01 8.61442983e-01 9.36790705e-01 -1.99418113e-01
1.59293711e-02 4.24263507e-01 3.82563025e-02 4.30362493e-01
-4.94127214e-01 -4.14877445e-01 -6.17126107e-01 -4.36246932e-01
6.70306444e-01 -6.76462591e-01 -2.16688633e-01 1.46849081e-01
-1.26088893e+00 1.10773325e+00 2.29991585e-01 3.34005237e-01
-6.98938668e-02 -1.17553391e-01 5.47338068e-01 2.72204041e-01
2.48182237e-01 1.05222547e+00 -1.90249100e-01 -6.65911257e-01
-2.12490469e-01 8.68445873e-01 1.49784088e+00 1.07322299e+00
4.44650143e-01 -1.85545862e-01 -5.20113766e-01 1.27212334e+00
-1.66860614e-02 5.61433360e-02 2.79873699e-01 -1.21861434e+00
6.76554501e-01 3.15446556e-01 3.92477095e-01 -9.63090479e-01
-4.21519816e-01 3.38951886e-01 -5.04027009e-01 -2.82929242e-01
7.15481102e-01 -5.71230710e-01 -2.04200476e-01 1.87223864e+00
2.77746618e-01 -2.11167425e-01 3.78351569e-01 9.16927338e-01
1.40032411e+00 1.02314699e+00 -2.75908168e-02 -9.28561669e-03
1.44541550e+00 -9.93259251e-01 -7.29775608e-01 -3.32885563e-01
6.38117671e-01 -9.05890524e-01 1.41750538e+00 1.97042506e-02
-1.17017055e+00 -3.41242224e-01 -7.02380180e-01 -1.94933787e-01
-1.81788489e-01 -3.13577324e-01 7.14005053e-01 4.65436429e-01
-9.29046154e-01 4.48950142e-01 -3.34484190e-01 -7.25886822e-01
-2.30343536e-01 -1.00722328e-01 -2.97881782e-01 7.68535435e-02
-1.53356111e+00 1.32374680e+00 2.28789851e-01 -9.79393199e-02
-8.40678871e-01 -3.38338226e-01 -8.78902435e-01 -4.60803397e-02
3.11589837e-01 -5.41203380e-01 1.91142905e+00 -6.03493214e-01
-1.84806263e+00 9.36718166e-01 5.14664203e-02 -3.04311931e-01
8.23145568e-01 -8.63559768e-02 -1.26982570e-01 1.77184537e-01
1.63221464e-01 7.62724102e-01 1.49318159e-01 -1.46728981e+00
-4.52251107e-01 -6.75316453e-02 5.72636425e-01 6.82588458e-01
9.45160463e-02 2.13581875e-01 -3.81161213e-01 -4.82210815e-01
-4.60942119e-01 -9.69285905e-01 -9.60228071e-02 -5.36376476e-01
-8.33151817e-01 -4.09302711e-01 7.26158917e-01 -4.56176221e-01
9.15842891e-01 -1.99107444e+00 -7.40565881e-02 -2.86324501e-01
1.22423977e-01 -1.11100143e-02 -2.27436319e-01 1.11309552e+00
3.22621346e-01 2.47865483e-01 1.25744548e-02 -3.74242574e-01
1.88057289e-01 2.55843312e-01 -4.43296403e-01 -7.13746203e-03
1.07778341e-01 7.75500357e-01 -1.15566576e+00 -4.42456961e-01
4.60902825e-02 1.61389604e-01 -4.36043769e-01 5.88228703e-01
-6.06497526e-01 5.48930526e-01 -4.24414188e-01 1.63810924e-01
1.62567705e-01 -1.41681820e-01 2.51059145e-01 7.53176883e-02
-1.57696798e-01 6.66853189e-01 -5.66065669e-01 1.39948463e+00
-4.73566920e-01 9.59893823e-01 7.62563720e-02 -2.71247804e-01
1.00789821e+00 5.36595523e-01 9.05496925e-02 -3.08095485e-01
8.59980732e-02 -1.25065923e-01 2.14341000e-01 -7.02361882e-01
1.14819872e+00 -1.70389295e-01 -5.29473782e-01 6.76105559e-01
-2.71181047e-01 -7.98361421e-01 3.32219064e-01 6.95140719e-01
1.26389885e+00 -2.14446634e-01 1.57421142e-01 7.97770694e-02
1.22744650e-01 3.12511504e-01 3.90111804e-01 1.02903581e+00
-1.75522506e-01 6.70539737e-01 7.93366492e-01 -1.93210870e-01
-1.03583467e+00 -1.01991594e+00 7.26513192e-02 1.33514977e+00
3.35267454e-01 -4.32839662e-01 -6.90100789e-01 -3.54989648e-01
-4.60395128e-01 1.23093748e+00 -4.38586414e-01 7.25096539e-02
-7.11924016e-01 -5.83668649e-01 8.89366627e-01 2.26865709e-01
7.04653740e-01 -1.40624774e+00 -5.78331888e-01 3.62523526e-01
-8.17419767e-01 -1.34647334e+00 -4.80333358e-01 -2.90587008e-01
-2.37345800e-01 -9.20995891e-01 -4.96364653e-01 -6.11855030e-01
1.28720254e-01 1.58293828e-01 1.40102875e+00 1.96001351e-01
-1.30766273e-01 5.46844065e-01 -7.04577267e-01 -1.65765688e-01
-1.14986706e+00 2.50328574e-02 -3.85744244e-01 -3.56117040e-01
9.06712040e-02 -4.08935934e-01 -3.24192882e-01 4.47229773e-01
-5.18257499e-01 4.68007594e-01 3.74507308e-02 9.12257671e-01
-4.21226799e-01 -1.50085136e-01 5.58673143e-01 -1.36297202e+00
1.22485065e+00 -5.96348882e-01 1.07188381e-01 1.64659858e-01
1.92967534e-01 -3.61969203e-01 6.82562530e-01 -3.87153983e-01
-1.53376865e+00 -3.59414220e-01 -1.11538373e-01 2.66182840e-01
-2.42927670e-01 1.12899445e-01 -1.22938126e-01 6.37406290e-01
8.94757628e-01 -4.76288348e-02 -1.03297114e-01 -2.31118396e-01
4.66363490e-01 7.53731787e-01 6.77284956e-01 -1.04365146e+00
6.52324080e-01 1.21804684e-01 -6.07023180e-01 -1.05715346e+00
-5.39677918e-01 -2.50739455e-01 -1.36147738e-01 -3.15337747e-01
8.54216218e-01 -8.11641991e-01 -9.75300372e-01 5.33160269e-01
-1.65791392e+00 -8.58784437e-01 -3.92114222e-02 1.42388016e-01
-5.45576394e-01 2.30758265e-01 -1.20369244e+00 -9.07420874e-01
-8.97799060e-02 -1.10720050e+00 6.37705803e-01 3.56397152e-01
-1.03074419e+00 -9.49124992e-01 9.44146961e-02 6.27526045e-01
2.67694831e-01 2.68365532e-01 1.15545297e+00 -9.90924418e-01
-3.44791740e-01 6.75265789e-02 -2.39013918e-02 -9.88264978e-02
1.60709023e-01 1.97310373e-01 -8.80296350e-01 8.43788311e-02
-8.73455629e-02 -8.03098202e-01 2.08342984e-01 -1.82768524e-01
4.30318862e-01 -7.53414631e-01 -3.05598248e-02 -7.63629004e-02
6.51985049e-01 4.53720808e-01 6.83171093e-01 2.56426603e-01
4.04480934e-01 1.03674924e+00 7.70378351e-01 7.61478066e-01
8.25068951e-01 8.40174973e-01 1.43050745e-01 -1.52598798e-01
-5.97394891e-02 -5.67497671e-01 4.40929562e-01 6.24438941e-01
8.78595859e-02 -5.61403811e-01 -8.01744342e-01 5.26525676e-01
-1.94345331e+00 -1.16189885e+00 -2.17749238e-01 1.63682544e+00
1.31164396e+00 -8.99097025e-02 4.43935782e-01 -4.68306601e-01
8.39174747e-01 4.99594986e-01 -2.58642644e-01 -7.24965513e-01
-2.78624147e-01 -2.07107589e-01 -2.11151481e-01 7.41643786e-01
-6.95518792e-01 1.24927366e+00 6.09380913e+00 3.39035630e-01
-6.97025120e-01 -1.62749261e-01 8.56275976e-01 1.55916825e-01
-3.72355074e-01 1.71184093e-01 -6.23915792e-01 3.65087479e-01
4.55294162e-01 -2.82106727e-01 5.88674545e-01 6.77157223e-01
2.88921505e-01 -2.50810683e-01 -1.47549367e+00 5.84298790e-01
5.69497384e-02 -1.28948486e+00 -5.41728735e-02 -3.88887525e-01
3.70318174e-01 -3.66309285e-01 -1.56619132e-01 6.18827999e-01
1.13206303e+00 -1.11267245e+00 5.24254322e-01 4.14457470e-02
3.04519325e-01 -5.36645055e-01 4.70649898e-01 5.63252032e-01
-5.20795941e-01 2.74692178e-01 -1.00203909e-01 -2.22257853e-01
3.04934919e-01 -1.08262010e-01 -1.73868966e+00 -2.80002262e-02
4.44087148e-01 1.14359677e-01 -4.01485443e-01 4.52181906e-01
-3.27403903e-01 4.81928736e-01 -1.36103645e-01 -4.97739702e-01
3.57558668e-01 -6.55378923e-02 7.44072437e-01 1.47650254e+00
-1.58556756e-02 6.71231925e-01 2.59527445e-01 9.50106621e-01
-1.16987579e-01 1.73761591e-01 -6.78517163e-01 -8.52076784e-02
1.01843631e+00 1.18668711e+00 -4.97048825e-01 -3.29715759e-01
-1.51885375e-01 9.21291828e-01 3.94453816e-02 2.98738778e-01
-7.01683521e-01 -9.71355960e-02 7.04508901e-01 7.17954561e-02
-3.62583637e-01 1.26215830e-01 -8.32960159e-02 -1.02061605e+00
-3.58289689e-01 -1.29927111e+00 2.45275766e-01 -1.20632958e+00
-1.49868631e+00 7.61484802e-01 1.61481559e-01 -7.26034880e-01
-7.86277235e-01 -1.27998501e-01 -8.23834658e-01 6.53093457e-01
-8.02174330e-01 -1.09575617e+00 -3.56299162e-01 5.41410089e-01
1.10262597e+00 -1.88114390e-01 1.19112766e+00 -2.95724273e-01
-5.24162650e-01 4.40018028e-01 -4.12197411e-01 5.94258845e-01
9.98977780e-01 -1.23518884e+00 6.25378549e-01 3.12452674e-01
2.20037960e-02 5.54770410e-01 1.19667768e+00 -5.33405900e-01
-1.06333506e+00 -4.96567816e-01 7.39998579e-01 -5.81817985e-01
7.61011362e-01 -3.99966151e-01 -9.61913228e-01 7.51886427e-01
7.06060648e-01 -7.89439797e-01 6.26186073e-01 3.70052218e-01
-4.05730814e-01 3.38344336e-01 -1.07694817e+00 1.13965929e+00
7.79421628e-01 -4.47007418e-01 -7.56816566e-01 7.13422954e-01
7.76466548e-01 -7.84006536e-01 -6.54813707e-01 -6.71943324e-03
3.22286546e-01 -1.23511887e+00 7.79448688e-01 -7.52672076e-01
9.26626205e-01 2.08254546e-01 -1.50017828e-01 -1.58515060e+00
-2.75485739e-02 -1.00072503e+00 4.14083570e-01 1.71549094e+00
4.39680964e-01 -3.86833429e-01 5.82752049e-01 1.24794936e+00
-1.07009694e-01 -4.06140566e-01 -5.31206012e-01 -3.77553910e-01
1.85303301e-01 -1.59994721e-01 7.39788413e-01 1.21544731e+00
4.74305660e-01 1.11690414e+00 -5.60585737e-01 -2.66706228e-01
1.63392313e-02 -4.70466539e-02 1.29041398e+00 -8.20790231e-01
-6.13181114e-01 -3.99579346e-01 1.71670273e-01 -1.14773631e+00
2.68764615e-01 -4.70823288e-01 4.22400922e-01 -1.35779405e+00
2.86489934e-01 -6.71985388e-01 7.10414708e-01 3.92594367e-01
-2.53034413e-01 -2.27490768e-01 5.98770916e-01 1.79909840e-01
-3.05717587e-01 4.21894491e-01 1.43678296e+00 -1.60064310e-01
-4.82649893e-01 1.31043032e-01 -7.03229427e-01 7.72842288e-01
9.26731944e-01 -4.24558781e-02 -6.43242776e-01 -3.35625142e-01
-2.08454914e-02 7.78204262e-01 3.81963223e-01 -4.03861642e-01
1.60082445e-01 -4.82706189e-01 2.76281703e-02 -3.62809092e-01
9.53064561e-01 -1.49623752e-01 5.58558963e-02 -9.95364599e-03
-7.70182014e-01 3.73912714e-02 4.83255014e-02 1.53963536e-01
-1.05680361e-01 -3.98981631e-01 6.76670253e-01 -6.08623683e-01
-6.28153980e-01 -2.39790678e-01 -8.85068357e-01 5.42335093e-01
7.78638542e-01 -1.20876849e-01 -8.38981092e-01 -1.22572351e+00
-4.58119005e-01 3.84540498e-01 5.64224839e-01 5.21422207e-01
4.75854218e-01 -9.45384383e-01 -9.99161482e-01 -6.04208291e-01
7.46127293e-02 8.82198215e-02 3.40966344e-01 1.57205105e-01
-5.30708790e-01 3.71384583e-02 -1.10295646e-01 -2.98712343e-01
-1.34947622e+00 -1.09691378e-02 2.82669097e-01 -3.45556110e-01
-8.09113383e-01 1.00565958e+00 4.55446035e-01 -4.79490370e-01
2.22205147e-01 2.72157997e-01 -3.48414361e-01 1.53111711e-01
4.60014284e-01 3.82912755e-02 -4.56651777e-01 -6.46429121e-01
-7.18942378e-03 -2.31592104e-01 -3.20180148e-01 -6.54399276e-01
1.24323893e+00 -1.93655536e-01 -1.79388359e-01 8.28782499e-01
8.54050279e-01 6.81818947e-02 -1.08536303e+00 -1.46109805e-01
-2.26227522e-01 -5.41879356e-01 -7.44187117e-01 -1.08198369e+00
-3.18546176e-01 5.80822885e-01 -2.97343016e-01 5.03436029e-01
5.55740297e-01 2.84538925e-01 9.45638597e-01 6.20002270e-01
2.89997220e-01 -1.12795472e+00 5.28241038e-01 8.63539100e-01
1.29972899e+00 -1.09973288e+00 -1.49798930e-01 -4.89953339e-01
-1.49171448e+00 9.77513194e-01 1.11134386e+00 3.32458168e-01
-1.82733655e-01 2.66236097e-01 4.12200361e-01 -2.63001353e-01
-1.31181884e+00 1.64738551e-01 -4.63054299e-01 7.32112706e-01
6.33684039e-01 2.58941948e-01 -2.10420806e-02 5.69731116e-01
-9.16841030e-01 -7.24339545e-01 1.09370768e+00 5.62741935e-01
-4.02796477e-01 -1.02538335e+00 -4.26924080e-01 7.17568025e-02
-2.35642064e-02 -7.52046481e-02 -1.23144698e+00 1.12403584e+00
-3.76583517e-01 1.39196146e+00 -7.83004537e-02 -3.63381356e-01
4.54570740e-01 5.67067340e-02 4.13702458e-01 -1.05144143e+00
-1.02774632e+00 -2.03848153e-01 9.61486995e-01 -1.11389339e-01
-5.05252033e-02 -5.31005740e-01 -1.30616903e+00 -7.92971671e-01
-1.31105408e-01 3.56583267e-01 1.58599362e-01 8.76446486e-01
-7.95668736e-02 1.72119051e-01 7.49791861e-01 -6.49459422e-01
-5.25735140e-01 -1.19112527e+00 -3.97687793e-01 7.90881693e-01
5.93340509e-02 -2.32973516e-01 -1.58202842e-01 -6.11262210e-02] | [12.725726127624512, 8.042787551879883] |
b1200ca2-6b06-4a1a-9731-6d10c5006776 | parallel-vertex-diffusion-for-unified-visual | 2303.07216 | null | https://arxiv.org/abs/2303.07216v2 | https://arxiv.org/pdf/2303.07216v2.pdf | Parallel Vertex Diffusion for Unified Visual Grounding | Unified visual grounding pursues a simple and generic technical route to leverage multi-task data with less task-specific design. The most advanced methods typically present boxes and masks as vertex sequences to model referring detection and segmentation as an autoregressive sequential vertex generation paradigm. However, generating high-dimensional vertex sequences sequentially is error-prone because the upstream of the sequence remains static and cannot be refined based on downstream vertex information, even if there is a significant location gap. Besides, with limited vertexes, the inferior fitting of objects with complex contours restricts the performance upper bound. To deal with this dilemma, we propose a parallel vertex generation paradigm for superior high-dimension scalability with a diffusion model by simply modifying the noise dimension. An intuitive materialization of our paradigm is Parallel Vertex Diffusion (PVD) to directly set vertex coordinates as the generation target and use a diffusion model to train and infer. We claim that it has two flaws: (1) unnormalized coordinate caused a high variance of loss value; (2) the original training objective of PVD only considers point consistency but ignores geometry consistency. To solve the first flaw, Center Anchor Mechanism (CAM) is designed to convert coordinates as normalized offset values to stabilize the training loss value. For the second flaw, Angle summation loss (ASL) is designed to constrain the geometry difference of prediction and ground truth vertexes for geometry-level consistency. Empirical results show that our PVD achieves state-of-the-art in both referring detection and segmentation, and our paradigm is more scalable and efficient than sequential vertex generation with high-dimension data. | ['Jie Chen', 'Chang Liu', 'Li Yuan', 'Xiangyang Ji', 'Peng Jin', 'Kehan Li', 'Zesen Cheng'] | 2023-03-13 | null | null | null | null | ['visual-grounding'] | ['computer-vision'] | [ 5.68474904e-02 3.01004171e-01 -6.62651658e-02 -6.83420822e-02
-9.02487814e-01 -5.61710060e-01 5.00233173e-01 -8.82569999e-02
-1.05562054e-01 3.37550849e-01 -7.35341534e-02 -2.91212440e-01
4.22536619e-02 -8.12718570e-01 -8.24004769e-01 -6.22223735e-01
2.08582297e-01 5.56378543e-01 6.47325814e-01 -2.39546955e-01
2.15510517e-01 5.70459425e-01 -1.24867594e+00 1.49520710e-01
1.11261499e+00 1.02091289e+00 2.40216717e-01 4.93514061e-01
-2.76150554e-01 7.15015605e-02 -5.09355366e-01 -6.44086242e-01
6.61136448e-01 -3.11382800e-01 -5.15661180e-01 1.32560298e-01
6.17750108e-01 -2.98621595e-01 -8.80607143e-02 1.08933258e+00
6.00014091e-01 -5.43037616e-02 8.02907526e-01 -1.35989773e+00
-8.63128304e-01 3.96922231e-01 -1.10697281e+00 -1.32211238e-01
3.35626230e-02 3.64361942e-01 9.72196221e-01 -9.89383876e-01
7.46105492e-01 1.17872071e+00 1.01411307e+00 6.00933373e-01
-1.37116945e+00 -4.38593954e-01 6.41180217e-01 -2.83680737e-01
-1.65106964e+00 -8.30788091e-02 9.26318049e-01 -5.43786287e-01
6.03347361e-01 2.44682118e-01 7.70380378e-01 1.19637072e+00
1.46970898e-01 7.67272353e-01 6.76628351e-01 -2.25907266e-01
1.39901847e-01 -1.91871859e-02 -1.58583045e-01 9.62597370e-01
3.15646380e-01 8.24755877e-02 -1.88485116e-01 -1.05471879e-01
1.30414891e+00 -1.75412953e-01 -2.54980475e-01 -7.75214851e-01
-1.27680683e+00 8.48250508e-01 5.76429248e-01 1.34304300e-01
-3.51774216e-01 1.49516046e-01 2.17552543e-01 1.97509676e-02
6.26571596e-01 4.07336324e-01 -2.81032145e-01 1.21830031e-01
-1.08267283e+00 3.94337058e-01 4.47584063e-01 1.24077153e+00
6.91197395e-01 1.47493377e-01 -4.14219737e-01 7.29939342e-01
4.14655894e-01 4.31144863e-01 2.18422934e-01 -9.52369452e-01
6.39919579e-01 5.85160792e-01 -1.87847093e-01 -1.27416837e+00
-4.97325718e-01 -5.43020904e-01 -8.43989968e-01 4.42244828e-01
7.44602203e-01 -1.75112620e-01 -1.27916396e+00 1.70842445e+00
5.04227042e-01 2.77272522e-01 -4.48757529e-01 1.18035579e+00
7.88782418e-01 3.87196481e-01 6.37804270e-02 3.18640172e-02
1.28305054e+00 -9.67272937e-01 -5.61456025e-01 -3.65388185e-01
5.65086305e-01 -6.57754481e-01 1.33964157e+00 1.28997773e-01
-1.13047040e+00 -4.46903169e-01 -1.10082054e+00 -2.92238325e-01
-1.69036299e-01 8.54082182e-02 8.00647140e-01 6.62161112e-01
-1.04307628e+00 4.97713357e-01 -9.04169321e-01 -3.00259353e-03
5.43498755e-01 1.40744507e-01 -1.04240321e-01 3.16043317e-01
-9.35887516e-01 5.02492607e-01 1.99171118e-02 1.94146037e-01
-5.02972722e-01 -1.07563674e+00 -8.47741544e-01 3.77038978e-02
5.71302533e-01 -1.16725039e+00 8.93702149e-01 -7.87817180e-01
-1.61619198e+00 8.58993232e-01 4.82364148e-02 -1.68981403e-01
1.15882671e+00 -3.48340012e-02 1.03038289e-01 7.85935111e-03
3.16992730e-01 7.10306585e-01 1.11662579e+00 -1.56266212e+00
-4.90471214e-01 -4.46231425e-01 -2.73318738e-01 2.44166777e-01
-9.02340561e-02 -3.74790698e-01 -9.93175745e-01 -9.51447189e-01
7.25108683e-01 -8.75972092e-01 -4.73541826e-01 3.85125905e-01
-5.58242381e-01 -2.27939673e-02 6.35918915e-01 -4.49598312e-01
1.31818914e+00 -2.12503624e+00 1.27414569e-01 4.94591385e-01
5.91694057e-01 6.98985383e-02 -1.12657748e-01 1.38317961e-02
7.38243163e-02 3.68604302e-01 -1.77366108e-01 -6.48525953e-01
-1.55800432e-01 4.10891548e-02 -3.70824933e-01 3.84979188e-01
3.07493448e-01 1.13643277e+00 -9.77772593e-01 -6.82648957e-01
2.27281138e-01 5.05508542e-01 -6.93638086e-01 -1.59318462e-01
-3.20568800e-01 2.76987195e-01 -6.89749599e-01 8.46624136e-01
8.56275499e-01 -5.25638521e-01 -7.22361058e-02 -5.59792817e-01
-1.35112479e-02 -3.51703584e-01 -1.56681132e+00 1.83607757e+00
-7.26684108e-02 2.26488337e-01 1.44207329e-01 -6.10512614e-01
8.63904893e-01 -8.81615654e-03 5.41041493e-01 -4.52026457e-01
7.71555007e-02 1.94274411e-01 -2.33717069e-01 -2.75750756e-01
5.59636354e-01 5.21315113e-02 1.25774711e-01 -3.71355824e-02
-2.91937619e-01 -3.70675504e-01 -1.74287498e-01 2.89106578e-01
9.04943407e-01 4.37258989e-01 -2.51648258e-02 -7.25155324e-02
1.48101980e-02 3.33227552e-02 7.81578600e-01 7.84574747e-01
-2.50660807e-01 1.25268042e+00 7.52250254e-01 -2.05372065e-01
-1.21257520e+00 -1.24353695e+00 -2.02946797e-01 6.89700007e-01
4.07605201e-01 -4.09562945e-01 -7.87603259e-01 -7.19010890e-01
1.65663823e-01 4.84272987e-01 -6.80092871e-01 4.17546928e-02
-5.82637250e-01 -6.84914947e-01 4.88551259e-01 7.94054985e-01
2.35623568e-01 -6.54786110e-01 -3.57423484e-01 1.48721814e-01
3.30764763e-02 -1.09260535e+00 -7.22517669e-01 -1.28790572e-01
-9.00978625e-01 -9.51372623e-01 -9.38507438e-01 -6.32642567e-01
8.87000322e-01 1.31171808e-01 9.69046533e-01 1.32382408e-01
-1.33792236e-01 6.33931383e-02 -1.84891120e-01 -1.91967189e-01
-2.38400981e-01 6.26406074e-02 -2.09597692e-01 -5.34313507e-02
-2.30743378e-01 -4.53496188e-01 -8.00523937e-01 3.86425734e-01
-6.73938990e-01 3.32068831e-01 5.23793995e-01 7.56742835e-01
9.10173118e-01 -2.88334966e-01 3.88887078e-01 -8.32071841e-01
5.86168170e-01 -2.42139637e-01 -6.82268322e-01 2.65496552e-01
-7.78602362e-01 2.34294832e-02 3.48187357e-01 -6.28888309e-01
-7.65316725e-01 1.73651055e-01 -1.51815876e-01 -9.88536358e-01
1.79442361e-01 1.97200701e-01 -1.46349743e-01 -1.88028827e-01
7.48839200e-01 -1.06006078e-01 1.27643168e-01 -4.10787821e-01
7.07331300e-01 1.57569900e-01 5.02050817e-01 -7.47829080e-01
7.60542691e-01 5.84771574e-01 1.41582862e-01 -5.21281064e-01
-4.46913332e-01 -1.12710409e-01 -6.26104593e-01 -2.72608072e-01
8.99729669e-01 -6.91025734e-01 -6.30558252e-01 3.84315848e-01
-1.21301854e+00 -4.74326700e-01 -2.52909452e-01 2.40102157e-01
-4.78739589e-01 4.87165123e-01 -6.73979938e-01 -7.26926446e-01
-3.19550425e-01 -1.29220605e+00 1.47951269e+00 -6.75600693e-02
-6.93212077e-02 -8.82335007e-01 -2.67030537e-01 6.96236119e-02
2.34282956e-01 5.87079704e-01 8.24496269e-01 -3.10447603e-01
-7.72580564e-01 -1.95843369e-01 -4.91606057e-01 1.19791813e-01
2.35392433e-02 3.07313472e-01 -7.91038632e-01 -1.42970830e-01
-1.44008741e-01 9.98334214e-02 6.53453648e-01 5.90885818e-01
1.32317924e+00 -1.12928106e-02 -6.46520078e-01 9.63306904e-01
1.32187581e+00 -5.43859079e-02 5.08189678e-01 2.00101018e-01
1.27467489e+00 5.25281549e-01 3.79476607e-01 2.75277644e-01
4.60621804e-01 8.04616392e-01 3.53526890e-01 -3.17741364e-01
-4.95037913e-01 -5.21637917e-01 -1.14354484e-01 3.38761747e-01
-2.11401448e-01 -2.65508711e-01 -1.04095590e+00 4.17587459e-01
-1.89279592e+00 -6.89427912e-01 -3.98232490e-01 2.23760581e+00
6.11032307e-01 2.96114266e-01 1.66546300e-01 -1.24361962e-01
7.24438667e-01 7.81017244e-02 -4.84252512e-01 7.72613585e-02
-1.07005715e-01 -2.37323031e-01 6.59688652e-01 6.42136455e-01
-9.17634666e-01 1.11861646e+00 6.04874086e+00 1.07503796e+00
-1.15067697e+00 8.08076113e-02 7.10468233e-01 -8.01688582e-02
-4.79038179e-01 -1.47593826e-01 -8.50198388e-01 5.73822021e-01
1.82530463e-01 1.14554934e-01 1.15520678e-01 8.63657534e-01
1.98367640e-01 1.63571700e-01 -1.03447556e+00 1.02325356e+00
-5.73417470e-02 -1.52145588e+00 2.66520917e-01 1.29122391e-01
7.37642646e-01 -8.38436782e-02 2.23351896e-01 1.16253182e-01
4.81109172e-02 -7.26229429e-01 1.11951423e+00 5.02346635e-01
9.19643223e-01 -4.90296304e-01 2.66228706e-01 1.78398103e-01
-1.24549234e+00 2.40515545e-01 -1.38722360e-01 2.50060201e-01
5.21559358e-01 5.81940711e-01 -6.30434752e-01 6.15250766e-01
5.43873668e-01 3.51139814e-01 -6.05897486e-01 9.45132613e-01
-1.44327059e-01 2.03823566e-01 -6.08262181e-01 2.35065028e-01
3.19866151e-01 -3.81112993e-01 7.88203359e-01 1.06750357e+00
3.07315648e-01 -1.23536520e-01 4.09647882e-01 1.21070373e+00
1.27670407e-01 7.32404925e-03 -5.67234576e-01 4.11199868e-01
5.55899680e-01 1.18156219e+00 -1.15612721e+00 -2.59104460e-01
-3.91843528e-01 9.56034243e-01 3.88820589e-01 5.24405420e-01
-1.04644728e+00 -7.62228742e-02 4.75716680e-01 5.27638435e-01
3.26162726e-01 -3.13239336e-01 -1.06170630e+00 -1.03407955e+00
2.46329412e-01 -5.63045025e-01 1.92808330e-01 -6.57062531e-01
-1.51335347e+00 6.22330427e-01 -2.54590251e-02 -1.27990222e+00
-2.91527789e-02 -4.94523525e-01 -4.44067001e-01 9.35608923e-01
-1.21505547e+00 -1.57078695e+00 -3.66416723e-01 4.30093288e-01
5.53817451e-01 2.20286071e-01 3.01424712e-01 4.36255842e-01
-6.28637850e-01 9.84708726e-01 -3.36963356e-01 2.24405423e-01
6.30451679e-01 -1.27619791e+00 7.25386858e-01 8.88859749e-01
5.21041192e-02 6.42241657e-01 6.08955622e-01 -1.07160735e+00
-1.21611214e+00 -1.11541402e+00 5.05554378e-01 -7.88986862e-01
6.83320880e-01 -5.50782084e-01 -1.16330338e+00 6.82542801e-01
-4.70410556e-01 3.79555047e-01 2.94332695e-03 -4.39673737e-02
-1.91776201e-01 2.50276893e-01 -1.08056188e+00 9.98585463e-01
1.30875814e+00 -3.27896416e-01 -2.41240159e-01 2.57237464e-01
9.05764759e-01 -9.54979539e-01 -8.10554087e-01 6.31476164e-01
4.41911012e-01 -7.76169956e-01 1.06370652e+00 -3.35528374e-01
3.01513463e-01 -5.11555374e-01 1.10671051e-01 -1.07195556e+00
-4.72239852e-01 -7.86041260e-01 -2.05897033e-01 1.29990041e+00
6.31247342e-01 -5.34358919e-01 1.11811852e+00 8.14705789e-01
-4.92636114e-01 -1.21601403e+00 -8.78809512e-01 -8.34545076e-01
1.54008135e-01 -6.69934690e-01 6.68961883e-01 8.88574719e-01
-5.06459355e-01 3.23048353e-01 -1.00964211e-01 3.54552180e-01
5.83440721e-01 1.51007408e-02 7.53106177e-01 -1.15792179e+00
-2.59995401e-01 -7.78696001e-01 -2.34192654e-01 -1.41775870e+00
-1.44813001e-01 -8.20638895e-01 3.04337610e-02 -1.56831503e+00
-1.94526047e-01 -8.80759299e-01 1.36600494e-01 4.27972078e-01
-4.15873706e-01 1.43668070e-01 3.15089494e-01 2.53268808e-01
-1.92787573e-01 4.87077892e-01 1.74668217e+00 7.49108791e-02
-5.44643164e-01 9.25028920e-02 -6.21003628e-01 8.31505358e-01
3.17326456e-01 -2.49465853e-01 -5.91913164e-01 -5.93394756e-01
3.19434494e-01 2.59555560e-02 6.59950316e-01 -6.68713927e-01
3.25419635e-01 -5.89088053e-02 3.98404092e-01 -6.28311455e-01
2.66072661e-01 -6.73904955e-01 1.30106926e-01 5.79287708e-02
1.38658229e-02 1.55150488e-01 1.00717403e-01 7.35517204e-01
2.13444069e-01 4.10722904e-02 5.92427492e-01 6.43749759e-02
-6.11069083e-01 6.28291130e-01 1.69681624e-01 1.75487429e-01
1.05723202e+00 -7.21316516e-01 -1.90040305e-01 -5.22242971e-02
-6.60645545e-01 3.23444247e-01 7.06140101e-01 5.07116079e-01
7.18554974e-01 -1.27092183e+00 -6.11054659e-01 4.50032771e-01
3.74497995e-02 4.80928779e-01 2.62329906e-01 8.26391399e-01
-5.23234069e-01 -6.78134710e-02 2.65186846e-01 -8.53615880e-01
-9.07252073e-01 6.87827885e-01 4.36446756e-01 -7.05040023e-02
-1.18064308e+00 1.08835971e+00 4.74255830e-01 -1.56920001e-01
2.53833175e-01 -5.83946943e-01 1.14173092e-01 7.59623721e-02
1.48996219e-01 2.59297490e-01 1.72034293e-01 -3.77062649e-01
-2.55445510e-01 8.12373579e-01 -8.52932036e-03 -1.27209544e-01
9.25660133e-01 -1.93447500e-01 3.36084157e-01 1.82752237e-01
9.53910887e-01 1.57718465e-01 -1.71715355e+00 3.26207541e-02
-1.95776239e-01 -5.55432141e-01 2.04347998e-01 -5.42958260e-01
-1.20237482e+00 6.19538605e-01 4.84257519e-01 2.24137351e-01
6.15879416e-01 5.63864037e-02 8.99240732e-01 -2.07426220e-01
3.19812596e-01 -1.10642517e+00 5.04505858e-02 2.94715822e-01
9.42724586e-01 -1.16588831e+00 -3.07664811e-03 -9.63847458e-01
-7.85906732e-01 8.77546489e-01 8.75753641e-01 -3.98124568e-02
5.21512330e-01 3.48843396e-01 1.31771639e-01 -5.29316604e-01
-2.50014275e-01 -1.17157348e-01 5.43004394e-01 6.13584757e-01
1.91622123e-01 -1.05550468e-01 -2.82578319e-01 6.23867512e-01
-2.19073489e-01 -3.58416408e-01 3.80651802e-02 6.02337301e-01
-9.09659788e-02 -9.10367429e-01 -4.02781427e-01 4.24557656e-01
-2.35040516e-01 -9.90467221e-02 -7.44756982e-02 8.75139296e-01
1.62277430e-01 5.62138498e-01 1.93576887e-01 -2.49880731e-01
6.46846235e-01 -2.01293215e-01 4.27249700e-01 -7.01545119e-01
-4.54773694e-01 3.14431250e-01 -2.53287613e-01 -6.83489680e-01
1.70090377e-01 -5.84440827e-01 -1.39273143e+00 -1.74082205e-01
-4.36785340e-01 -2.59030044e-01 6.40815914e-01 7.01798558e-01
5.92227578e-01 7.93191731e-01 3.65356356e-01 -1.10628510e+00
-6.65206611e-01 -5.40152669e-01 -4.28467184e-01 5.35817325e-01
2.76473999e-01 -9.07763958e-01 -3.95818144e-01 -1.47536308e-01] | [8.11263656616211, -3.2543728351593018] |
5483abce-a9fa-4634-9e4a-1eff8746c7dd | conformal-prediction-with-temporal-quantile | 2205.09940 | null | https://arxiv.org/abs/2205.09940v2 | https://arxiv.org/pdf/2205.09940v2.pdf | Conformal Prediction with Temporal Quantile Adjustments | We develop Temporal Quantile Adjustment (TQA), a general method to construct efficient and valid prediction intervals (PIs) for regression on cross-sectional time series data. Such data is common in many domains, including econometrics and healthcare. A canonical example in healthcare is predicting patient outcomes using physiological time-series data, where a population of patients composes a cross-section. Reliable PI estimators in this setting must address two distinct notions of coverage: cross-sectional coverage across a cross-sectional slice, and longitudinal coverage along the temporal dimension for each time series. Recent works have explored adapting Conformal Prediction (CP) to obtain PIs in the time series context. However, none handles both notions of coverage simultaneously. CP methods typically query a pre-specified quantile from the distribution of nonconformity scores on a calibration set. TQA adjusts the quantile to query in CP at each time $t$, accounting for both cross-sectional and longitudinal coverage in a theoretically-grounded manner. The post-hoc nature of TQA facilitates its use as a general wrapper around any time series regression model. We validate TQA's performance through extensive experimentation: TQA generally obtains efficient PIs and improves longitudinal coverage while preserving cross-sectional coverage. | ['Jimeng Sun', 'Shubhendu Trivedi', 'Zhen Lin'] | 2022-05-20 | null | null | null | null | ['predicting-patient-outcomes', 'econometrics', 'prediction-intervals', 'time-series-regression'] | ['medical', 'miscellaneous', 'miscellaneous', 'time-series'] | [ 1.41713127e-01 9.63952485e-03 -5.43696642e-01 -5.87284565e-01
-1.29039800e+00 -5.84784150e-01 2.33747497e-01 5.95151663e-01
-1.07935801e-01 7.92562723e-01 4.39114332e-01 -5.00514507e-01
-7.05334961e-01 -8.54182899e-01 -8.41134906e-01 -6.39195561e-01
-6.37549639e-01 4.17423040e-01 6.35549054e-02 4.01929840e-02
6.67190030e-02 4.05272454e-01 -9.83030856e-01 1.37394384e-01
9.80907083e-01 9.36986804e-01 -5.87719560e-01 4.47006524e-01
3.31691414e-01 1.75640389e-01 -2.95662016e-01 -4.98614073e-01
2.66260058e-01 -4.29662079e-01 -4.72049206e-01 -5.01393199e-01
1.06667235e-01 5.51845618e-02 2.37498492e-01 6.59089744e-01
4.79664743e-01 -2.14978382e-02 9.72084701e-01 -1.42195797e+00
-9.84217897e-02 6.11408949e-01 -8.43822420e-01 2.10176244e-01
2.30442122e-01 -1.39168337e-01 8.90295982e-01 -4.90159839e-01
4.32455629e-01 8.92469466e-01 1.47685111e+00 1.75969169e-01
-1.69641280e+00 -6.72013700e-01 -9.42829326e-02 -5.91644049e-01
-1.25114560e+00 -2.03954861e-01 4.92687643e-01 -7.26131558e-01
4.87934351e-01 4.27238345e-01 7.57878900e-01 9.40065026e-01
1.05930293e+00 2.22019136e-01 1.11998379e+00 -1.65908739e-01
3.51496309e-01 -1.51684746e-01 2.12205887e-01 9.64249298e-02
1.68363318e-01 4.51926380e-01 -4.01790261e-01 -7.00198233e-01
9.27219748e-01 3.15808207e-01 -1.28843382e-01 -4.02585328e-01
-1.28702986e+00 9.39038217e-01 8.44962150e-02 2.83088684e-02
-5.50512910e-01 -6.75168931e-02 7.44604230e-01 2.37643003e-01
8.60614598e-01 3.87731254e-01 -8.60813498e-01 -6.37430372e-03
-1.10650361e+00 5.26383817e-01 5.96318722e-01 9.88828003e-01
1.34606645e-01 -4.12698656e-01 -6.35635376e-01 5.93792975e-01
-1.97663397e-01 6.06761694e-01 3.08560789e-01 -9.50686872e-01
4.12851363e-01 2.68045992e-01 2.13943198e-01 -7.59440839e-01
-8.40821385e-01 -6.31358922e-01 -1.09374785e+00 -1.97504640e-01
6.52499497e-01 -2.55830050e-01 -4.17452455e-01 2.25114989e+00
5.98336756e-01 8.10539797e-02 -1.12469867e-01 4.58598703e-01
-2.96221767e-02 3.46173078e-01 3.04600149e-01 -9.10254717e-01
1.38901436e+00 -1.60901457e-01 -4.04522508e-01 4.10358161e-01
7.08548665e-01 -4.25046027e-01 9.25596237e-01 3.56707752e-01
-1.23303521e+00 -2.80415803e-01 -6.67608142e-01 2.10857049e-01
-1.11177582e-02 -2.81078011e-01 2.83544123e-01 4.15949464e-01
-7.59510636e-01 8.01010013e-01 -8.66307259e-01 -1.51589572e-01
4.25060183e-01 1.38104767e-01 -2.12259337e-01 6.05330952e-02
-1.29261839e+00 6.61861062e-01 -1.40171692e-01 -4.42666024e-01
-3.62294942e-01 -1.66455805e+00 -6.85571730e-01 6.63917838e-03
9.10985097e-03 -8.33373666e-01 1.17709863e+00 -6.14743412e-01
-8.31661403e-01 5.90078175e-01 -1.83000132e-01 -6.47620857e-01
8.66194367e-01 1.05349272e-01 -4.56295788e-01 -1.58620551e-02
5.21675229e-01 1.89037994e-01 5.68918228e-01 -7.75578916e-01
-4.18482244e-01 -7.41647184e-01 -4.36116368e-01 -1.87993720e-01
9.80400816e-02 -2.25110814e-01 1.07043935e-02 -9.78328407e-01
1.92382142e-01 -7.34724164e-01 -4.13162440e-01 -1.09082460e-01
-3.43667001e-01 -2.40256891e-01 -7.13125840e-02 -6.46781385e-01
1.52982891e+00 -1.98718381e+00 -2.10834101e-01 4.64834839e-01
1.80166334e-01 -6.56329691e-01 2.09905267e-01 6.37068748e-01
-5.13870716e-01 6.05020970e-02 -6.97305858e-01 -1.79139167e-01
-1.12994231e-01 -1.26803204e-01 -5.57670832e-01 8.10045600e-01
1.28458187e-01 9.24184084e-01 -8.29220116e-01 -5.27659535e-01
-9.20477286e-02 3.25371921e-01 -7.18080103e-01 -1.89460758e-02
-5.10494187e-02 5.28947294e-01 -2.84230232e-01 4.68427062e-01
6.02754653e-01 -2.65794486e-01 7.85281658e-02 7.11775348e-02
-2.48117611e-01 6.65375069e-02 -8.36082518e-01 1.46536791e+00
-1.44319400e-01 2.03639343e-01 -4.84739542e-01 -1.19553518e+00
8.17369699e-01 5.42143106e-01 1.33273780e+00 -7.51187027e-01
-7.62906745e-02 1.41971141e-01 -3.00946057e-01 -3.86733949e-01
9.99608338e-02 -8.64257097e-01 -4.36366290e-01 4.35641408e-01
-3.88978630e-01 1.67711124e-01 -1.96319550e-01 -2.44540498e-01
1.12740827e+00 8.33087489e-02 7.60477901e-01 -7.69736111e-01
1.35072693e-01 1.92206457e-01 8.41789782e-01 6.22346163e-01
-2.56567448e-01 8.62875581e-01 9.25064087e-01 -3.98132503e-01
-1.30947518e+00 -1.39932787e+00 -9.99700904e-01 7.10343361e-01
-2.64023602e-01 -1.21077314e-01 -7.88267851e-01 -5.21250367e-01
3.39266181e-01 7.11823702e-01 -1.08051741e+00 -1.82630792e-01
-4.26320791e-01 -8.96801472e-01 3.52973521e-01 7.58600473e-01
-9.10466760e-02 -7.78326333e-01 -9.44282830e-01 3.97697031e-01
-1.56237572e-01 -6.49253488e-01 -9.05331492e-01 1.48182034e-01
-1.26989818e+00 -1.21092105e+00 -8.70186388e-01 -2.94877440e-01
4.88045454e-01 -9.56065655e-02 1.29564619e+00 -4.04020250e-01
-2.51926631e-02 4.81865436e-01 -7.58717731e-02 -7.14208245e-01
-1.88183457e-01 3.47041711e-03 1.79739982e-01 -7.72395208e-02
2.07013503e-01 -7.22701192e-01 -9.91237044e-01 4.41259265e-01
-7.34253347e-01 -5.73741235e-02 1.27284631e-01 8.60740960e-01
1.03852487e+00 -1.62997186e-01 1.10230196e+00 -1.08813798e+00
5.07250190e-01 -9.71596777e-01 -6.41069829e-01 2.81638920e-01
-9.12273467e-01 -3.95961851e-02 5.10892451e-01 -6.30276859e-01
-4.60419118e-01 -3.41836751e-01 7.83432424e-02 -4.05049562e-01
2.31684372e-01 8.55587304e-01 3.43660772e-01 6.29758418e-01
8.23147595e-01 -3.04088388e-02 2.13665113e-01 -8.87185335e-02
-1.22354925e-02 2.20422819e-01 5.77049792e-01 -7.41667628e-01
3.69430512e-01 4.45942014e-01 3.33268225e-01 -3.19962412e-01
-7.70832002e-01 -5.99273145e-01 -5.95829666e-01 -8.83312747e-02
7.83354938e-01 -7.17826247e-01 -9.72526610e-01 -1.71582758e-01
-6.53186977e-01 -4.89037126e-01 -8.23794544e-01 6.36020243e-01
-9.72352028e-01 -1.35560662e-01 -1.68957159e-01 -8.90592515e-01
-4.21857536e-01 -6.68203354e-01 1.26782143e+00 -1.37720391e-01
-4.95344758e-01 -1.24585593e+00 5.29280782e-01 -2.16032550e-01
2.25060180e-01 9.55920637e-01 1.12297857e+00 -5.43467402e-01
1.14153616e-01 -4.26412702e-01 5.35714813e-02 -1.01991914e-01
3.83335203e-02 -6.40759543e-02 -7.88622200e-01 -4.30190057e-01
1.07766829e-01 1.98508665e-01 5.14983535e-01 1.23121488e+00
1.33596349e+00 -4.32518989e-01 -5.42870700e-01 5.88288546e-01
1.27007270e+00 2.53884733e-01 4.86893862e-01 -2.51010042e-02
3.63394544e-02 8.26186538e-01 7.43669689e-01 7.26191282e-01
3.78916562e-01 6.04731917e-01 -2.62521878e-02 -1.95696622e-01
5.89971423e-01 -3.06721509e-01 7.71528333e-02 4.71013784e-01
7.80527741e-02 2.49836370e-01 -9.58131492e-01 6.72694981e-01
-1.72452366e+00 -1.06843996e+00 -1.35932550e-01 2.69933105e+00
1.07095301e+00 1.38544530e-01 6.35908306e-01 2.91710764e-01
4.78833735e-01 -3.23283076e-01 -8.50154400e-01 -2.39536345e-01
5.09382561e-02 2.14583173e-01 6.78764224e-01 3.51987541e-01
-9.00474250e-01 -2.36036219e-02 7.00728989e+00 5.74968934e-01
-9.47189689e-01 1.21165782e-01 1.13595533e+00 -5.30781075e-02
-5.28987050e-01 -1.33840233e-01 -4.18552041e-01 5.00283897e-01
1.42477250e+00 -4.92171943e-01 2.08530910e-02 5.65488875e-01
3.80708367e-01 -8.90496075e-02 -1.23846388e+00 5.57498217e-01
-4.66992140e-01 -1.28395343e+00 -4.64767039e-01 1.81127369e-01
7.07969785e-01 -1.27737746e-01 1.98501781e-01 3.09898108e-01
8.01450014e-02 -1.17589116e+00 6.50246978e-01 8.50976348e-01
1.34351885e+00 -9.00331438e-01 5.92170298e-01 3.56111884e-01
-1.06848156e+00 1.42204063e-02 -7.09041283e-02 2.87938952e-01
2.24655062e-01 1.00692475e+00 -6.54063940e-01 7.04785407e-01
6.84743226e-01 6.95930243e-01 -1.19766653e-01 9.43694472e-01
6.17219985e-01 7.77570784e-01 -2.92043626e-01 5.21989882e-01
-2.73573875e-01 -4.34053093e-01 3.26039463e-01 1.01786983e+00
6.73038006e-01 4.03001517e-01 -1.12127505e-01 7.45702982e-01
2.65018165e-01 2.23900765e-01 -6.65280282e-01 3.60361993e-01
5.45191407e-01 6.73930049e-01 -5.31496882e-01 -2.58019984e-01
-4.53598827e-01 1.94776282e-01 -1.21087627e-02 3.32293570e-01
-1.00091732e+00 -5.52020483e-02 6.37784243e-01 6.66172326e-01
2.45443638e-02 7.37427175e-02 -8.86068463e-01 -8.79158556e-01
1.08446181e-01 -7.01546550e-01 1.03341234e+00 -5.75286448e-01
-1.65448606e+00 3.57997775e-01 5.78192890e-01 -1.66693604e+00
-2.69242734e-01 -2.14391962e-01 -3.71415168e-01 9.39935923e-01
-1.19467199e+00 -8.87583613e-01 2.32495693e-03 6.34525776e-01
2.39323005e-01 3.49713206e-01 9.90333080e-01 8.49516019e-02
-4.58248705e-01 7.83414304e-01 1.53575987e-01 -1.66695774e-01
8.20439339e-01 -1.30306196e+00 2.41024733e-01 4.70313549e-01
-3.60824466e-01 6.96400285e-01 7.75813282e-01 -8.27877343e-01
-9.24782813e-01 -1.14993501e+00 9.12962735e-01 -6.74512506e-01
5.68866849e-01 -2.55744401e-02 -1.06210625e+00 7.95771480e-01
-3.28764260e-01 1.80017814e-01 8.30218732e-01 4.69373286e-01
-3.44797671e-01 -5.27472556e-01 -1.35912371e+00 4.69349980e-01
7.94529855e-01 -3.76954556e-01 -2.89565980e-01 2.52371043e-01
8.10180783e-01 -4.34886664e-01 -1.76966870e+00 8.23766232e-01
9.18734252e-01 -1.00226974e+00 9.61986303e-01 -7.48246968e-01
5.34494877e-01 1.97035968e-02 -1.48174688e-01 -1.14702272e+00
-1.80115625e-01 -7.61237144e-01 1.55240566e-01 9.57760751e-01
4.62389112e-01 -9.51639414e-01 3.31466496e-01 7.79518187e-01
-1.60504490e-01 -1.15150785e+00 -1.14579439e+00 -8.04081082e-01
6.42044604e-01 -7.39500940e-01 9.01454091e-01 1.04874790e+00
4.09228444e-01 -8.92514288e-02 -8.00037533e-02 1.78660199e-01
7.31198490e-01 2.92460620e-01 4.95196342e-01 -1.42575145e+00
-1.85236260e-01 -4.58198249e-01 -8.35887641e-02 -3.31038654e-01
-3.15269262e-01 -7.14175403e-01 1.21165654e-02 -1.07333863e+00
3.67474288e-01 -8.19461167e-01 -4.66077030e-01 1.34767786e-01
-1.39173433e-01 -1.45883458e-02 -4.38600212e-01 1.95511386e-01
1.11961495e-02 3.85337770e-01 1.16465676e+00 2.29247108e-01
-5.09631395e-01 4.45185959e-01 -6.34716988e-01 4.02302921e-01
8.71315956e-01 -7.70155907e-01 -6.41465068e-01 3.69932503e-01
1.83793336e-01 9.73688841e-01 2.74120629e-01 -6.77932024e-01
-8.84483010e-02 -5.34411728e-01 4.34343636e-01 -7.07196593e-01
-1.19453490e-01 -7.11001337e-01 6.11878514e-01 5.27135968e-01
-5.36791980e-01 5.50973117e-01 3.43211681e-01 6.76821113e-01
9.45402682e-02 3.67912352e-01 8.61305356e-01 3.38300914e-01
4.49854732e-01 5.52930474e-01 2.70898174e-02 4.27635729e-01
1.09313726e+00 -1.32784694e-01 -1.73115671e-01 -2.14909792e-01
-7.16467619e-01 3.20952505e-01 2.57439017e-01 6.71774596e-02
2.83383310e-01 -1.48897350e+00 -7.96249747e-01 3.59749526e-01
3.51450264e-01 2.28173919e-02 5.48130870e-01 1.54309309e+00
-4.05265316e-02 3.64117920e-01 -6.41420931e-02 -9.15491939e-01
-7.29709983e-01 8.39600563e-01 4.89371777e-01 -3.69076431e-01
-9.59279776e-01 2.48282596e-01 6.00754619e-01 -1.91738859e-01
4.08183075e-02 -7.36299336e-01 -4.65733148e-02 1.21073321e-01
4.92995739e-01 4.35027748e-01 -2.06202120e-02 -2.57938802e-01
-2.55092978e-01 7.44032979e-01 4.06541497e-01 -3.92752141e-01
1.31135702e+00 -9.85669494e-02 1.27292767e-01 1.00187743e+00
1.25019193e+00 -4.55254242e-02 -1.49590969e+00 -1.01979606e-01
1.40432119e-01 -5.68053387e-02 -4.86957282e-01 -6.63362026e-01
-7.03441679e-01 5.59275985e-01 2.65266269e-01 4.53254938e-01
1.33531451e+00 -4.88673849e-03 4.86576378e-01 -5.33030450e-01
3.46865207e-01 -8.03176999e-01 -2.72579700e-01 -3.00463568e-02
1.09204519e+00 -9.65849996e-01 9.14470181e-02 -2.27709159e-01
-6.92811608e-01 7.96923757e-01 2.42169090e-02 -2.73574948e-01
1.05003369e+00 3.06015819e-01 -2.01881230e-01 -1.27196789e-01
-8.56402636e-01 4.20034498e-01 4.74831045e-01 4.33456749e-01
5.46729147e-01 3.81585747e-01 -4.65925515e-01 9.88938987e-01
-4.82396603e-01 -3.05266380e-02 2.53105491e-01 5.99082291e-01
8.32711607e-02 -6.32337034e-01 -3.57415289e-01 7.42225230e-01
-5.09929776e-01 6.54123127e-02 5.33290446e-01 1.01000249e+00
-7.45682120e-02 6.60381854e-01 3.77318919e-01 -1.56898826e-01
6.64610445e-01 3.04212898e-01 3.04634064e-01 -1.90738499e-01
-4.31371391e-01 3.42189938e-01 -2.45598271e-01 -6.35374904e-01
-3.85455817e-01 -1.32755029e+00 -1.27655637e+00 -4.65595126e-01
-2.42571142e-02 2.04210475e-01 3.03527415e-01 9.14738715e-01
2.66716301e-01 5.80099344e-01 8.65025043e-01 -2.29886577e-01
-6.83369398e-01 -9.26087081e-01 -6.51243508e-01 2.69485116e-01
7.42783070e-01 -6.61020815e-01 -2.06577867e-01 7.92764872e-02] | [7.792013168334961, 4.791163444519043] |
41f79d20-ae5f-4dfb-a944-21ccecc32f58 | formal-language-recognition-by-hard-attention | 2204.06618 | null | https://arxiv.org/abs/2204.06618v1 | https://arxiv.org/pdf/2204.06618v1.pdf | Formal Language Recognition by Hard Attention Transformers: Perspectives from Circuit Complexity | This paper analyzes three formal models of Transformer encoders that differ in the form of their self-attention mechanism: unique hard attention (UHAT); generalized unique hard attention (GUHAT), which generalizes UHAT; and averaging hard attention (AHAT). We show that UHAT and GUHAT Transformers, viewed as string acceptors, can only recognize formal languages in the complexity class AC$^0$, the class of languages recognizable by families of Boolean circuits of constant depth and polynomial size. This upper bound subsumes Hahn's (2020) results that GUHAT cannot recognize the DYCK languages or the PARITY language, since those languages are outside AC$^0$ (Furst et al., 1984). In contrast, the non-AC$^0$ languages MAJORITY and DYCK-1 are recognizable by AHAT networks, implying that AHAT can recognize languages that UHAT and GUHAT cannot. | ['Robert Frank', 'Dana Angluin', 'Yiding Hao'] | 2022-04-13 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 2.04477429e-01 8.98269951e-01 1.99704729e-02 -3.77727523e-02
-6.51793480e-01 -1.04586875e+00 4.08947587e-01 2.01626904e-02
-9.23256427e-02 8.11101377e-01 -4.47792234e-03 -9.13299382e-01
-4.96010818e-02 -1.19696987e+00 -1.00940824e+00 -6.30460978e-01
-1.96601793e-01 6.98224247e-01 3.00185323e-01 -2.96618313e-01
-4.28303406e-02 5.72881401e-01 -1.56612086e+00 3.29819709e-01
4.77622718e-01 9.44359303e-01 -1.70400113e-01 1.05656612e+00
-2.58847233e-02 9.61941123e-01 -1.67536005e-01 -6.05318308e-01
2.69783854e-01 -7.40470171e-01 -1.11877835e+00 -7.41201043e-01
5.48708439e-01 -3.38932723e-02 -5.06367862e-01 1.26393890e+00
1.96906939e-01 -3.32668185e-01 5.91161847e-01 -1.18324685e+00
-1.43387341e+00 1.30314147e+00 3.95150259e-02 3.72207671e-01
4.47816908e-01 4.99943122e-02 1.73611450e+00 -3.95158708e-01
1.91347614e-01 1.28818536e+00 6.51486218e-01 1.02361381e+00
-1.16307557e+00 -7.83601582e-01 1.84004709e-01 6.33916631e-02
-1.48902845e+00 -5.94053805e-01 3.11555922e-01 -2.64211267e-01
1.73072612e+00 6.48526728e-01 7.36877739e-01 6.76470280e-01
4.83714789e-01 6.58898652e-01 9.91543114e-01 -7.67593682e-01
2.26866260e-01 -3.50622945e-02 3.02857906e-01 1.13082480e+00
5.34946859e-01 -2.71129124e-02 -1.31291866e-01 -1.27057925e-01
7.47348130e-01 -2.10643172e-01 -3.67578328e-01 3.01368702e-02
-7.07477212e-01 7.72256672e-01 2.21541926e-01 7.58376777e-01
-1.98638067e-03 4.79451716e-01 6.30533248e-02 9.40406501e-01
-2.35918730e-01 6.81762516e-01 -2.60503232e-01 1.03472270e-01
-1.88848183e-01 -1.60872355e-01 1.04475343e+00 1.42212892e+00
7.39052057e-01 3.71312320e-01 -1.28955198e-02 9.20158103e-02
1.65814068e-02 8.97118151e-01 3.54512751e-01 -7.30319321e-01
2.98724025e-01 3.91203016e-01 -2.29910508e-01 -3.52990270e-01
-3.00238822e-02 -1.65700302e-01 -7.49243140e-01 -2.36895368e-01
3.83996785e-01 -7.06095546e-02 -8.62799287e-01 2.18553185e+00
-5.88111997e-01 -4.66174722e-01 2.16271937e-01 4.35253352e-01
6.38825238e-01 9.52014267e-01 1.52871723e-03 -3.26707989e-01
1.30247104e+00 -6.95553958e-01 -5.14255702e-01 -3.18750799e-01
5.81423819e-01 -2.35085964e-01 1.08242202e+00 2.68535882e-01
-1.65525568e+00 -3.79826039e-01 -1.08086455e+00 -1.84431940e-01
-4.37983364e-01 -6.48738325e-01 8.78154635e-01 9.42263424e-01
-1.77208161e+00 9.56127718e-02 -5.85349977e-01 -3.55685830e-01
2.34254509e-01 9.35776532e-01 -2.76987821e-01 -4.16101031e-02
-1.30186212e+00 7.57447064e-01 1.39233395e-01 2.56383885e-02
-1.32700932e+00 1.84564386e-02 -1.01761007e+00 6.53710723e-01
1.04404919e-01 -5.01519561e-01 1.32322931e+00 -1.46332073e+00
-1.40303040e+00 8.56663525e-01 -1.55207515e-01 -8.51649642e-01
-1.53848961e-01 2.79707700e-01 -4.48254824e-01 1.37420729e-01
-3.07718605e-01 3.33169907e-01 3.76141191e-01 -8.47695410e-01
-6.11684561e-01 -3.08388799e-01 5.02888024e-01 -1.27812892e-01
-2.35067874e-01 8.24775323e-02 8.65156800e-02 -2.58199796e-02
6.61203265e-02 -7.55420148e-01 1.38243720e-01 -3.89816761e-01
-5.11551142e-01 -3.76633048e-01 5.27475715e-01 -1.05461828e-01
1.05643463e+00 -2.16020966e+00 2.71569073e-01 3.85973930e-01
5.67231655e-01 2.04027712e-01 -2.89751649e-01 5.60332716e-01
-3.63071054e-01 3.29189837e-01 -1.92594856e-01 7.86360949e-02
5.42197764e-01 5.41212022e-01 -3.71856928e-01 3.56621176e-01
2.66814679e-01 1.09800088e+00 -6.93597198e-01 -3.28779370e-01
-2.61983365e-01 6.65837005e-02 -8.80817294e-01 1.98381364e-01
-4.10047710e-01 -4.45230097e-01 -3.48262668e-01 7.70476818e-01
2.76156873e-01 -3.50703984e-01 3.53541851e-01 3.54134262e-01
-1.88084498e-01 3.42906177e-01 -5.57316363e-01 7.30404437e-01
-1.73523977e-01 8.26616883e-01 2.98338473e-01 -1.00129199e+00
7.52067029e-01 5.76387286e-01 -2.86314562e-02 -8.04363191e-01
3.84037673e-01 3.53014767e-01 3.89067352e-01 -1.30217150e-01
3.16598445e-01 -5.60342610e-01 -3.28219146e-01 5.56935906e-01
2.05567256e-01 5.31187057e-02 5.96007379e-03 1.96976855e-01
1.48349440e+00 -6.80642188e-01 3.98055390e-02 -6.14529371e-01
5.08236289e-01 -4.77497727e-01 6.94515169e-01 1.22085071e+00
-3.01380724e-01 4.26143765e-01 8.94132316e-01 -3.53434801e-01
-1.13739550e+00 -1.31599534e+00 8.40834603e-02 1.09064364e+00
-7.12648556e-02 -4.47342992e-01 -9.77516413e-01 -2.13050306e-01
-2.48300806e-01 7.24079669e-01 -7.43267417e-01 -3.68628740e-01
-4.96908128e-01 -3.25167030e-01 8.67172539e-01 7.60211170e-01
5.09100318e-01 -1.45675874e+00 -4.45799351e-01 -4.67432663e-02
1.38869360e-01 -6.20866895e-01 -4.16827321e-01 8.92617047e-01
-4.44268912e-01 -8.19249392e-01 -5.53103447e-01 -1.05795920e+00
9.17840600e-01 -3.85869801e-01 1.00406420e+00 1.37250632e-01
-5.29178008e-02 5.68608880e-01 -6.99044913e-02 -3.03618789e-01
-4.83297199e-01 2.03160748e-01 -9.19566974e-02 -2.33657897e-01
5.81814766e-01 -5.83560884e-01 -1.06787987e-01 -8.83424133e-02
-6.73405886e-01 -2.35075310e-01 5.31798065e-01 9.15977299e-01
4.45675790e-01 1.76062062e-01 3.00105155e-01 -1.02088714e+00
4.20660973e-01 -3.34847718e-01 -8.88668895e-01 4.54380870e-01
-5.86114168e-01 3.72495592e-01 1.10157228e+00 -2.46075615e-01
-4.39564109e-01 -6.22483380e-02 -2.96737969e-01 -2.31258348e-01
2.68862903e-01 2.68277109e-01 -4.29515719e-01 -3.39782059e-01
3.62644345e-01 3.93226713e-01 -4.08168346e-01 -3.66868302e-02
-3.05741243e-02 5.62564790e-01 4.70466137e-01 -7.28266597e-01
6.12663031e-01 -9.13598090e-02 -1.21852003e-01 -6.17876053e-01
-4.54806060e-01 1.52302593e-01 -4.08997312e-02 2.36491963e-01
9.18524265e-01 -5.96554220e-01 -1.03444147e+00 2.49819562e-01
-9.09776807e-01 -4.93415982e-01 -6.76158130e-01 2.85454076e-02
-7.97775030e-01 2.35103846e-01 -1.21615350e+00 -1.12674522e+00
-4.88825887e-01 -1.05328965e+00 7.65454292e-01 1.18803792e-01
-2.15782464e-01 -7.66777754e-01 -1.25712350e-01 -2.14062050e-01
4.45548356e-01 -7.96465352e-02 1.74187660e+00 -7.40541101e-01
-8.74313056e-01 -3.23563755e-01 -2.82527447e-01 5.16799748e-01
-4.00688171e-01 -2.31592208e-02 -1.03058612e+00 -3.80165815e-01
-6.09881617e-02 -4.85079348e-01 7.33658850e-01 2.31885239e-01
1.06267416e+00 -7.53102422e-01 -1.45618096e-01 5.12511432e-01
1.48654377e+00 7.96936512e-01 8.51795733e-01 4.51958068e-02
3.15641671e-01 -9.93951410e-02 -6.20100856e-01 -1.87415823e-01
3.31894279e-01 1.50636462e-02 3.76134843e-01 -3.42528932e-02
6.47743046e-02 -4.40129846e-01 6.81132972e-01 1.28779757e+00
-1.70815364e-01 -6.68586552e-01 -8.39541733e-01 7.34818041e-01
-1.39806283e+00 -1.17876208e+00 4.67799120e-02 2.00550485e+00
7.98004925e-01 6.49509132e-01 -4.99646887e-02 4.98202294e-01
7.90479124e-01 -8.27122703e-02 -1.54068887e-01 -1.39470494e+00
-5.67985058e-01 1.01154792e+00 6.24979138e-01 8.17042768e-01
-4.42161888e-01 8.78747225e-01 7.59048176e+00 4.40156400e-01
-6.87349617e-01 4.37064841e-02 5.73415875e-01 2.94195026e-01
-8.78071129e-01 3.09665710e-01 -7.45065808e-01 3.51571947e-01
1.41132319e+00 -2.41399392e-01 7.76282549e-01 5.79531968e-01
-8.73962820e-01 2.26341501e-01 -1.57921231e+00 5.38136601e-01
6.05420023e-02 -1.05364025e+00 1.30895570e-01 7.57630616e-02
5.83007872e-01 -1.00540459e-01 3.37631434e-01 5.99745512e-01
8.86633515e-01 -1.19542682e+00 7.66584277e-01 3.28531265e-01
9.58419979e-01 -7.99481869e-01 6.82438195e-01 1.93760395e-01
-1.01394999e+00 -5.64937294e-01 -5.26883781e-01 -2.93304861e-01
-3.52693111e-01 2.42982898e-02 -3.99035096e-01 1.83823239e-02
5.41101575e-01 1.06022611e-01 -3.69901210e-01 6.12019897e-01
-9.33701396e-02 7.34896958e-01 -4.36967760e-01 -2.51456559e-01
4.65032279e-01 1.03675134e-01 1.54990569e-01 1.06508219e+00
6.03541851e-01 6.08749211e-01 -2.21272618e-01 8.47244740e-01
-1.22145012e-01 -5.10150254e-01 -6.65123045e-01 -5.94469130e-01
4.84411895e-01 6.11755431e-01 -5.18595994e-01 -4.48161691e-01
-4.01126832e-01 7.21156776e-01 2.81367511e-01 1.38438702e-01
-5.63828945e-01 -9.47136343e-01 4.15348172e-01 1.98395327e-01
6.27697051e-01 2.40693688e-01 1.22811466e-01 -1.11881959e+00
-4.74696219e-01 -1.00409448e+00 6.77746177e-01 -1.06504798e+00
-1.07901883e+00 9.52463865e-01 -3.90382499e-01 -4.98017460e-01
-5.42873368e-02 -7.05597579e-01 -5.29787004e-01 8.10506344e-01
-1.07209420e+00 -7.79149532e-01 2.60407448e-01 8.50984693e-01
-2.15686504e-02 -1.13735288e-01 1.25110257e+00 2.15801746e-01
-5.85666716e-01 1.02671838e+00 -1.22932315e-01 3.21520984e-01
-2.39933088e-01 -1.47167826e+00 4.21790838e-01 7.62766898e-01
1.47030577e-01 8.14387739e-01 5.27550042e-01 -1.18178867e-01
-1.92202771e+00 -7.68136501e-01 1.56979430e+00 -3.20678681e-01
5.84005415e-01 -6.97083592e-01 -6.11951292e-01 1.40566802e+00
5.04049301e-01 2.05880584e-04 5.70159495e-01 9.74726900e-02
-7.23733664e-01 -2.03665778e-01 -1.04791605e+00 5.88974178e-01
1.14604247e+00 -1.14146996e+00 -6.27609253e-01 2.16334939e-01
8.17568898e-01 -6.58544078e-02 -7.10908473e-01 1.19824618e-01
5.98403513e-01 -1.03895307e+00 3.87614071e-01 -7.20128179e-01
2.97513902e-01 9.87928808e-02 -3.29435021e-01 -8.07498336e-01
-8.33490133e-01 -8.69140804e-01 3.08062863e-02 8.95819128e-01
6.68993711e-01 -1.04629505e+00 5.21113396e-01 2.89976895e-01
-4.58117932e-01 -3.65066290e-01 -1.19513655e+00 -6.49300814e-01
4.75666821e-01 -2.11149111e-01 9.25665498e-01 7.78622210e-01
6.00182474e-01 5.08549929e-01 -7.15095699e-02 -2.59116758e-02
2.63680398e-01 1.87342614e-01 4.81032245e-02 -1.06801939e+00
-6.63717031e-01 -6.61518633e-01 -6.66432619e-01 -8.20841372e-01
2.15185538e-01 -1.06427252e+00 1.77866846e-01 -1.21753323e+00
3.44889998e-01 -3.60253692e-01 -6.23512745e-01 8.49534333e-01
5.92206776e-01 1.32721201e-01 -7.31725916e-02 -2.06096143e-01
-5.15646994e-01 2.01923728e-01 1.04057395e+00 -3.59737784e-01
4.80073661e-01 -2.34383062e-01 -1.12775874e+00 4.41717565e-01
5.76315761e-01 -4.01515216e-01 -4.83331144e-01 -6.58695638e-01
5.46681225e-01 3.13117832e-01 1.33160025e-01 -9.99098957e-01
2.22151294e-01 2.21195165e-02 1.27267182e-01 -1.17489807e-01
2.35812351e-01 -7.83105731e-01 1.56108841e-01 6.59962773e-01
-7.26814568e-01 6.53808117e-01 1.80261880e-01 2.52090394e-01
-9.02646570e-04 -1.75793752e-01 7.44248331e-01 -3.47988993e-01
-1.98362052e-01 2.10644796e-01 -8.80145371e-01 1.77194491e-01
8.56241822e-01 -2.92481929e-01 -6.96204662e-01 -6.58152342e-01
-8.93685281e-01 4.04744633e-02 3.68420839e-01 -2.32017905e-01
5.49329162e-01 -1.11506093e+00 -2.11597055e-01 8.01197827e-01
2.08701305e-02 -6.46157339e-02 -6.72614574e-02 4.57255244e-01
-4.23625469e-01 7.83752441e-01 -4.31706786e-01 -7.80819580e-02
-8.04716349e-01 6.20614588e-01 6.27126634e-01 -1.67859167e-01
-2.99712241e-01 1.22034919e+00 4.75322843e-01 -1.12074472e-01
2.00192004e-01 -6.76608503e-01 2.99798816e-01 -5.51513910e-01
3.08694810e-01 -2.92512923e-01 -1.15937360e-01 -5.12543261e-01
-3.62827659e-01 3.74893069e-01 -2.02056259e-01 5.46688698e-02
1.06799817e+00 2.93358594e-01 -5.69055319e-01 4.18499023e-01
9.69528973e-01 -3.92373167e-02 -4.88177270e-01 2.56714120e-04
-3.15853596e-01 1.26532093e-01 -1.91388607e-01 -6.19292915e-01
-1.01298976e+00 1.16808069e+00 2.65434355e-01 9.27478492e-01
1.23036349e+00 1.77302271e-01 8.28251481e-01 6.08053088e-01
5.25978804e-01 -7.22880006e-01 -4.74850446e-01 8.57230723e-01
5.05645752e-01 -2.98377216e-01 -6.92425370e-01 -5.13766557e-02
-3.30038592e-02 9.11609948e-01 7.25976050e-01 -1.73018351e-01
4.36207771e-01 7.09447503e-01 -5.06315589e-01 -1.33705214e-01
-1.45168960e+00 -3.09610248e-01 -2.36306772e-01 5.11956453e-01
3.35496575e-01 1.26519769e-01 -4.96866964e-02 7.32084394e-01
-5.23943186e-01 -1.43881455e-01 4.71441478e-01 9.06825066e-01
-7.50794530e-01 -9.33662891e-01 -8.52648169e-02 4.27470803e-01
-5.39810359e-01 -4.57828313e-01 -7.31351137e-01 7.14084446e-01
1.75920188e-01 6.30420864e-01 4.56981212e-01 -5.07881105e-01
2.38948271e-01 2.01027632e-01 8.47451568e-01 -7.35951424e-01
-6.64821029e-01 -1.04317605e-01 2.27535918e-01 -1.04123227e-01
1.82253182e-01 -3.33517164e-01 -1.05648935e+00 -7.54810929e-01
-5.17463505e-01 6.94899440e-01 -2.13943794e-01 7.52199471e-01
1.04669318e-01 3.78648281e-01 4.35005277e-01 1.03335865e-01
-6.52511299e-01 -5.81518590e-01 -8.00162554e-01 -1.24590464e-01
7.27051973e-01 -7.04209656e-02 -6.65694475e-01 -9.94911864e-02] | [9.455772399902344, 7.069998741149902] |
58cb6605-a4c1-4bfc-9dbc-dc30e2efb1bf | fast-and-robust-template-matching-with | null | null | https://www.sciencedirect.com/science/article/abs/pii/S0031320319303322 | https://www.sciencedirect.com/science/article/abs/pii/S0031320319303322 | Fast and robust template matching with majority neighbour similarity and annulus projection transformation | In the paper, a novel fast and robust template matching method named A-MNS based on Majority Neighbour Similarity (MNS) and the annulus projection transformation (APT) is proposed. Its essence is the MNS, a useful, rotation-invariant, low computational cost and robust similarity measurement. The proposed method is theoretically demonstrated and experimentally evaluated as being able to estimate the rotation angle of the target object, overcome challenges such as background clutter, occlusion, arbitrary rotation transformation, and non-rigid deformation, while performing fast matching. Empirical results evaluated on the up-to-date benchmark show that A-MNS is 4.419 times faster than DDIS (the state-ofthe-art) and is also competitive in terms of its matching accuracy. | ['Zhou Jinyun', 'Yang Ruan', 'Runming Yan', 'Kaiyuan Deng', 'Liang Lei', 'Jinxiang Lai'] | 2020-02-10 | null | null | null | pattern-recognition-2020-2 | ['template-matching'] | ['computer-vision'] | [ 1.45754397e-01 -6.00397110e-01 7.15692118e-02 -6.97208717e-02
-5.66119015e-01 -5.71157515e-01 7.84846783e-01 -1.67514876e-01
-3.13587189e-01 2.07978517e-01 -5.88658266e-02 2.44914796e-02
-2.66812414e-01 -4.90439415e-01 -4.04878110e-01 -6.12749815e-01
1.81160927e-01 6.38879240e-01 6.12952352e-01 -1.41351476e-01
6.54342592e-01 1.27858877e+00 -1.55072534e+00 -3.84293795e-01
4.76229191e-01 1.17453325e+00 4.48061898e-02 3.06827128e-01
4.39170778e-01 -5.15145957e-02 -5.34767509e-01 -6.13905489e-01
8.79398048e-01 3.77576873e-02 -7.15886295e-01 1.19633675e-02
7.87138879e-01 -2.88693421e-02 -4.72009510e-01 8.34916234e-01
8.22785735e-01 3.97604287e-01 6.93299770e-01 -8.64685237e-01
-2.45497853e-01 -1.88756436e-01 -7.21298933e-01 1.01702958e-01
6.24356091e-01 -1.44028991e-01 5.64534426e-01 -1.24394131e+00
9.76811051e-01 1.19901335e+00 8.42385292e-01 1.40593439e-01
-1.00779974e+00 -5.30956030e-01 -5.17093360e-01 3.29157919e-01
-1.75296760e+00 -2.41833657e-01 6.17537677e-01 -2.34388113e-01
8.64212513e-01 7.08288133e-01 3.81672710e-01 6.33550286e-01
3.74148905e-01 1.70119941e-01 1.12015247e+00 -6.97167397e-01
5.49446791e-02 -3.56220007e-01 -1.04923859e-01 4.31716174e-01
3.90135378e-01 4.55691129e-01 -3.24638575e-01 -5.39928317e-01
9.26664054e-01 -5.97396679e-02 -1.17770739e-01 -8.12608123e-01
-1.45931947e+00 4.24834669e-01 4.26422179e-01 3.13658208e-01
-3.77727538e-01 -2.32860371e-01 3.08855295e-01 3.54474217e-01
1.41407356e-01 3.01729977e-01 -2.91159928e-01 -2.02989623e-01
-6.91527426e-01 2.33517990e-01 7.38308668e-01 9.05147076e-01
4.39793497e-01 4.10134811e-03 -1.20147757e-01 7.53924787e-01
3.06476831e-01 9.72632289e-01 4.89734679e-01 -7.61076510e-01
2.76582628e-01 5.91728389e-01 1.60365030e-01 -1.34930658e+00
-5.94570756e-01 -5.02613187e-01 -8.82374465e-01 3.49368513e-01
2.88997024e-01 4.79202062e-01 -6.75557673e-01 1.19346225e+00
6.19463861e-01 3.47273141e-01 7.13249743e-02 1.02720392e+00
8.05932164e-01 2.10039571e-01 -3.58269036e-01 -1.03287205e-01
1.31279814e+00 -8.44115615e-01 -3.45601082e-01 -5.25902025e-02
1.42291009e-01 -1.74412107e+00 3.87070149e-01 2.12717637e-01
-9.29933846e-01 -9.33514833e-01 -1.25215507e+00 3.38038921e-01
-6.89034238e-02 1.44187465e-01 5.61060071e-01 6.22893989e-01
-7.31467664e-01 7.24398673e-01 -5.76011181e-01 -5.99174440e-01
-1.22271210e-01 5.72562575e-01 -7.97212303e-01 -1.30327955e-01
-8.54827285e-01 1.28783846e+00 1.83564603e-01 -6.16614847e-03
1.52253322e-02 -5.16723871e-01 -6.68590546e-01 -2.38315612e-01
1.07471749e-01 -7.99042106e-01 7.14761257e-01 -5.11242330e-01
-1.77630377e+00 1.10381293e+00 -2.30285689e-01 -3.11347246e-01
7.54125834e-01 -8.54973644e-02 -5.72087407e-01 1.49561167e-01
-1.41117558e-01 3.85989130e-01 9.08055246e-01 -8.35061789e-01
-2.08190441e-01 -6.19835138e-01 -4.70739186e-01 2.99674392e-01
1.13806553e-01 4.05816436e-01 -6.18043900e-01 -9.47411418e-01
7.75646925e-01 -1.22506499e+00 -1.40154868e-01 1.71388403e-01
4.06770259e-02 -1.17439434e-01 1.09726703e+00 -6.08423054e-01
8.64561200e-01 -2.19110036e+00 -9.98690203e-02 7.21470296e-01
-2.92487562e-01 7.01337636e-01 -1.22752383e-01 4.53914762e-01
-2.88803130e-01 -6.16316199e-01 1.07702352e-01 4.47662249e-02
-1.78391382e-01 -3.53002101e-02 -2.16873765e-01 1.01940405e+00
-2.01317668e-01 7.86692441e-01 -5.68705678e-01 -4.02837634e-01
6.53870583e-01 5.78988612e-01 1.18478864e-01 1.70051798e-01
6.34333789e-01 3.31341773e-01 -1.21236630e-01 8.26252937e-01
1.23175669e+00 2.23857880e-01 4.93794233e-02 -6.85218155e-01
-1.43432245e-01 -2.08429679e-01 -1.79996622e+00 1.49548054e+00
-1.09426938e-01 6.01832151e-01 -2.51605451e-01 -7.82891691e-01
1.71073008e+00 2.59413362e-01 5.37234604e-01 -8.19978952e-01
1.92491174e-01 5.37938595e-01 4.01340090e-02 -4.65780944e-02
4.31355953e-01 4.00454015e-01 3.46470565e-01 1.87959135e-01
-3.13555360e-01 -1.65203363e-01 7.53060728e-02 -2.53883451e-01
9.59766686e-01 2.99561501e-01 6.73238218e-01 -4.56412464e-01
1.05268669e+00 -2.45742172e-01 4.96048719e-01 4.06669676e-01
-1.23201631e-01 8.64328623e-01 -3.20784211e-01 -6.39698029e-01
-1.08635283e+00 -1.13731980e+00 -4.37255472e-01 3.51313233e-01
6.42319262e-01 -2.09358618e-01 -4.83103275e-01 -1.70835704e-01
2.46337712e-01 -3.92887443e-02 -2.48705789e-01 -8.25584009e-02
-9.90325749e-01 -3.57557952e-01 1.99995607e-01 6.19533241e-01
6.11970663e-01 -6.49616241e-01 -5.79269350e-01 2.46510804e-01
-6.94452077e-02 -1.36905348e+00 -7.23114908e-01 -4.40310866e-01
-1.01214588e+00 -1.13011885e+00 -8.79774213e-01 -7.55746007e-01
8.56493175e-01 6.75871432e-01 1.02450180e+00 5.80712454e-03
-6.65147066e-01 1.74638554e-01 -1.56352296e-01 -1.11405917e-01
-1.84940904e-01 -1.89553544e-01 2.84585595e-01 7.24773556e-02
3.89536500e-01 -5.78106046e-01 -7.08096325e-01 1.23500764e+00
-5.28414488e-01 -2.48799965e-01 8.04847419e-01 6.91682398e-01
7.60402322e-01 -3.92718166e-01 4.47850823e-02 -1.64206609e-01
1.15629695e-01 2.85089284e-01 -1.08194125e+00 3.55630308e-01
-7.34559000e-01 4.77505326e-02 2.23794818e-01 -5.85435271e-01
-8.34048271e-01 3.55000973e-01 1.41520575e-02 -5.37549257e-01
-1.96335688e-02 -1.02171548e-01 -4.15355600e-02 -1.12271130e+00
6.97490335e-01 9.35898647e-02 2.10382998e-01 -5.56179345e-01
6.37916997e-02 6.42464459e-01 1.07405770e+00 -6.24808133e-01
1.39840543e+00 7.26444900e-01 7.10098147e-01 -7.67556906e-01
-1.79834902e-01 -1.08954191e+00 -9.14293468e-01 -1.03747696e-01
2.55391300e-01 -6.66897535e-01 -8.62994611e-01 6.87400818e-01
-1.15856397e+00 5.06246805e-01 -1.72006816e-01 7.65129447e-01
-6.29119575e-01 9.04762089e-01 -1.63779035e-01 -4.28317040e-01
-9.23400462e-01 -1.06743383e+00 9.22488153e-01 4.48051006e-01
-1.72673672e-01 -7.05556989e-01 1.13860801e-01 4.11722153e-01
6.53018236e-01 4.62997735e-01 2.31052801e-01 -6.92251444e-01
-6.05065525e-01 -6.99244142e-01 -4.37329412e-01 4.19418477e-02
1.86011150e-01 2.59374171e-01 -5.80206215e-01 -5.88193178e-01
-4.08823229e-02 2.41843835e-01 1.92877382e-01 8.79952013e-02
4.64290231e-01 1.20534435e-01 -4.38664168e-01 7.76480854e-01
1.48994195e+00 2.08127856e-01 9.29660857e-01 7.32977629e-01
2.84853309e-01 2.69811988e-01 1.23123145e+00 4.31403875e-01
1.60843581e-02 1.45044422e+00 3.47017616e-01 -1.91351533e-01
-3.39794606e-01 3.29749376e-01 1.09850712e-01 7.56908357e-01
-4.38994169e-01 4.14517581e-01 -6.92991674e-01 5.61301336e-02
-1.83384550e+00 -9.68266428e-01 -3.46067727e-01 2.64793229e+00
2.94242889e-01 -1.37168039e-02 -1.34031801e-02 2.50421196e-01
1.09046543e+00 -1.62960678e-01 -2.10632980e-01 -3.49431723e-01
-4.54710931e-01 6.25340044e-01 7.75276661e-01 4.73710090e-01
-1.08543563e+00 6.47144496e-01 6.60290241e+00 9.27594364e-01
-1.14890265e+00 -2.20953196e-01 -1.64345410e-02 5.48740029e-01
4.04725730e-01 5.36265671e-02 -8.28181863e-01 1.57512188e-01
5.20929277e-01 -2.02425122e-01 2.71586359e-01 8.42732370e-01
-2.61598736e-01 -7.42564127e-02 -7.70621061e-01 1.22739208e+00
4.22619700e-01 -1.19022739e+00 -1.61821246e-01 -8.42736661e-02
6.52515590e-01 2.30864366e-03 -1.55284628e-01 -2.18542010e-01
-3.51041555e-01 -6.23365343e-01 5.09949803e-01 4.97830123e-01
7.20227182e-01 -7.04402447e-01 9.84518588e-01 -7.20755234e-02
-1.68367553e+00 3.14180851e-01 -5.34971178e-01 3.39014441e-01
-2.33755574e-01 2.30804190e-01 -6.16893172e-01 1.12812352e+00
5.45156419e-01 3.39647233e-01 -7.10686028e-01 1.53466356e+00
7.03314766e-02 -1.43671319e-01 -6.66957200e-01 2.43629649e-01
-3.33453938e-02 -3.98202986e-01 7.06909060e-01 1.16710055e+00
4.44040626e-01 -1.96021143e-02 1.05979212e-01 1.67063043e-01
3.54728341e-01 3.60557795e-01 -2.59460747e-01 6.11190677e-01
5.05341113e-01 1.43909657e+00 -8.96154940e-01 -8.88471827e-02
-1.51000381e-01 1.02348340e+00 -2.80883700e-01 5.23783043e-02
-7.98054039e-01 -6.31361842e-01 5.59236825e-01 7.91840032e-02
5.83241463e-01 -1.46877989e-01 -7.28518888e-02 -8.76345158e-01
4.15506989e-01 -9.21741545e-01 3.80630732e-01 -6.90786719e-01
-1.03106105e+00 7.38255143e-01 -1.04971617e-01 -1.92662001e+00
-3.50502700e-01 -7.95760214e-01 -5.04084766e-01 1.04528689e+00
-1.30508387e+00 -1.21636248e+00 -6.30136371e-01 7.82646954e-01
1.98246032e-01 -4.36297953e-01 9.36955690e-01 5.41276097e-01
-2.92754415e-02 7.95181453e-01 2.66269445e-01 -1.16848648e-02
1.06563914e+00 -7.14731276e-01 8.08527410e-01 7.58424461e-01
1.59176677e-01 6.94392562e-01 8.52817893e-01 -4.14322317e-01
-1.62828505e+00 -5.36636233e-01 9.49915648e-01 -1.95471466e-01
5.63522339e-01 6.37272224e-02 -8.21669817e-01 1.31845891e-01
-2.11249560e-01 5.74994087e-01 2.90969163e-01 -3.17170024e-01
-7.96142757e-01 -5.01779020e-01 -1.48009944e+00 3.43078554e-01
1.00307822e+00 -3.11873108e-01 -8.66813302e-01 2.69945771e-01
4.69260253e-02 -9.88445103e-01 -1.18745983e+00 8.37897003e-01
8.31563413e-01 -1.06560528e+00 1.53967822e+00 2.12462600e-02
-6.11277044e-01 -5.65055490e-01 -2.44787022e-01 -6.92220867e-01
-4.89773214e-01 -1.02688015e+00 -4.42769416e-02 1.06757951e+00
-2.24162653e-01 -9.70301926e-01 7.91845381e-01 4.07188743e-01
3.13106552e-02 -6.54163539e-01 -1.36278355e+00 -1.31586814e+00
-6.14208817e-01 3.70639488e-02 6.47416055e-01 9.21202838e-01
-4.38625216e-01 -1.15402468e-01 -2.74555027e-01 2.53159612e-01
7.70811081e-01 4.40106362e-01 1.13181043e+00 -1.56384194e+00
-1.60503581e-01 -3.90827149e-01 -1.31959367e+00 -8.81602585e-01
-1.42226323e-01 -4.21888858e-01 -4.36640412e-01 -1.12048471e+00
6.67457804e-02 -3.43290389e-01 -1.63391456e-01 -3.57865659e-03
-2.65559908e-02 5.51399708e-01 1.59325525e-01 5.53019226e-01
-2.59104908e-01 2.73277741e-02 1.09560943e+00 1.29587874e-02
2.00593784e-01 4.14122343e-01 5.55988811e-02 6.46387935e-01
5.55081606e-01 -4.00763571e-01 1.93656579e-01 1.16109513e-01
-2.91465998e-01 -1.04863837e-01 1.93463027e-01 -1.14480507e+00
3.33734304e-01 3.54035124e-02 5.67038536e-01 -1.09741485e+00
4.94249523e-01 -1.09233618e+00 6.92211330e-01 8.98920596e-01
4.94465441e-01 5.60873568e-01 1.28276497e-01 1.58026993e-01
-3.45371872e-01 -2.49424681e-01 9.64268029e-01 3.37835521e-01
-6.98426187e-01 2.71546274e-01 2.99687654e-01 -2.39185736e-01
1.18084669e+00 -8.16787481e-01 -3.98554683e-01 1.56271495e-02
-3.17492247e-01 -4.33112442e-01 7.38890111e-01 4.89857197e-01
5.79060674e-01 -1.64519620e+00 -8.57791901e-01 3.70563775e-01
2.85636693e-01 -4.61613417e-01 1.06075063e-01 7.23280251e-01
-8.59317362e-01 6.30584061e-01 -3.95953000e-01 -8.37357342e-01
-1.88213253e+00 5.15785873e-01 1.74774349e-01 -3.13326329e-01
-6.17065132e-01 4.17918146e-01 -4.11922187e-01 -4.20318395e-01
1.80804670e-01 9.96337235e-02 1.05728574e-01 -3.30230772e-01
4.43201989e-01 7.77093112e-01 5.52128136e-01 -1.06790400e+00
-7.28657544e-01 1.46155202e+00 1.56046480e-01 1.34402975e-01
9.79761839e-01 1.21830314e-01 -1.86830729e-01 -3.43858421e-01
1.14225125e+00 3.04079577e-02 -7.58962870e-01 -5.71874261e-01
1.21175706e-01 -1.06676948e+00 -3.69966507e-01 -4.40481395e-01
-1.07022893e+00 3.17646831e-01 1.10496163e+00 -2.26455584e-01
9.89264965e-01 -2.50420421e-01 7.15498805e-01 6.11094952e-01
4.27358299e-01 -9.71780539e-01 -8.54399502e-02 4.23295259e-01
1.12715554e+00 -1.07071280e+00 5.49946904e-01 -6.32487357e-01
-2.02463642e-01 1.48439991e+00 5.74347317e-01 -3.61918420e-01
5.35193741e-01 3.14455807e-01 2.71178395e-01 2.32783005e-01
-5.68548329e-02 -7.18939379e-02 8.69421482e-01 7.54496992e-01
2.58236051e-01 -2.31846780e-01 -6.09790206e-01 -2.63305724e-01
-9.76004601e-02 -2.92671412e-01 -1.67299613e-01 9.68654335e-01
-1.55412257e-01 -1.42991984e+00 -9.55066741e-01 1.41325248e-02
-4.16534871e-01 2.59893328e-01 -3.61809134e-01 8.80163252e-01
-1.87512964e-01 5.88066578e-01 1.02160759e-01 -4.00123268e-01
7.53099382e-01 -4.45733964e-01 8.72678876e-01 1.80405393e-01
-8.61505032e-01 1.67950854e-01 -2.26893574e-01 -8.03289890e-01
-7.05577493e-01 -6.42175734e-01 -8.84385049e-01 -4.42846864e-01
-6.20682240e-01 5.72671965e-02 1.02665615e+00 6.88663185e-01
4.81611252e-01 -1.70509741e-01 1.01691175e+00 -9.67454612e-01
-8.07717741e-01 -7.44508922e-01 -9.83195975e-02 7.19120681e-01
-2.57643521e-01 -8.43306720e-01 -3.67032558e-01 -3.02584141e-01] | [8.16873550415039, -2.308897018432617] |
d782f6b5-d4fa-41a3-a7ff-5862cccee069 | exploring-navigation-maps-for-learning-based | 2302.06195 | null | https://arxiv.org/abs/2302.06195v1 | https://arxiv.org/pdf/2302.06195v1.pdf | Exploring Navigation Maps for Learning-Based Motion Prediction | The prediction of surrounding agents' motion is a key for safe autonomous driving. In this paper, we explore navigation maps as an alternative to the predominant High Definition (HD) maps for learning-based motion prediction. Navigation maps provide topological and geometrical information on road-level, HD maps additionally have centimeter-accurate lane-level information. As a result, HD maps are costly and time-consuming to obtain, while navigation maps with near-global coverage are freely available. We describe an approach to integrate navigation maps into learning-based motion prediction models. To exploit locally available HD maps during training, we additionally propose a model-agnostic method for knowledge distillation. In experiments on the publicly available Argoverse dataset with navigation maps obtained from OpenStreetMap, our approach shows a significant improvement over not using a map at all. Combined with our method for knowledge distillation, we achieve results that are close to the original HD map-reliant models. Our publicly available navigation map API for Argoverse enables researchers to develop and evaluate their own approaches using navigation maps. | ['Klaus Dietmayer', 'Thomas Monninger', 'Franz Gritschneder', 'Julian Jordan', 'Julian Schmidt'] | 2023-02-13 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [-3.68751973e-01 3.83490682e-01 -3.59874576e-01 -4.39140975e-01
-8.53722811e-01 -5.89184344e-01 8.76363397e-01 4.20548469e-02
-6.87422395e-01 9.53798890e-01 2.12084576e-01 -6.96221292e-01
-1.25647083e-01 -1.48995960e+00 -8.99188936e-01 -3.49453002e-01
-2.50744641e-01 6.98394895e-01 9.85564053e-01 -5.70993364e-01
3.77642572e-01 5.49244225e-01 -1.88895452e+00 -7.68606961e-02
1.03779554e+00 6.08561039e-01 6.11428022e-01 9.47552085e-01
-6.57450482e-02 8.75579715e-01 1.87000990e-01 -1.99035574e-02
3.58336508e-01 2.75939226e-01 -8.12846899e-01 -5.71860611e-01
5.19866824e-01 -4.27315563e-01 -7.25519896e-01 4.47478205e-01
2.25589633e-01 3.39596152e-01 5.45286357e-01 -1.33913696e+00
9.25997049e-02 1.95633486e-01 6.65145144e-02 2.27843061e-01
3.12986434e-01 3.37124318e-01 5.40229321e-01 -7.42840767e-01
9.03495908e-01 8.62488866e-01 9.79966402e-01 1.05373584e-01
-1.00261116e+00 -4.49947625e-01 2.07441732e-01 7.33044982e-01
-1.64908779e+00 -4.22792345e-01 5.01636028e-01 -6.45267367e-01
1.21526265e+00 2.55807191e-01 7.35485911e-01 7.99875617e-01
5.99633396e-01 6.63349032e-01 9.25907433e-01 -4.63707820e-02
3.98725271e-01 2.56520361e-01 -1.09752953e-01 9.10049796e-01
6.21569119e-02 2.68786579e-01 -5.91904581e-01 8.53293985e-02
5.77503204e-01 -3.99856299e-01 -8.95745903e-02 -8.31888437e-01
-1.21532261e+00 7.86752939e-01 7.29057968e-01 -2.31385529e-01
-4.18898851e-01 3.09863687e-01 3.12114097e-02 -3.38295698e-02
4.81216699e-01 3.00938845e-01 -5.58455288e-01 -6.48297608e-01
-9.16301310e-01 7.91114986e-01 8.33599448e-01 1.08282959e+00
1.42522502e+00 -9.08737108e-02 2.23517522e-01 5.22251368e-01
2.38553137e-01 6.83100462e-01 1.38276562e-01 -1.36628532e+00
7.07516968e-01 3.48618031e-01 5.50183654e-01 -1.34910607e+00
-8.92744064e-01 -1.44019395e-01 -4.64428484e-01 4.75340307e-01
3.52812201e-01 -8.93433467e-02 -1.02379036e+00 1.42924011e+00
3.33103329e-01 4.50008303e-01 1.83655381e-01 6.19001150e-01
7.07160294e-01 8.16241026e-01 2.29866542e-02 4.73795444e-01
7.38991678e-01 -1.27808416e+00 -4.50012833e-01 -5.80843091e-01
1.15909553e+00 -2.28296429e-01 8.55609298e-01 1.16581209e-01
-4.99664992e-01 -5.39490879e-01 -1.09838045e+00 -1.55321762e-01
-1.06982780e+00 -1.50044069e-01 7.85126567e-01 5.56112468e-01
-1.54282904e+00 5.78760982e-01 -1.08346045e+00 -5.47846496e-01
3.13059896e-01 9.71186757e-02 -5.37905872e-01 -1.34114310e-01
-1.27983379e+00 1.58205044e+00 5.38344145e-01 -7.93495253e-02
-6.72404110e-01 -7.89267778e-01 -1.18852603e+00 -5.19250095e-01
1.03085838e-01 -4.92417425e-01 1.10851228e+00 -3.56374048e-02
-1.39986944e+00 4.71284479e-01 -3.38832289e-01 -7.15037704e-01
7.13454008e-01 -1.14406452e-01 -4.86942172e-01 -8.35364982e-02
5.24188221e-01 1.24914157e+00 6.61739558e-02 -1.31380486e+00
-1.20540226e+00 1.03452154e-01 2.48192102e-01 3.53569806e-01
2.53305674e-01 -7.40099430e-01 -6.11406505e-01 -1.39764650e-02
2.52928108e-01 -1.16812479e+00 -6.02479994e-01 1.01396963e-01
-2.90497303e-01 1.81917951e-01 9.52449143e-01 -8.47499490e-01
1.21168745e+00 -1.67097902e+00 -2.35388488e-01 3.46041262e-01
8.04481506e-02 3.22624967e-02 -2.07085609e-02 4.66035962e-01
5.90915084e-01 4.55004461e-02 -2.89647371e-01 -2.18354315e-01
1.01605959e-01 5.86149752e-01 -1.89951301e-01 2.21867353e-01
-1.19420610e-01 1.03943777e+00 -1.06686795e+00 -4.94246811e-01
7.91081965e-01 3.71399999e-01 -5.39147675e-01 -2.67175257e-01
-2.35743865e-01 5.96441150e-01 -5.06748140e-01 3.55075091e-01
8.68534982e-01 2.22159013e-01 -8.43972564e-02 2.61043310e-01
-7.90320754e-01 5.35533428e-01 -8.87159050e-01 1.78736150e+00
-6.23987973e-01 1.27239716e+00 -3.76187295e-01 -4.44745898e-01
9.43752050e-01 -2.21825719e-01 1.98316768e-01 -1.04773891e+00
-4.72190589e-01 2.81943470e-01 -4.83261317e-01 -3.63481909e-01
1.06666219e+00 3.70979339e-01 -1.69448778e-01 1.37059420e-01
-3.66040885e-01 -4.12494361e-01 -4.49914159e-03 1.14015497e-01
1.24722397e+00 6.95954740e-01 1.17223494e-01 -4.13561821e-01
3.83371711e-01 7.70820498e-01 4.18905377e-01 8.14383388e-01
-2.22959027e-01 5.27818441e-01 1.91880181e-01 -9.16579008e-01
-1.19015205e+00 -1.15869009e+00 -2.74361819e-01 7.73514986e-01
5.29563904e-01 -6.53092563e-01 -5.54766715e-01 -5.70106745e-01
6.42533228e-02 1.00786495e+00 -6.63350344e-01 1.97526097e-01
-7.33140111e-01 -5.99961817e-01 5.63829958e-01 6.32746458e-01
7.89994478e-01 -5.75775802e-01 -8.63120317e-01 3.52624595e-01
-4.30124164e-01 -1.07552028e+00 2.61636615e-01 5.10915071e-02
-6.00427508e-01 -8.55219185e-01 -2.84387589e-01 -5.21640778e-01
3.99345875e-01 4.71548408e-01 9.80343044e-01 -2.49957636e-01
2.61044741e-01 3.08352768e-01 -2.46898070e-01 -4.12468076e-01
-2.91352570e-01 4.32034016e-01 -5.82289249e-02 -5.40947318e-01
3.91886175e-01 -6.34120524e-01 -6.87909901e-01 5.98132670e-01
-1.75026968e-01 6.51717067e-01 4.15666103e-01 1.81435689e-01
5.90026140e-01 -7.26627260e-02 4.38629389e-01 -5.16916692e-01
-9.85053107e-02 -8.33114624e-01 -7.82647967e-01 -2.24818140e-01
-7.58758247e-01 1.67770684e-01 2.25617006e-01 1.95887834e-01
-1.00985110e+00 4.29907113e-01 -4.27612424e-01 1.06746539e-01
-3.39317352e-01 5.99141240e-01 3.15642767e-02 -4.81571287e-01
7.08879471e-01 2.96187311e-01 -1.33298844e-01 -2.65272170e-01
5.90078115e-01 3.46090615e-01 6.29321814e-01 -8.32527727e-02
9.76604462e-01 5.99110842e-01 7.19914511e-02 -8.27702343e-01
-3.24796230e-01 -3.36829960e-01 -9.95589733e-01 -4.76843923e-01
8.53224099e-01 -1.10944176e+00 -3.46435547e-01 3.49799097e-01
-9.42119718e-01 -9.53614235e-01 5.96795790e-02 4.69400525e-01
-9.33365226e-01 6.86959475e-02 -1.22218974e-01 -6.34861469e-01
3.07247728e-01 -9.01846111e-01 8.86633813e-01 2.07870245e-01
-1.61097929e-01 -1.26649821e+00 5.61977744e-01 2.38474593e-01
7.08158016e-01 4.76446450e-01 4.77568805e-01 -1.82907388e-01
-1.29047072e+00 -3.23581368e-01 -1.75621629e-01 -4.77609515e-01
-5.08983806e-02 -2.63107359e-01 -9.74309921e-01 1.93280816e-01
-7.01715052e-01 7.07289949e-02 1.01867533e+00 3.41572404e-01
7.47803569e-01 -2.87813514e-01 -8.91053498e-01 6.46968067e-01
1.37637329e+00 1.00105628e-01 1.01631403e+00 1.20381105e+00
8.74045193e-01 8.09755981e-01 1.22549391e+00 2.10099205e-01
1.52638340e+00 1.00237262e+00 6.07633173e-01 -1.12983204e-01
8.92451126e-03 -6.41952872e-01 2.48575702e-01 4.66394454e-01
-2.47357443e-01 -2.00399950e-01 -1.48951697e+00 8.36346924e-01
-2.33729625e+00 -9.40367281e-01 -7.01901972e-01 2.20487142e+00
3.32633346e-01 1.67876795e-01 6.66653737e-03 -3.15439641e-01
2.38716051e-01 3.11913341e-01 -3.98885101e-01 -1.66642174e-01
-1.34491846e-01 -3.21348846e-01 1.34069133e+00 1.10467517e+00
-1.40981364e+00 1.42096388e+00 6.65964222e+00 7.27490902e-01
-8.38390708e-01 2.09047437e-01 1.42043844e-01 7.16843903e-02
-4.90951389e-01 2.13237137e-01 -1.05847073e+00 4.88064557e-01
1.39067185e+00 -5.15423641e-02 -2.13626903e-02 1.02062392e+00
5.34239948e-01 -7.51167774e-01 -6.16843879e-01 6.46236002e-01
-3.72856796e-01 -1.92879677e+00 -3.34367663e-01 2.96032697e-01
8.12653005e-01 6.69078112e-01 -6.54650033e-02 5.50903559e-01
5.88341415e-01 -9.88927960e-01 7.99026132e-01 8.30099285e-01
3.86614650e-01 -9.53761280e-01 7.95424938e-01 6.96133256e-01
-1.51310587e+00 1.65247962e-01 -3.73835862e-01 -2.96785116e-01
5.45286179e-01 1.18992388e-01 -1.27553248e+00 7.39075065e-01
7.70154238e-01 8.98285747e-01 -8.62409472e-01 1.26659381e+00
-2.49005601e-01 2.98594832e-01 -4.75871742e-01 1.48846641e-01
5.46831012e-01 -3.56692038e-02 4.58719194e-01 1.25217092e+00
4.44690466e-01 -2.14742616e-01 2.51555115e-01 5.40597260e-01
6.15681708e-01 -1.41111746e-01 -1.20035863e+00 6.66636705e-01
6.59688056e-01 9.78507400e-01 -6.38421237e-01 -3.09312999e-01
-1.51825458e-01 6.79312885e-01 3.44161779e-01 3.00024569e-01
-9.73901689e-01 -2.60840505e-01 9.18053687e-01 3.92687172e-01
2.98129708e-01 -8.16682637e-01 -3.05722564e-01 -6.58159316e-01
-1.54385194e-01 3.93918753e-02 -1.49220020e-01 -9.32126880e-01
-3.26553375e-01 4.36841428e-01 2.70478100e-01 -1.57059753e+00
-7.11215854e-01 -4.94978130e-01 -5.05040526e-01 9.11678255e-01
-1.97925282e+00 -1.30299187e+00 -6.94436669e-01 1.28540412e-01
3.80682617e-01 4.26331386e-02 8.69990408e-01 1.95961758e-01
-8.47899616e-02 -3.10407970e-02 3.41185272e-01 -2.28919461e-01
3.62511933e-01 -1.28346050e+00 1.14833534e+00 8.33035886e-01
-1.11943232e-02 2.62104198e-02 9.05617774e-01 -8.74073029e-01
-1.15962577e+00 -1.54522836e+00 9.35724556e-01 -1.08414698e+00
6.13016427e-01 -2.33197421e-01 -9.50420380e-01 9.57822204e-01
-1.15266286e-01 -2.34465450e-01 2.79436857e-01 -1.16360676e-03
9.95439887e-02 -1.51130324e-02 -8.19234014e-01 7.72840261e-01
1.20655191e+00 -3.74057502e-01 -2.79345363e-01 2.69626379e-01
6.26759350e-01 -5.36800027e-01 -6.34983003e-01 5.42265236e-01
6.64579749e-01 -1.22391593e+00 9.66898203e-01 1.11685216e-01
1.54435709e-01 -6.80921555e-01 -2.06414387e-01 -1.36580050e+00
-4.28144068e-01 -1.67958781e-01 3.14807594e-02 6.40072227e-01
8.57574224e-01 -8.60676408e-01 1.25070262e+00 6.85631990e-01
-5.07782936e-01 -4.79627490e-01 -1.01682043e+00 -8.92073929e-01
1.11680202e-01 -9.38660026e-01 5.64552486e-01 7.07386613e-01
-6.21662363e-02 -2.07807824e-01 -4.23430949e-01 5.77813625e-01
4.14674550e-01 -2.78996199e-01 1.40935719e+00 -1.18814218e+00
2.50255644e-01 -2.88688064e-01 -9.24129248e-01 -1.26558161e+00
2.26566195e-01 -7.06910491e-01 4.24180299e-01 -2.23494935e+00
-3.01118195e-01 -9.30179596e-01 1.24219440e-01 5.81759572e-01
1.69771060e-01 3.13960344e-01 -2.43842170e-01 2.07799524e-01
-6.03856802e-01 5.86885750e-01 7.86727250e-01 -2.08685808e-02
-4.82638210e-01 -6.67962208e-02 -3.57627384e-02 7.46699512e-01
1.01118124e+00 -3.30635905e-01 -6.03613734e-01 -3.80674392e-01
3.64766449e-01 -5.51403351e-02 5.26926160e-01 -1.42416334e+00
6.69992387e-01 -3.16297561e-01 1.07486352e-01 -1.35916686e+00
7.39366710e-01 -4.61003035e-01 5.06690860e-01 3.94218773e-01
1.69642687e-01 -3.56499897e-03 6.04763627e-01 7.43942201e-01
-1.32833228e-01 8.02693889e-02 2.75366753e-01 5.37027493e-02
-1.63406670e+00 2.96299219e-01 -9.97311532e-01 -4.44030344e-01
1.14460361e+00 -7.30177939e-01 -6.16219997e-01 -6.36792481e-01
-5.53183198e-01 6.96824789e-01 7.96701670e-01 6.00820959e-01
6.93029761e-01 -1.23988378e+00 -5.69395363e-01 2.29165986e-01
4.06786352e-01 1.20748632e-01 3.91018718e-01 8.92660618e-01
-1.08154941e+00 8.38608861e-01 -4.87489671e-01 -6.07340574e-01
-9.30366874e-01 2.39266247e-01 3.24815661e-01 4.57194727e-03
-9.04670537e-01 5.96902907e-01 8.14647004e-02 -1.07135439e+00
-2.61609674e-01 -3.64361763e-01 -3.20720732e-01 -2.00531110e-01
4.34626341e-01 6.62258089e-01 3.72631811e-02 -9.21762705e-01
-5.43820560e-01 5.79904795e-01 3.73069078e-01 -5.39055824e-01
1.16720307e+00 -5.91355026e-01 3.66757452e-01 6.09794319e-01
8.53583753e-01 -6.15674034e-02 -1.67207038e+00 1.53560162e-01
3.09719682e-01 -6.25167072e-01 2.60997027e-01 -6.58083260e-01
-5.83539665e-01 6.40895367e-01 6.42289817e-01 -3.29763256e-02
4.11664158e-01 -8.61779898e-02 6.24896407e-01 6.86478436e-01
1.07865846e+00 -1.23264432e+00 -4.91215914e-01 8.26724946e-01
6.34373605e-01 -1.44521797e+00 -9.52845961e-02 -4.98286545e-01
-6.56221688e-01 9.09051478e-01 7.61168599e-01 1.08855270e-01
7.61232793e-01 1.15105845e-01 1.30354494e-01 -4.50758226e-02
-7.60516167e-01 -4.36645240e-01 1.09779567e-01 1.17792749e+00
-3.10711414e-01 2.49415562e-01 2.45818958e-01 8.22252780e-02
-7.07271576e-01 -4.11123075e-02 7.27040350e-01 1.13336980e+00
-1.03716052e+00 -9.99069691e-01 -2.81674773e-01 3.01877409e-01
4.72092539e-01 3.09766480e-03 5.31039014e-02 1.04272318e+00
6.75622895e-02 9.08190608e-01 3.08586299e-01 -7.57288456e-01
2.09727302e-01 -7.29750395e-02 1.34630427e-01 -3.09354663e-01
2.99411088e-01 -7.68610418e-01 7.24901617e-01 -1.02883446e+00
-2.02935442e-01 -5.96445203e-01 -1.37059307e+00 -7.23679781e-01
1.91720217e-01 -1.54408403e-02 1.00840390e+00 9.52489197e-01
7.25989521e-01 1.55615747e-01 3.10905993e-01 -1.43546414e+00
5.65347910e-01 -5.85855722e-01 -1.74808562e-01 -4.45873648e-01
4.07967210e-01 -8.94055009e-01 -8.81484225e-02 -1.57777309e-01] | [5.705243110656738, 0.6820877194404602] |
0b41348e-d03f-4532-9719-4623647c6d97 | faster-bounding-box-annotation-for-object | 1807.03142 | null | http://arxiv.org/abs/1807.03142v1 | http://arxiv.org/pdf/1807.03142v1.pdf | Faster Bounding Box Annotation for Object Detection in Indoor Scenes | This paper proposes an approach for rapid bounding box annotation for object
detection datasets. The procedure consists of two stages: The first step is to
annotate a part of the dataset manually, and the second step proposes
annotations for the remaining samples using a model trained with the first
stage annotations. We experimentally study which first/second stage split
minimizes to total workload. In addition, we introduce a new fully labeled
object detection dataset collected from indoor scenes. Compared to other indoor
datasets, our collection has more class categories, different backgrounds,
lighting conditions, occlusion and high intra-class differences. We train deep
learning based object detectors with a number of state-of-the-art models and
compare them in terms of speed and accuracy. The fully annotated dataset is
released freely available for the research community. | ['Jukka Peltomäki', 'Jussi Puura', 'Heikki Huttunen', 'Bishwo Adhikari'] | 2018-07-03 | null | null | null | null | ['object-detection-in-indoor-scenes'] | ['computer-vision'] | [ 2.48811170e-01 -4.43828963e-02 3.37544046e-02 -6.61323726e-01
-5.24522007e-01 -6.12558603e-01 1.97419569e-01 2.77921855e-01
-7.43243515e-01 6.64051473e-01 -2.50877798e-01 -1.10116780e-01
4.63022202e-01 -6.03990734e-01 -6.86591446e-01 -4.04971749e-01
3.42153125e-02 6.36461079e-01 1.04932666e+00 3.50733221e-01
1.20533735e-01 7.73445725e-01 -1.53036833e+00 5.38997710e-01
3.24544698e-01 1.18754423e+00 3.82733434e-01 9.98627603e-01
-2.30845243e-01 8.28124642e-01 -6.76801264e-01 -3.36741239e-01
5.23328841e-01 -6.52013123e-02 -8.75247598e-01 2.48236984e-01
8.30326736e-01 -5.26469767e-01 -7.47558400e-02 9.48884189e-01
5.56553364e-01 6.35874942e-02 5.10289431e-01 -1.22414148e+00
-1.47984013e-01 2.71014482e-01 -5.10891914e-01 4.34564620e-01
1.42173678e-01 5.22192530e-02 6.50889516e-01 -1.17336869e+00
6.15421712e-01 1.11700130e+00 7.78607070e-01 6.60141468e-01
-1.24830127e+00 -7.41242886e-01 3.70472968e-01 -3.33245099e-02
-1.46309233e+00 -3.45519483e-01 5.23985207e-01 -7.59213269e-01
7.81956613e-01 1.89870670e-01 5.31506419e-01 9.28854346e-01
-1.64546356e-01 8.00552666e-01 1.00307155e+00 -4.21169519e-01
3.73815566e-01 4.93168116e-01 4.90118980e-01 6.92016661e-01
3.74627650e-01 5.10508567e-02 4.05288637e-02 -9.83681381e-02
4.35532570e-01 1.40012324e-01 7.44698942e-02 -7.28805244e-01
-9.75064993e-01 6.02564156e-01 5.28394461e-01 -2.97208522e-02
1.71897374e-02 1.70149669e-01 5.40676355e-01 -2.66784638e-01
3.95147115e-01 1.30030438e-01 -6.87092543e-01 2.58398414e-01
-9.16227102e-01 1.67010278e-01 8.04373264e-01 1.37351859e+00
8.44792426e-01 -4.84353185e-01 -2.87396818e-01 6.90549850e-01
4.77533013e-01 2.36827344e-01 8.45519826e-03 -7.10993588e-01
4.39267039e-01 6.58646226e-01 3.84663075e-01 -6.20321810e-01
-5.70484459e-01 -2.24065393e-01 -3.83453369e-01 2.01247558e-01
5.59687853e-01 -5.78345172e-02 -1.09410334e+00 1.10573912e+00
6.22310996e-01 -7.83172473e-02 -4.19842333e-01 8.72391522e-01
9.98746455e-01 3.19077939e-01 2.07760006e-01 1.24601208e-01
1.44248974e+00 -1.31718194e+00 -5.80369592e-01 -5.58559060e-01
4.61087108e-01 -1.01270366e+00 9.18190479e-01 2.80663073e-01
-7.65706599e-01 -8.95843327e-01 -1.00723076e+00 -2.61645198e-01
-6.11683667e-01 8.20076108e-01 5.13245225e-01 8.99897873e-01
-7.43205786e-01 2.12898180e-01 -9.73948419e-01 -6.23131752e-01
8.59408200e-01 3.16414416e-01 -2.20232364e-02 -2.41137780e-02
-4.47875410e-01 6.41372561e-01 6.60019219e-01 2.84802765e-02
-1.07281518e+00 -4.30143803e-01 -6.86131537e-01 -2.32259393e-01
4.58511591e-01 -2.22801134e-01 1.50439584e+00 -6.52166128e-01
-1.07145309e+00 1.03860605e+00 -1.68961510e-01 -3.39596570e-01
9.76185858e-01 -4.99790639e-01 5.21260165e-02 -2.80070066e-01
1.07892305e-01 8.82153571e-01 5.75874090e-01 -1.48211205e+00
-1.03169096e+00 -2.45572150e-01 1.70346230e-01 -2.36651123e-01
-6.72236681e-02 4.11580831e-01 -6.75764263e-01 -4.37929600e-01
-6.32376820e-02 -8.94329906e-01 -4.03063595e-01 5.61404049e-01
-4.46284801e-01 -8.59617069e-02 1.28681695e+00 -3.96270812e-01
1.09059882e+00 -2.19463110e+00 -4.60900605e-01 -7.87025839e-02
1.41132638e-01 3.42034161e-01 1.18850976e-01 8.17415044e-02
1.50808617e-01 4.44641486e-02 -1.05649225e-01 -8.65614712e-01
-1.01806305e-01 2.18543455e-01 -3.55631202e-01 5.21393418e-01
2.30107054e-01 4.39976573e-01 -9.14503634e-01 -7.39811957e-01
4.73075181e-01 2.98181713e-01 -4.34207469e-01 3.61681551e-01
-2.93816686e-01 4.06311125e-01 -2.25341439e-01 9.92302835e-01
8.80772293e-01 -9.17407721e-02 -5.42744882e-02 -3.58574331e-01
-2.51363963e-01 1.85205042e-01 -1.58725941e+00 1.52388275e+00
-1.47358686e-01 9.75038528e-01 1.11990467e-01 -5.56535184e-01
9.26730394e-01 -7.56450221e-02 1.81686863e-01 -2.04331934e-01
4.33722883e-01 1.50204256e-01 -1.11105487e-01 -4.60103571e-01
6.18451357e-01 4.53267157e-01 2.02210426e-01 1.99951246e-01
9.26290005e-02 -7.92651623e-02 4.82936084e-01 8.04407075e-02
1.02134109e+00 3.59515578e-01 3.58794391e-01 -1.94271177e-01
4.21706766e-01 1.50331736e-01 4.02702898e-01 9.11925197e-01
-7.27474689e-01 6.19602680e-01 2.33666286e-01 -9.63368833e-01
-1.25759673e+00 -7.56261230e-01 -4.45815563e-01 1.20873392e+00
3.68559897e-01 -3.35786402e-01 -8.19126189e-01 -9.90189195e-01
-1.06346700e-02 4.22859579e-01 -7.43761778e-01 3.33388329e-01
-4.80585128e-01 -7.48423457e-01 4.46777403e-01 1.00370157e+00
7.78675020e-01 -9.39245224e-01 -1.08351517e+00 -8.84026438e-02
7.74490610e-02 -1.35390341e+00 -2.19605222e-01 4.68500018e-01
-6.83918893e-01 -1.26558220e+00 -2.58812517e-01 -6.76857889e-01
8.68120432e-01 3.44614863e-01 1.20127404e+00 1.81638628e-01
-8.59580219e-01 1.70024902e-01 -4.57359880e-01 -9.03199017e-01
-1.58208802e-01 2.76802648e-02 -1.23886190e-01 -1.41606331e-01
4.98738915e-01 2.01387599e-01 -6.62792146e-01 6.03611112e-01
-5.77149093e-01 1.82983428e-02 4.57492262e-01 3.63047212e-01
8.92044187e-01 -1.05498374e-01 -1.58416629e-01 -8.82856846e-01
-1.05242208e-01 -2.24489868e-01 -1.13003767e+00 1.74702018e-01
-9.79536995e-02 -3.07178169e-01 2.44749606e-01 -4.63662893e-01
-1.01210642e+00 7.67529428e-01 -4.33748327e-02 -2.56190389e-01
-6.97023332e-01 -3.92960310e-01 -1.89202026e-01 -2.26169005e-01
8.04811120e-01 -3.00695360e-01 -7.22547412e-01 -7.32487619e-01
2.53306180e-01 6.53943777e-01 4.28131670e-01 -5.33622563e-01
7.96592832e-01 6.67453647e-01 -2.50183165e-01 -6.81639075e-01
-1.14833248e+00 -8.96264970e-01 -1.31043541e+00 -2.58842409e-01
1.00313950e+00 -8.37767422e-01 -5.49497366e-01 4.72710520e-01
-1.41764450e+00 -5.98429501e-01 -3.63073558e-01 3.62118185e-01
-1.62479341e-01 1.17254788e-02 -5.01366198e-01 -1.09295940e+00
-8.69665742e-02 -1.12149918e+00 1.38724220e+00 2.86733806e-01
-3.52812856e-02 -6.18201077e-01 4.10891883e-02 3.25034142e-01
2.02924117e-01 3.19250673e-01 2.02839866e-01 -9.16112304e-01
-8.31610739e-01 -3.71342659e-01 -6.93104267e-01 3.28423917e-01
-3.83007303e-02 2.35650077e-01 -1.30394769e+00 -3.94926131e-01
-3.97977710e-01 -4.86097783e-01 9.60556805e-01 1.77075610e-01
1.49197066e+00 4.56860987e-03 -7.85375118e-01 5.60774565e-01
1.45348537e+00 2.97612697e-01 3.99642766e-01 4.44210768e-01
7.52051413e-01 4.22991693e-01 9.21079695e-01 3.79225671e-01
1.66537225e-01 6.68766618e-01 5.24068058e-01 -2.70737529e-01
-2.81866252e-01 7.62985200e-02 1.31690018e-02 6.17018938e-02
-3.71662267e-02 -2.77566731e-01 -1.27715683e+00 4.92323577e-01
-1.76347482e+00 -7.46137381e-01 -3.52197826e-01 2.09233308e+00
5.93004644e-01 3.72039974e-01 3.88100624e-01 1.36164829e-01
7.31009901e-01 -4.76251543e-02 -3.85210127e-01 -1.38893440e-01
2.87619531e-01 1.64874699e-02 7.27242470e-01 3.15984458e-01
-1.76465464e+00 8.84531140e-01 6.92664671e+00 4.95813489e-01
-8.37072551e-01 2.66057014e-01 6.53388083e-01 -5.66057451e-02
6.26364291e-01 -6.93591610e-02 -1.38338304e+00 4.41300929e-01
5.86589754e-01 5.35402060e-01 -4.18710299e-02 1.55129099e+00
-1.49106398e-01 -3.94131780e-01 -1.27645028e+00 6.16951287e-01
4.98704500e-02 -1.05981636e+00 -4.36163604e-01 -7.79538676e-02
7.86984980e-01 3.10766339e-01 -4.82807308e-01 2.95981407e-01
4.31778133e-01 -6.06708646e-01 1.11347961e+00 2.79693872e-01
6.68182373e-01 -4.60064769e-01 1.00172722e+00 2.63796419e-01
-1.58654249e+00 -2.02770233e-01 -5.60952723e-01 -1.77981883e-01
-1.54618233e-01 5.50078571e-01 -1.02735341e+00 6.70818165e-02
1.21116865e+00 2.77012199e-01 -1.01095140e+00 1.56638074e+00
-3.18012685e-01 4.50665921e-01 -3.08704823e-01 -1.68929771e-01
-1.36118412e-01 1.20275557e-01 2.04019085e-01 1.73957384e+00
-1.14707775e-01 -4.35896628e-02 8.39262247e-01 9.26357031e-01
-7.03893825e-02 -4.17266749e-02 -3.25550616e-01 3.57679427e-01
5.95123410e-01 1.58057106e+00 -1.38762498e+00 -5.31326115e-01
-4.52629805e-01 8.74500334e-01 4.07795429e-01 1.27586171e-01
-1.10510850e+00 -4.00408596e-01 3.84605169e-01 2.24004388e-01
3.66277963e-01 -2.83544898e-01 -3.55212986e-01 -8.47316623e-01
-4.16419515e-03 -2.86202043e-01 3.40818673e-01 -7.06647456e-01
-1.06444383e+00 5.37693739e-01 2.45290995e-01 -9.49590087e-01
4.13190812e-01 -9.40506756e-01 -4.24316406e-01 4.63258684e-01
-1.37978506e+00 -1.21858883e+00 -1.06535006e+00 2.64456123e-01
9.34986532e-01 2.79024631e-01 7.14430034e-01 6.37466371e-01
-9.44600940e-01 4.10194874e-01 -1.50429621e-01 7.24702060e-01
6.83518767e-01 -1.34990335e+00 5.22117674e-01 1.03574669e+00
3.56141701e-02 2.66777098e-01 4.55574989e-01 -6.07291520e-01
-7.65368581e-01 -1.56532907e+00 5.89968264e-01 -7.89464176e-01
3.81738812e-01 -8.74821424e-01 -6.53366864e-01 9.07205045e-01
-1.24559984e-01 7.07120657e-01 6.01388872e-01 -9.88516584e-02
-2.75287092e-01 -1.43297121e-01 -1.22897279e+00 7.82744586e-02
1.13521469e+00 -1.87298045e-01 -2.98360050e-01 6.30279303e-01
6.51079237e-01 -9.37900007e-01 -4.58257854e-01 5.64408302e-01
5.58784127e-01 -9.70773816e-01 9.37180161e-01 -3.46916229e-01
1.21503323e-01 -6.47460818e-01 -3.52871776e-01 -6.28177702e-01
-2.26436377e-01 -2.04557464e-01 -2.39387944e-01 1.38015902e+00
2.89650142e-01 -2.36220375e-01 8.68675351e-01 6.14704907e-01
-1.90774009e-01 -6.39037371e-01 -5.65120578e-01 -8.72822642e-01
-4.43316877e-01 -5.01175523e-01 2.99910009e-01 5.31845391e-01
-7.17583001e-01 9.61975101e-03 -5.29660210e-02 2.66970485e-01
6.10177875e-01 9.21432897e-02 1.21967447e+00 -1.29329073e+00
-3.37894373e-02 -2.50433803e-01 -4.96839404e-01 -8.32454622e-01
-3.14355433e-01 -5.15448272e-01 4.57200795e-01 -1.59006929e+00
5.04868805e-01 -7.27850616e-01 -8.17990229e-02 7.20363140e-01
-2.24009171e-01 5.83829641e-01 1.25956357e-01 2.06247523e-01
-9.95499313e-01 2.15182733e-02 6.24560416e-01 -7.93068931e-02
-1.74787134e-01 1.85414732e-01 -9.29948315e-03 1.02959347e+00
8.28828454e-01 -7.33232498e-01 -8.91887024e-02 -3.13822597e-01
-2.21944332e-01 -7.15065181e-01 6.02303624e-01 -1.54065549e+00
1.21846765e-01 -1.52517334e-01 8.49723339e-01 -1.03304827e+00
2.34550714e-01 -1.16333747e+00 -2.14536473e-01 6.02069139e-01
-8.29220191e-02 -1.95926622e-01 5.61610222e-01 5.05282521e-01
9.83308479e-02 -2.93187231e-01 1.07081139e+00 -1.83880329e-01
-8.67005467e-01 2.37947151e-01 -1.66117787e-01 -6.95202649e-02
1.46431196e+00 -4.81558830e-01 -3.46018434e-01 3.34459543e-01
-6.15249276e-01 1.01570010e-01 5.31330466e-01 4.05420244e-01
3.48195702e-01 -1.08898246e+00 -4.63790566e-01 2.53432781e-01
2.86416769e-01 4.30425435e-01 -2.15960875e-01 5.68981826e-01
-8.43986332e-01 3.09809774e-01 -1.28883019e-01 -9.11252737e-01
-1.47193038e+00 7.86434054e-01 3.65284503e-01 -1.79754674e-01
-3.14639807e-01 1.22373605e+00 2.10803702e-01 -4.53731388e-01
6.65917754e-01 -5.87894320e-01 7.73092210e-02 1.49664328e-01
6.52908444e-01 5.91558993e-01 2.03456253e-01 -5.67206860e-01
-7.46467590e-01 5.20850062e-01 -1.60651073e-01 2.80362636e-01
9.46781576e-01 9.82367545e-02 9.73173752e-02 4.61738586e-01
1.02287674e+00 -5.15837893e-02 -1.39462972e+00 -1.20410994e-01
1.01638071e-01 -8.19340467e-01 -8.00559968e-02 -7.39953578e-01
-8.46631408e-01 8.63232553e-01 9.67217743e-01 4.39877287e-02
9.96272981e-01 1.46870717e-01 4.27942634e-01 4.85450655e-01
4.36989069e-01 -1.23394251e+00 2.78604150e-01 5.74822605e-01
7.11288333e-01 -1.53162122e+00 2.79289573e-01 -7.20274508e-01
-2.41315603e-01 1.07976258e+00 1.10362077e+00 -1.58253148e-01
5.59231460e-01 7.37221539e-01 1.00380123e-01 -1.07984111e-01
-2.92505771e-01 -3.52550685e-01 3.38964492e-01 6.12278044e-01
5.09596705e-01 -4.97685522e-02 -7.48091936e-02 3.08321029e-01
1.19781867e-02 -9.21691135e-02 1.68612853e-01 1.15503871e+00
-6.98182404e-01 -9.09951925e-01 -6.07213199e-01 2.47322902e-01
-3.94889951e-01 3.13211977e-01 -7.24203169e-01 9.58428025e-01
6.36321485e-01 9.31962132e-01 2.24860489e-01 -2.51065373e-01
4.23437953e-01 -1.04195729e-01 3.55408400e-01 -8.87836516e-01
-5.21613300e-01 1.67919137e-02 6.21361434e-02 -7.22404301e-01
-4.55832362e-01 -5.95376790e-01 -1.25218141e+00 6.74979985e-02
-7.76639223e-01 -2.50279546e-01 9.08290625e-01 6.45534694e-01
1.54672712e-01 7.47013330e-01 2.30522782e-01 -1.36888278e+00
-1.10552296e-01 -1.01163256e+00 -2.75815520e-02 3.56277078e-01
2.85527557e-01 -7.57750690e-01 -1.47834614e-01 3.14541399e-01] | [9.047628402709961, 0.41503024101257324] |
69aae883-82a6-47d8-9ed4-4793dd2d299c | predicting-solar-flares-using-a-long-short | 1905.07095 | null | https://arxiv.org/abs/1905.07095v1 | https://arxiv.org/pdf/1905.07095v1.pdf | Predicting Solar Flares Using a Long Short-Term Memory Network | We present a long short-term memory (LSTM) network for predicting whether an active region (AR) would produce a gamma-class flare within the next 24 hours. We consider three gamma classes, namely >=M5.0 class, >=M class, and >=C class, and build three LSTM models separately, each corresponding to a gamma class. Each LSTM model is used to make predictions of its corresponding gamma-class flares. The essence of our approach is to model data samples in an AR as time series and use LSTMs to capture temporal information of the data samples. Each data sample has 40 features including 25 magnetic parameters obtained from the Space-weather HMI Active Region Patches (SHARP) and related data products as well as 15 flare history parameters. We survey the flare events that occurred from 2010 May to 2018 May, using the GOES X-ray flare catalogs provided by the National Centers for Environmental Information (NCEI), and select flares with identified ARs in the NCEI flare catalogs. These flare events are used to build the labels (positive vs. negative) of the data samples. Experimental results show that (i) using only 14-22 most important features including both flare history and magnetic parameters can achieve better performance than using all the 40 features together; (ii) our LSTM network outperforms related machine learning methods in predicting the labels of the data samples. To our knowledge, this is the first time that LSTMs have been used for solar flare prediction. | ['Hao Liu', 'Chang Liu', 'Jason T. L. Wang', 'Haimin Wang'] | 2019-05-17 | null | null | null | null | ['solar-flare-prediction'] | ['time-series'] | [ 1.75513074e-01 -5.44278145e-01 -3.97658885e-01 -4.91431564e-01
-8.99874389e-01 -5.63702881e-01 5.68966269e-01 2.27513593e-02
-7.32781217e-02 8.63081813e-01 1.06081307e-01 -3.40119809e-01
-2.88955152e-01 -1.17039120e+00 -7.44517565e-01 -7.50133872e-01
-3.94596517e-01 4.30272430e-01 -1.19451033e-02 -3.31740677e-01
2.00097457e-01 8.92170846e-01 -1.60778928e+00 3.92384142e-01
9.62015986e-01 1.71228969e+00 2.14305133e-01 5.93861639e-01
-3.77021879e-02 1.15470803e+00 -8.51693988e-01 4.66894627e-01
-9.77422819e-02 -4.70357358e-01 -4.56718266e-01 -4.71631497e-01
4.86208826e-01 -5.31205274e-02 -2.66269714e-01 9.08028483e-01
3.85895252e-01 3.66264641e-01 6.92225695e-01 -1.11110103e+00
-3.79161388e-01 5.26626229e-01 -5.29092968e-01 8.00451100e-01
4.28757742e-02 -1.36083260e-01 6.67972207e-01 -5.39787650e-01
3.46555978e-01 8.34725320e-01 9.56553400e-01 2.40552537e-02
-7.10604608e-01 -1.07906234e+00 -2.09406197e-01 5.90299726e-01
-1.19827831e+00 -2.26983026e-01 7.50834167e-01 -8.73894930e-01
1.54889750e+00 6.49330616e-02 8.08388591e-01 9.62700725e-01
9.81000602e-01 4.84263539e-01 1.09213078e+00 -1.30221188e-01
2.95697212e-01 -5.81525505e-01 7.26941347e-01 4.26192909e-01
-1.13903493e-01 7.02753723e-01 -4.55752820e-01 -3.62318903e-01
4.25098151e-01 1.25700235e-01 -3.23402256e-01 4.54379201e-01
-9.73992288e-01 1.09242916e+00 9.63275194e-01 6.94485426e-01
-6.23196602e-01 1.52809456e-01 9.33569297e-02 4.47756797e-01
5.26622951e-01 1.73452139e-01 -6.11032367e-01 1.37393549e-01
-9.67504799e-01 3.11704397e-01 3.65272462e-01 3.12331229e-01
1.02473140e+00 5.64737439e-01 -1.12097047e-01 7.06007063e-01
9.25287157e-02 1.36758173e+00 4.60986346e-01 -4.59433585e-01
2.02372402e-01 5.06052732e-01 2.22631112e-01 -9.75089610e-01
-8.44129562e-01 -4.54195231e-01 -9.51635599e-01 9.36617032e-02
-1.22155339e-01 -4.72009867e-01 -1.38783026e+00 1.66079843e+00
-3.05450827e-01 1.66872472e-01 2.81434536e-01 6.18363798e-01
9.27452981e-01 1.05932713e+00 -9.13034603e-02 -4.21071470e-01
9.58542168e-01 -4.60593641e-01 -9.45287049e-01 -5.34783602e-01
6.04599833e-01 -2.85414129e-01 4.44304407e-01 3.04207683e-01
-5.93499362e-01 -5.70055664e-01 -1.17997539e+00 4.37944531e-01
-6.62847161e-01 8.31222683e-02 7.87696004e-01 -2.34354064e-02
-6.72442317e-01 4.97738451e-01 -9.96481478e-01 -3.67803186e-01
6.69061467e-02 1.48231328e-01 1.74197450e-01 4.59139079e-01
-1.59786081e+00 8.08613896e-01 3.51290196e-01 4.43397045e-01
-9.05082822e-01 -5.04958987e-01 -7.78947830e-01 -2.25644093e-02
-1.44003242e-01 -4.52718347e-01 1.13775039e+00 -1.23754215e+00
-8.67632806e-01 4.68046933e-01 -3.04576188e-01 -8.85282338e-01
-3.76936197e-01 7.92984366e-02 -1.24253607e+00 -8.32841098e-02
-4.58685495e-03 4.88824338e-01 8.12695503e-01 -9.92886603e-01
-1.09492731e+00 -4.39608961e-01 -2.28280559e-01 -2.34140933e-01
1.98763043e-01 1.94149137e-01 3.75078022e-01 -5.03423035e-01
2.18861088e-01 -1.02076089e+00 1.56071559e-01 -5.70826769e-01
-2.02339321e-01 -3.74190986e-01 8.55524898e-01 -7.86070943e-01
1.31843841e+00 -2.04341459e+00 -2.48453662e-01 1.65716574e-01
-2.02755168e-01 1.67356610e-01 1.76505476e-01 2.08931550e-01
-3.05431098e-01 -1.23006314e-01 -4.00557429e-01 3.76167297e-01
-1.69977471e-01 3.96954507e-01 -9.92378116e-01 2.61354595e-01
1.06790271e-02 8.23239088e-01 -4.89003986e-01 2.77500212e-01
2.29023144e-01 4.17259932e-01 1.02272682e-01 3.18032801e-02
-5.58041632e-01 3.97254169e-01 -4.77296650e-01 7.85455644e-01
8.21597874e-01 9.23035741e-02 -2.42208764e-01 -1.93929106e-01
-4.84880179e-01 3.84876393e-02 -5.58562577e-01 1.12415755e+00
-3.58193249e-01 6.52506411e-01 -3.63953829e-01 -6.82684779e-01
1.37652218e+00 2.21953958e-01 1.62945315e-01 -1.20723772e+00
1.99448109e-01 4.26664293e-01 -6.21852390e-02 -5.14492512e-01
3.39122266e-01 -4.44575369e-01 -2.72834957e-01 3.34440976e-01
8.95915180e-03 7.39685968e-02 1.61443874e-01 -3.36929172e-01
1.14175498e+00 -1.67769223e-01 4.91566174e-02 -3.62556905e-01
2.03703016e-01 2.33759880e-02 9.32946444e-01 8.01179230e-01
2.52467860e-02 2.77844131e-01 1.36777699e-01 -8.35759223e-01
-3.90903294e-01 -9.53108609e-01 -4.55652624e-01 9.96439159e-01
-2.77169228e-01 1.82208475e-02 -3.02010477e-02 -3.82143319e-01
-2.26458944e-02 1.03296888e+00 -9.22050893e-01 -2.39818737e-01
-5.79127192e-01 -1.17501450e+00 5.21753192e-01 8.77383947e-01
6.89702988e-01 -1.64177370e+00 -8.28547657e-01 8.66242275e-02
-3.70843709e-01 -5.11730850e-01 3.48129198e-02 7.49333084e-01
-7.42425025e-01 -1.03821528e+00 -2.45776381e-02 -8.01472306e-01
8.68805721e-02 -8.06241855e-02 1.11320221e+00 -3.61002594e-01
-2.29926348e-01 -1.27583802e-01 -4.30616945e-01 -5.31006694e-01
8.17351043e-03 -8.30725208e-02 -4.69211638e-02 -9.20901299e-02
6.25539243e-01 -6.40898168e-01 -2.98050940e-01 4.49646600e-02
-7.86734998e-01 -3.23173612e-01 2.81504333e-01 5.52093506e-01
8.51143897e-01 2.89229244e-01 9.82859850e-01 -4.12379146e-01
8.60602334e-02 -6.10623896e-01 -6.57049000e-01 1.70359790e-01
-2.63510972e-01 -1.57435268e-01 6.83684468e-01 -2.55490661e-01
-1.31584978e+00 -1.01303831e-01 -2.40772456e-01 -3.00553411e-01
-2.43355855e-01 9.20223951e-01 1.34223193e-01 -7.23011568e-02
7.19056606e-01 -8.52565542e-02 -8.86039853e-01 -4.61830765e-01
-1.53838366e-01 7.35681474e-01 8.12156618e-01 -7.20220581e-02
3.46195489e-01 4.45942581e-01 -1.56058565e-01 -8.47824395e-01
-1.37363517e+00 -1.51208624e-01 -4.05187070e-01 -3.10085654e-01
9.99141395e-01 -9.10899103e-01 -2.41411895e-01 9.46451366e-01
-7.68862844e-01 -3.91210437e-01 -6.37396052e-02 7.37304568e-01
-4.58159924e-01 -3.74378085e-01 -5.11857450e-01 -8.49518597e-01
-7.35738814e-01 -5.59310019e-01 9.29544151e-01 5.33233345e-01
1.45426765e-01 -1.00078309e+00 4.75843996e-01 -2.49880347e-02
6.36891961e-01 6.48772895e-01 1.19491649e+00 -3.81950438e-01
-2.34131724e-01 -1.58920899e-01 -7.77041167e-03 1.58978879e-01
1.23538159e-01 1.37298897e-01 -1.03948128e+00 -4.68693912e-01
5.41724563e-01 -4.70251709e-01 1.44945931e+00 8.70559216e-01
9.04882371e-01 -6.96768165e-02 -5.39764285e-01 8.40265214e-01
1.50579035e+00 1.05818224e+00 7.81910360e-01 2.97827095e-01
3.95652086e-01 3.31210569e-02 8.40197265e-01 4.38072145e-01
1.16440721e-01 3.44244093e-01 4.12439167e-01 -1.69956416e-01
1.33978546e-01 -7.04279318e-02 6.05947495e-01 5.04194736e-01
4.34887141e-01 -4.82471406e-01 -1.07001734e+00 7.05816388e-01
-1.71900833e+00 -1.19512975e+00 -3.66324693e-01 2.27437806e+00
4.03998703e-01 -1.82538349e-02 -3.74767482e-01 1.20455667e-01
5.98186970e-01 6.00483656e-01 -7.21377671e-01 -1.91242576e-01
-4.44464386e-01 5.89451551e-01 5.25374055e-01 5.56068361e-01
-1.26816142e+00 7.75156677e-01 6.39551878e+00 5.46688259e-01
-1.69536734e+00 -5.15605956e-02 2.59783655e-01 -2.15858027e-01
-3.10848773e-01 1.04021780e-01 -9.58477676e-01 4.44177508e-01
1.32208920e+00 -8.24063271e-02 4.00305390e-01 2.33510718e-01
1.92796201e-01 -4.37674999e-01 -6.31712496e-01 8.22419345e-01
3.66193414e-01 -1.09690261e+00 -1.40067279e-01 -3.19895715e-01
8.39415133e-01 7.56822944e-01 -1.32710472e-01 6.22135639e-01
2.23800465e-01 -1.09875607e+00 4.40053105e-01 1.33926177e+00
5.76188445e-01 -1.06442988e+00 7.49684453e-01 3.35488766e-01
-1.19903576e+00 -3.13546032e-01 -4.35736358e-01 -5.96296340e-02
3.84540975e-01 9.90418077e-01 -3.23557347e-01 6.27901554e-01
1.18942380e+00 9.05959606e-01 -7.22737432e-01 8.46162975e-01
3.91773172e-02 9.90601718e-01 -6.13511860e-01 3.78633887e-01
2.61642247e-01 -3.77549499e-01 4.82224047e-01 7.27655649e-01
8.13117445e-01 -3.38647217e-02 2.77308196e-01 7.21971691e-01
1.59185350e-01 -3.93069655e-01 -5.70094347e-01 -2.58268237e-01
3.24729681e-01 1.24091494e+00 -3.39015633e-01 -2.88383037e-01
-9.17751193e-02 3.56991589e-01 -9.69956145e-02 5.19611597e-01
-9.98305798e-01 -6.89450085e-01 2.27859899e-01 -6.94864914e-02
4.20786738e-01 5.97637333e-02 -2.37488355e-02 -8.97177637e-01
-2.10049376e-01 -4.61831808e-01 7.30887055e-01 -1.40315580e+00
-1.30794191e+00 9.88975465e-01 -6.31708503e-02 -1.01218665e+00
-6.44763231e-01 -2.58946806e-01 -1.00537395e+00 8.67433488e-01
-1.63710189e+00 -1.32907724e+00 -4.81598377e-01 5.11179328e-01
1.01515189e-01 -2.32423887e-01 8.60274613e-01 2.47917488e-01
-4.13116455e-01 -4.29438241e-02 4.35917497e-01 -1.64684847e-01
6.64252818e-01 -1.03813219e+00 3.46085638e-01 6.58967674e-01
-1.69819981e-01 4.00467105e-02 6.72797143e-01 -9.47947979e-01
-1.00796628e+00 -1.58306444e+00 1.31345928e+00 -1.53343156e-01
9.79257345e-01 -8.93488452e-02 -1.04490042e+00 1.25857759e+00
1.23786122e-01 -1.31620705e-01 4.13789749e-01 1.09876823e-02
1.96782932e-01 -2.84656495e-01 -1.09689498e+00 -1.18411019e-01
6.95724785e-01 -5.97929180e-01 -7.45635152e-01 5.96448004e-01
4.35291588e-01 -2.56789833e-01 -1.02601326e+00 1.28713822e+00
1.55011296e-01 -1.09893262e+00 4.69007105e-01 -1.94318384e-01
-5.31133153e-02 -4.57811981e-01 -5.01947761e-01 -1.47277808e+00
-4.81425732e-01 -7.28308931e-02 -1.96942747e-01 1.10980761e+00
3.37361693e-01 -7.58876622e-01 2.94314861e-01 -3.40331465e-01
-4.10084426e-01 -5.11550903e-01 -9.35046613e-01 -6.39817595e-01
3.20114404e-01 -5.86525261e-01 7.12648749e-01 8.51277649e-01
-5.42995691e-01 3.44031274e-01 -2.23487139e-01 5.01817286e-01
1.70906529e-01 8.69815648e-01 1.66961983e-01 -1.56135643e+00
-1.60844103e-02 -1.50098084e-02 -1.28119782e-01 -5.58270097e-01
3.80870134e-01 -9.85518336e-01 2.64617149e-02 -1.63144290e+00
1.91649031e-02 -4.94796664e-01 -5.05163312e-01 1.02817476e+00
3.15163851e-01 2.22613022e-01 -2.67411381e-01 1.45790577e-01
-7.27583319e-02 7.60602951e-01 3.98965329e-01 -2.13416025e-01
-3.57903004e-01 8.41900334e-02 -1.76690832e-01 8.22062194e-01
9.97641683e-01 -3.86404276e-01 -1.31753489e-01 -4.52243835e-01
4.52116281e-01 5.73749542e-01 3.18659365e-01 -1.70644486e+00
6.78351149e-02 -3.10903013e-01 6.31443024e-01 -1.38269269e+00
3.02193403e-01 -3.83694828e-01 5.24491727e-01 3.84476423e-01
-3.91051965e-03 -6.67526647e-02 4.88305479e-01 2.69458830e-01
-4.49980736e-01 -2.23264545e-01 9.52014863e-01 2.63408542e-01
-6.75708830e-01 3.86382490e-01 -6.92273319e-01 -2.72596359e-01
7.16720402e-01 2.62089193e-01 -6.78842247e-01 -3.19220454e-01
-8.96750808e-01 3.66683543e-01 2.93260515e-01 7.69587457e-01
2.59560674e-01 -1.36409557e+00 -5.20535052e-01 7.37380683e-01
2.36219153e-01 -4.50323731e-01 7.28115439e-01 6.67703032e-01
-1.93227783e-01 6.21268332e-01 -3.00707787e-01 -6.30594432e-01
-9.59210098e-01 3.44579488e-01 8.65192413e-01 -1.94417506e-01
-5.92007399e-01 4.26117659e-01 2.78970480e-01 -4.70048010e-01
2.25688308e-03 -6.26913369e-01 -6.67845964e-01 5.39495409e-01
7.08592713e-01 1.45349398e-01 5.47626436e-01 -8.01906943e-01
-2.65490413e-01 6.45867884e-01 1.87587619e-01 1.81657732e-01
1.42505074e+00 2.54320294e-01 -2.74182975e-01 1.05420816e+00
8.79125059e-01 -3.78505349e-01 -8.27474892e-01 -1.97158143e-01
-1.68453813e-01 -2.54030563e-02 6.07553661e-01 -1.30631649e+00
-1.43092680e+00 8.70190144e-01 1.27915907e+00 6.06234223e-02
1.38094234e+00 -6.45189732e-02 1.16288507e+00 5.94354093e-01
3.73460829e-01 -9.22144294e-01 -4.72389638e-01 1.12898886e+00
1.03100109e+00 -6.65160954e-01 -5.38217545e-01 4.83481169e-01
-5.61517656e-01 9.63207901e-01 4.79980379e-01 -2.36060470e-01
7.99755812e-01 2.86426723e-01 1.11597784e-01 -6.03034437e-01
-1.00549793e+00 -3.31437558e-01 6.42204732e-02 3.41260344e-01
7.45792091e-02 2.31044218e-01 -1.23314805e-01 6.78637743e-01
-3.48159373e-01 1.79134086e-01 4.07892376e-01 9.56604302e-01
-1.00479281e+00 -6.67116582e-01 -6.17574155e-01 1.09311652e+00
-1.01679340e-01 6.55299798e-02 -3.53632718e-01 3.85148436e-01
3.29645753e-01 1.04772413e+00 5.96600056e-01 -4.23746079e-01
1.79733098e-01 2.87714154e-01 2.96907067e-01 -1.30780444e-01
-6.78612351e-01 1.07011106e-03 3.23077440e-02 -2.16053322e-01
-3.99439394e-01 -5.68506598e-01 -1.64863038e+00 -1.94142565e-01
-2.58923739e-01 2.20084801e-01 5.03742516e-01 1.06821024e+00
2.80178487e-01 5.92105150e-01 8.98333967e-01 -9.19897377e-01
-6.21980987e-02 -1.28987515e+00 -1.14968514e+00 5.61790876e-02
4.90507126e-01 -9.21736538e-01 -8.01723301e-01 -3.53430301e-01] | [6.602529525756836, 2.7492852210998535] |
2add17f4-7ca3-46ba-a8c9-abbb7f7b1c36 | nudgeseg-zero-shot-object-segmentation-by | 2109.13859 | null | https://arxiv.org/abs/2109.13859v1 | https://arxiv.org/pdf/2109.13859v1.pdf | NudgeSeg: Zero-Shot Object Segmentation by Repeated Physical Interaction | Recent advances in object segmentation have demonstrated that deep neural networks excel at object segmentation for specific classes in color and depth images. However, their performance is dictated by the number of classes and objects used for training, thereby hindering generalization to never seen objects or zero-shot samples. To exacerbate the problem further, object segmentation using image frames rely on recognition and pattern matching cues. Instead, we utilize the 'active' nature of a robot and their ability to 'interact' with the environment to induce additional geometric constraints for segmenting zero-shot samples. In this paper, we present the first framework to segment unknown objects in a cluttered scene by repeatedly 'nudging' at the objects and moving them to obtain additional motion cues at every step using only a monochrome monocular camera. We call our framework NudgeSeg. These motion cues are used to refine the segmentation masks. We successfully test our approach to segment novel objects in various cluttered scenes and provide an extensive study with image and motion segmentation methods. We show an impressive average detection rate of over 86% on zero-shot objects. | ['Yiannis Aloimonos', 'Cornelia Fermüller', 'Chethan M. Parameshwara', 'Nitin J. Sanket', 'Chahat Deep Singh'] | 2021-09-22 | null | null | null | null | ['motion-segmentation'] | ['computer-vision'] | [ 6.50636196e-01 -8.80318135e-02 -2.56063677e-02 -4.39548820e-01
-4.08965886e-01 -7.84674525e-01 2.33672485e-01 -1.81580663e-01
-7.79388011e-01 4.83723313e-01 -7.01393664e-01 -4.30409983e-02
8.74286890e-02 -5.37104309e-01 -7.32563436e-01 -8.32265496e-01
-2.81619430e-02 5.70696712e-01 7.84375072e-01 1.27400666e-01
3.64006996e-01 8.40900242e-01 -1.78862596e+00 1.47047520e-01
5.35880506e-01 7.02595532e-01 7.01745689e-01 1.03383064e+00
-2.47606367e-01 4.46957529e-01 -7.09188759e-01 1.61912560e-01
5.13171077e-01 -3.79251212e-01 -9.34222639e-01 9.07206416e-01
4.88360733e-01 -5.70028067e-01 -1.73223004e-01 1.15397584e+00
9.34685841e-02 4.30994779e-01 6.03438675e-01 -1.17013454e+00
-3.11947614e-01 2.31973097e-01 -6.18099034e-01 3.63548756e-01
3.70859176e-01 4.20054764e-01 6.41818643e-01 -9.70706046e-01
8.66413653e-01 1.00897300e+00 3.02597225e-01 8.54309857e-01
-1.26445031e+00 -3.38234037e-01 4.36904103e-01 9.17158723e-02
-1.21265984e+00 -7.04853952e-01 7.80908227e-01 -5.24661481e-01
8.11145663e-01 2.43853569e-01 7.50263691e-01 7.25630105e-01
-1.44739136e-01 1.19371319e+00 6.42443001e-01 -4.11657810e-01
5.54762661e-01 6.88796714e-02 2.78348625e-01 5.78073800e-01
2.58430719e-01 1.99372038e-01 -8.09870362e-02 2.96184689e-01
1.04532206e+00 -4.60489318e-02 -3.09119731e-01 -6.60054088e-01
-1.10370362e+00 5.22969246e-01 4.74316835e-01 3.59196573e-01
-1.97595909e-01 2.34539852e-01 -1.02227822e-01 -8.47858563e-02
9.69244093e-02 5.85460663e-01 -5.13386428e-01 1.95972890e-01
-9.56669509e-01 7.01465309e-02 6.40744150e-01 1.06851256e+00
9.77269769e-01 1.54911295e-01 2.29914516e-01 7.02495694e-01
1.55614868e-01 3.29358101e-01 2.40558118e-01 -1.40709174e+00
1.84781045e-01 4.71277952e-01 3.29343110e-01 -8.64818215e-01
-5.56287467e-01 -1.02875732e-01 -3.72144014e-01 5.70018113e-01
7.18069375e-01 -1.58080846e-01 -1.54114461e+00 1.36773515e+00
4.72281188e-01 -1.68039903e-01 5.72907692e-03 1.24153817e+00
5.95319092e-01 5.84347606e-01 -1.14282183e-01 -1.42458662e-01
9.61219609e-01 -1.09535801e+00 -3.82389128e-01 -6.72597885e-01
6.09463632e-01 -6.43689811e-01 9.34807122e-01 6.35932446e-01
-1.11781347e+00 -6.72925115e-01 -1.03917122e+00 1.02293983e-01
-4.40485537e-01 -3.20551656e-02 7.46476293e-01 6.12036705e-01
-6.79432929e-01 5.52368939e-01 -1.02524936e+00 -3.46353859e-01
6.15184426e-01 6.15018427e-01 -3.61859173e-01 -2.24560410e-01
-4.28690016e-01 5.53108633e-01 6.74184144e-01 3.00944299e-01
-1.03040183e+00 -2.93765754e-01 -8.19722831e-01 -2.90166408e-01
6.97667420e-01 -3.69639307e-01 1.32083535e+00 -1.33123946e+00
-1.37835503e+00 9.30962980e-01 -6.28502816e-02 -2.25368991e-01
4.92657244e-01 -2.20259055e-01 -6.18039705e-02 5.43848097e-01
8.44364986e-02 1.25941050e+00 8.20227206e-01 -1.59641707e+00
-9.04927671e-01 -2.95252264e-01 2.91715860e-01 3.27719212e-01
2.92162836e-01 -5.99694550e-02 -7.87475884e-01 -2.72668093e-01
5.08347869e-01 -9.22204971e-01 -6.09683573e-01 1.10756978e-01
-5.10570168e-01 1.08456664e-01 1.15822685e+00 -1.33643642e-01
5.53772271e-01 -2.14706945e+00 1.29037350e-01 -7.17926119e-03
8.95046145e-02 3.24323654e-01 -2.02280119e-01 -7.63399825e-02
1.08414002e-01 -7.80143663e-02 -5.20775676e-01 -4.09779280e-01
-3.52447599e-01 5.74249685e-01 5.97935393e-02 5.56783795e-01
3.42415720e-01 9.69061434e-01 -8.79373252e-01 -5.85462630e-01
5.52742600e-01 3.65868839e-03 -5.47777116e-01 2.10803717e-01
-4.42271680e-01 6.89930379e-01 -3.79178971e-01 8.75587046e-01
6.83853805e-01 -1.93806916e-01 -1.63256660e-01 1.99834555e-01
-1.45381048e-01 -3.22364986e-01 -1.36341691e+00 1.73073792e+00
5.04243607e-03 8.33150029e-01 3.47967327e-01 -9.80032027e-01
6.35409653e-01 -1.03921838e-01 5.85529685e-01 -4.61783886e-01
2.59912491e-01 1.43326491e-01 1.45907253e-01 -6.90245330e-01
6.23890460e-01 2.17649899e-02 8.03753212e-02 3.86092752e-01
3.64690495e-04 -3.23978364e-01 3.84563595e-01 1.06275924e-01
7.30364501e-01 2.75558472e-01 -9.58855823e-02 -6.32552952e-02
2.34968528e-01 3.87182206e-01 5.33720851e-01 1.11359823e+00
-4.83350694e-01 9.90650952e-01 1.21230911e-02 -4.49700862e-01
-8.77039015e-01 -1.08686948e+00 -1.11891158e-01 1.03413260e+00
7.25881577e-01 4.20864046e-01 -7.90862083e-01 -5.90850830e-01
-3.77637446e-01 5.92653275e-01 -5.76002002e-01 1.15049817e-01
-7.46399403e-01 -8.22513282e-01 1.43714592e-01 6.96185410e-01
5.26770473e-01 -1.19077814e+00 -1.42032683e+00 4.12657320e-01
-5.93622513e-02 -1.23246193e+00 -2.28663757e-01 6.23983502e-01
-8.52943122e-01 -1.22726166e+00 -7.89578438e-01 -9.20444191e-01
1.00042593e+00 6.06162488e-01 8.50804329e-01 3.20028029e-02
-7.81737983e-01 5.18533409e-01 -3.99947643e-01 -2.30859250e-01
-1.53531715e-01 -1.28003627e-01 -8.70615095e-02 -6.04147837e-02
2.75311559e-01 -2.62541890e-01 -7.06371784e-01 4.07077134e-01
-1.14254272e+00 2.01722860e-01 6.51847541e-01 4.76712763e-01
4.04734880e-01 -8.41279551e-02 2.32492313e-01 -6.69961393e-01
-3.35866846e-02 -1.04072392e-01 -7.70784259e-01 -2.43142042e-02
8.73491168e-04 -1.88162804e-01 2.80856043e-01 -7.68521667e-01
-9.15906131e-01 7.62421012e-01 1.24051847e-01 -5.12694180e-01
-7.07404733e-01 -2.88187899e-02 -1.25950918e-01 -2.10473269e-01
5.68568408e-01 1.75109148e-01 -2.45388165e-01 -3.12890321e-01
4.12328809e-01 3.89211446e-01 8.01614761e-01 -4.08061236e-01
6.22332871e-01 7.40270972e-01 -1.76614851e-01 -1.17734528e+00
-6.88685417e-01 -7.07546651e-01 -1.12131882e+00 -4.35128748e-01
1.22073495e+00 -5.25041044e-01 -4.98612404e-01 5.22170842e-01
-1.22991455e+00 -5.41494668e-01 -4.29715633e-01 4.36508805e-01
-6.36406124e-01 3.89234483e-01 -4.59444612e-01 -1.05527914e+00
1.39023632e-01 -1.19046807e+00 1.06481516e+00 4.65244770e-01
-3.01098805e-02 -7.23009944e-01 -2.74166733e-01 1.97399408e-01
1.51147703e-02 2.86660731e-01 5.85904717e-01 -5.41012168e-01
-1.11976707e+00 -1.87782079e-01 -3.37546408e-01 3.38047981e-01
1.53230265e-01 5.26477173e-02 -1.05110860e+00 -7.33762160e-02
-9.74872150e-03 -1.83397427e-01 8.21368635e-01 5.85182071e-01
1.13477314e+00 1.04018331e-01 -5.89780748e-01 5.79102099e-01
1.46141720e+00 8.44527543e-01 5.08873403e-01 2.58144289e-01
8.73851657e-01 7.83367634e-01 7.41472900e-01 1.21650778e-01
-1.73457041e-01 5.33441842e-01 3.86049807e-01 -2.59323806e-01
3.73579003e-02 2.55888015e-01 -4.99760844e-02 2.90374726e-01
-6.59128278e-03 -2.77410358e-01 -1.03850770e+00 7.40800440e-01
-1.74692369e+00 -7.39638746e-01 -2.82612830e-01 1.99537396e+00
4.43980873e-01 3.51175159e-01 1.08270511e-01 6.53861910e-02
7.94797242e-01 -1.98187992e-01 -7.23991394e-01 -1.56140968e-01
-1.23750135e-01 -4.43076454e-02 5.63337862e-01 7.10720420e-01
-1.28698087e+00 1.30326176e+00 6.71009159e+00 3.61631393e-01
-9.61275637e-01 -3.32374752e-01 5.33452213e-01 -6.54446408e-02
1.58137962e-01 -2.73474287e-02 -8.55093122e-01 1.00029506e-01
3.20298374e-01 3.13483179e-01 3.56163055e-01 8.32898259e-01
7.90101439e-02 -6.95726931e-01 -1.36959136e+00 8.96011889e-01
4.75008450e-02 -9.58303392e-01 -1.38053834e-01 -3.77219580e-02
8.33108902e-01 -4.60349023e-02 -1.82298627e-02 -1.22456729e-01
1.81744337e-01 -9.14948761e-01 7.11488783e-01 3.56454879e-01
2.45602980e-01 -5.72038472e-01 3.66767943e-01 4.88899499e-01
-1.01172698e+00 -1.74466535e-01 -5.59093833e-01 -1.94976717e-01
2.84136623e-01 2.93591440e-01 -9.37599301e-01 2.11562589e-01
4.87246782e-01 4.16809320e-01 -6.17305458e-01 1.24350548e+00
1.27242178e-01 5.30500785e-02 -5.48365653e-01 -4.39160466e-02
5.94742835e-01 -2.14079857e-01 5.51795363e-01 1.14132726e+00
5.42809293e-02 4.47503984e-01 3.89442116e-01 9.48675215e-01
3.64044517e-01 -3.52803975e-01 -5.01008213e-01 -2.29638424e-02
1.20288379e-01 1.15172255e+00 -1.47761595e+00 -4.44831461e-01
-2.38803506e-01 1.28046989e+00 -2.80904304e-02 6.39170051e-01
-7.76019394e-01 -5.14864743e-01 3.56747359e-01 -1.73173189e-01
6.70609057e-01 -6.44657731e-01 -2.78050363e-01 -1.07484353e+00
-1.27524301e-01 -4.48237717e-01 -3.39674279e-02 -9.14424241e-01
-8.42706978e-01 4.81644213e-01 1.53708383e-01 -1.01308608e+00
-2.39653274e-01 -8.90823364e-01 -2.85623133e-01 4.58009034e-01
-1.17541289e+00 -7.96331286e-01 -3.90733451e-01 5.44000089e-01
9.09974396e-01 2.19698623e-01 4.92052943e-01 1.82862565e-01
-5.53058863e-01 -1.25668589e-02 4.82731201e-02 2.69668847e-01
1.75588131e-01 -1.17429411e+00 2.65440464e-01 1.02754605e+00
2.71350354e-01 3.08094770e-01 7.54425049e-01 -5.34447849e-01
-1.24369562e+00 -9.11855996e-01 1.65093929e-01 -5.38070500e-01
2.22276762e-01 -5.81809700e-01 -9.74532247e-01 7.00555801e-01
-9.99714956e-02 7.27189258e-02 3.33164036e-01 -4.01007473e-01
1.36242241e-01 2.30274677e-01 -1.04321086e+00 6.57232761e-01
1.16220093e+00 1.33654764e-02 -6.59922481e-01 2.36092612e-01
7.05002666e-01 -4.03210640e-01 -2.15398446e-01 5.12395203e-01
5.59927106e-01 -1.01246834e+00 1.00678456e+00 -6.17988884e-01
2.14012265e-01 -4.16515410e-01 -2.17285797e-01 -7.79645503e-01
-4.87517305e-02 -3.01536143e-01 1.80788651e-01 8.47891986e-01
3.45674455e-01 -2.85108417e-01 1.21669686e+00 8.70778680e-01
-2.48694971e-01 -4.73785132e-01 -7.45970547e-01 -7.70967901e-01
-1.14686914e-01 -6.43294871e-01 -9.98234525e-02 7.25141644e-01
-4.46451485e-01 1.29635260e-01 -7.70038515e-02 3.31957906e-01
7.28097498e-01 2.78136462e-01 8.74723017e-01 -1.09908390e+00
-4.24523443e-01 -3.53806645e-01 -4.65949386e-01 -1.43018508e+00
-8.32674131e-02 -4.27177489e-01 6.95353448e-01 -1.38354623e+00
8.71561691e-02 -4.59606588e-01 9.73218828e-02 2.64273375e-01
-6.52195960e-02 5.10854900e-01 2.70260423e-01 2.23691836e-01
-7.27650642e-01 1.23698831e-01 1.28514838e+00 -1.49512768e-01
-6.31930053e-01 -1.51104853e-02 -2.28859752e-01 9.56871629e-01
7.93087006e-01 -4.15194988e-01 -3.63290578e-01 -6.48820996e-01
-2.30334908e-01 6.19941298e-03 4.36446249e-01 -1.27894938e+00
3.64919901e-01 -4.28951114e-01 8.05175722e-01 -8.92171681e-01
6.22572899e-01 -8.62601280e-01 5.64794950e-02 6.33989990e-01
-1.48791805e-01 -3.74696881e-01 3.10897976e-01 6.46246850e-01
4.71776538e-02 -7.06877828e-01 1.11885107e+00 -5.21043301e-01
-1.25813973e+00 1.48253292e-01 -8.22228968e-01 -1.60739914e-01
1.30603147e+00 -8.09849679e-01 -3.50201391e-02 -1.12939529e-01
-1.17247391e+00 1.97915763e-01 5.66753507e-01 2.70289272e-01
8.64468873e-01 -7.69076347e-01 -1.59430310e-01 4.72722590e-01
6.29572123e-02 4.58118170e-01 1.71971157e-01 6.01378739e-01
-8.14046025e-01 3.40180218e-01 -2.62584388e-01 -1.10006332e+00
-1.13177860e+00 7.71022797e-01 4.69541162e-01 4.28449929e-01
-6.95052147e-01 9.86579597e-01 5.89570463e-01 -1.17456436e-01
3.41831356e-01 -5.26541889e-01 6.10121563e-02 -8.13715607e-02
3.87365818e-01 2.83712119e-01 -2.57886291e-01 -4.69334692e-01
-2.52229869e-01 7.64369309e-01 -2.59698004e-01 -3.90638530e-01
1.09981024e+00 -3.86258513e-01 2.26274282e-01 7.24756718e-01
1.17832732e+00 -3.52158159e-01 -1.72763526e+00 -9.05360747e-03
1.82222992e-01 -6.41500175e-01 -1.65340632e-01 -5.74956954e-01
-9.52904820e-01 9.58092093e-01 6.30981028e-01 1.99520275e-01
1.02457964e+00 2.02733442e-01 5.72594345e-01 5.58682799e-01
4.41483021e-01 -1.10675299e+00 4.09724891e-01 5.22543609e-01
3.94555718e-01 -1.35098457e+00 -1.98359698e-01 -6.00238025e-01
-4.80212897e-01 1.14945471e+00 8.18467975e-01 -2.09576726e-01
2.71311700e-01 2.23532647e-01 2.34138280e-01 -4.61411588e-02
-2.55476475e-01 -5.56337714e-01 1.93031833e-01 7.59783745e-01
-1.14143707e-01 -1.67148963e-01 1.55856356e-01 6.96849227e-02
1.70901671e-01 -1.53739914e-01 7.73364842e-01 1.20962775e+00
-1.01328540e+00 -7.41356730e-01 -4.93604124e-01 1.31688982e-01
-2.32287914e-01 2.83346981e-01 -6.20110214e-01 9.86696899e-01
2.41289943e-01 9.64454234e-01 3.09733123e-01 -1.28286302e-01
4.07676585e-02 -2.49883533e-02 7.56309807e-01 -9.50464904e-01
-9.76232514e-02 3.36158693e-01 -9.32280198e-02 -5.58478415e-01
-7.13258088e-01 -6.89544678e-01 -1.38355792e+00 3.32879156e-01
-4.73097116e-01 -1.47008687e-01 7.21012294e-01 1.05178857e+00
-6.91160336e-02 5.50701678e-01 4.48834687e-01 -1.46795106e+00
5.93029195e-03 -5.51981926e-01 -5.38294494e-01 2.61405796e-01
5.44005990e-01 -5.69374681e-01 -3.49844903e-01 3.64709049e-01] | [8.974493026733398, -0.45785605907440186] |
a22b2647-9fe7-4a31-9f86-dda75122bffa | leveraging-tripartite-interaction-information | 2106.03415 | null | https://arxiv.org/abs/2106.03415v1 | https://arxiv.org/pdf/2106.03415v1.pdf | Leveraging Tripartite Interaction Information from Live Stream E-Commerce for Improving Product Recommendation | Recently, a new form of online shopping becomes more and more popular, which combines live streaming with E-Commerce activity. The streamers introduce products and interact with their audiences, and hence greatly improve the performance of selling products. Despite of the successful applications in industries, the live stream E-commerce has not been well studied in the data science community. To fill this gap, we investigate this brand-new scenario and collect a real-world Live Stream E-Commerce (LSEC) dataset. Different from conventional E-commerce activities, the streamers play a pivotal role in the LSEC events. Hence, the key is to make full use of rich interaction information among streamers, users, and products. We first conduct data analysis on the tripartite interaction data and quantify the streamer's influence on users' purchase behavior. Based on the analysis results, we model the tripartite information as a heterogeneous graph, which can be decomposed to multiple bipartite graphs in order to better capture the influence. We propose a novel Live Stream E-Commerce Graph Neural Network framework (LSEC-GNN) to learn the node representations of each bipartite graph, and further design a multi-task learning approach to improve product recommendation. Extensive experiments on two real-world datasets with different scales show that our method can significantly outperform various baseline approaches. | ['JinFeng Yi', 'Qi Liu', 'Dongsheng Li', 'Shanshan Feng', 'Dong-Dong Chen', 'Zhuoxuan Jiang', 'Sanshi Yu'] | 2021-06-07 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-6.28823936e-02 -3.29134315e-01 -5.64103067e-01 -5.18479824e-01
-7.67027587e-03 -4.25380021e-01 2.46587649e-01 3.81653130e-01
-2.68769432e-02 -5.76377846e-02 5.01287401e-01 -1.80816889e-01
-3.64443481e-01 -1.36750090e+00 -8.85761499e-01 -5.78400671e-01
-3.63616377e-01 4.02874231e-01 2.00609207e-01 -7.83143401e-01
-1.33601576e-01 -1.91393778e-01 -1.10942829e+00 6.14730537e-01
6.70783281e-01 1.12219882e+00 3.72967985e-03 1.72606856e-01
-2.70251870e-01 6.95492685e-01 -2.28160024e-01 -1.00923240e+00
2.96975493e-01 -4.79576826e-01 -2.54892141e-01 2.53189772e-01
8.18120316e-03 -2.69767404e-01 -7.29223311e-01 1.15932798e+00
3.18183392e-01 1.85581386e-01 2.54260391e-01 -1.56539595e+00
-8.82714927e-01 1.38829553e+00 -9.46064413e-01 3.94513100e-01
1.79281294e-01 -7.15454891e-02 1.63135993e+00 -4.80196029e-01
6.20653212e-01 1.29168642e+00 5.95394433e-01 -2.44392902e-01
-9.25694823e-01 -9.05991793e-01 6.64185047e-01 3.59976888e-01
-1.01773584e+00 1.34765238e-01 1.30028665e+00 -6.01808913e-02
6.27923727e-01 4.23506042e-03 1.14020979e+00 1.32141936e+00
1.87926903e-01 1.32189691e+00 2.46153429e-01 2.71592855e-01
-9.92159173e-02 2.52417419e-02 3.49176645e-01 4.31589872e-01
3.94379556e-01 -1.45252973e-01 -6.99958146e-01 -3.73945646e-02
8.95793796e-01 7.43004799e-01 -8.19499716e-02 -3.83057088e-01
-1.08470082e+00 1.10373437e+00 8.75386298e-01 3.50558609e-01
-7.32779682e-01 -9.69488919e-03 8.39761615e-01 5.64522445e-01
5.94964325e-01 -2.10336417e-01 -3.37203503e-01 3.31878737e-02
-2.16562316e-01 2.34941781e-01 1.09358680e+00 1.15655267e+00
2.86656499e-01 -1.64747730e-01 3.00091416e-01 9.26112771e-01
8.76020789e-01 2.76358783e-01 2.65996933e-01 -2.55353481e-01
8.49308968e-01 1.01275945e+00 -3.84089172e-01 -1.70323098e+00
-2.51518637e-01 -8.59301805e-01 -1.36907959e+00 -7.73453653e-01
3.66903484e-01 -7.51992613e-02 -1.39916390e-01 1.19910872e+00
2.99184859e-01 3.12229991e-01 -3.41570348e-01 1.06649661e+00
9.68802392e-01 1.12337494e+00 4.14814502e-02 -3.66637945e-01
1.29068625e+00 -1.02633739e+00 -7.75841594e-01 3.74614261e-02
4.27179158e-01 -4.87070650e-01 7.72582889e-01 8.27512026e-01
-7.80125618e-01 -5.15887856e-01 -1.13310719e+00 3.44666362e-01
-2.19036877e-01 -5.48286855e-01 1.24294519e+00 4.48753715e-01
-3.17557126e-01 4.23527151e-01 -3.89365494e-01 -7.37865344e-02
5.53902447e-01 2.55425066e-01 1.49301859e-02 -3.55027139e-01
-1.80859411e+00 -6.01684675e-02 3.00562292e-01 2.43883431e-01
-4.51793700e-01 -5.93879700e-01 -7.87951052e-01 2.88466036e-01
7.02881515e-01 -5.20159960e-01 1.03681970e+00 -1.03481007e+00
-1.06586337e+00 2.51324326e-01 3.48319024e-01 -4.06854868e-01
2.83272177e-01 8.27170461e-02 -8.72345507e-01 -7.87572339e-02
-2.74427742e-01 -8.29835087e-02 6.39876246e-01 -1.23750889e+00
-9.85171199e-01 -6.76712155e-01 3.33843172e-01 1.86732635e-01
-8.00536573e-01 -6.66623339e-02 -6.95361435e-01 -6.82469368e-01
6.70237318e-02 -6.19609416e-01 -2.63085306e-01 -5.21471918e-01
-1.13205478e-01 -5.25319517e-01 6.90716326e-01 -3.54435295e-01
1.47904205e+00 -2.21198750e+00 6.12304732e-02 4.45150375e-01
6.64790511e-01 1.14724979e-01 -3.54223013e-01 7.41179585e-01
1.86463118e-01 -4.81454320e-02 3.29985559e-01 -2.69371625e-02
1.69696704e-01 3.32820863e-01 -1.57517120e-01 2.34865576e-01
-3.02854687e-01 1.23640156e+00 -1.04515743e+00 -1.64369896e-01
-1.23316601e-01 4.69710290e-01 -5.59907258e-01 -5.48537001e-02
-2.69970357e-01 2.18954254e-02 -8.74738038e-01 6.88356280e-01
8.36841762e-01 -7.34668970e-01 2.66789496e-01 -4.02474701e-01
4.90377665e-01 5.65934144e-02 -1.32491624e+00 1.36386752e+00
-3.61275017e-01 -9.80681404e-02 3.44982386e-01 -1.29129016e+00
9.91996884e-01 7.64953066e-03 9.36093569e-01 -1.12633896e+00
3.06108117e-01 -4.38592061e-02 7.42832497e-02 -5.53941011e-01
4.56461519e-01 -1.08480282e-01 -5.57364151e-02 5.41130900e-01
2.00714961e-01 6.27211452e-01 3.61573458e-01 6.30355895e-01
9.11384702e-01 -2.38748029e-01 2.85335749e-01 1.06960461e-01
3.04965466e-01 -2.32680172e-01 6.69027627e-01 3.51403266e-01
-3.74579549e-01 1.07075863e-01 6.81933999e-01 -6.85638070e-01
-6.86847031e-01 -9.28103864e-01 2.11630881e-01 1.52793992e+00
7.21825600e-01 -6.32312477e-01 -2.14786902e-01 -8.88309836e-01
4.60276604e-01 4.32098627e-01 -5.53550184e-01 -4.01206762e-01
-5.73097587e-01 -9.29976821e-01 -3.08686472e-03 4.74744618e-01
5.24558842e-01 -1.21986771e+00 3.36477965e-01 5.56817770e-01
-2.85567790e-01 -9.19525266e-01 -8.53524625e-01 -2.02814490e-01
-9.29760575e-01 -1.00078225e+00 -5.21514118e-01 -8.91133845e-01
1.87273875e-01 8.02092254e-01 1.38652134e+00 -2.90156472e-02
1.67795390e-01 4.02801931e-01 -7.68984318e-01 -4.04207766e-01
-2.03279749e-01 1.96271956e-01 -1.40071362e-01 7.49942005e-01
9.02101636e-01 -7.73476303e-01 -8.27248454e-01 6.53872490e-01
-1.22880292e+00 -1.39849186e-01 5.98603725e-01 6.81975722e-01
4.95901853e-01 5.35793364e-01 1.04025483e+00 -1.47586036e+00
1.00292385e+00 -1.31306767e+00 -3.03905517e-01 5.12979068e-02
-8.97742987e-01 -6.38544500e-01 6.94588006e-01 -6.26328826e-01
-9.16936994e-01 -2.37785190e-01 -2.91767512e-02 -4.55547750e-01
2.04796523e-01 1.31615376e+00 -1.98346496e-01 2.94836998e-01
2.44639650e-01 1.36058033e-01 -2.06100374e-01 -3.97505403e-01
3.90491426e-01 6.94064140e-01 4.86968160e-02 -1.23974755e-01
4.94125485e-01 3.11480761e-01 -3.01788568e-01 -5.62739491e-01
-5.84364712e-01 -1.01221526e+00 -2.82291919e-01 -5.15922129e-01
4.03594792e-01 -8.83485913e-01 -1.18138099e+00 5.72104871e-01
-1.03293633e+00 1.32568210e-01 -9.25102159e-02 5.24538636e-01
-9.04252380e-03 3.34952980e-01 -1.05495834e+00 -6.32060349e-01
-3.36754203e-01 -7.18027592e-01 8.08421195e-01 3.02227020e-01
3.47235590e-01 -1.20690846e+00 -9.07289758e-02 6.95179999e-01
3.09761524e-01 -1.03860952e-01 7.42828071e-01 -8.95315349e-01
-8.78729820e-01 -3.31238598e-01 -5.23303211e-01 9.36610699e-02
2.66922805e-02 -2.85916507e-01 -3.61812204e-01 -2.93808430e-01
9.05504227e-02 -4.41192612e-02 9.61249053e-01 4.39528942e-01
8.02267790e-01 -3.85967404e-01 -2.51993507e-01 5.21168411e-01
1.29075229e+00 5.08766830e-01 4.16390449e-01 1.97904378e-01
1.06024492e+00 7.43739247e-01 5.67700505e-01 7.96741068e-01
9.85820770e-01 2.60982126e-01 7.62691677e-01 8.35153386e-02
3.39884549e-01 -8.09160948e-01 3.71462345e-01 1.64616144e+00
-2.28741303e-01 -6.86240375e-01 -4.32149112e-01 4.18298870e-01
-2.25466251e+00 -9.87566411e-01 -4.06887203e-01 1.60064065e+00
2.91261077e-01 2.27007493e-01 5.40830851e-01 -5.11826649e-02
8.51320505e-01 4.93760109e-01 -8.23236287e-01 7.81849027e-02
-5.32266917e-03 -3.73359859e-01 4.83819544e-01 -1.28773987e-01
-1.01615763e+00 4.86086249e-01 5.16539764e+00 9.62903678e-01
-7.48093247e-01 9.51208547e-02 7.07074821e-01 -1.59125626e-01
-6.97747111e-01 -4.89674389e-01 -7.89120674e-01 5.81218958e-01
8.21704507e-01 -4.96730864e-01 7.58396566e-01 7.79188693e-01
1.64177105e-01 5.64351976e-01 -8.16509604e-01 1.21013510e+00
2.29618624e-01 -1.16128457e+00 1.65426001e-01 2.49246106e-01
5.55518150e-01 1.30215734e-01 -4.71233726e-02 6.75265551e-01
5.00778139e-01 -4.57154512e-01 4.08065289e-01 2.50731260e-01
1.19083248e-01 -9.39735174e-01 9.14096653e-01 4.37526852e-01
-1.72952926e+00 -3.16650063e-01 -5.20101011e-01 -3.59350294e-02
5.92141807e-01 7.91520000e-01 -2.00432047e-01 9.07860279e-01
8.54500353e-01 1.52504957e+00 -1.22041129e-01 8.52462769e-01
2.03583807e-01 9.97597873e-01 -1.90036029e-01 -4.81821746e-01
2.92692006e-01 -8.72219503e-01 3.63719553e-01 9.98473644e-01
2.37284690e-01 2.76420116e-01 3.71423423e-01 7.48839319e-01
-5.68300307e-01 5.74432850e-01 -7.16675222e-01 -5.62366903e-01
-3.16327326e-02 1.51175380e+00 -8.59969914e-01 -1.08964920e-01
-9.71329331e-01 7.26435721e-01 1.37647271e-01 4.05490607e-01
-7.63342142e-01 -3.59335124e-01 2.81981945e-01 2.19725236e-01
5.82087696e-01 -9.44239125e-02 1.01022519e-01 -1.32416129e+00
1.96154982e-01 -9.48498487e-01 7.39247680e-01 -2.53329724e-01
-2.08409786e+00 3.50794911e-01 -2.93000430e-01 -1.24966514e+00
8.12010244e-02 -3.40581506e-01 -6.86615229e-01 3.12291831e-01
-1.29745412e+00 -1.16628754e+00 -1.90569714e-01 7.23880231e-01
5.33890784e-01 -3.09797019e-01 1.96558043e-01 9.38698351e-01
-6.13372803e-01 5.24001002e-01 1.14775516e-01 2.10541904e-01
3.82816821e-01 -9.07577455e-01 4.97620195e-01 4.31703776e-01
4.62768435e-01 6.20805800e-01 8.03226829e-02 -8.88652444e-01
-2.02238750e+00 -1.12579429e+00 5.72292089e-01 -1.96787901e-02
1.09399760e+00 -5.49702346e-01 -8.13714027e-01 7.52185941e-01
2.20706150e-01 -6.51956126e-02 8.90157223e-01 6.78303301e-01
-4.72541600e-01 -6.28689647e-01 -7.56505489e-01 2.97509700e-01
1.49925876e+00 -2.80780256e-01 -3.86971086e-01 4.43053186e-01
1.09997833e+00 1.25963971e-01 -1.18152201e+00 3.47068071e-01
5.37701368e-01 -9.29361343e-01 9.74953234e-01 -6.05165064e-01
6.24450803e-01 1.42534837e-01 -2.80046207e-03 -1.59432435e+00
-6.37277961e-01 -5.36730051e-01 -3.51302445e-01 1.31781530e+00
3.49314481e-01 -8.42251897e-01 9.88907039e-01 2.03029707e-01
1.69058993e-01 -8.54579985e-01 -4.20594126e-01 -3.62719536e-01
-4.50565577e-01 -5.36377728e-01 9.66533244e-01 1.27096140e+00
4.15105999e-01 8.48203778e-01 -5.17123282e-01 -1.11904263e-01
6.22902334e-01 5.97151875e-01 6.28721714e-01 -1.37801838e+00
-7.02888310e-01 -5.80835402e-01 -3.37736160e-01 -1.55348039e+00
-3.43240261e-01 -1.10798359e+00 -3.77219498e-01 -1.45654607e+00
2.98816830e-01 -3.25017273e-01 -7.96510696e-01 2.98970025e-02
-2.26931591e-02 6.78468049e-02 2.12055936e-01 3.65552068e-01
-1.05465186e+00 7.41103292e-01 1.81510520e+00 -4.61661160e-01
-2.98145384e-01 4.88250107e-01 -1.12541246e+00 3.52090359e-01
3.46799225e-01 -2.39182711e-01 -7.83452272e-01 -2.12646589e-01
1.05743206e+00 3.97946775e-01 -2.04753414e-01 -2.41466761e-01
3.86966288e-01 2.47207228e-02 5.67370094e-02 -5.95542073e-01
-1.60886005e-01 -1.09982848e+00 6.94376007e-02 2.73413897e-01
-3.79268110e-01 1.79302820e-03 -3.67356867e-01 1.25262344e+00
-5.47863185e-01 1.03585951e-01 -2.36631511e-03 -1.32335782e-01
-5.98118186e-01 1.08638299e+00 -3.79370302e-02 -2.01277971e-01
9.77371395e-01 6.60875514e-02 -2.57189602e-01 -8.27176452e-01
-5.96845686e-01 6.15568042e-01 -3.70997608e-01 8.74127150e-01
5.17733753e-01 -1.62681019e+00 -8.59595418e-01 3.27681988e-01
2.15368614e-01 -6.80308929e-03 5.42070806e-01 6.32923901e-01
-1.26197964e-01 6.52061403e-02 -2.53413096e-02 -4.53966707e-01
-7.87447989e-01 9.42468762e-01 -9.88682210e-02 -4.67423230e-01
-5.46955347e-01 6.74091339e-01 2.01600134e-01 -5.14103293e-01
2.99006581e-01 -1.71721190e-01 -8.20716262e-01 4.90464061e-01
4.50159073e-01 4.72992748e-01 -1.16466366e-01 -6.41341448e-01
2.01519370e-01 1.27764300e-01 -3.69097054e-01 4.74932611e-01
1.80397058e+00 -3.39353144e-01 -2.36235727e-02 5.87305546e-01
1.35866821e+00 -2.03230903e-01 -9.10950720e-01 -7.15977788e-01
-3.02476674e-01 -6.20988309e-01 2.12498128e-01 -4.11583722e-01
-1.96152747e+00 6.24864638e-01 9.03024673e-02 1.19163799e+00
1.15226519e+00 7.31079057e-02 1.54890645e+00 2.96140492e-01
6.97448552e-01 -8.49730074e-01 8.83936286e-02 2.89500594e-01
6.91721737e-01 -1.30575681e+00 -2.33018205e-01 -8.03714216e-01
-8.29251230e-01 9.53931868e-01 5.87528944e-01 -3.08921814e-01
1.14820981e+00 -5.56799434e-02 -1.19710388e-02 -5.52514076e-01
-6.16978407e-01 8.66755396e-02 1.44543707e-01 2.82462895e-01
3.18574369e-01 2.24116281e-01 -3.02436233e-01 1.22394586e+00
1.36280507e-01 1.97540089e-01 2.78058946e-02 4.68858510e-01
-1.72175229e-01 -1.10150123e+00 1.89342961e-01 1.00130582e+00
-2.38894954e-01 1.21393256e-01 -2.02014282e-01 5.02022147e-01
-1.49190165e-02 9.84509647e-01 3.18789564e-04 -9.23969924e-01
6.65183425e-01 -4.24953103e-01 1.28870845e-01 -2.26546511e-01
-9.49443877e-01 3.88622969e-01 -6.99584037e-02 -6.31722689e-01
-4.63119805e-01 -6.93167984e-01 -1.03795254e+00 -6.70219779e-01
-4.03370112e-01 2.04246864e-01 3.95643979e-01 6.98194921e-01
3.86884719e-01 8.36380601e-01 9.94673610e-01 -6.71730816e-01
-5.09240150e-01 -8.25368643e-01 -1.30503309e+00 9.50552881e-01
-2.30376482e-01 -5.41650057e-01 -3.21288794e-01 -1.31415501e-01] | [10.17721939086914, 5.632208824157715] |
d9de235a-ed88-4caf-85f0-11b20f35c006 | videoretalking-audio-based-lip | 2211.14758 | null | https://arxiv.org/abs/2211.14758v1 | https://arxiv.org/pdf/2211.14758v1.pdf | VideoReTalking: Audio-based Lip Synchronization for Talking Head Video Editing In the Wild | We present VideoReTalking, a new system to edit the faces of a real-world talking head video according to input audio, producing a high-quality and lip-syncing output video even with a different emotion. Our system disentangles this objective into three sequential tasks: (1) face video generation with a canonical expression; (2) audio-driven lip-sync; and (3) face enhancement for improving photo-realism. Given a talking-head video, we first modify the expression of each frame according to the same expression template using the expression editing network, resulting in a video with the canonical expression. This video, together with the given audio, is then fed into the lip-sync network to generate a lip-syncing video. Finally, we improve the photo-realism of the synthesized faces through an identity-aware face enhancement network and post-processing. We use learning-based approaches for all three steps and all our modules can be tackled in a sequential pipeline without any user intervention. Furthermore, our system is a generic approach that does not need to be retrained to a specific person. Evaluations on two widely-used datasets and in-the-wild examples demonstrate the superiority of our framework over other state-of-the-art methods in terms of lip-sync accuracy and visual quality. | ['Nannan Wang', 'Jue Wang', 'Xuan Wang', 'Mingrui Zhu', 'Fei Yin', 'Menghan Xia', 'Yong Zhang', 'Xiaodong Cun', 'Kun Cheng'] | 2022-11-27 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 5.03534317e-01 -5.28420806e-02 2.26104781e-01 -5.13579786e-01
-7.43948281e-01 -5.23701549e-01 4.11051542e-01 -4.06521827e-01
-2.31127903e-01 4.38584507e-01 1.80395797e-01 1.78982750e-01
4.40712959e-01 -4.33201313e-01 -7.51601994e-01 -6.47641063e-01
4.31732178e-01 2.27430165e-02 4.24557999e-02 -4.80368249e-02
2.00383738e-02 6.75477684e-01 -2.12487793e+00 5.68379700e-01
6.00462198e-01 1.24234891e+00 -7.32478797e-02 9.75173652e-01
4.02111262e-02 6.25080228e-01 -5.70680857e-01 -7.56429493e-01
2.27954239e-01 -5.67195415e-01 -4.27481800e-01 4.64559436e-01
8.06571305e-01 -3.74157369e-01 -1.02106169e-01 1.14135587e+00
8.47712815e-01 3.63963516e-03 3.47911179e-01 -1.49640918e+00
-2.24765927e-01 3.05353582e-01 -3.52546126e-01 -5.63637555e-01
9.41148043e-01 4.85401988e-01 5.71265340e-01 -1.20813501e+00
8.26399565e-01 1.56868863e+00 5.10590076e-01 9.02206481e-01
-1.45309174e+00 -9.42900836e-01 -1.21361032e-01 2.83939540e-01
-1.57245576e+00 -1.29989982e+00 7.31020451e-01 -3.94568324e-01
5.19884050e-01 1.81469932e-01 8.92362952e-01 1.12475634e+00
-1.93000540e-01 4.98992026e-01 8.76932323e-01 -4.30680156e-01
1.18968725e-01 1.64459258e-01 -7.04971194e-01 5.92666209e-01
-5.53907633e-01 6.52741492e-02 -9.04344738e-01 2.01815024e-01
4.19570953e-01 -3.57497305e-01 -5.07865608e-01 -1.91122741e-01
-1.02074921e+00 5.40765226e-01 5.46262115e-02 1.22199871e-01
-4.73256767e-01 8.45170766e-02 4.96840656e-01 2.62067974e-01
5.00589848e-01 6.93362877e-02 -1.83746204e-01 -1.54542580e-01
-1.37192404e+00 2.83149034e-01 7.03319371e-01 7.64408112e-01
7.05974400e-01 9.89310890e-02 -1.88140839e-01 8.89666438e-01
2.20371351e-01 5.09973824e-01 1.10323474e-01 -1.27610457e+00
2.25623295e-01 2.19011515e-01 5.27146906e-02 -9.09317434e-01
-2.17208520e-01 1.46365687e-01 -6.76417768e-01 6.73622608e-01
3.40170056e-01 -4.26490098e-01 -4.87884521e-01 2.07223439e+00
5.35234988e-01 4.98682380e-01 8.27755108e-02 8.43232393e-01
9.44377720e-01 8.17616343e-01 1.84555203e-02 -5.10449171e-01
1.44527221e+00 -9.13185716e-01 -1.08583367e+00 9.65315998e-02
1.84170544e-01 -1.01728916e+00 9.35386300e-01 4.83573318e-01
-1.43305326e+00 -7.54447877e-01 -7.97345281e-01 -1.44788817e-01
4.49446738e-02 3.67988586e-01 7.80036673e-02 7.27442563e-01
-1.55602527e+00 4.81770396e-01 -2.29333073e-01 -3.15941066e-01
3.43927175e-01 4.51555252e-01 -8.24444354e-01 1.65612876e-01
-9.45083141e-01 5.31467855e-01 1.98436350e-01 -7.21213520e-02
-6.70977473e-01 -9.38366354e-01 -1.11414385e+00 1.29626216e-02
4.94145870e-01 -8.18595886e-01 1.25894010e+00 -1.71485043e+00
-2.27019739e+00 1.02517152e+00 -5.41083276e-01 -7.47732520e-02
7.16577232e-01 -4.49876487e-02 -5.07070482e-01 3.70701313e-01
-1.70321822e-01 1.15209937e+00 1.44788849e+00 -1.25397980e+00
-6.66783035e-01 -1.87873155e-01 -8.54512751e-02 3.86769265e-01
-1.82554618e-01 4.89969671e-01 -1.03152013e+00 -5.97065151e-01
-3.35227877e-01 -9.41571355e-01 5.12182534e-01 5.76191127e-01
-3.08200091e-01 -1.03856616e-01 9.54879463e-01 -7.99869478e-01
9.78650749e-01 -2.39354944e+00 3.30456734e-01 -3.13841067e-02
-4.72193025e-02 4.10338521e-01 -3.30452949e-01 7.96643943e-02
-3.10360461e-01 -1.03182010e-01 -1.58104658e-01 -8.98560107e-01
-1.28765762e-01 -2.83680737e-01 3.23434803e-03 4.30762202e-01
4.01546240e-01 6.49876475e-01 -9.34479415e-01 -7.08864987e-01
3.60171050e-01 9.31125164e-01 -8.15864503e-01 5.73492229e-01
-1.99462309e-01 5.28728306e-01 3.61427277e-01 5.12403607e-01
8.02951932e-01 4.97482032e-01 2.61675864e-01 -6.05740964e-01
-1.40863791e-01 -1.76188782e-01 -1.33648503e+00 1.79569304e+00
-7.93636978e-01 8.84923160e-01 6.05772555e-01 -4.50328380e-01
8.36188018e-01 5.88672876e-01 4.46012795e-01 -4.08457488e-01
1.15050301e-01 2.51100272e-01 -3.34093451e-01 -8.06588769e-01
4.26104635e-01 -2.13533118e-01 2.04810694e-01 3.88427019e-01
3.97257209e-01 -3.58658075e-01 1.52088761e-01 8.36548768e-03
6.16043806e-01 1.87985331e-01 3.11100930e-02 1.15493648e-01
9.97078598e-01 -9.64039326e-01 3.79580706e-01 -1.34636220e-02
-5.95871024e-02 7.31581032e-01 5.92066586e-01 -1.68042555e-02
-1.02807999e+00 -1.02739072e+00 2.47067660e-01 1.20252097e+00
-2.26076141e-01 -5.88935256e-01 -1.30902457e+00 -4.43066448e-01
-3.63863945e-01 4.80141342e-01 -5.82738280e-01 -1.12152919e-02
-3.75397801e-01 4.75070029e-02 7.07382798e-01 1.20936185e-01
4.48383570e-01 -1.42165089e+00 -4.27805215e-01 3.49004902e-02
-5.29158592e-01 -1.38370931e+00 -1.06568301e+00 -5.24843335e-01
-1.38258949e-01 -9.22174692e-01 -9.47478592e-01 -7.17528284e-01
7.51351476e-01 -3.62788141e-02 7.98424780e-01 1.77458063e-01
-1.65738672e-01 2.86736906e-01 -5.30077815e-02 -1.73253015e-01
-8.48381162e-01 -3.00929099e-01 2.59295672e-01 1.02031720e+00
-1.49867013e-01 -5.98571658e-01 -5.50822616e-01 1.77015781e-01
-9.77590382e-01 2.36510426e-01 1.56636059e-01 6.90188229e-01
3.49731117e-01 -9.09006298e-02 4.80826914e-01 -2.96239316e-01
5.38901508e-01 1.32904819e-03 -6.06036603e-01 1.47327185e-01
-7.47759710e-04 -2.19727144e-01 6.87383890e-01 -6.83659732e-01
-1.31074917e+00 4.37517226e-01 -4.94734585e-01 -7.31885552e-01
-1.16441883e-01 -1.37405619e-01 -7.48708189e-01 -8.60651582e-02
3.73568147e-01 1.11728624e-01 3.28645974e-01 -1.83754504e-01
6.21916771e-01 9.06818509e-01 1.06625187e+00 -2.72935182e-01
6.94306672e-01 5.02911329e-01 -5.48034385e-02 -1.09462488e+00
-3.67692679e-01 -8.87515694e-02 -6.42369211e-01 -7.58421481e-01
1.01370370e+00 -9.09780979e-01 -1.31481600e+00 1.05339193e+00
-1.48046184e+00 -3.50767821e-01 -2.67590404e-01 2.12715343e-01
-9.13120925e-01 3.24551433e-01 -3.65129322e-01 -8.47880065e-01
-3.61502290e-01 -1.40894830e+00 1.27723479e+00 3.39757293e-01
-2.21395150e-01 -5.79582870e-01 -9.84426960e-02 3.07102084e-01
2.67612040e-01 1.13966271e-01 5.76303124e-01 -1.29876256e-01
-2.55540788e-01 3.97482365e-02 -1.57522380e-01 5.00711441e-01
1.25063330e-01 5.51920831e-01 -1.34464693e+00 -1.93586633e-01
-3.26377511e-01 -4.24304217e-01 4.55289841e-01 2.06672177e-01
1.08301985e+00 -4.75412458e-01 1.07579902e-01 9.72195685e-01
9.98482168e-01 3.32331248e-02 8.02090526e-01 -3.42974097e-01
3.83558959e-01 9.88856494e-01 3.90772879e-01 4.13267344e-01
2.25358695e-01 1.27669811e+00 3.27582121e-01 -3.00062805e-01
-4.31873024e-01 -5.25164485e-01 8.16232383e-01 3.70397538e-01
5.98448254e-02 -1.47699699e-01 -3.74171048e-01 3.68374169e-01
-1.60490108e+00 -1.22024488e+00 1.54084250e-01 2.11688519e+00
9.53452170e-01 -3.42672646e-01 3.81394982e-01 4.03840125e-01
1.02930534e+00 -8.72631650e-03 -4.18177843e-01 -5.15026510e-01
-1.19367801e-01 3.01311553e-01 -1.96273342e-01 7.33552217e-01
-8.52531374e-01 1.09453940e+00 5.17190027e+00 9.17917788e-01
-1.72735667e+00 7.49492273e-02 5.26747167e-01 -3.80421966e-01
-1.60705119e-01 -2.81132460e-01 -5.08515298e-01 4.22513068e-01
8.26939285e-01 -1.74096242e-01 6.04257166e-01 6.44264638e-01
5.76702416e-01 -7.49175623e-02 -1.22747290e+00 1.46306860e+00
4.95929092e-01 -1.34877610e+00 9.65717733e-02 -2.68571049e-01
4.64997083e-01 -5.87752283e-01 5.94790466e-02 -5.06403856e-02
-3.34462315e-01 -9.85490263e-01 1.18953049e+00 6.48756385e-01
1.54080272e+00 -9.02585745e-01 4.23581302e-01 -1.00345470e-01
-1.26079476e+00 8.59502777e-02 1.52810127e-01 4.90685105e-01
4.96150374e-01 2.35198438e-01 -6.10101044e-01 3.22058946e-01
7.62956679e-01 5.39064765e-01 -2.68058240e-01 7.29725420e-01
-4.24967378e-01 1.39153227e-01 -1.87400177e-01 3.98501575e-01
-3.28800499e-01 8.76751542e-02 7.10893869e-01 1.45063603e+00
4.67773855e-01 3.91361723e-03 -2.22673655e-01 8.45604777e-01
-4.12695915e-01 3.46000224e-01 -5.10822475e-01 1.72408193e-01
4.57772046e-01 1.53718829e+00 -3.49744141e-01 -3.79945278e-01
-1.48171455e-01 1.32041919e+00 -6.52553961e-02 2.38726720e-01
-8.70701253e-01 -4.40548897e-01 9.91139174e-01 3.41619015e-01
3.02973390e-01 3.84706259e-01 3.25933248e-01 -1.01288640e+00
1.59466729e-01 -1.22064722e+00 -7.41841123e-02 -1.25296366e+00
-8.51747870e-01 9.30599093e-01 -2.64139801e-01 -1.07363904e+00
-5.85075855e-01 -3.25728685e-01 -6.31721497e-01 9.09879923e-01
-1.42397785e+00 -1.23682141e+00 -5.29229224e-01 1.16022444e+00
6.10608459e-01 -1.24151938e-01 8.13758731e-01 5.04453778e-01
-4.22806859e-01 9.42598999e-01 -4.90042537e-01 -4.00542244e-02
1.12424338e+00 -6.71815097e-01 2.00526178e-01 8.29696715e-01
-1.62321091e-01 9.20071527e-02 6.61883771e-01 -3.48635286e-01
-1.17833304e+00 -1.11337006e+00 9.84795511e-01 8.57466683e-02
3.03975612e-01 -6.69656813e-01 -7.45978832e-01 2.38915190e-01
3.07307512e-01 9.46838558e-02 3.92618597e-01 -5.51615179e-01
-3.32454890e-01 -4.20376122e-01 -1.33699989e+00 6.97503150e-01
9.73751545e-01 -7.39413440e-01 -1.05670765e-01 -2.48499140e-02
6.19069755e-01 -3.52113128e-01 -5.71827412e-01 2.89035171e-01
8.52448463e-01 -1.27102828e+00 7.96845436e-01 -2.71684617e-01
4.01277989e-01 -4.47201639e-01 4.26644608e-02 -1.28480959e+00
2.78598696e-01 -1.36319065e+00 2.67451763e-01 1.68787003e+00
1.86307952e-01 -2.36102045e-01 3.88819337e-01 4.22477692e-01
1.38765782e-01 -4.20777380e-01 -9.46924448e-01 -4.19059962e-01
-4.99455661e-01 -6.86218977e-01 7.32462883e-01 5.94184518e-01
4.10331152e-02 3.23616266e-01 -6.20967031e-01 -6.29455270e-03
4.63425159e-01 -2.59878963e-01 1.24186909e+00 -9.73227799e-01
-2.81087514e-02 -3.32726210e-01 -3.93963426e-01 -8.41880798e-01
6.48815572e-01 -7.09762990e-01 2.55422860e-01 -1.02957928e+00
-1.97213814e-02 1.16381288e-01 5.15048921e-01 4.11834568e-01
-3.34769003e-02 5.21616995e-01 7.14292288e-01 -9.15872827e-02
-3.13241184e-01 5.26459754e-01 9.53716218e-01 -1.37466025e-02
-3.21037054e-01 3.59623581e-02 -4.33555722e-01 8.49906445e-01
3.84477854e-01 -3.22641402e-01 -3.18733454e-01 -7.19071031e-02
6.76660538e-02 4.36701566e-01 4.17722136e-01 -1.15033853e+00
2.94048190e-01 1.03036523e-01 2.80994058e-01 -1.97204903e-01
8.25321198e-01 -7.86284029e-01 3.96783173e-01 2.28441343e-01
-3.12931180e-01 2.76813079e-02 3.52207065e-01 2.02224329e-02
-1.71874404e-01 -6.83234110e-02 1.33625507e+00 3.20979118e-01
-3.80726337e-01 3.30198467e-01 -4.28452998e-01 -7.24604130e-02
1.07701361e+00 -2.86380500e-01 9.53931808e-02 -9.78408515e-01
-8.39301586e-01 -2.19705597e-01 6.64033413e-01 5.37809372e-01
7.50361085e-01 -1.37261295e+00 -9.97577488e-01 5.35390556e-01
-5.81266172e-02 -3.15349013e-01 4.99725163e-01 7.49753118e-01
-3.57462466e-01 -1.68315396e-01 -3.24571878e-01 -7.88492024e-01
-1.83565736e+00 4.97891366e-01 6.28433824e-01 8.19800645e-02
-3.83930087e-01 7.20109999e-01 2.98533410e-01 -1.75566569e-01
3.91011119e-01 7.43373111e-02 -8.42945129e-02 3.42147052e-01
8.40478361e-01 2.40519315e-01 2.81168427e-02 -1.11066878e+00
-2.97898471e-01 8.11624885e-01 5.12534916e-01 -5.68626523e-01
1.16652095e+00 -2.37166330e-01 -1.04151867e-01 1.72168434e-01
1.37946248e+00 3.57470423e-01 -1.45640504e+00 1.93767436e-02
-5.76425612e-01 -5.23257554e-01 -2.04681709e-01 -6.71479404e-01
-1.44097626e+00 9.73649383e-01 6.26101613e-01 -3.22315276e-01
1.58337915e+00 -2.26267979e-01 5.92744350e-01 -7.07591102e-02
2.49947291e-02 -1.01644969e+00 3.43928754e-01 1.82421625e-01
1.23795271e+00 -8.38331759e-01 -4.22210842e-01 -5.27817249e-01
-7.64997005e-01 1.30970275e+00 3.72500151e-01 3.93388867e-01
5.74843764e-01 4.86996502e-01 3.28612238e-01 2.33299062e-01
-7.91314781e-01 -1.34783909e-01 2.38873884e-01 8.93668234e-01
3.18183482e-01 -2.80323744e-01 1.18768826e-01 3.10926467e-01
-4.34444606e-01 4.23463523e-01 4.29827392e-01 1.21643767e-01
4.52418327e-02 -9.04700220e-01 -5.50070286e-01 -2.70389318e-01
-4.99010980e-01 -7.47910812e-02 -4.74551409e-01 4.59645510e-01
3.84970635e-01 1.22832096e+00 1.06924273e-01 -3.45560968e-01
4.51213837e-01 3.52105200e-01 5.35364032e-01 -3.55140835e-01
-7.02672184e-01 2.61170775e-01 4.07339372e-02 -8.41250360e-01
-4.55504090e-01 -5.97828627e-01 -1.04628074e+00 -5.32886565e-01
-1.45839840e-01 -1.45925358e-01 7.73311794e-01 6.16920710e-01
5.36534250e-01 3.13857079e-01 8.28411818e-01 -1.37736821e+00
9.45603196e-03 -7.98222482e-01 -2.95891494e-01 6.64913833e-01
3.65729600e-01 -4.16592598e-01 -4.42544281e-01 5.86875856e-01] | [13.217886924743652, -0.43536219000816345] |
adca0a64-263e-4d1c-98b9-d9b8d6333fde | a-new-local-radon-descriptor-for-content | 2007.15523 | null | https://arxiv.org/abs/2007.15523v1 | https://arxiv.org/pdf/2007.15523v1.pdf | A new Local Radon Descriptor for Content-Based Image Search | Content-based image retrieval (CBIR) is an essential part of computer vision research, especially in medical expert systems. Having a discriminative image descriptor with the least number of parameters for tuning is desirable in CBIR systems. In this paper, we introduce a new simple descriptor based on the histogram of local Radon projections. We also propose a very fast convolution-based local Radon estimator to overcome the slow process of Radon projections. We performed our experiments using pathology images (KimiaPath24) and lung CT patches and test our proposed solution for medical image processing. We achieved superior results compared with other histogram-based descriptors such as LBP and HoG as well as some pre-trained CNNs. | ['Morteza Babaie', 'Meghana D. Kumar', 'Hany Kashani', 'Hamid. R. Tizhoosh'] | 2020-07-30 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [-2.86811893e-03 -4.56901550e-01 1.24979526e-01 -4.33221281e-01
-1.14140892e+00 1.87921431e-02 3.04091007e-01 4.39929008e-01
-9.83880341e-01 5.31623542e-01 1.98460206e-01 -1.93010986e-01
-3.98193032e-01 -9.43512321e-01 -1.92134142e-01 -8.92313361e-01
-6.17386252e-02 2.92209268e-01 7.28696227e-01 -6.50535971e-02
4.68385220e-01 8.10328424e-01 -1.25303972e+00 4.31096435e-01
1.98005185e-01 1.25409949e+00 4.50229704e-01 8.75825703e-01
-1.57387141e-04 7.95818388e-01 -3.23253661e-01 2.69930929e-01
1.85833827e-01 -3.05239111e-01 -8.52247000e-01 -5.49769811e-02
3.14983457e-01 -3.96300763e-01 -3.87612551e-01 9.38016355e-01
9.12955940e-01 4.52422917e-01 8.73060584e-01 -5.44639528e-01
-8.53320658e-01 -8.16787928e-02 -5.45695662e-01 6.78757966e-01
1.69094056e-01 -2.46209845e-01 6.11742198e-01 -8.36286962e-01
6.15142286e-01 1.06394327e+00 8.58720958e-01 3.06343138e-01
-6.28510952e-01 -2.14814276e-01 -7.03685641e-01 6.55492902e-01
-1.74190879e+00 -5.92634492e-02 5.95454693e-01 -1.20349295e-01
8.45402896e-01 4.26326513e-01 4.12076920e-01 3.13733637e-01
7.57390499e-01 5.74229121e-01 1.24659121e+00 -8.78780246e-01
7.73701519e-02 6.63234070e-02 2.58529007e-01 1.19749773e+00
7.28746057e-02 -7.10360557e-02 6.93233218e-03 -4.22707826e-01
1.10190368e+00 5.23991883e-01 -4.93867882e-02 -8.62497389e-02
-1.03242683e+00 1.00781250e+00 7.46749282e-01 9.04641032e-01
-3.32204312e-01 4.94511873e-01 2.61093676e-01 6.70509636e-02
2.21934855e-01 1.48547158e-01 -9.94431600e-02 3.63423944e-01
-8.86374712e-01 -4.25480008e-02 5.00082970e-01 5.75692177e-01
5.84108651e-01 -5.79177618e-01 -5.78538477e-01 1.11320007e+00
2.18363106e-01 4.76465017e-01 1.00359225e+00 -6.89426541e-01
-2.63463348e-01 2.83611000e-01 -2.35511750e-01 -1.28538895e+00
-5.18054545e-01 -5.01445942e-02 -9.41933811e-01 3.30386870e-02
2.60018170e-01 5.72871745e-01 -1.12311959e+00 9.40204740e-01
7.17297718e-02 -1.22468814e-01 -3.38047445e-01 9.97452021e-01
9.76559341e-01 5.18739998e-01 -2.94144470e-02 2.54458413e-02
1.67376256e+00 -1.01203227e+00 -5.89356005e-01 4.08019781e-01
5.54843366e-01 -1.05731404e+00 8.86098325e-01 1.74865440e-01
-8.60615373e-01 -6.09151125e-01 -7.48826623e-01 -3.89489383e-01
-5.74876189e-01 2.53917545e-01 4.58259851e-01 6.31970406e-01
-1.27257895e+00 6.29984260e-01 -7.24235475e-01 -6.98971748e-01
1.83910936e-01 2.69425988e-01 -6.51463926e-01 -2.36439273e-01
-8.64665627e-01 1.03036094e+00 1.84889391e-01 8.46999064e-02
-4.45927560e-01 -1.36551946e-01 -6.96691036e-01 4.72857095e-02
-1.20217413e-01 -4.04133081e-01 1.06185925e+00 -4.04307306e-01
-1.27680695e+00 1.01134694e+00 -1.02682687e-01 -2.67937928e-01
3.90153080e-01 1.73958302e-01 -1.57156765e-01 6.09851301e-01
-1.79929584e-01 5.77043712e-01 7.03668296e-01 -7.49956250e-01
-4.86795008e-01 -2.84409016e-01 -4.08949763e-01 -6.53982861e-03
-2.85166621e-01 2.84760118e-01 -4.53786612e-01 -5.02729893e-01
3.78072500e-01 -8.30960214e-01 -4.45337206e-01 4.79848862e-01
-6.69883192e-02 -2.86910832e-01 8.60218644e-01 -7.71636188e-01
9.22973871e-01 -2.18731284e+00 -6.04127049e-01 4.59659547e-01
6.78569376e-02 2.29005128e-01 8.49073566e-03 3.28001291e-01
3.57389189e-02 4.10745405e-02 1.77120008e-02 3.93665545e-02
-3.74808282e-01 1.79396436e-01 2.51224428e-01 9.18076277e-01
-4.58435379e-02 8.20039630e-01 -4.67865437e-01 -1.31449091e+00
4.94265765e-01 7.52427399e-01 -3.75684053e-01 -1.07599363e-01
5.25825381e-01 3.02210212e-01 -6.03058636e-01 8.10708106e-01
7.66098201e-01 -3.79084766e-01 -1.81703493e-01 -3.90315056e-01
1.20026935e-02 -2.23106414e-01 -8.42212677e-01 1.51225352e+00
-5.36760390e-01 6.45970285e-01 -2.48278394e-01 -1.11403334e+00
9.33972120e-01 3.42523634e-01 7.84510612e-01 -9.84680712e-01
3.91707987e-01 2.58607060e-01 -2.30998829e-01 -5.93852282e-01
3.27032238e-01 -2.08573565e-01 1.78232193e-01 3.30859900e-01
9.07026082e-02 -9.89005342e-02 1.11380786e-01 -1.88649237e-01
1.32745647e+00 -3.66806567e-01 7.75393307e-01 -6.02346182e-01
9.22007740e-01 -4.37642932e-02 9.08993930e-02 8.68003964e-01
-6.07893884e-01 1.09273040e+00 -5.57639189e-02 -8.08101952e-01
-1.04440284e+00 -9.45700228e-01 -7.28473604e-01 8.31802368e-01
9.29956362e-02 -2.82536144e-03 -4.96613413e-01 -4.80535209e-01
-1.51253968e-01 -7.66879395e-02 -6.93195581e-01 5.12383170e-02
-8.46682668e-01 -6.99743390e-01 4.84435856e-01 4.88791376e-01
6.73984468e-01 -1.40975356e+00 -7.72338808e-01 8.19047540e-02
-2.09407434e-02 -8.53482604e-01 -6.82471216e-01 2.72073805e-01
-9.33845818e-01 -1.16515267e+00 -1.14151084e+00 -1.32853484e+00
9.22801971e-01 4.45944041e-01 7.35460341e-01 4.71595645e-01
-1.08415234e+00 3.48460436e-01 -5.02479911e-01 -1.69714078e-01
-1.22220017e-01 -1.50498956e-01 -4.49567288e-01 -5.11773765e-01
4.03688282e-01 -3.57553065e-02 -1.01513636e+00 3.01635027e-01
-1.06918788e+00 -4.75489289e-01 8.69824827e-01 1.23330402e+00
1.02034032e+00 1.74419418e-01 -1.26134396e-01 -7.30006814e-01
6.60151899e-01 1.60989344e-01 -3.69998246e-01 4.95873749e-01
-4.48623925e-01 1.24917202e-01 4.87416983e-01 -2.11702034e-01
-9.21816707e-01 4.79609929e-02 -4.40385818e-01 -2.16412202e-01
-1.40330806e-01 3.68538082e-01 5.79392910e-01 -5.54547191e-01
7.39479959e-01 3.36067349e-01 1.06969558e-01 -4.48291928e-01
-2.11810204e-03 9.00980711e-01 3.85824025e-01 -3.73978317e-01
4.61987227e-01 6.62224174e-01 3.39253873e-01 -1.18716979e+00
-5.27486086e-01 -1.31782877e+00 -5.47121465e-01 9.34507102e-02
1.05543935e+00 -5.52147388e-01 -8.34845722e-01 2.25484043e-01
-1.09141791e+00 9.41608250e-02 -2.43470773e-01 6.90751314e-01
-7.17819989e-01 4.99956191e-01 -1.05118608e+00 -5.81567764e-01
-7.66062915e-01 -1.21446800e+00 1.12972105e+00 2.28275880e-01
2.50774205e-01 -1.11850441e+00 4.48319227e-01 8.71354640e-02
8.28649700e-01 2.15241164e-02 7.86050379e-01 -6.40160739e-01
-3.41751575e-01 -4.79182631e-01 -7.10051954e-01 3.35195571e-01
2.34393805e-01 -1.40254781e-01 -8.58226418e-01 -2.79385239e-01
2.56486863e-01 -2.00563505e-01 1.18784750e+00 6.93274736e-01
1.63885164e+00 -3.39419581e-02 -4.35476005e-01 6.87940300e-01
1.90226293e+00 3.46150964e-01 9.03134942e-01 4.51626301e-01
4.34891433e-01 2.65306890e-01 6.11985803e-01 2.15586752e-01
-8.18284377e-02 6.20875239e-01 -1.25993088e-01 -4.21297073e-01
-3.45851004e-01 1.03204548e-01 -3.13495219e-01 9.77453411e-01
-1.46108374e-01 4.77064997e-02 -8.30876410e-01 3.90960515e-01
-1.71568131e+00 -9.08283353e-01 -5.79617508e-02 2.06587386e+00
7.27910519e-01 -3.69055063e-01 -3.33669752e-01 2.10640058e-02
6.86218083e-01 -1.24363303e-01 1.68555006e-01 -4.67705756e-01
1.97204694e-01 9.71174479e-01 8.81303728e-01 3.99751514e-01
-1.57606423e+00 6.67660713e-01 7.14015293e+00 1.21109116e+00
-1.26258504e+00 4.36694682e-01 5.88769019e-01 6.11488283e-01
2.70103723e-01 -3.08340549e-01 -5.90582550e-01 9.53438431e-02
5.29294729e-01 1.53291702e-01 -1.31036505e-01 1.06925976e+00
-6.45666495e-02 -4.07824188e-01 -6.29437447e-01 1.09705830e+00
2.86131114e-01 -1.17248237e+00 -1.76735148e-01 1.30621269e-01
4.80140358e-01 -3.90965044e-02 8.08161870e-02 -1.35945737e-01
-1.64218068e-01 -1.02164102e+00 1.03553787e-01 6.82364702e-01
5.77147543e-01 -6.17891073e-01 1.17667532e+00 -4.82579274e-03
-1.13768876e+00 5.65695167e-02 -9.97476161e-01 5.18632054e-01
-3.15971702e-01 4.85300124e-01 -1.03479254e+00 3.59473854e-01
8.50988150e-01 4.35399264e-01 -7.35171974e-01 1.48106778e+00
3.04007173e-01 1.38153329e-01 -3.11770171e-01 -1.43919826e-01
2.39411935e-01 -6.23309389e-02 -3.87032330e-02 1.67068839e+00
6.14128351e-01 1.41583428e-01 1.89696431e-01 4.54686105e-01
1.64717346e-01 5.74858546e-01 -6.59252822e-01 2.92885840e-01
-1.52045205e-01 1.53490591e+00 -1.15510607e+00 -3.73524904e-01
-2.65007824e-01 8.74236465e-01 -8.16782042e-02 -2.35504974e-02
-5.53807199e-01 -5.82600713e-01 -5.65337911e-02 7.18596950e-02
4.03180242e-01 -1.99520484e-01 1.15685523e-01 -9.18074548e-01
-3.34998220e-01 -4.86679673e-01 5.83631158e-01 -6.01981461e-01
-1.39778757e+00 6.82342172e-01 -3.93519439e-02 -1.40472651e+00
1.15566645e-02 -9.67387438e-01 -4.59529907e-01 7.44997859e-01
-1.63372016e+00 -1.26395762e+00 -4.20106143e-01 9.30821717e-01
3.56653571e-01 -7.20982701e-02 9.82706428e-01 4.34102714e-01
-2.33949050e-02 4.33367997e-01 3.91169041e-01 3.08628291e-01
1.07172680e+00 -1.16516662e+00 -2.24176288e-01 4.69373852e-01
-7.63731543e-03 7.33820498e-01 3.84796053e-01 -2.79803663e-01
-1.07347083e+00 -7.51958430e-01 8.70013535e-01 -4.32238616e-02
3.62654150e-01 2.51142085e-01 -6.92331672e-01 3.22590739e-01
1.71672195e-01 5.93287051e-01 6.37998760e-01 -3.79267752e-01
-8.39421451e-02 -2.92140603e-01 -1.51113260e+00 5.88872842e-02
4.75573033e-01 -5.96826971e-01 -4.80127692e-01 7.05361903e-01
2.78168589e-01 -2.96652585e-01 -9.20688212e-01 3.16009879e-01
6.77701771e-01 -1.08413923e+00 1.37723637e+00 -1.91779882e-02
1.28201216e-01 -1.39796183e-01 -2.55136073e-01 -8.10524702e-01
-5.35060942e-01 3.22108239e-01 7.53974319e-01 4.95008498e-01
-9.96634811e-02 -5.51417589e-01 6.79413736e-01 1.91321656e-01
-5.18754683e-02 -5.61231852e-01 -8.13308358e-01 -7.62944341e-01
-1.48616200e-02 -2.83991941e-03 -5.71741350e-02 5.46804726e-01
-3.99138257e-02 -3.86415601e-01 -1.47563338e-01 -3.06289226e-01
6.10806882e-01 -3.78267393e-02 2.26333857e-01 -8.52829874e-01
-3.41511965e-01 -4.12756652e-01 -8.99049163e-01 -4.28697884e-01
-2.79872656e-01 -8.72581482e-01 4.40912217e-01 -1.63042116e+00
6.80669308e-01 -4.75116968e-01 -9.12813902e-01 3.87833327e-01
7.21343085e-02 7.20778763e-01 1.37429442e-02 3.20599079e-01
-5.90103686e-01 3.63125280e-02 1.49205506e+00 -1.89395443e-01
1.08021274e-01 -9.43636075e-02 -1.09619826e-01 6.03408575e-01
7.34745860e-01 -6.76281810e-01 -4.18152474e-02 6.16697669e-02
-3.36054206e-01 1.27564855e-02 3.78203899e-01 -1.10322607e+00
4.53628957e-01 1.70560703e-02 6.75048590e-01 -6.79470360e-01
3.24693322e-01 -8.42891514e-01 -1.37609109e-01 8.94356608e-01
-3.51584345e-01 1.19746253e-01 -1.65513575e-01 3.20337415e-01
-6.70911252e-01 -6.10951185e-01 1.09844506e+00 -6.40701413e-01
-7.28933752e-01 3.25523436e-01 -3.34666044e-01 -6.20880902e-01
9.83434856e-01 -1.16525263e-01 -4.50131595e-01 -1.07735679e-01
-7.24090934e-01 -3.47389311e-01 2.59160429e-01 -7.38664567e-02
1.04479301e+00 -1.25700414e+00 -3.69216472e-01 1.20372362e-01
2.15078384e-01 -3.41658324e-01 2.49641642e-01 1.04324675e+00
-1.48737431e+00 8.20742190e-01 -5.77093661e-01 -5.76892793e-01
-1.42751956e+00 6.86568081e-01 3.46194267e-01 -6.11078739e-01
-7.61339009e-01 6.26954257e-01 1.17842488e-01 -2.53916204e-01
7.09802434e-02 -4.25574452e-01 -3.40464264e-01 -2.14216143e-01
6.81939781e-01 4.01220977e-01 4.50161278e-01 -6.19815946e-01
-6.36260450e-01 9.89120245e-01 -1.45883590e-01 -7.23776892e-02
1.25322890e+00 9.00239721e-02 -4.12014604e-01 -5.78163750e-03
1.80092287e+00 -2.50697732e-01 -3.35597575e-01 -2.51648724e-01
1.29191115e-01 -7.21733689e-01 4.84894663e-01 -4.04131293e-01
-1.00647497e+00 1.00677550e+00 1.24990821e+00 5.10140099e-02
1.06221354e+00 8.22418109e-02 9.11919117e-01 7.48209596e-01
4.83391523e-01 -1.14450622e+00 3.46311599e-01 2.07477599e-01
8.06380570e-01 -1.48514652e+00 3.38354737e-01 -1.54224992e-01
-3.45284879e-01 1.42853999e+00 1.21536270e-01 -6.74166620e-01
1.26260746e+00 1.01568267e-01 3.03051919e-01 -2.52752870e-01
-3.01269710e-01 -5.00016868e-01 5.16749084e-01 5.11718571e-01
7.22519100e-01 -6.32550940e-02 -8.68133187e-01 -1.64034367e-01
3.89880896e-01 2.26247415e-01 1.06975034e-01 1.28106725e+00
-7.59062111e-01 -1.20222664e+00 -6.52610660e-01 4.93969768e-01
-9.36236024e-01 5.54425921e-03 1.46837145e-01 8.39488745e-01
-8.44594166e-02 5.01331031e-01 1.17854826e-01 -3.38993073e-02
6.77423552e-02 -1.95060015e-01 6.98835552e-01 -4.37715828e-01
-5.26319683e-01 4.91911918e-01 -4.25139338e-01 -6.51155889e-01
-5.33826649e-01 -2.73482025e-01 -1.19605255e+00 -1.17911384e-01
-3.96136761e-01 -1.65868670e-01 9.17474806e-01 7.16887653e-01
-2.35599324e-01 3.90546441e-01 5.18117785e-01 -6.14951253e-01
-5.46061158e-01 -9.86941457e-01 -9.35140491e-01 4.25516129e-01
2.49012128e-01 -4.92949516e-01 -1.82870343e-01 1.74546279e-02] | [14.281717300415039, -1.4742181301116943] |
ae52fee2-8c32-4827-918f-eb66457a6395 | sc2-pcr-a-second-order-spatial-compatibility | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chen_SC2-PCR_A_Second_Order_Spatial_Compatibility_for_Efficient_and_Robust_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chen_SC2-PCR_A_Second_Order_Spatial_Compatibility_for_Efficient_and_Robust_CVPR_2022_paper.pdf | SC2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registration | In this paper, we present a second order spatial compatibility (SC^2) measure based method for efficient and robust point cloud registration (PCR), called SC^2-PCR. Firstly, we propose a second order spatial compatibility (SC^2) measure to compute the similarity between correspondences. It considers the global compatibility instead of local consistency, allowing for more distinctive clustering between inliers and outliers at early stage. Based on this measure, our registration pipeline employs a global spectral technique to find some reliable seeds from the initial correspondences. Then we design a two-stage strategy to expand each seed to a consensus set based on the SC^2 measure matrix. Finally, we feed each consensus set to a weighted SVD algorithm to generate a candidate rigid transformation and select the best model as the final result. Our method can guarantee to find a certain number of outlier-free consensus sets using fewer samplings, making the model estimation more efficient and robust. In addition, the proposed SC^2 measure is general and can be easily plugged into deep learning based frameworks. Extensive experiments are carried out to investigate the performance of our method. | ['Wenbing Tao', 'Fan Yang', 'Kun Sun', 'Zhi Chen'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['point-cloud-registration'] | ['computer-vision'] | [-1.34820849e-01 -3.86209369e-01 1.55790344e-01 -2.39781782e-01
-8.77608716e-01 -2.82004267e-01 6.04317188e-01 2.46289372e-01
-2.54198551e-01 1.53531879e-01 3.94669138e-02 2.05444321e-01
-3.00955623e-01 -7.40787327e-01 -5.96727133e-01 -7.18635917e-01
-8.57406110e-02 5.07458389e-01 4.77711439e-01 -7.30478242e-02
4.37832177e-01 8.55164766e-01 -1.35068893e+00 -1.89096689e-01
1.08764458e+00 9.04199779e-01 3.70000660e-01 -6.54316545e-02
1.20912731e-01 9.02353749e-02 -2.21738681e-01 -3.66031006e-02
6.61952317e-01 -2.48406738e-01 -5.85288823e-01 1.01582520e-01
4.79612648e-01 -3.07686478e-02 2.18514472e-01 1.21563864e+00
5.40482581e-01 4.06742424e-01 5.40085196e-01 -1.06293166e+00
-1.91233605e-01 2.06699923e-01 -8.79642308e-01 -5.40785789e-02
4.82997537e-01 -1.75182391e-02 9.37991798e-01 -1.43604016e+00
5.76365232e-01 1.19799936e+00 9.01933253e-01 -1.69006050e-01
-1.12856114e+00 -7.23283052e-01 -9.99216456e-03 1.95964009e-01
-1.77308977e+00 -4.35481161e-01 1.01549590e+00 -4.47520047e-01
3.92672509e-01 3.42126101e-01 6.99825168e-01 3.98157507e-01
2.74499966e-04 3.72361720e-01 9.46122229e-01 -2.83695787e-01
3.73079747e-01 -3.38615090e-01 3.20812874e-02 6.56117082e-01
2.64600217e-01 -3.76066603e-02 -4.00365293e-01 -4.53184158e-01
9.41543579e-01 1.04955114e-01 -3.37999314e-01 -5.17477810e-01
-1.46104872e+00 6.82178020e-01 7.45654643e-01 4.05790806e-01
-5.50699770e-01 -5.33820279e-02 2.17669442e-01 6.16391785e-02
3.12765688e-01 3.01870555e-01 -1.20741419e-01 1.49808615e-01
-1.16844881e+00 1.52529508e-01 4.03814495e-01 7.31554806e-01
1.04697239e+00 -1.92743003e-01 -3.34373452e-02 9.45889771e-01
5.80708206e-01 3.95539910e-01 3.38293374e-01 -9.65301871e-01
2.54373163e-01 6.92653835e-01 8.23623836e-02 -1.65889299e+00
-4.14517641e-01 -3.73630941e-01 -1.00413811e+00 1.82100594e-01
5.85948192e-02 2.21016034e-01 -6.21676445e-01 1.30866015e+00
5.45464933e-01 7.03853905e-01 -2.49203309e-01 1.14176822e+00
5.52987754e-01 5.74472725e-01 -4.17664737e-01 -3.67573112e-01
1.13318336e+00 -6.31120741e-01 -3.71943712e-01 -7.63640627e-02
3.76274854e-01 -1.05762386e+00 7.62520969e-01 1.23736627e-01
-7.98947215e-01 -6.13533437e-01 -9.80443716e-01 2.03302979e-01
1.35652900e-01 2.37562224e-01 2.48350874e-01 1.55486599e-01
-1.10261083e+00 9.14694369e-01 -1.02119875e+00 -3.83018970e-01
9.42856297e-02 4.46178019e-01 -4.81015235e-01 -4.05578651e-02
-7.92869747e-01 7.57577062e-01 4.05419558e-01 3.21232408e-01
-2.33251482e-01 -4.37399566e-01 -8.50740373e-01 -1.92649618e-01
1.56774789e-01 -5.63154578e-01 5.20403147e-01 -8.20200801e-01
-1.42487001e+00 8.09327900e-01 -5.63472331e-01 -1.70378666e-02
5.79076529e-01 -2.26182882e-02 -1.58064440e-01 1.56994145e-02
3.78474176e-01 3.45962495e-01 7.80233860e-01 -1.58764970e+00
-4.44619387e-01 -2.40818858e-01 -5.12013078e-01 2.17569619e-01
3.15738916e-02 2.50144303e-01 -8.22499931e-01 -7.13951290e-01
9.67417181e-01 -1.11009681e+00 -4.28196311e-01 1.76208653e-02
-3.00365090e-01 -3.46529722e-01 5.50163448e-01 -7.11938262e-01
1.10939002e+00 -2.31807828e+00 2.46929750e-01 1.03392696e+00
2.47820660e-01 1.28076717e-01 -2.32676193e-01 3.39951336e-01
-1.33862585e-01 -2.12145239e-01 -3.34364086e-01 -4.50153559e-01
-2.05007061e-01 -9.42923129e-02 -3.90497819e-02 8.56494188e-01
1.95069119e-01 4.76383507e-01 -7.88606763e-01 -5.50326347e-01
3.22324395e-01 2.78185070e-01 -4.52364862e-01 3.23712863e-02
3.25525612e-01 6.20571196e-01 -4.50098097e-01 6.80376887e-01
1.17580688e+00 9.27630998e-03 -5.81264384e-02 -4.99675453e-01
-2.67381042e-01 -1.15304455e-01 -1.82840252e+00 1.63058150e+00
-1.97832093e-01 2.74688363e-01 8.43151100e-03 -7.40470052e-01
1.46035028e+00 -8.69482458e-02 8.81581426e-01 -4.00010645e-01
2.57023722e-01 5.47817826e-01 -5.21545075e-02 -9.60961208e-02
4.60706472e-01 4.38715592e-02 1.79271996e-01 2.24090800e-01
-2.89612681e-01 8.18087682e-02 -9.34912562e-02 -5.84501997e-02
8.51871073e-01 1.42630890e-01 3.48276675e-01 -5.29043078e-01
7.95531273e-01 -1.66620716e-01 1.00804031e+00 4.42857265e-01
-2.48436511e-01 9.72189844e-01 -2.27544643e-02 -4.97120112e-01
-9.76571620e-01 -9.40252900e-01 -2.03707635e-01 3.30696851e-01
6.26003027e-01 -4.44132835e-01 -5.17898917e-01 -2.67840117e-01
5.33272661e-02 1.19981274e-01 -2.44270757e-01 -1.48406133e-01
-8.76779854e-01 -4.61202085e-01 1.26738809e-02 3.67320359e-01
5.50726950e-01 -7.66349077e-01 -2.34643891e-01 1.11017600e-01
-2.61736006e-01 -8.13409209e-01 -6.08410716e-01 -3.20055932e-01
-1.01468837e+00 -9.73868966e-01 -5.63150346e-01 -9.60410953e-01
9.96873558e-01 5.19955158e-01 6.58022523e-01 4.35508877e-01
1.91669866e-01 -3.91127281e-02 -4.12289888e-01 7.73143694e-02
-1.97514519e-01 -9.34225470e-02 4.68222648e-01 2.95758635e-01
2.35235453e-01 -6.82435453e-01 -6.24130249e-01 5.44274390e-01
-6.35397434e-01 -2.28744641e-01 5.13294339e-01 5.99701762e-01
9.51246321e-01 8.19128156e-02 1.47495925e-01 -1.56915441e-01
6.49415791e-01 -2.33913586e-01 -8.29232156e-01 1.14541702e-01
-3.84045392e-01 5.33722751e-02 6.47570968e-01 -3.89541119e-01
-5.53998113e-01 4.34911191e-01 -1.78641275e-01 -8.20881963e-01
1.41547665e-01 7.31418192e-01 -1.77599177e-01 -5.27916253e-01
4.92877096e-01 2.79243052e-01 1.83100477e-01 -5.41828096e-01
2.63221025e-01 4.92499769e-01 6.02489114e-01 -6.29845679e-01
1.32773018e+00 5.66984594e-01 1.21444941e-03 -7.12364674e-01
-3.65017235e-01 -7.97917843e-01 -1.02639914e+00 -2.37752944e-01
5.77006102e-01 -9.14163828e-01 -7.37896264e-01 4.35438216e-01
-1.18864667e+00 2.69515932e-01 4.22932953e-02 6.37116015e-01
-4.03348207e-01 8.20341051e-01 -3.01912397e-01 -6.36359215e-01
-4.24527287e-01 -1.44617832e+00 1.25003445e+00 5.93203027e-03
-2.24738970e-01 -7.74660945e-01 2.92872816e-01 2.74733771e-02
1.41528144e-01 4.84500945e-01 3.74055237e-01 -6.59293652e-01
-6.10507786e-01 -3.71818781e-01 -1.98657990e-01 9.69884321e-02
2.00976312e-01 2.71261066e-01 -4.46195543e-01 -5.82626998e-01
1.43953070e-01 3.34198505e-01 5.98789096e-01 2.78060049e-01
9.15994942e-01 -1.39394045e-01 -4.55757201e-01 8.76076877e-01
1.55014145e+00 -3.03845033e-02 6.48097873e-01 7.02445269e-01
8.70412529e-01 1.73325896e-01 8.88642013e-01 4.57007170e-01
3.52313906e-01 9.59108651e-01 3.62565041e-01 -2.10113689e-01
1.27254143e-01 -9.53863859e-02 3.34702402e-01 1.34696615e+00
-3.55148643e-01 5.12665868e-01 -1.02800095e+00 4.23410803e-01
-2.01409912e+00 -8.31852019e-01 -3.48806947e-01 2.51384091e+00
7.29287803e-01 -5.43300323e-02 2.36962046e-02 4.38033380e-02
1.07863092e+00 7.73968361e-03 -2.73010403e-01 5.11469394e-02
3.31539065e-02 1.43406823e-01 4.41394985e-01 5.34395397e-01
-1.11764562e+00 9.55690384e-01 5.65229177e+00 9.29214299e-01
-1.21459460e+00 -1.51309565e-01 1.08228423e-01 3.36072117e-01
-2.42436230e-01 3.28669190e-01 -5.10808527e-01 5.89152515e-01
3.70318115e-01 -1.65786907e-01 3.43784064e-01 8.52625191e-01
4.64127272e-01 7.49232769e-02 -8.10284257e-01 1.17200887e+00
-3.28704566e-02 -1.34924018e+00 9.72990412e-03 8.52316692e-02
6.35937393e-01 1.25038862e-01 -3.10273170e-01 -2.33760208e-01
1.74823046e-01 -5.60580492e-01 7.26270497e-01 6.62292421e-01
5.74994624e-01 -8.12620163e-01 7.66020417e-01 3.50925177e-01
-1.57433450e+00 9.49223340e-02 -7.30927229e-01 3.33222032e-01
1.20691448e-01 7.60520458e-01 -4.65696424e-01 1.01704621e+00
7.62389481e-01 9.16343868e-01 -5.60120046e-01 1.41874349e+00
7.23251402e-02 1.37549922e-01 -6.60701036e-01 3.10353398e-01
3.00344639e-03 -7.56199539e-01 7.62918651e-01 8.66225362e-01
7.32380033e-01 3.26863602e-02 4.82585490e-01 7.93224216e-01
1.73380494e-01 3.73279601e-01 -3.14918935e-01 6.47725999e-01
9.31100070e-01 1.37670195e+00 -9.29821789e-01 4.69168508e-03
-1.64471850e-01 1.07610071e+00 3.35207373e-01 7.10551217e-02
-7.67636478e-01 -2.75431126e-01 5.48200130e-01 7.33383298e-02
1.61879122e-01 -5.93926907e-01 -2.61034191e-01 -1.37379980e+00
2.75593370e-01 -1.00908852e+00 1.52835414e-01 -5.82863152e-01
-1.29499924e+00 5.84108829e-01 -6.33704737e-02 -1.81535220e+00
1.44088864e-01 -2.94249982e-01 -1.06204665e+00 8.20376515e-01
-1.29007602e+00 -1.06598723e+00 -6.52332008e-01 6.58766508e-01
2.80699402e-01 -1.38807967e-01 4.98969764e-01 4.05589134e-01
-7.04298615e-01 4.58863676e-01 2.31661305e-01 1.50781304e-01
8.49747002e-01 -9.17426229e-01 3.99977565e-01 1.18203330e+00
2.30338052e-02 9.75823760e-01 5.73282182e-01 -8.19255710e-01
-1.33760905e+00 -1.04679024e+00 6.56697810e-01 -2.31376380e-01
5.31917512e-01 -1.29137188e-02 -1.04291201e+00 5.10393143e-01
-2.12362707e-01 2.66743749e-02 3.68014783e-01 -1.08042605e-01
-1.26058295e-01 -2.82364458e-01 -1.11130130e+00 6.66167259e-01
8.94337833e-01 -2.25162849e-01 -5.78743994e-01 2.64390498e-01
5.73973119e-01 -4.57467586e-01 -1.10752571e+00 6.87365234e-01
2.98989356e-01 -8.86625528e-01 1.08248317e+00 1.35712162e-01
-3.32727768e-02 -9.17368710e-01 -9.32730436e-02 -1.18860972e+00
-8.07649195e-01 -6.14700437e-01 2.70133138e-01 1.32237065e+00
1.06257886e-01 -7.15597630e-01 4.85769570e-01 2.58009493e-01
-2.42712140e-01 -6.30553305e-01 -1.17028272e+00 -8.97771239e-01
-1.42448589e-01 -2.56860971e-01 8.81150186e-01 1.27287865e+00
-2.93021739e-01 -1.40255302e-01 -1.94478214e-01 6.09622121e-01
8.10177028e-01 2.62223691e-01 9.41096127e-01 -1.59079003e+00
9.23660174e-02 -4.23077673e-01 -6.88095391e-01 -8.71520102e-01
8.43899623e-02 -9.51512575e-01 2.16843650e-01 -1.47439253e+00
1.02786936e-01 -7.98680842e-01 -2.46350989e-01 3.42955738e-01
-3.32115591e-01 2.14236945e-01 2.41719231e-01 9.94021535e-01
-4.21241432e-01 6.11271679e-01 1.00468338e+00 1.61415219e-01
-3.45849961e-01 9.82948244e-02 -4.17049319e-01 7.74011135e-01
6.62370384e-01 -3.70314837e-01 1.15418963e-01 -2.41962537e-01
5.64540504e-03 -2.41060168e-01 2.29445294e-01 -1.26824749e+00
4.07141715e-01 -3.08475673e-01 3.32373798e-01 -7.20563114e-01
6.89574704e-02 -1.10603559e+00 4.49204683e-01 3.82092118e-01
2.22615182e-01 2.29129210e-01 -4.83910292e-02 4.90180671e-01
-2.85873115e-01 -1.72882676e-01 9.33154702e-01 1.22473523e-01
-7.42789149e-01 4.78525519e-01 1.57632649e-01 -5.28386533e-01
1.06681585e+00 -4.87519681e-01 1.39571488e-01 -6.47205561e-02
-6.04486883e-01 3.69935274e-01 9.16526437e-01 3.24808449e-01
7.94635594e-01 -1.81106865e+00 -7.89098024e-01 3.01838100e-01
1.53856769e-01 2.62598515e-01 -7.48452693e-02 1.13086772e+00
-8.44508231e-01 -1.24965303e-01 -9.73967910e-02 -1.06357133e+00
-1.23337388e+00 2.98500389e-01 3.46982807e-01 2.33189408e-02
-6.92076802e-01 6.03348613e-01 -2.71107048e-01 -5.46131372e-01
-5.22111170e-02 -2.24524423e-01 -1.22913919e-01 -6.08203150e-02
2.63201475e-01 4.85054076e-01 1.04775451e-01 -1.11420071e+00
-7.31907189e-01 1.16912735e+00 2.61971980e-01 7.65026212e-02
1.42910016e+00 -4.64608222e-02 -6.62711561e-01 1.64159499e-02
1.28497803e+00 3.38536561e-01 -1.00679779e+00 -3.68154585e-01
2.18912855e-01 -7.40380347e-01 -3.56040969e-02 -1.20348055e-02
-1.22760594e+00 2.81843275e-01 6.91782832e-01 4.56127455e-04
1.15756345e+00 -4.93389070e-02 7.93755651e-01 2.03635097e-01
3.55027288e-01 -9.94073272e-01 -6.14803992e-02 4.42235976e-01
1.01418877e+00 -1.10133433e+00 3.05944473e-01 -6.09917641e-01
-4.39931154e-01 1.18824339e+00 4.85552460e-01 -5.93914807e-01
6.58915997e-01 -3.36762667e-02 1.03524737e-01 -2.61076123e-01
4.56832238e-02 -1.02830708e-01 5.23086071e-01 3.53024572e-01
1.73715428e-01 -1.26587659e-01 -5.05116642e-01 1.81065753e-01
-3.80760252e-01 -1.27550036e-01 2.25546539e-01 7.02785671e-01
-6.32396102e-01 -1.21830773e+00 -7.07356393e-01 1.94358110e-01
4.61000949e-02 1.53932244e-01 -2.31231183e-01 4.38301116e-01
8.88807103e-02 8.25676918e-01 1.33916348e-01 -5.99631727e-01
4.74244475e-01 -2.92192787e-01 1.60198644e-01 -5.00894606e-01
-4.15369093e-01 4.43856388e-01 -4.01710808e-01 -7.99859703e-01
-5.31978786e-01 -7.77452886e-01 -1.30666983e+00 -4.21803236e-01
-6.23998165e-01 7.43085071e-02 5.83902121e-01 9.66636598e-01
5.79140186e-01 -7.08763525e-02 1.07268739e+00 -1.23977995e+00
-4.29160118e-01 -6.68163180e-01 -3.98919374e-01 5.41164875e-01
9.45911929e-02 -7.76913941e-01 -3.81923437e-01 -1.94555238e-01] | [7.722474575042725, -2.935042142868042] |
d45b4e8c-6735-44a4-ab09-5cc3b0002f25 | efficient-traffic-sign-recognition-with-scale | 1805.12289 | null | http://arxiv.org/abs/1805.12289v1 | http://arxiv.org/pdf/1805.12289v1.pdf | Efficient Traffic-Sign Recognition with Scale-aware CNN | The paper presents a Traffic Sign Recognition (TSR) system, which can fast
and accurately recognize traffic signs of different sizes in images. The system
consists of two well-designed Convolutional Neural Networks (CNNs), one for
region proposals of traffic signs and one for classification of each region. In
the proposal CNN, a Fully Convolutional Network (FCN) with a dual multi-scale
architecture is proposed to achieve scale invariant detection. In training the
proposal network, a modified "Online Hard Example Mining" (OHEM) scheme is
adopted to suppress false positives. The classification network fuses
multi-scale features as representation and adopts an "Inception" module for
efficiency. We evaluate the proposed TSR system and its components with
extensive experiments. Our method obtains $99.88\%$ precision and $96.61\%$
recall on the Swedish Traffic Signs Dataset (STSD), higher than
state-of-the-art methods. Besides, our system is faster and more lightweight
than state-of-the-art deep learning networks for traffic sign recognition. | ['Zheng Liu', 'Yuchen Yang', 'Shuo Liu', 'Qiuyuan Wang', 'Wei Ma'] | 2018-05-31 | null | null | null | null | ['traffic-sign-recognition'] | ['computer-vision'] | [ 9.01998207e-02 -3.17668885e-01 -3.85913104e-01 -5.20538390e-01
-4.94116485e-01 1.28120303e-01 3.55294138e-01 -6.52272522e-01
-5.55266023e-01 5.00372350e-01 -4.21428084e-01 -7.44924784e-01
-2.60990039e-02 -7.36122668e-01 -6.23111963e-01 -5.43486357e-01
1.76269725e-01 3.24653983e-01 8.88163447e-01 -2.26720437e-01
2.82433331e-01 9.90484059e-01 -1.93945742e+00 5.83018959e-01
8.23199272e-01 1.50259614e+00 -1.30593836e-01 7.02560008e-01
-3.53640318e-01 1.04490054e+00 -6.76833212e-01 -4.90531772e-01
6.72345638e-01 -2.10693367e-02 -2.34174192e-01 -4.12287377e-02
9.98040795e-01 -4.35227782e-01 -9.30095017e-01 9.06298220e-01
4.36291367e-01 3.62084582e-02 6.46887004e-01 -1.40634072e+00
-4.88682449e-01 4.59432676e-02 -7.52213776e-01 3.29726905e-01
-5.68093598e-01 4.27344352e-01 7.93622911e-01 -8.85683477e-01
4.23033744e-01 9.96841133e-01 7.02799797e-01 6.89414501e-01
-5.81326246e-01 -1.11273146e+00 5.62900007e-02 7.68217444e-01
-1.35225964e+00 -4.19859409e-01 7.39745021e-01 -1.03464797e-01
9.37581837e-01 2.90737629e-01 5.24997592e-01 5.82185030e-01
4.29329351e-02 1.10678053e+00 1.13386607e+00 -4.14067924e-01
4.00496311e-02 1.39582843e-01 5.91674685e-01 1.06533551e+00
6.79005325e-01 3.38504434e-01 -1.85647175e-01 2.65740424e-01
7.94610381e-01 8.28469023e-02 2.30129808e-01 -4.78762798e-02
-7.32362986e-01 5.10808051e-01 6.42454863e-01 2.65770495e-01
-6.00838602e-01 3.84276569e-01 5.23130238e-01 2.09691763e-01
1.64010618e-02 -3.16946536e-01 -3.83378834e-01 8.52832124e-02
-9.47676539e-01 4.33035381e-03 5.23155451e-01 7.43956149e-01
5.95532298e-01 4.54848379e-01 -4.16046292e-01 1.05493903e+00
3.25646043e-01 7.79856741e-01 4.07069772e-01 -6.76664770e-01
5.56605339e-01 9.32483554e-01 -1.57266781e-01 -8.30495894e-01
-3.19841623e-01 -4.16268378e-01 -1.00696826e+00 7.55087554e-01
6.26956880e-01 1.95553806e-02 -1.56820214e+00 9.36317503e-01
1.38382673e-01 4.04259503e-01 -9.12583396e-02 9.66164410e-01
1.00964713e+00 4.75368440e-01 2.11884435e-02 4.69333410e-01
1.52830458e+00 -9.89884377e-01 -3.90658081e-01 -2.14155182e-01
5.24011075e-01 -5.75719893e-01 5.75226367e-01 5.22849500e-01
-7.89314628e-01 -9.00318503e-01 -9.63962197e-01 -7.06044808e-02
-8.06952536e-01 1.06019974e+00 6.08142376e-01 9.45839226e-01
-8.60042751e-01 1.50771454e-01 -3.58184963e-01 -3.48026514e-01
7.71897197e-01 4.31769878e-01 -2.13176891e-01 -1.23689190e-01
-7.57230699e-01 1.08795393e+00 6.73670098e-02 6.63341343e-01
-5.60018301e-01 -2.81988621e-01 -4.25499499e-01 -6.58509359e-02
3.17236096e-01 -1.49193123e-01 9.97317970e-01 -1.07021916e+00
-1.33442640e+00 1.06431067e+00 -2.52330035e-01 -6.86127365e-01
7.53667891e-01 3.53404284e-02 -8.98045599e-01 1.82293743e-01
-1.87208518e-01 6.52087450e-01 9.60883021e-01 -9.25145268e-01
-1.13118398e+00 -1.78345870e-02 -3.06877732e-01 -4.60900217e-01
1.47779271e-01 3.60341579e-01 -5.83114266e-01 -5.42554319e-01
6.65023625e-02 -7.43807614e-01 -7.11405575e-02 5.39245307e-01
-1.17747508e-01 -4.53549922e-01 1.33152294e+00 -7.58316338e-01
1.23307741e+00 -1.98631537e+00 -9.13862586e-01 7.19611406e-01
3.23311001e-01 1.27886653e+00 -3.96231651e-01 -2.89798141e-01
-1.55040324e-01 -2.48096332e-01 1.78277358e-01 -4.75080870e-02
9.07031223e-02 3.87132138e-01 -1.45831317e-01 5.59236944e-01
3.14352810e-01 1.04193175e+00 -2.71048218e-01 -6.11006737e-01
6.88311398e-01 3.41232300e-01 -1.37701169e-01 -3.01528782e-01
2.99488723e-01 -3.67438883e-01 -4.69693720e-01 1.13693833e+00
1.02276826e+00 -8.43191966e-02 -1.06826954e-01 -5.26201248e-01
-4.03012276e-01 -4.49227951e-02 -1.36459172e+00 5.78414381e-01
-2.81020641e-01 1.14113069e+00 6.80627627e-03 -1.32578814e+00
1.52806413e+00 -1.02664880e-01 2.06374034e-01 -1.11076033e+00
3.89833212e-01 5.61626256e-01 1.89392418e-01 -5.57319880e-01
5.35150766e-01 3.60682786e-01 3.72400790e-01 1.96263999e-01
-1.80566773e-01 5.84256589e-01 4.51259941e-01 -6.06460311e-02
1.19309843e+00 -1.72925681e-01 2.23484654e-02 6.63550198e-03
9.65656996e-01 -1.70200586e-01 7.05518246e-01 9.28715348e-01
-7.39129007e-01 1.33757100e-01 2.86733598e-01 -9.38228369e-01
-9.13423002e-01 -8.92592132e-01 -3.43916453e-02 8.82387221e-01
8.28084275e-02 2.92938083e-01 -3.02650571e-01 -9.88725841e-01
3.87521267e-01 3.71728122e-01 -6.32303894e-01 7.27679431e-02
-1.18361592e+00 -4.07306701e-01 9.70389247e-01 1.10471129e+00
1.16031766e+00 -1.06709778e+00 -6.54920816e-01 2.58095982e-03
2.24867240e-01 -1.11394191e+00 -2.34827444e-01 -3.72197814e-02
-6.67886913e-01 -1.42679083e+00 -7.43623912e-01 -1.03228807e+00
7.36673355e-01 4.84216273e-01 5.44006884e-01 3.31230402e-01
-5.81964731e-01 -8.07433426e-02 -3.32995802e-01 -5.01371026e-01
-3.70828331e-01 -1.98072836e-01 -2.42053285e-01 5.40035307e-01
7.95645416e-01 -8.67459029e-02 -7.35326111e-01 7.03805745e-01
-4.72913682e-01 -1.98822692e-01 1.26729977e+00 7.50560641e-01
3.11200351e-01 -3.47575903e-01 6.01715565e-01 -4.27379996e-01
3.23598027e-01 2.16702878e-01 -9.43560660e-01 5.19917488e-01
-5.85318744e-01 -9.76631641e-02 4.66641128e-01 -4.27678138e-01
-9.06148732e-01 2.04918846e-01 -3.50288190e-02 -5.68512678e-01
-5.22203147e-01 -3.11304361e-01 1.60087004e-01 -6.19539499e-01
5.39415836e-01 4.53387588e-01 4.97626606e-04 -5.18176317e-01
3.73240411e-01 1.06864202e+00 7.26007879e-01 -1.54868767e-01
8.61799061e-01 5.62076628e-01 2.34509841e-01 -1.05105591e+00
-4.02372986e-01 -6.30773008e-01 -6.17580652e-01 -7.34854519e-01
6.74562395e-01 -6.18576884e-01 -8.86266291e-01 8.53566587e-01
-9.21420932e-01 -1.92433700e-01 -3.38449031e-01 5.02495110e-01
-6.35611638e-02 2.51967818e-01 -4.36925203e-01 -9.57317948e-01
-4.46520150e-01 -9.62466300e-01 9.09202099e-01 5.39819896e-01
3.15191001e-01 -4.63069469e-01 -4.05462563e-01 5.10128081e-01
8.34150255e-01 -5.55425361e-02 3.70210737e-01 -7.64958024e-01
-9.62163746e-01 -7.97662139e-01 -1.19459808e+00 8.40653419e-01
-1.06259950e-01 2.83819228e-01 -9.60918307e-01 1.70522809e-01
-7.82271504e-01 -9.87508595e-02 1.51302195e+00 6.26137555e-01
1.08246803e+00 -1.41803294e-01 -5.35118461e-01 4.12401348e-01
1.15902340e+00 4.66658771e-01 1.00045347e+00 3.67928326e-01
4.86960232e-01 2.88040340e-01 5.03406465e-01 7.44711831e-02
3.81860495e-01 4.11686957e-01 8.03718716e-02 -3.87836993e-01
-6.03213251e-01 8.50177854e-02 3.32464784e-01 2.76862472e-01
-2.21175134e-01 -1.24648042e-01 -7.33845651e-01 5.17511249e-01
-1.78731346e+00 -1.20978677e+00 -4.86665577e-01 1.77688360e+00
5.29583842e-02 3.01846713e-01 3.54257554e-01 4.67453033e-01
1.04200530e+00 1.54607967e-01 -4.77890134e-01 -6.57397091e-01
-2.46378213e-01 5.14866948e-01 1.05510068e+00 -3.97296548e-02
-1.25965416e+00 9.05835867e-01 5.69525337e+00 1.02230167e+00
-1.45960617e+00 -7.02967867e-02 4.87336636e-01 3.07698160e-01
5.36562979e-01 -3.43496799e-01 -1.04038668e+00 2.91008413e-01
7.16405272e-01 5.59469521e-01 7.89028332e-02 1.10214186e+00
2.81292558e-01 -1.03633299e-01 -4.38051075e-01 9.57995057e-01
1.33637309e-01 -1.55101025e+00 -3.86932679e-02 -3.59230265e-02
5.09562850e-01 3.46748352e-01 3.72719355e-02 4.66928959e-01
2.38356218e-01 -6.69427931e-01 6.54242218e-01 7.31918693e-01
9.55908537e-01 -6.87306523e-01 9.28671956e-01 4.59322184e-02
-1.70942223e+00 -5.67948937e-01 -3.76288623e-01 2.63785690e-01
2.07699656e-01 4.16552871e-01 -5.08586764e-01 3.00780833e-01
5.50951421e-01 4.92418945e-01 -8.29728842e-01 1.51387739e+00
-1.73869580e-01 7.63912916e-01 -4.80375946e-01 -3.76537651e-01
4.94970828e-01 -1.15308568e-01 3.26339513e-01 1.55920327e+00
1.20761663e-01 -1.30000845e-01 -5.86124063e-02 6.76629424e-01
-9.38457549e-02 -5.76444343e-02 -4.47082371e-01 3.76830786e-01
3.93261760e-01 1.25781763e+00 -8.45569015e-01 -8.90126288e-01
-4.17483956e-01 6.05207384e-01 -9.71472338e-02 2.09276006e-01
-1.09045672e+00 -9.97467518e-01 4.36867744e-01 4.87941541e-02
8.20637286e-01 -5.62121347e-02 -3.81838799e-01 -9.21011746e-01
2.89106965e-01 -6.07776463e-01 4.93369728e-01 -7.60125518e-01
-1.15543747e+00 2.73437858e-01 -3.55121940e-01 -1.59938908e+00
4.50824767e-01 -1.16308653e+00 -1.01316941e+00 5.83172023e-01
-1.95468760e+00 -1.38002884e+00 -5.10950267e-01 5.07874310e-01
4.46232915e-01 -5.02252817e-01 2.08231285e-02 7.69538820e-01
-8.77097368e-01 1.03984118e+00 -4.70435284e-02 7.13156164e-01
5.49245834e-01 -8.38706195e-01 4.80756640e-01 1.05755639e+00
-1.80869460e-01 1.96077958e-01 -1.16661593e-01 -5.85707784e-01
-1.15792668e+00 -1.56577766e+00 1.05364907e+00 3.11508309e-02
5.09948790e-01 3.38977464e-02 -6.71513438e-01 4.44734693e-01
-3.67945552e-01 4.27337587e-01 3.54166552e-02 -4.58146095e-01
-7.41257548e-01 -8.82103741e-01 -1.35243726e+00 4.68081623e-01
9.93700385e-01 -2.65014172e-01 -5.13017833e-01 -1.06961466e-01
-6.11740910e-02 -1.56381369e-01 -2.29600966e-01 5.17468750e-01
1.04027534e+00 -7.30945826e-01 9.65545654e-01 -3.81818682e-01
-1.39220521e-01 -7.20040739e-01 -7.08034709e-02 -4.25379902e-01
-1.58860117e-01 -2.72998184e-01 -2.52812713e-01 9.07029569e-01
4.83478367e-01 -9.97163355e-01 1.04744709e+00 4.52523649e-01
-4.19714928e-01 -5.76477230e-01 -1.28820264e+00 -1.03643608e+00
-4.14067298e-01 -6.85498595e-01 4.15758759e-01 2.32617259e-01
-6.96954727e-01 -5.18280789e-02 -3.37402791e-01 1.76658541e-01
7.72553563e-01 -7.69142881e-02 9.42553997e-01 -1.17886007e+00
2.13154703e-01 -8.78208756e-01 -9.12906528e-01 -1.12590003e+00
-7.87807256e-02 -5.66514730e-01 -4.67132777e-03 -1.40203965e+00
-2.50371605e-01 -4.48530912e-01 -5.53508162e-01 7.77830064e-01
3.52265745e-01 5.63247204e-01 2.04411238e-01 -1.22188196e-01
-8.05605471e-01 1.71983182e-01 1.16209590e+00 -2.21864402e-01
-6.16227612e-02 2.02419803e-01 -2.92897403e-01 7.67658055e-01
8.41892898e-01 -1.64065987e-01 2.71513492e-01 -1.42261349e-02
-4.72308308e-01 -4.86919045e-01 7.86859572e-01 -1.20977139e+00
4.91645753e-01 6.03823625e-02 5.02612591e-01 -1.28767860e+00
1.91156715e-01 -7.49290049e-01 -4.19826359e-01 8.44215393e-01
-1.75970495e-01 -1.04389504e-01 2.40076393e-01 4.48802203e-01
-8.94165933e-02 1.39081433e-01 1.17973053e+00 1.77387759e-01
-1.15741539e+00 1.45435542e-01 -6.30619407e-01 -4.08364266e-01
9.57455277e-01 -8.05513442e-01 -6.19002163e-01 2.08592694e-02
-2.43967459e-01 1.79288536e-01 -2.37217829e-01 5.86453736e-01
9.58320558e-01 -1.35367882e+00 -6.21168911e-01 5.14970720e-01
1.66776463e-01 -4.46801841e-01 3.83342415e-01 9.63643849e-01
-8.62261653e-01 5.74523211e-01 -2.78669387e-01 -4.59334671e-01
-1.48344028e+00 2.17364710e-02 5.63806295e-01 -1.81281239e-01
-7.21879005e-01 5.57519913e-01 -4.39625442e-01 -2.55382776e-01
5.87772191e-01 -6.89877152e-01 -1.80844948e-01 -4.17093299e-02
7.51298964e-01 9.81562555e-01 2.21703276e-01 -7.77592063e-01
-5.45219302e-01 9.59654212e-01 -1.68764591e-01 2.93169767e-01
1.15241170e+00 3.11816782e-01 2.33790427e-01 -3.70186687e-01
9.81416702e-01 -4.73439008e-01 -1.04484785e+00 -2.51371145e-01
-7.96257183e-02 -3.54747444e-01 2.90677138e-02 -9.93612349e-01
-1.45070255e+00 8.73127997e-01 1.06777453e+00 -1.90440208e-01
1.13020277e+00 -3.38095844e-01 1.06952965e+00 8.24650645e-01
-4.90860976e-02 -1.51136649e+00 -1.24197654e-01 6.49985492e-01
6.84944153e-01 -1.28142023e+00 -2.59166688e-01 6.54391125e-02
-5.77306390e-01 1.45794868e+00 9.20759797e-01 -5.48991501e-01
8.12584162e-01 1.57616749e-01 2.94008940e-01 -2.40672946e-01
-6.29015148e-01 -8.14776361e-01 6.84673786e-01 6.58610344e-01
-2.28886276e-01 1.33702204e-01 -4.90629524e-01 2.14749381e-01
3.94536436e-01 4.33971673e-01 1.55792236e-01 8.89136195e-01
-8.10463190e-01 -7.39255846e-01 -4.07910436e-01 7.66193330e-01
-9.39596146e-02 2.62180656e-01 -4.76465493e-01 1.09267545e+00
5.90775311e-01 7.80837059e-01 1.91801399e-01 -6.54349446e-01
6.37841582e-01 1.66278109e-01 2.54368186e-01 1.13592155e-01
-5.71881473e-01 -2.75962472e-01 3.26943099e-01 -8.01933944e-01
-2.49867216e-01 -5.02717853e-01 -1.11234021e+00 -3.56990278e-01
-6.31188095e-01 -3.24479550e-01 7.50904620e-01 8.86301279e-01
3.87926251e-01 5.34128547e-01 6.04333282e-01 -6.22159898e-01
-4.25677180e-01 -8.51518810e-01 -4.94636565e-01 8.72439295e-02
2.88682371e-01 -5.66630721e-01 -3.37621346e-02 -1.84882775e-01] | [7.984241008758545, -0.8076330423355103] |
dd624ff6-dbf3-4f1c-a49c-86e4eb862ec5 | image-feature-information-extraction-for | 2106.07929 | null | https://arxiv.org/abs/2106.07929v5 | https://arxiv.org/pdf/2106.07929v5.pdf | Image Feature Information Extraction for Interest Point Detection: A Review | Interest point detection is one of the most fundamental and critical problems in computer vision and image processing. In this paper, we carry out a comprehensive review on image feature information (IFI) extraction techniques for interest point detection. To systematically introduce how the existing interest point detection methods extract IFI from an input image, we propose a taxonomy of the IFI extraction techniques for interest point detection. According to this taxonomy, we discuss different types of IFI extraction techniques for interest point detection. Furthermore, we identify the main unresolved issues related to the existing IFI extraction techniques for interest point detection and any interest point detection methods that have not been discussed before. The existing popular datasets and evaluation standards are provided and the performances for eighteen state-of-the-art approaches are evaluated and discussed. Moreover, future research directions on IFI extraction techniques for interest point detection are elaborated. | ['Changming Sun', 'Yongsheng Gao', 'Weichuan Zhang', 'Tian Gao', 'Junfeng Jing'] | 2021-06-15 | null | null | null | null | ['interest-point-detection'] | ['computer-vision'] | [ 1.45543233e-01 -3.82604927e-01 -4.85856265e-01 1.14004351e-01
-5.67047775e-01 -3.34695935e-01 5.99766135e-01 -1.17922418e-01
-1.01203747e-01 2.54791349e-01 -2.60179311e-01 1.18980847e-01
-2.50782698e-01 -6.49224937e-01 -1.99976042e-01 -4.76783991e-01
-3.89895111e-01 4.06535976e-02 6.61442995e-01 -7.08241463e-02
1.14860022e+00 1.17280650e+00 -1.87999797e+00 6.39808327e-02
5.50145268e-01 1.30405831e+00 1.50952101e-01 5.95108569e-01
1.36024430e-02 2.94564664e-01 -8.38463068e-01 2.26396590e-01
4.31945205e-01 -1.73807189e-01 -7.27271736e-01 3.15758288e-01
7.54524946e-01 -3.88720661e-01 -2.92228103e-01 8.93487811e-01
5.23853123e-01 2.26698384e-01 1.27843559e+00 -1.68803740e+00
-2.01894701e-01 -9.70507488e-02 -1.08020473e+00 7.88058698e-01
2.49360129e-01 -2.81349063e-01 6.33312702e-01 -1.62713981e+00
8.23219061e-01 7.49834955e-01 5.67646384e-01 1.34361148e-01
-5.83749652e-01 -5.53069115e-01 -9.96231362e-02 6.50157630e-01
-1.58533740e+00 -2.75767118e-01 1.11505902e+00 -3.50064844e-01
9.19182062e-01 5.00693858e-01 8.03328276e-01 2.13067845e-01
5.46269238e-01 1.11683118e+00 6.33573830e-01 -8.51887524e-01
-7.39440024e-02 -5.15017696e-02 4.58739817e-01 8.18076432e-01
3.75639200e-01 3.62216562e-01 -4.48177457e-01 -2.02692434e-01
1.25646222e+00 -1.58280998e-01 -3.80527020e-01 -8.83161128e-01
-1.44861460e+00 8.65901411e-01 6.00235462e-01 5.05130768e-01
-1.60381794e-01 -2.80222028e-01 4.00427133e-01 5.64896548e-03
3.15017790e-01 5.28250337e-01 -1.59264848e-01 5.70733964e-01
-1.26175630e+00 1.53988898e-01 3.94581527e-01 1.15944266e+00
6.47134900e-01 -1.09292567e-01 -2.50913799e-01 9.04388726e-01
4.17229325e-01 3.52366418e-01 4.05964464e-01 -7.82540858e-01
2.03194201e-01 6.99555635e-01 -2.57325824e-02 -1.32782006e+00
-5.54293931e-01 -1.45711422e-01 -4.38183725e-01 7.99133480e-01
1.88032165e-01 1.46433026e-01 -7.64234006e-01 3.62743795e-01
4.43711758e-01 4.40058023e-01 -1.78018719e-01 7.60601044e-01
1.56264484e+00 7.56277442e-01 -3.04235727e-01 -2.11485714e-01
1.77646887e+00 -9.87010181e-01 -4.12305474e-01 -3.77761275e-02
2.31601268e-01 -1.01007080e+00 6.02608979e-01 3.27366024e-01
-1.03991008e+00 -5.97691178e-01 -1.09375286e+00 -2.50898629e-01
-2.96697944e-01 5.76820016e-01 5.29366016e-01 2.88788587e-01
-6.43433094e-01 -1.52968150e-02 -4.24566418e-01 -5.78344285e-01
3.59516323e-01 4.96277094e-01 -3.65640044e-01 3.74042064e-01
-7.91456401e-01 1.02814543e+00 2.53895223e-01 -1.31070197e-01
-3.09937298e-01 -8.15897465e-01 -7.96410620e-01 -2.52045095e-01
4.03323740e-01 -7.95623600e-01 1.12542927e+00 -3.92171234e-01
-1.21197855e+00 1.13278437e+00 -3.44737828e-01 -3.31409097e-01
1.51931241e-01 -1.33406416e-01 -4.80166376e-01 6.67185724e-01
2.34023005e-01 6.55703545e-01 1.07149041e+00 -1.31493592e+00
-1.60104823e+00 -2.37069443e-01 -5.76529689e-02 4.05884743e-01
2.39762932e-01 4.12569970e-01 -4.44932580e-01 -7.31439233e-01
4.20144975e-01 -4.72277075e-01 4.44111079e-02 4.62617964e-01
-3.45977247e-01 -7.45698571e-01 1.43864858e+00 -5.68867847e-02
1.13193715e+00 -1.99446189e+00 -4.22421664e-01 4.86047328e-01
3.83999348e-01 1.79872304e-01 1.89171329e-01 3.85230571e-01
-2.61355013e-01 -1.21200874e-01 1.02126017e-01 3.10970277e-01
-4.82497901e-01 -4.08432633e-01 -1.85598582e-01 8.74129236e-01
-1.01377241e-01 8.86212766e-01 -5.13278425e-01 -9.75016594e-01
1.11639941e+00 5.02019882e-01 -5.90692163e-02 -8.24138522e-02
6.90225363e-01 2.20884215e-02 -5.80045104e-01 1.12341726e+00
8.20053637e-01 -8.21535066e-02 -6.93432391e-01 -7.99956083e-01
-5.45306563e-01 -1.64033175e-01 -1.44908226e+00 1.16038883e+00
-1.67918101e-01 8.36527526e-01 -1.95565239e-01 -9.32425141e-01
9.87787724e-01 2.91715264e-01 8.47735703e-01 -1.73712492e-01
1.88644290e-01 2.90765405e-01 -2.42958978e-01 -3.09907407e-01
5.97961724e-01 2.70978101e-02 8.70904773e-02 -1.54730314e-02
1.00655124e-01 -4.08279419e-01 2.99219489e-01 -1.13826416e-01
5.01869082e-01 -2.36355707e-01 1.32531631e+00 -3.59568715e-01
1.33417237e+00 4.32828635e-01 1.00123413e-01 6.91874802e-01
-5.34156621e-01 8.95550966e-01 -5.68287931e-02 -5.44794679e-01
-5.09370208e-01 -1.00143266e+00 -7.14118361e-01 6.71265960e-01
6.94016159e-01 -2.35842556e-01 -5.27937472e-01 -9.76871848e-01
-1.23973802e-01 1.40643045e-01 -6.51317418e-01 3.25029314e-01
-6.69449329e-01 -4.89518374e-01 1.96392089e-01 3.93051773e-01
6.54252172e-01 -1.22346330e+00 -9.98347521e-01 -1.62638783e-01
-3.09393644e-01 -8.54449272e-01 -4.47224468e-01 -2.86739409e-01
-1.10463583e+00 -1.47676849e+00 -1.03776014e+00 -1.24779892e+00
1.08365464e+00 9.71838236e-01 1.15531695e+00 2.61349589e-01
-7.47320950e-01 5.72301924e-01 -3.67925167e-01 -5.16724825e-01
3.06189090e-01 -5.69228269e-02 -2.17013255e-01 -3.05092543e-01
5.20355880e-01 8.50319415e-02 -9.12275851e-01 6.47975385e-01
-4.90517437e-01 -4.69132364e-01 5.84209502e-01 8.68338168e-01
7.06683636e-01 1.73105314e-01 3.50096077e-02 -4.39580113e-01
4.03770208e-01 -1.04506291e-01 -7.67837644e-01 7.99160898e-02
-3.80586982e-01 -3.56133759e-01 -3.46239358e-02 -4.65477258e-02
-9.41455901e-01 1.43187299e-01 -1.19298406e-01 -2.03414053e-01
-4.90800768e-01 4.33141962e-02 3.33800241e-02 -9.93880749e-01
8.08488667e-01 2.78808773e-01 -6.30580634e-02 -2.42804155e-01
5.07093191e-01 7.71343529e-01 6.13433540e-01 -1.37951374e-01
7.85739183e-01 9.01433647e-01 2.28140727e-01 -1.42404532e+00
-5.15969276e-01 -1.31912851e+00 -1.09199047e+00 -4.25667226e-01
3.56945723e-01 -4.15001184e-01 -4.28656936e-01 3.95549715e-01
-1.11468756e+00 5.73957622e-01 -5.48747063e-01 4.92961645e-01
-9.51034248e-01 7.73791909e-01 -3.06911409e-01 -4.66364294e-01
-7.66573310e-01 -1.12756777e+00 1.11290467e+00 5.10993719e-01
-1.53183311e-01 -9.48546112e-01 4.48921230e-03 1.55594826e-01
-7.34859407e-02 1.88956767e-01 5.04939675e-01 -3.30692530e-01
-4.56021994e-01 -6.00236893e-01 -4.41431701e-01 -1.78096533e-01
3.45180273e-01 1.99815333e-01 -6.08799696e-01 -2.48554230e-01
1.85142085e-01 4.94305313e-01 5.30479670e-01 1.02926469e+00
8.80116403e-01 -1.75154358e-02 -1.09808481e+00 7.35669672e-01
1.47390914e+00 4.40685749e-01 8.99743795e-01 9.08808768e-01
3.36269140e-01 7.42080390e-01 1.17556739e+00 2.32016936e-01
-8.95330235e-02 8.56381476e-01 3.09797674e-01 -1.91621721e-01
-3.89115274e-01 1.33300364e-01 -1.54486284e-01 3.51254553e-01
-5.33630908e-01 -1.05040614e-02 -7.01083124e-01 4.60479796e-01
-1.29951048e+00 -1.10731220e+00 -7.16335952e-01 2.01110101e+00
9.69706550e-02 -2.89806068e-01 4.29082751e-01 5.93148649e-01
8.03851128e-01 -8.23883638e-02 -3.45901310e-01 2.14907676e-01
-1.40688578e-02 1.83707699e-01 5.09198725e-01 1.78135410e-01
-1.73065758e+00 7.59045601e-01 7.78187847e+00 9.24542665e-01
-1.33342433e+00 -2.66041636e-01 1.86351120e-01 3.82199973e-01
5.97163200e-01 -2.44270280e-01 -1.13958347e+00 1.04419790e-01
1.10944599e-01 -4.25939173e-01 -5.94359457e-01 1.14141226e+00
3.62305753e-02 -4.15890396e-01 -6.55547082e-01 1.54685879e+00
4.00855780e-01 -1.36729479e+00 7.98956603e-02 -4.90981266e-02
4.54890013e-01 -9.24425647e-02 1.67620614e-01 -2.87022620e-01
-6.28513753e-01 -5.95992923e-01 3.48887116e-01 1.59825608e-01
7.90525615e-01 -9.37426865e-01 7.41188169e-01 1.73488125e-01
-1.62895334e+00 2.59383135e-02 -6.01431787e-01 3.17314297e-01
2.24054053e-01 5.29631674e-01 -8.17541361e-01 6.47045672e-01
6.75534725e-01 1.00571561e+00 -7.03372836e-01 1.90566742e+00
-2.02219948e-01 1.31189972e-01 -5.05475044e-01 -1.88900309e-03
1.56609118e-01 -3.18772569e-02 8.73549044e-01 1.22278857e+00
6.27133965e-01 8.02662224e-02 -2.13502944e-01 6.72390938e-01
4.48739082e-01 5.07301033e-01 -8.68278205e-01 6.27226412e-01
4.90623146e-01 1.53099155e+00 -1.33972168e+00 -3.29725742e-01
-6.81227505e-01 7.95860529e-01 -3.05162102e-01 9.65299979e-02
-5.21732211e-01 -9.49046493e-01 4.43331659e-01 5.72763324e-01
2.36884981e-01 -2.56900758e-01 -4.02329385e-01 -1.07620168e+00
-3.12654942e-01 -4.00556386e-01 6.21678889e-01 -5.85339010e-01
-1.25377321e+00 2.78551936e-01 6.27788663e-01 -2.11102343e+00
-2.88461715e-01 -9.00362074e-01 -7.93108582e-01 5.89130044e-01
-1.62228298e+00 -1.13197649e+00 -3.70584756e-01 6.39119327e-01
1.16522312e+00 -2.70396203e-01 3.24193865e-01 -1.18469104e-01
-1.92709789e-01 3.22999239e-01 2.68540978e-02 2.63714045e-01
5.20820618e-01 -8.99157941e-01 6.51593626e-01 9.80508745e-01
3.54506403e-01 6.98111355e-01 4.02966648e-01 -4.44225699e-01
-1.21740949e+00 -6.71011925e-01 6.92506015e-01 -4.47321862e-01
2.02099159e-01 1.16540618e-01 -8.74139190e-01 5.25648952e-01
9.53561217e-02 5.06692588e-01 4.82773274e-01 -3.57535869e-01
2.09428951e-01 1.66810185e-01 -1.40874839e+00 4.61803675e-01
9.37636673e-01 8.80286843e-02 -1.09062886e+00 -1.97139066e-02
-2.37423241e-01 -4.97144073e-01 -6.99552774e-01 7.23684371e-01
6.46084785e-01 -9.76500809e-01 1.85929441e+00 1.81118757e-01
-2.15727761e-01 -7.34116495e-01 2.07792476e-01 -9.67237651e-01
-7.51754224e-01 2.63406858e-02 -3.20040375e-01 7.43330717e-01
-1.34119898e-01 -5.98929524e-01 1.08165586e+00 -3.27176531e-03
-6.39087614e-03 -6.15366936e-01 -9.59886730e-01 -7.36853600e-01
-1.39489770e-01 -1.54106632e-01 2.18038142e-01 5.47899425e-01
2.90514559e-01 2.88692027e-01 -2.34763157e-02 9.93803144e-02
1.06737304e+00 2.97021508e-01 8.76074612e-01 -1.62751865e+00
5.69419920e-01 -5.78556657e-01 -9.86593783e-01 -1.33434427e+00
-2.68320620e-01 -5.64044774e-01 -1.70902878e-01 -1.74742115e+00
1.90971997e-02 -3.26617718e-01 -1.10125504e-01 2.13187989e-02
-9.02681276e-02 7.46340394e-01 3.09492975e-01 6.24368370e-01
-1.99849039e-01 -9.18569565e-02 1.20898890e+00 -9.55465659e-02
-2.16246337e-01 4.17669028e-01 -4.89154160e-01 1.15445685e+00
6.71872914e-01 -1.15773559e-01 -4.90844369e-01 2.36784890e-01
-3.71724725e-01 -2.55873561e-01 3.52351636e-01 -1.31782341e+00
4.36467141e-01 -2.23070562e-01 6.51582956e-01 -1.67570257e+00
2.92105824e-01 -1.00430179e+00 -2.50024915e-01 4.81195480e-01
3.41319412e-01 -1.38954461e-01 1.23791985e-01 2.65543461e-01
-4.40731108e-01 -6.93751931e-01 1.13512862e+00 -2.72055149e-01
-1.56055212e+00 2.36728221e-01 -5.60196698e-01 -1.07802778e-01
1.69249654e+00 -9.24236417e-01 -7.74455070e-02 -2.28607208e-02
-4.49573368e-01 -3.39090712e-02 2.54711807e-01 5.59942842e-01
1.31831121e+00 -1.36375749e+00 -7.14157939e-01 6.34424508e-01
4.08290267e-01 -1.45494625e-01 1.91545814e-01 8.12454879e-01
-8.03969860e-01 7.62123764e-01 -5.21521091e-01 -8.16362917e-01
-1.86439717e+00 7.59978354e-01 4.06216413e-01 1.71914846e-01
-9.63241756e-01 7.01156318e-01 6.66988969e-01 5.21185212e-02
-7.48738050e-02 -4.79394674e-01 -9.54901457e-01 9.71143618e-02
7.11043060e-01 7.96595037e-01 2.77629912e-01 -1.14109731e+00
-5.78567982e-01 1.62649632e+00 -2.89282352e-02 2.72921354e-01
7.64656544e-01 -2.98279554e-01 2.38063708e-02 3.00854027e-01
1.07500637e+00 -9.83002335e-02 -6.28716826e-01 7.79093802e-02
-5.07632606e-02 -9.15986717e-01 2.40008920e-01 -3.79446089e-01
-9.22159612e-01 8.78351331e-01 8.12588871e-01 1.82765692e-01
1.19194531e+00 1.75486609e-01 4.51837033e-01 -1.48054520e-02
8.60415697e-01 -8.72200489e-01 -2.25449830e-01 3.70472968e-01
1.27082276e+00 -1.19867408e+00 4.88482922e-01 -1.20563352e+00
-2.02581465e-01 1.52780282e+00 4.49065179e-01 -5.63268483e-01
1.07195520e+00 2.64904171e-01 -1.95565391e-02 -3.50654215e-01
-3.59679759e-02 -3.78392279e-01 9.62490082e-01 1.20479643e+00
4.44524109e-01 -3.76246154e-01 -6.44563377e-01 1.94242150e-02
-7.68507421e-02 1.09744281e-01 2.29021057e-01 8.85520458e-01
-8.06134224e-01 -9.39642072e-01 -1.06049192e+00 6.38071120e-01
-4.39737439e-01 2.08946705e-01 -3.26623976e-01 1.24904001e+00
1.31547183e-01 7.47767806e-01 2.94199377e-01 -7.63488263e-02
7.93962479e-01 -5.12876153e-01 5.35505652e-01 -5.81420422e-01
-2.62831032e-01 8.34055543e-02 -3.31296414e-01 -3.30640435e-01
-5.02389371e-01 -7.31815279e-01 -1.41476536e+00 3.74455377e-02
-8.62482011e-01 2.10361660e-01 5.29908597e-01 5.30515671e-01
1.48044005e-01 1.48250982e-01 5.36141574e-01 -1.33599079e+00
-2.49905437e-02 -6.43988192e-01 -7.60990560e-01 -2.50012372e-02
3.52413207e-01 -1.30960786e+00 -5.98208368e-01 1.71533883e-01] | [8.791410446166992, -1.479999303817749] |
ae1b4b92-fee4-415c-a0ea-276bd3187408 | nuclei-glands-instance-segmentation-in | 2208.12460 | null | https://arxiv.org/abs/2208.12460v1 | https://arxiv.org/pdf/2208.12460v1.pdf | Nuclei & Glands Instance Segmentation in Histology Images: A Narrative Review | Instance segmentation of nuclei and glands in the histology images is an important step in computational pathology workflow for cancer diagnosis, treatment planning and survival analysis. With the advent of modern hardware, the recent availability of large-scale quality public datasets and the community organized grand challenges have seen a surge in automated methods focusing on domain specific challenges, which is pivotal for technology advancements and clinical translation. In this survey, 126 papers illustrating the AI based methods for nuclei and glands instance segmentation published in the last five years (2017-2022) are deeply analyzed, the limitations of current approaches and the open challenges are discussed. Moreover, the potential future research direction is presented and the contribution of state-of-the-art methods is summarized. Further, a generalized summary of publicly available datasets and a detailed insights on the grand challenges illustrating the top performing methods specific to each challenge is also provided. Besides, we intended to give the reader current state of existing research and pointers to the future directions in developing methods that can be used in clinical practice enabling improved diagnosis, grading, prognosis, and treatment planning of cancer. To the best of our knowledge, no previous work has reviewed the instance segmentation in histology images focusing towards this direction. | ['Muhammad Moazam Fraz', 'Arshi Perviaz', 'Esha Sadia Nasir'] | 2022-08-26 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [ 3.64015907e-01 3.55367452e-01 -4.48454857e-01 -2.37410083e-01
-1.24164522e+00 -4.31471020e-01 2.62915611e-01 6.80648804e-01
-3.20296228e-01 7.50384569e-01 1.37359604e-01 -1.80319443e-01
-2.86370337e-01 -4.90913957e-01 9.74207297e-02 -1.35034597e+00
-1.52713478e-01 5.95630884e-01 2.14883342e-01 -7.89310187e-02
1.97054371e-01 7.20936120e-01 -1.00889623e+00 4.86968935e-01
5.72025537e-01 7.51009703e-01 2.74180084e-01 8.67550552e-01
-1.24126747e-01 3.52007866e-01 -2.83327341e-01 -4.37048763e-01
-1.23131111e-01 -3.71990591e-01 -1.08788586e+00 1.84940681e-01
2.87993133e-01 1.20380290e-01 -7.89941028e-02 9.62386608e-01
7.29846656e-01 -3.97605926e-01 7.59208262e-01 -9.08703208e-01
-5.27217202e-02 5.28880715e-01 -4.63942230e-01 3.89714390e-01
-1.48067132e-01 1.95089415e-01 6.15107656e-01 -4.39703375e-01
1.12771940e+00 4.03458804e-01 7.71679342e-01 1.05451012e+00
-8.46642733e-01 -6.71475306e-02 -1.27547562e-01 1.26673982e-01
-1.31174695e+00 -1.29759550e-01 2.84194708e-01 -4.00740564e-01
9.48326945e-01 7.53596723e-01 7.38605440e-01 7.43258238e-01
5.87780118e-01 8.11417878e-01 1.10537922e+00 -2.50671715e-01
2.51286775e-01 2.23504856e-01 1.18300632e-01 7.08508849e-01
2.43659183e-01 -3.85642052e-01 -1.96870282e-01 -1.55672818e-01
2.72021890e-01 3.71018648e-02 -3.53584796e-01 -1.27466843e-01
-1.57941532e+00 5.25113404e-01 6.22144938e-01 7.77709484e-01
-3.77047504e-03 3.14532556e-02 8.44200313e-01 -2.56153703e-01
5.09194314e-01 7.71244094e-02 -4.01576340e-01 1.63132370e-01
-1.27271867e+00 5.59351668e-02 5.60909986e-01 5.75884879e-01
-1.85438991e-02 -4.61299181e-01 -2.86076188e-01 5.62221169e-01
3.58400613e-01 6.51346073e-02 7.83630013e-01 -7.12121487e-01
-2.96054304e-01 7.53544688e-01 -4.68186438e-01 -3.30801904e-01
-1.05809617e+00 -6.04314387e-01 -1.24925864e+00 -7.27852583e-02
5.86663723e-01 2.12918177e-01 -1.06079459e+00 8.33631516e-01
5.05397201e-01 2.19469875e-01 -9.65576321e-02 7.90187716e-01
1.26115203e+00 -2.20282488e-02 2.17879936e-01 -3.74814183e-01
1.87407601e+00 -8.44443619e-01 -8.72644484e-01 2.09741771e-01
1.07685935e+00 -7.92751968e-01 3.56736392e-01 2.69922823e-01
-1.05822039e+00 1.31567433e-01 -6.45855725e-01 -2.36269489e-01
-7.33280778e-01 2.65332460e-01 8.26222718e-01 7.87170053e-01
-1.25233865e+00 4.99608368e-01 -1.33418190e+00 -1.09024894e+00
9.04333472e-01 7.27742553e-01 -4.36080813e-01 -6.46940544e-02
-5.19866467e-01 1.02811396e+00 8.06996077e-02 8.86552408e-02
-7.57711172e-01 -1.18494129e+00 -4.38506097e-01 -5.52430689e-01
9.34838206e-02 -8.90645385e-01 1.32598221e+00 -2.77847141e-01
-1.23786020e+00 1.62960303e+00 -3.16325486e-01 -4.35119539e-01
5.84207118e-01 7.29621649e-01 -4.57216054e-02 2.85345972e-01
-1.32303968e-01 8.81593525e-01 -1.10041767e-01 -8.19042265e-01
-1.02669132e+00 -6.28630877e-01 -4.72808093e-01 1.46254227e-01
-3.11618358e-01 1.11533247e-01 -6.48450911e-01 -4.43798572e-01
6.94061369e-02 -9.39730048e-01 -7.34464884e-01 3.98756057e-01
-3.71818662e-01 -1.91600248e-01 6.70763969e-01 -6.63527966e-01
1.05671930e+00 -1.94548810e+00 2.40846142e-01 2.38772020e-01
3.83737177e-01 -3.56517104e-03 3.29996824e-01 3.38747889e-01
2.92424671e-02 3.49067539e-01 -3.57230902e-01 -5.09317994e-01
-2.91368276e-01 1.65248930e-01 3.09378475e-01 9.98778045e-01
-3.18338662e-01 1.17830515e+00 -9.37878430e-01 -6.58192813e-01
2.71223038e-01 4.74270284e-01 -4.53077815e-02 -2.14052483e-01
9.94102005e-03 7.47801721e-01 -2.23353967e-01 1.16460812e+00
5.56255400e-01 -3.69476080e-01 3.77502680e-01 -2.67300993e-01
-1.07445613e-01 1.54427066e-02 -5.08154809e-01 1.57004488e+00
-3.45518216e-02 4.41203117e-01 4.38731104e-01 -7.62507439e-01
4.85978931e-01 6.52683318e-01 9.17692602e-01 -1.05887979e-01
3.57162088e-01 5.09711206e-01 1.73643589e-01 -3.23076934e-01
8.21195841e-02 -2.39839762e-01 2.33948633e-01 -3.32551263e-02
-4.23052832e-02 -5.02176285e-01 4.68274236e-01 2.27691591e-01
1.19328451e+00 -2.72542894e-01 5.98306179e-01 -3.21273476e-01
8.95437717e-01 4.46477145e-01 3.28606904e-01 2.20110565e-01
-7.86815107e-01 6.91607535e-01 5.41928649e-01 -3.94796968e-01
-7.95474589e-01 -5.60490251e-01 -5.29329062e-01 4.71835107e-01
-1.84812874e-01 -4.14906621e-01 -8.26787949e-01 -9.27324653e-01
-1.15205057e-01 5.85070951e-03 -1.09275031e+00 3.52977365e-01
-5.14024019e-01 -1.45484209e+00 6.36658132e-01 3.98462355e-01
2.10634410e-01 -7.36074507e-01 -4.11173046e-01 -3.28573585e-02
-2.83004522e-01 -1.10508645e+00 8.43775123e-02 7.08156601e-02
-1.36850655e+00 -1.23859739e+00 -1.17771399e+00 -1.07277572e+00
1.10292339e+00 1.22309349e-01 9.60095644e-01 4.44396913e-01
-9.78615761e-01 2.21839905e-01 -1.16720751e-01 -5.02943575e-01
-4.42815304e-01 1.40668973e-01 -4.62193400e-01 -4.73480552e-01
2.81256795e-01 -5.85125051e-02 -8.10992777e-01 9.19234380e-02
-9.67612922e-01 2.35669166e-01 5.99633396e-01 8.13081205e-01
1.08288276e+00 -4.85170707e-02 1.60753459e-01 -1.17280519e+00
5.65372333e-02 -4.03402567e-01 -2.69186407e-01 1.91347867e-01
-3.89232993e-01 -5.24154603e-01 2.09336519e-01 2.47988462e-01
-6.98495090e-01 2.19528839e-01 -4.81160909e-01 4.08288687e-01
-2.87865520e-01 4.29289460e-01 3.42156500e-01 -5.39603293e-01
5.99928260e-01 -1.27464920e-01 2.23731473e-01 -2.35048354e-01
-1.13939948e-01 4.76646006e-01 3.28600347e-01 4.74813357e-02
4.11947012e-01 1.13316226e+00 6.87022209e-01 -5.68708122e-01
-8.44570458e-01 -9.32060957e-01 -6.54490292e-01 -3.33688408e-01
8.62721264e-01 -3.62309247e-01 -5.19900978e-01 6.82813764e-01
-7.24102914e-01 -4.30596232e-01 -3.07347834e-01 1.06785499e-01
-5.03357768e-01 3.39009404e-01 -1.01404893e+00 -4.03707772e-01
-7.36414611e-01 -1.52104986e+00 1.19502139e+00 5.32960474e-01
-2.29050145e-01 -1.40466213e+00 1.60510764e-01 5.47549427e-01
3.46874475e-01 5.91068208e-01 8.77619743e-01 -5.56024671e-01
-3.95492345e-01 -4.37313974e-01 -2.23595072e-02 -2.14642614e-01
2.23676160e-01 4.95228738e-01 -9.26673293e-01 -3.55896384e-01
-3.78089666e-01 1.05982430e-01 8.24725211e-01 7.15500891e-01
1.47375131e+00 1.85155660e-01 -1.20148456e+00 6.84746861e-01
1.52535152e+00 -1.22846998e-01 7.24336028e-01 3.63809764e-01
2.52821445e-01 6.88856065e-01 7.68293142e-01 -6.05002008e-02
3.42303187e-01 4.78496730e-01 5.92855394e-01 -3.38990301e-01
-4.56365258e-01 4.27322626e-01 -3.86510223e-01 6.48346722e-01
-2.06738114e-01 -2.55040318e-01 -1.25586414e+00 9.20675159e-01
-1.44289303e+00 -6.05715394e-01 -5.53778887e-01 1.87553334e+00
7.55115330e-01 -1.23169847e-01 1.49842130e-03 4.13083613e-01
5.77119410e-01 -4.19919670e-01 -2.83813834e-01 -1.61489338e-01
9.99519136e-03 1.54726967e-01 5.52638650e-01 2.84217089e-01
-1.10203326e+00 5.55922329e-01 7.04219484e+00 1.08090365e+00
-1.16305125e+00 1.93228796e-01 1.11457133e+00 9.48535129e-02
3.42402570e-02 -2.13173822e-01 -9.64489996e-01 2.26375416e-01
9.41960216e-01 2.19121911e-02 -4.66181003e-02 5.60831845e-01
2.49185577e-01 -4.12276626e-01 -8.95308793e-01 7.29748309e-01
2.60553006e-02 -1.74784875e+00 -3.95537347e-01 5.37059546e-01
7.24572301e-01 1.87618017e-01 8.60765129e-02 -1.83073834e-01
-1.18833818e-01 -9.97719109e-01 -1.02698006e-01 5.98627388e-01
8.51946235e-01 -4.31660414e-01 1.36057889e+00 2.04622194e-01
-9.35277939e-01 2.26508021e-01 -2.53450751e-01 4.83136833e-01
6.17373772e-02 7.91911721e-01 -1.29715419e+00 8.84073079e-01
5.27938783e-01 8.07271898e-01 -8.88975680e-01 1.51534832e+00
1.01273589e-01 5.82291126e-01 -2.29138628e-01 -7.25975186e-02
1.27411723e-01 2.51412362e-01 3.53521496e-01 1.54227924e+00
2.93948978e-01 1.81238666e-01 1.44263532e-03 8.18292648e-02
1.14841960e-01 2.03038335e-01 3.78621966e-02 -2.55688161e-01
-5.77915683e-02 2.15363359e+00 -1.48595703e+00 -2.70717680e-01
-2.64088333e-01 7.33448625e-01 -4.93157916e-02 -1.98145628e-01
-6.21354222e-01 -1.24366544e-02 3.48481804e-01 2.51866162e-01
-1.94396600e-01 1.37013480e-01 -5.95526874e-01 -7.41139054e-01
-6.42393529e-01 -6.82477832e-01 9.35757160e-01 -4.55562949e-01
-1.36558509e+00 3.75650734e-01 -2.71876216e-01 -1.13338196e+00
2.20001400e-01 -9.06062365e-01 -5.78134596e-01 4.98666346e-01
-1.64461422e+00 -1.16653264e+00 -4.60782886e-01 5.96636087e-02
5.33450484e-01 1.29145533e-01 1.17238784e+00 5.67613095e-02
-5.88057220e-01 5.49419224e-01 3.03497940e-01 5.29049337e-02
8.50430965e-01 -1.47586524e+00 -1.11680754e-01 1.67212278e-01
-4.99082088e-01 4.41449404e-01 4.90864098e-01 -5.88501751e-01
-1.30102849e+00 -1.05043054e+00 9.78169203e-01 -4.10168260e-01
6.45865500e-01 -5.32040857e-02 -5.17310977e-01 4.40346897e-01
2.32909307e-01 3.62534225e-01 1.46730030e+00 -4.04961646e-01
5.55778325e-01 -1.04436874e-01 -1.72701454e+00 5.78804612e-01
4.24208850e-01 1.23435572e-01 7.25479573e-02 8.17856193e-01
-3.94013189e-02 -9.91526306e-01 -1.12425530e+00 6.41918600e-01
4.82392758e-01 -8.15150619e-01 7.37101674e-01 -1.86379820e-01
5.57257272e-02 -3.47426504e-01 2.86422282e-01 -9.48921263e-01
-2.83284426e-01 -2.07708567e-01 8.23766887e-02 8.45155776e-01
2.99066693e-01 -3.50609124e-01 1.29221034e+00 3.49667937e-01
-4.70264018e-01 -1.42058301e+00 -1.10676003e+00 -6.56512007e-02
3.57030243e-01 -4.04392406e-02 1.80077583e-01 5.63942730e-01
3.54090780e-01 -4.29397345e-01 3.79467130e-01 -4.38565128e-02
6.02869809e-01 -3.25933605e-01 4.17075247e-01 -1.04641819e+00
1.67183980e-01 -6.65522814e-01 -8.83894384e-01 -3.74357462e-01
-2.22001672e-01 -1.21111965e+00 -2.30635464e-01 -1.97926044e+00
5.83626032e-01 -1.14849783e-01 -3.08190078e-01 4.35933918e-01
-1.28048211e-01 9.48607266e-01 -6.53372556e-02 8.66005942e-02
-5.28320253e-01 -2.86322236e-01 1.56058371e+00 -3.21394563e-01
3.22591871e-01 7.35174417e-02 -7.95308590e-01 8.12579691e-01
9.45989966e-01 -3.94616574e-01 8.53983387e-02 1.30668685e-01
7.08349794e-02 -1.27836570e-01 3.93740505e-01 -1.03940523e+00
4.33434844e-01 -1.94810048e-01 6.18963301e-01 -9.97203350e-01
1.63417622e-01 -7.99384654e-01 1.45628452e-01 8.86276364e-01
-3.10410839e-02 -2.11200342e-01 1.95848882e-01 3.39363605e-01
-1.13405608e-01 -3.37383837e-01 1.01181567e+00 -5.43068290e-01
-3.10185701e-01 3.59695762e-01 -5.92436492e-01 -4.52211291e-01
1.60160339e+00 -4.70485181e-01 -3.11777472e-01 2.03755841e-01
-1.04917014e+00 5.49166620e-01 4.37960356e-01 -2.18956113e-01
1.79821402e-01 -8.97141218e-01 -8.66541982e-01 -3.80988449e-01
1.14061236e-01 2.59377718e-01 7.73220718e-01 1.65651882e+00
-8.94230723e-01 7.45029271e-01 3.69883142e-02 -8.62514198e-01
-1.74987543e+00 2.80269951e-01 8.57645988e-01 -8.75923634e-01
-4.09274578e-01 1.01757574e+00 7.71914795e-02 -3.97651851e-01
1.82937115e-01 -3.06128889e-01 -5.35425723e-01 -1.67854086e-01
6.44286633e-01 5.13441086e-01 9.09820437e-01 -5.17084181e-01
-7.65325487e-01 3.63174766e-01 -3.67030203e-01 2.46624768e-01
1.19796121e+00 -2.53526032e-01 -6.32367492e-01 3.48785460e-01
9.01283145e-01 -4.24156748e-02 -6.35234356e-01 2.60720700e-01
-4.32641096e-02 2.64786072e-02 3.23877394e-01 -1.23800528e+00
-1.25515497e+00 8.43753278e-01 6.95934713e-01 -6.04674667e-02
1.30066216e+00 1.41866192e-01 7.06078649e-01 -2.52263635e-01
2.76221812e-01 -7.96733081e-01 -3.12857985e-01 3.24144542e-01
4.28254664e-01 -9.83223259e-01 3.75967741e-01 -1.01372051e+00
-4.69131351e-01 1.40096056e+00 3.92905861e-01 8.12949315e-02
6.81012630e-01 6.34380400e-01 3.23340207e-01 -3.00304264e-01
-6.87418103e-01 -1.59021482e-01 3.33246052e-01 6.92623258e-01
9.21924174e-01 1.14905126e-01 -7.31716573e-01 4.78591233e-01
-1.68476745e-01 4.36211154e-02 5.30726671e-01 9.52306092e-01
-2.40291923e-01 -1.42463958e+00 -3.60839397e-01 6.22985959e-01
-9.86513197e-01 1.40244514e-01 -5.81689596e-01 7.18084276e-01
3.05576861e-01 4.82953727e-01 -1.57689437e-01 2.63217539e-01
2.11260263e-02 -1.40760884e-01 6.93450153e-01 -6.63679779e-01
-8.94458592e-01 2.40382046e-01 -2.90723234e-01 -1.68004066e-01
-4.84546185e-01 -8.05486262e-01 -1.65765607e+00 -2.65865982e-01
-2.69174248e-01 8.75441283e-02 9.86940801e-01 8.77374053e-01
-4.11275737e-02 8.83317828e-01 4.55461591e-02 -6.67811453e-01
-3.68921012e-02 -7.56853819e-01 -5.12540162e-01 -1.98466912e-01
7.11984560e-02 -3.60652298e-01 -2.69613415e-01 2.01651797e-01] | [15.030198097229004, -3.051969051361084] |
10817c12-c97e-4d68-a986-a763a92463fb | computer-aided-road-inspection-systems-and | 2203.02355 | null | https://arxiv.org/abs/2203.02355v1 | https://arxiv.org/pdf/2203.02355v1.pdf | Computer-Aided Road Inspection: Systems and Algorithms | Road damage is an inconvenience and a safety hazard, severely affecting vehicle condition, driving comfort, and traffic safety. The traditional manual visual road inspection process is pricey, dangerous, exhausting, and cumbersome. Also, manual road inspection results are qualitative and subjective, as they depend entirely on the inspector's personal experience. Therefore, there is an ever-increasing need for automated road inspection systems. This chapter first compares the five most common road damage types. Then, 2-D/3-D road imaging systems are discussed. Finally, state-of-the-art machine vision and intelligence-based road damage detection algorithms are introduced. | ['Mohammud Junaid Bocus', 'Li Wang', 'Sicen Guo', 'Rui Fan'] | 2022-03-04 | null | null | null | null | ['road-damage-detection'] | ['computer-vision'] | [-7.60952979e-02 -7.61106461e-02 -6.70951456e-02 5.07963672e-02
-2.33605281e-01 -3.13946545e-01 1.44772470e-01 1.57473102e-01
-2.70721287e-01 6.62321508e-01 -3.47279370e-01 -6.41771019e-01
-1.86753109e-01 -9.06952083e-01 -6.28929809e-02 -7.99088061e-01
2.79761493e-01 8.17108825e-02 5.50649822e-01 -2.71143258e-01
5.01920938e-01 8.62527132e-01 -1.98711026e+00 -6.53576791e-01
1.35920644e+00 9.18765247e-01 3.06760639e-01 5.02923310e-01
1.07054263e-01 3.69450897e-01 -3.19673449e-01 -4.90474224e-01
7.27044651e-03 -1.59207344e-01 -6.86223924e-01 3.07242006e-01
2.03002691e-01 -4.93819624e-01 -3.59399945e-01 9.90305781e-01
2.16918081e-01 5.68639301e-02 1.02550614e+00 -1.00391304e+00
-6.59864426e-01 -6.26625240e-01 -6.60443664e-01 2.97084957e-01
5.20398580e-02 4.59870756e-01 4.39855039e-01 -8.92296910e-01
2.01649472e-01 9.79902267e-01 5.01462877e-01 1.53832182e-01
-7.46884406e-01 -2.07295805e-01 -1.27716005e-01 7.19388783e-01
-1.24952161e+00 -3.00465018e-01 9.33507323e-01 -7.27673471e-01
9.10837352e-01 2.87527114e-01 5.74874580e-01 3.15706611e-01
5.38805246e-01 3.57890040e-01 1.18686318e+00 -1.79305136e-01
9.18973982e-02 2.99056172e-01 2.20537543e-01 6.38812363e-01
6.81066990e-01 4.23671544e-01 2.17125893e-01 3.50179881e-01
4.28288430e-01 -1.34714087e-03 -1.86045378e-01 -4.55730885e-01
-3.26035053e-01 6.38050735e-01 2.89485902e-01 2.19310716e-01
-4.51438725e-01 -3.02038670e-01 4.61278349e-01 2.03709304e-01
2.12658763e-01 -2.13269400e-03 1.55213356e-01 -4.03887123e-01
-5.15136778e-01 -2.26953149e-01 1.55839980e-01 2.49896750e-01
9.00397778e-01 1.19007565e-01 4.04466182e-01 9.44356978e-01
6.47142470e-01 8.74017477e-01 -1.09025024e-01 -9.80393708e-01
8.74205120e-03 4.67517316e-01 1.06018193e-01 -1.36113441e+00
-2.95501143e-01 -1.45391330e-01 -9.13489938e-01 9.71334815e-01
8.07619765e-02 2.56938666e-01 -9.24674034e-01 6.83042526e-01
2.74353713e-01 -6.24276102e-01 -1.47018567e-01 8.04549396e-01
8.09010386e-01 3.67158324e-01 2.43572265e-01 -1.37528449e-01
1.17114377e+00 -7.24821031e-01 -1.00251734e+00 -4.06361222e-01
1.42667696e-01 -6.71031058e-01 1.28047180e+00 5.73149025e-01
-1.15715706e+00 -6.42981529e-01 -1.12734854e+00 -1.45234391e-01
-4.87255573e-01 2.63735682e-01 2.46152550e-01 9.66369987e-01
-7.38585532e-01 2.18161210e-01 -5.70764959e-01 -3.29259187e-01
4.27504927e-01 -1.17041267e-01 -4.38684762e-01 -3.99899125e-01
-9.75683928e-01 1.58308625e+00 -4.78905700e-02 4.60407853e-01
-8.47987771e-01 -2.88497984e-01 -9.72363174e-01 -2.47286767e-01
3.58025849e-01 -1.89209074e-01 1.13163531e+00 3.66781726e-02
-1.33055127e+00 9.36224818e-01 -3.74859512e-01 1.26004726e-01
4.46835756e-01 -2.45657295e-01 -8.10845375e-01 3.68760556e-01
2.97015514e-02 -3.14671285e-02 7.68794358e-01 -1.66145074e+00
-6.68668032e-01 -3.04674357e-01 -2.37645239e-01 2.00893879e-01
1.41835272e-01 9.05029997e-02 -6.22606613e-02 3.27220783e-02
-1.69981476e-02 -5.15544057e-01 -1.01346746e-01 2.00774461e-01
-3.77281725e-01 -3.33860517e-01 1.28362453e+00 -9.70559120e-01
1.22305202e+00 -2.29813957e+00 -6.30305529e-01 2.24895641e-01
2.53468484e-01 7.15350032e-01 3.48010272e-01 4.51275676e-01
9.16335806e-02 2.02238247e-01 -3.10713530e-01 -2.67148241e-02
-2.09093839e-01 2.65186787e-01 1.05765961e-01 5.47647119e-01
1.63393229e-01 4.58311439e-01 -7.70522714e-01 -4.65973347e-01
8.28126788e-01 4.68088388e-01 3.42504084e-01 9.74087492e-02
7.35367298e-01 3.74057680e-01 -2.53112555e-01 9.37398255e-01
7.94181287e-01 2.82199055e-01 -5.68115890e-01 -9.79215056e-02
-6.36583924e-01 2.12402433e-01 -7.16791213e-01 6.18922949e-01
-5.70963442e-01 9.12971616e-01 1.87077336e-02 -9.20597196e-01
1.19932163e+00 3.03288609e-01 2.50310320e-02 -9.95032191e-01
6.76616803e-02 4.42386836e-01 -2.49434471e-01 -1.23879349e+00
5.09015918e-01 -2.67116159e-01 2.31997088e-01 1.18303411e-01
-6.46608770e-01 -3.90744507e-01 2.63535112e-01 -1.11232966e-01
7.26948619e-01 -2.65605003e-01 2.69276202e-01 3.30147923e-05
5.58486462e-01 -8.46692733e-03 3.86015087e-01 1.36468217e-01
-4.86463696e-01 4.49252665e-01 4.67808932e-01 -3.69918823e-01
-9.59829628e-01 -1.41085076e+00 -4.27769125e-01 3.14914435e-01
4.76120442e-01 4.61661071e-01 -5.68931282e-01 -3.69881630e-01
-2.92870775e-02 7.84120500e-01 -2.64840782e-01 -4.42864269e-01
-2.33497486e-01 -3.07115912e-01 2.06605509e-01 3.78823459e-01
7.56520092e-01 -9.42046642e-01 -6.39991939e-01 -3.35692987e-02
-1.60435244e-01 -8.56205583e-01 -4.42968868e-02 -4.12437052e-01
-6.80108666e-01 -1.19630945e+00 -5.29583633e-01 -8.09011996e-01
9.89388049e-01 9.13375378e-01 9.76478994e-01 4.00682062e-01
-6.47001922e-01 2.91753918e-01 -1.75140157e-01 -4.15821284e-01
-3.81565511e-01 -3.70330691e-01 -6.98361695e-02 -1.49947375e-01
3.10262501e-01 -4.04058129e-01 -7.67575204e-01 7.89139926e-01
-4.93143260e-01 -4.27235186e-01 6.50722086e-01 4.28674757e-01
4.71417218e-01 6.88724697e-01 8.02084267e-01 -3.59288305e-01
5.50526440e-01 -1.79369792e-01 -6.09610796e-01 5.52408993e-02
-8.07584941e-01 -4.38986927e-01 4.15585101e-01 1.57698870e-01
-1.48063862e+00 -2.40847185e-01 -2.54976004e-01 -9.41944495e-02
-7.56331861e-01 3.62678587e-01 -4.49167669e-01 -1.92627490e-01
7.11236358e-01 -1.19204998e-01 1.75425246e-01 -5.33486545e-01
1.99556611e-02 9.22884583e-01 9.37612474e-01 9.21934918e-02
1.06675994e+00 4.92559463e-01 1.70480996e-01 -1.46736276e+00
-5.01974046e-01 -6.26384377e-01 -6.74527526e-01 -9.86884117e-01
9.66070116e-01 -3.33467901e-01 -6.56745613e-01 9.88942921e-01
-1.06678212e+00 -2.37658754e-01 -5.90825977e-04 4.32056904e-01
-1.95131376e-01 9.11557555e-01 -2.72622287e-01 -1.08311117e+00
-2.30469078e-01 -1.08910859e+00 6.97140932e-01 4.02212530e-01
-1.95379574e-02 -8.37739885e-01 7.61277005e-02 8.31154644e-01
2.98749506e-01 2.92275995e-01 9.76343930e-01 4.76496041e-01
-3.61762255e-01 -5.52167654e-01 -6.31503522e-01 6.36735976e-01
4.85309571e-01 4.69231993e-01 -1.04327071e+00 1.75102979e-01
-1.30320549e-01 -4.87425998e-02 7.26581931e-01 4.09022868e-01
7.50466704e-01 -6.73434837e-03 -3.53521496e-01 4.11017574e-02
1.34481251e+00 6.13753140e-01 1.24788666e+00 3.01793545e-01
6.28751636e-01 1.10480344e+00 1.15563369e+00 -7.23416209e-02
4.37050998e-01 3.29001993e-01 7.01280832e-01 -5.24706125e-01
-1.44366607e-01 -1.04534663e-01 1.39383137e-01 4.59654897e-01
-4.27547157e-01 -9.77700129e-02 -1.05774558e+00 9.25313115e-01
-1.37189758e+00 -1.09244525e+00 -8.70357692e-01 2.36426854e+00
4.45673794e-01 2.01521367e-01 7.20819086e-02 7.48757064e-01
8.70877206e-01 -3.29071552e-01 -3.48210037e-01 -6.79205775e-01
2.23034039e-01 -2.41165042e-01 5.81638336e-01 4.81625319e-01
-1.02213180e+00 4.39980060e-01 6.78649998e+00 6.53207660e-01
-9.06969786e-01 -1.29620461e-02 4.85579640e-01 4.68522310e-01
-2.02713102e-01 -1.42525285e-01 -4.25066233e-01 4.18081373e-01
5.96456707e-01 5.05275391e-02 -2.66778022e-02 6.13203287e-01
7.69741178e-01 -9.22365606e-01 -3.03847134e-01 8.00696611e-01
-2.21547738e-01 -6.65648282e-01 -3.95650744e-01 2.00871661e-01
2.74897039e-01 -2.47667775e-01 8.50447714e-02 -1.41695693e-01
1.87838133e-02 -9.36056018e-01 3.17462623e-01 6.32138491e-01
8.80142868e-01 -1.06381881e+00 8.91334236e-01 4.26672399e-02
-1.16637611e+00 8.12071040e-02 -3.73182327e-01 -1.12338208e-01
8.00976992e-01 9.20217216e-01 -4.47828114e-01 3.39926124e-01
1.00047851e+00 3.62583548e-01 -5.35998762e-01 1.46714306e+00
-7.11201012e-01 4.18252796e-01 -1.27916727e-02 1.96317971e-01
1.78796336e-01 -3.29769224e-01 4.00181621e-01 8.29048693e-01
1.97380632e-01 1.37500288e-02 -1.41759321e-01 4.99795258e-01
6.15261734e-01 -2.66356170e-01 -1.13228190e+00 3.22613299e-01
3.81914914e-01 1.33416855e+00 -7.22994268e-01 -1.05107032e-01
-6.25586510e-01 6.58256531e-01 -2.28317171e-01 3.10897648e-01
-6.90839291e-01 -8.53803873e-01 7.38823712e-01 4.37946051e-01
1.50684430e-03 -3.78520668e-01 -5.67834795e-01 -2.84490585e-01
1.01144500e-01 -1.74749389e-01 1.03259673e-02 -7.96954751e-01
-1.16646779e+00 1.36673674e-01 1.35806307e-01 -1.31751370e+00
1.75269380e-01 -6.60920322e-01 -8.90032828e-01 8.98627937e-01
-1.91311967e+00 -6.85640037e-01 -6.67006791e-01 1.03896640e-01
3.94368321e-01 1.23872429e-01 2.95412928e-01 5.36667526e-01
-7.75074005e-01 3.35602880e-01 -1.05966933e-01 -8.96527916e-02
3.43523026e-01 -8.91583264e-01 2.62478024e-01 9.13429439e-01
-8.91032815e-01 1.04776762e-01 6.83592319e-01 -6.26567721e-01
-9.85551596e-01 -9.00900364e-01 1.04195642e+00 -2.05727052e-02
4.79839295e-01 3.85095596e-01 -1.29913342e+00 -7.34748683e-05
2.85094410e-01 -2.77191788e-01 4.65138644e-01 -1.87913209e-01
9.90135372e-02 -2.26720780e-01 -1.47901714e+00 6.72595561e-01
7.20519245e-01 -5.70343375e-01 -5.21544933e-01 -4.63636816e-02
2.80357711e-02 5.46276793e-02 -7.54847229e-01 3.23482484e-01
4.57252115e-01 -1.05734348e+00 9.97306764e-01 3.40999961e-01
9.13197100e-02 -5.59386075e-01 5.85460842e-01 -1.02736926e+00
-4.35072720e-01 -2.80939907e-01 5.14444649e-01 1.04881954e+00
4.20791507e-01 -9.20359075e-01 3.87578994e-01 7.65177488e-01
-7.45486200e-01 -4.90509480e-01 -9.31179821e-01 -1.04271483e+00
-3.00152123e-01 -4.22741413e-01 1.71946213e-01 5.20956397e-01
2.89154768e-01 3.72801214e-01 -2.46024638e-01 3.29453588e-01
8.13963294e-01 -4.36207116e-01 5.96610785e-01 -1.51520979e+00
3.71087253e-01 -6.80040717e-01 -5.38466871e-01 -5.86555779e-01
-2.58828908e-01 -8.94119665e-02 2.52564669e-01 -2.01090121e+00
-1.01861201e-01 -1.66794077e-01 2.10851192e-01 5.15750885e-01
-1.75446987e-01 2.38039911e-01 -3.72140616e-01 1.63115442e-01
-2.88741980e-02 5.98710716e-01 1.42762494e+00 -2.35445872e-01
-6.01070598e-02 3.06912839e-01 -4.62558836e-01 7.47409403e-01
1.21753967e+00 -2.27249786e-02 -6.40313923e-01 -2.89875388e-01
4.19581771e-01 -1.69570684e-01 5.06937861e-01 -9.36934292e-01
3.55612524e-02 -2.74547309e-01 -1.92090526e-01 -1.07909167e+00
2.61622161e-01 -9.46308792e-01 -7.99276978e-02 5.78743398e-01
4.88664120e-01 1.78430211e-02 1.64049506e-01 3.78659904e-01
-3.63766372e-01 -6.03098691e-01 1.06845331e+00 1.43746972e-01
-7.88534224e-01 -8.74174573e-03 -1.09727788e+00 -4.44677919e-01
1.19601595e+00 -8.84111822e-01 -4.57934856e-01 -3.21925402e-01
-4.11573678e-01 3.16442072e-01 5.28562307e-01 2.15489402e-01
1.05675387e+00 -9.86177027e-01 -4.28906918e-01 1.77197948e-01
1.44086570e-01 1.62390187e-01 7.26069212e-01 1.07342505e+00
-6.85182273e-01 3.52363974e-01 -4.04822856e-01 -5.26682377e-01
-1.35662293e+00 6.81701481e-01 1.29988834e-01 1.05846308e-01
-6.58780575e-01 1.46552384e-01 1.45184249e-01 2.88305283e-02
-5.21760806e-02 -9.68037397e-02 -5.36280453e-01 -3.33517715e-02
4.79799569e-01 1.18092287e+00 1.49617180e-01 -6.93025589e-01
-1.82590440e-01 1.07336473e+00 2.00039223e-01 6.72303662e-02
6.69636607e-01 -6.38152957e-01 -7.31486678e-02 3.58950645e-01
8.00899923e-01 -3.08651775e-01 -1.16422272e+00 2.78875351e-01
-2.13623837e-01 -6.15518868e-01 3.24621022e-01 -5.76587141e-01
-9.54325795e-01 1.25540161e+00 6.32213473e-01 6.76993012e-01
1.36399579e+00 -2.01948155e-02 9.14318264e-01 2.18355894e-01
3.53625536e-01 -1.44769835e+00 -9.49999243e-02 3.23894233e-01
8.71890247e-01 -1.30162203e+00 1.05303712e-02 -7.50024438e-01
-7.46655941e-01 8.40670884e-01 6.54652715e-01 3.77651483e-01
6.68713212e-01 1.26099549e-02 3.42937559e-01 -4.24687803e-01
-1.11718111e-01 -4.55538720e-01 3.10549706e-01 1.16511643e+00
1.83668256e-01 -3.94743383e-02 -4.93842125e-01 -6.24199621e-02
1.63042113e-01 -2.77753323e-01 4.44148868e-01 8.07592213e-01
-1.04270542e+00 -8.89428556e-01 -6.02443635e-01 4.46572572e-01
-1.32517546e-01 4.26036566e-01 -2.18348786e-01 8.36880147e-01
-5.73321432e-03 1.35443246e+00 2.52228486e-03 -4.03955013e-01
6.06811643e-01 -3.38156730e-01 2.68389490e-02 -3.00311118e-01
1.53658211e-01 -2.73511201e-01 4.53160077e-01 -1.91642165e-01
-2.41483673e-01 -8.03957522e-01 -1.37994623e+00 -5.78359067e-01
-2.79913545e-01 -1.68359831e-01 7.72548676e-01 9.58520830e-01
1.41508237e-01 2.25497007e-01 9.56924558e-01 -6.58267975e-01
6.82968041e-03 -7.89968550e-01 -7.32181728e-01 -7.46294111e-02
4.48720366e-01 -1.17322612e+00 -4.18767869e-01 -4.81229909e-02] | [7.417540073394775, 1.0203731060028076] |
c9b83210-9f9e-4643-8020-bec52f1709cc | keyphrase-prediction-with-pre-trained | 2004.10462 | null | https://arxiv.org/abs/2004.10462v1 | https://arxiv.org/pdf/2004.10462v1.pdf | Keyphrase Prediction With Pre-trained Language Model | Recently, generative methods have been widely used in keyphrase prediction, thanks to their capability to produce both present keyphrases that appear in the source text and absent keyphrases that do not match any source text. However, the absent keyphrases are generated at the cost of the performance on present keyphrase prediction, since previous works mainly use generative models that rely on the copying mechanism and select words step by step. Besides, the extractive model that directly extracts a text span is more suitable for predicting the present keyphrase. Considering the different characteristics of extractive and generative methods, we propose to divide the keyphrase prediction into two subtasks, i.e., present keyphrase extraction (PKE) and absent keyphrase generation (AKG), to fully exploit their respective advantages. On this basis, a joint inference framework is proposed to make the most of BERT in two subtasks. For PKE, we tackle this task as a sequence labeling problem with the pre-trained language model BERT. For AKG, we introduce a Transformer-based architecture, which fully integrates the present keyphrase knowledge learned from PKE by the fine-tuned BERT. The experimental results show that our approach can achieve state-of-the-art results on both tasks on benchmark datasets. | ['Zheng Lin', 'Rui Liu', 'Weiping Wang'] | 2020-04-22 | null | null | null | null | ['keyphrase-generation'] | ['natural-language-processing'] | [ 2.22389638e-01 -1.77817777e-01 -3.63773227e-01 2.33406171e-01
-8.52449894e-01 -7.73035765e-01 1.13124096e+00 3.67257893e-01
-4.45568264e-01 7.61089802e-01 4.21715260e-01 -4.17864352e-01
1.14737280e-01 -1.02843583e+00 -7.44343340e-01 -6.98528945e-01
2.36066848e-01 2.90561408e-01 2.79170990e-01 -3.21743786e-01
3.44447434e-01 3.78369063e-01 -1.49066079e+00 4.24320906e-01
8.85057807e-01 9.72718298e-01 4.12723273e-01 6.80836141e-01
-4.80865479e-01 1.11276698e+00 -5.75791895e-01 -5.22812486e-01
1.58815160e-01 -6.26008868e-01 -5.51881671e-01 -2.52301488e-02
9.47033912e-02 -4.92276073e-01 -7.60411024e-01 8.17395091e-01
2.11258158e-01 -9.36633907e-03 7.16086805e-01 -1.38951480e+00
-4.08678293e-01 1.20574558e+00 -4.54012960e-01 1.85142919e-01
2.47719988e-01 -1.57158911e-01 1.61332977e+00 -1.00796187e+00
5.31086087e-01 9.33537126e-01 2.02679500e-01 9.40776095e-02
-9.52868104e-01 -7.70403028e-01 2.40517572e-01 3.60841990e-01
-1.48601961e+00 -2.69282520e-01 9.65725124e-01 -2.13865757e-01
9.38240230e-01 2.46135592e-01 8.23848546e-01 1.09739554e+00
4.60374564e-01 1.34555304e+00 1.20061171e+00 -3.85896772e-01
-1.20907277e-02 -1.59266908e-02 7.52778584e-03 5.48017621e-01
1.58052683e-01 2.92719901e-02 -6.90530419e-01 -2.27207333e-01
5.13426781e-01 2.73145616e-01 -3.29153001e-01 -4.18920182e-02
-1.54105663e+00 7.79625833e-01 2.96462402e-02 4.00542289e-01
-7.05129266e-01 -2.41452339e-03 3.87459636e-01 3.23154062e-01
4.45209831e-01 4.84644383e-01 -5.28959990e-01 -1.94628716e-01
-1.20558035e+00 6.23876274e-01 1.02614999e+00 1.04792452e+00
9.35028613e-01 -2.86646754e-01 -5.62624931e-01 5.54394007e-01
1.80930063e-01 3.18720788e-01 5.78286290e-01 1.34599507e-01
5.25858223e-01 7.39205599e-01 -1.74701251e-02 -1.05883610e+00
-1.60083622e-01 -5.33522189e-01 -8.43202472e-01 -6.51539505e-01
-5.02334759e-02 -1.02834985e-01 -1.14645445e+00 1.36979246e+00
1.35295153e-01 3.41490597e-01 1.08727686e-01 2.51525998e-01
8.82006943e-01 1.15127897e+00 -1.06448263e-01 -3.97054583e-01
1.53186655e+00 -1.14286625e+00 -8.80513072e-01 -2.63512850e-01
4.24186796e-01 -1.08722520e+00 7.32912242e-01 3.18743587e-01
-6.12978816e-01 -4.29429770e-01 -1.07028127e+00 -8.23341832e-02
-5.21644711e-01 4.68857080e-01 3.63793284e-01 3.29576135e-01
-5.67571640e-01 4.33909982e-01 -5.61039388e-01 -7.09354579e-02
-6.60043433e-02 -2.06930012e-01 2.16330010e-02 1.33091956e-01
-1.51879513e+00 5.60006797e-01 1.15588045e+00 -3.25061381e-02
-1.07716572e+00 -8.37971509e-01 -5.46256900e-01 1.88438535e-01
1.02663553e+00 -6.79831088e-01 1.20761764e+00 -3.85484546e-01
-1.40393341e+00 5.33892930e-01 -1.77090853e-01 -6.31399214e-01
4.54692483e-01 -4.05771703e-01 -5.71511745e-01 2.51717567e-01
6.95507750e-02 3.45237494e-01 1.24887526e+00 -1.03000426e+00
-1.14485908e+00 5.56072332e-02 8.53823498e-02 1.94324583e-01
-4.69653517e-01 1.00102294e-02 -6.81885779e-01 -1.33687401e+00
-1.58242688e-01 -9.03700829e-01 1.60327554e-02 -7.79569030e-01
-7.90875673e-01 -4.49718773e-01 6.90717816e-01 -9.95966613e-01
1.96963000e+00 -1.67549372e+00 1.62461907e-01 2.16941059e-01
3.39612156e-01 2.36293972e-01 3.50595750e-02 1.09010375e+00
1.69275895e-01 9.32614952e-02 5.93639910e-04 -9.79983434e-02
1.85089841e-01 3.26206023e-03 -1.08017683e+00 -1.78807288e-01
4.86076474e-02 1.16606390e+00 -9.05174255e-01 -6.78809404e-01
-2.81758737e-02 8.96935761e-02 -1.67210430e-01 5.78414083e-01
-5.72947323e-01 1.24259992e-02 -6.72682345e-01 3.98712844e-01
4.15714562e-01 -2.79334843e-01 1.61554635e-01 -3.85726810e-01
-2.51250684e-01 5.94892740e-01 -1.18268657e+00 1.41508043e+00
-5.48337936e-01 1.75641119e-01 -3.06061298e-01 -7.01349676e-01
8.87670577e-01 4.87404495e-01 5.02870202e-01 -3.33810389e-01
6.51618317e-02 4.57809329e-01 -1.22806989e-01 6.31753430e-02
1.01022434e+00 6.28117621e-02 -1.65008307e-01 7.67272174e-01
1.91739187e-01 -3.11282814e-01 5.39147019e-01 6.70135140e-01
9.41925287e-01 1.03439786e-01 6.17253840e-01 1.17224921e-02
6.96050942e-01 -1.03786074e-01 2.72265673e-01 7.96823025e-01
4.71316457e-01 3.47201198e-01 5.36755979e-01 -2.74850070e-01
-9.92589414e-01 -6.42097950e-01 3.11657906e-01 1.10222268e+00
7.28647783e-02 -1.15409410e+00 -5.17862678e-01 -1.03009450e+00
-2.06569761e-01 7.68620133e-01 -5.40703177e-01 -2.12766245e-01
-6.79041803e-01 -7.54698873e-01 5.63687503e-01 3.30424070e-01
3.28794301e-01 -1.04456568e+00 -3.60325485e-01 5.24185598e-01
-4.61092919e-01 -1.29919851e+00 -6.80600464e-01 3.12126249e-01
-4.14597839e-01 -8.14469278e-01 -1.07215619e+00 -5.86426079e-01
4.24264431e-01 3.85829121e-01 1.14981878e+00 -4.04682793e-02
8.06100294e-02 2.78795779e-01 -8.31953168e-01 -4.89461094e-01
-5.97735643e-01 6.26696527e-01 -2.65474290e-01 3.27918142e-01
1.57998398e-01 -5.07898569e-01 -3.59800309e-01 3.86783332e-02
-1.19516349e+00 6.64860785e-01 1.08163512e+00 8.74442220e-01
6.96601272e-01 3.76942277e-01 3.68912518e-01 -8.53997111e-01
8.00662160e-01 -4.46311712e-01 -5.51439166e-01 4.93191272e-01
-1.01386333e+00 3.94733846e-01 1.05366862e+00 -5.48467100e-01
-9.40287590e-01 -2.94885248e-01 -1.26313433e-01 -3.07906717e-01
1.36662379e-01 7.69461274e-01 -1.54794589e-01 3.83638471e-01
1.94512233e-01 1.19501615e+00 -4.79671866e-01 -5.59830010e-01
4.70312238e-01 5.08106351e-01 2.48739302e-01 -6.22690022e-01
1.24739754e+00 2.05494389e-01 -9.73700210e-02 -7.35886097e-01
-1.02727604e+00 -5.61140776e-01 -3.91701788e-01 -2.10103542e-02
4.69577163e-01 -1.06466830e+00 -2.03978226e-01 7.54118562e-01
-1.05898046e+00 1.16265610e-01 -1.51413962e-01 2.51638770e-01
-4.63654399e-01 5.93240142e-01 -5.86341798e-01 -5.90626836e-01
-8.66143525e-01 -1.02701783e+00 1.21812308e+00 1.56554073e-01
-7.55512118e-02 -7.69911587e-01 4.60876301e-02 -1.21113956e-01
-7.46860076e-03 3.65407555e-03 1.18248320e+00 -1.16972899e+00
-8.30306709e-01 -3.18487495e-01 -1.64809331e-01 2.07896486e-01
3.21150631e-01 -2.71935351e-02 -6.94488108e-01 -2.84105927e-01
-2.60783076e-01 -2.77848631e-01 1.15035307e+00 -5.26689999e-02
1.01270497e+00 -6.92970574e-01 -4.43696171e-01 4.41371441e-01
1.17778468e+00 1.51855469e-01 5.48779368e-01 3.87647539e-01
9.23638701e-01 4.30525124e-01 7.22420692e-01 5.05640447e-01
6.08204365e-01 5.42861342e-01 3.40660103e-02 2.13615507e-01
-6.96665645e-02 -1.05036092e+00 4.93537992e-01 9.96904373e-01
5.12248427e-02 -4.09461617e-01 -6.92106307e-01 5.46778142e-01
-1.83427310e+00 -9.10103798e-01 1.49708062e-01 1.91899002e+00
1.19585133e+00 3.45221728e-01 1.40923157e-01 2.82670796e-01
4.84774470e-01 7.24084079e-01 -1.81252420e-01 1.03190564e-01
1.15319863e-02 3.53334486e-01 4.81995940e-01 9.84128565e-02
-1.09983683e+00 1.20543754e+00 4.81312847e+00 1.50923991e+00
-9.71859038e-01 -2.20781878e-01 3.18941206e-01 3.15392017e-01
-6.57512188e-01 2.80970991e-01 -1.35277069e+00 6.21664882e-01
7.51695871e-01 -4.30236340e-01 3.99064481e-01 7.10698545e-01
-5.53830639e-02 9.15546529e-03 -1.04931808e+00 9.31214929e-01
1.53466943e-03 -1.31317234e+00 5.86071253e-01 4.19979505e-02
7.40561903e-01 -3.79029095e-01 -1.66177392e-01 4.80727971e-01
2.51999706e-01 -6.40182555e-01 9.51411486e-01 5.28653264e-01
3.33213687e-01 -8.91165495e-01 5.26969850e-01 6.98970318e-01
-1.59439123e+00 1.49089143e-01 -1.77316681e-01 3.56700271e-01
1.39503211e-01 8.52131307e-01 -1.04051232e+00 1.00312221e+00
3.93829167e-01 6.10759377e-01 -5.50918221e-01 7.77794421e-01
-7.41568983e-01 7.08332539e-01 -1.49687007e-01 -4.41307247e-01
4.48537827e-01 -2.02844724e-01 6.69794500e-01 1.35261750e+00
4.06982243e-01 -7.48629346e-02 4.50500488e-01 7.78218746e-01
-1.48391470e-01 4.53233123e-01 -1.44607052e-01 -7.30605721e-01
6.09741271e-01 1.59392321e+00 -9.86693680e-01 -6.78797066e-01
-2.64234424e-01 1.00338840e+00 -1.07108124e-01 2.91839093e-01
-7.23646164e-01 -7.09238470e-01 3.51703078e-01 -9.97629911e-02
6.52758658e-01 -2.56748259e-01 1.29984424e-01 -1.42862952e+00
1.03221521e-01 -1.22379708e+00 4.22980577e-01 -4.15569752e-01
-9.82343674e-01 6.84288919e-01 2.23533824e-01 -1.18575180e+00
-6.71288431e-01 -2.96441585e-01 -4.60638553e-01 8.81537259e-01
-1.75153685e+00 -1.55059886e+00 -6.89464957e-02 3.87412727e-01
7.66442776e-01 -1.11047506e-01 5.58122993e-01 -5.18652494e-04
-3.85378689e-01 5.14564037e-01 1.63529292e-02 3.15853298e-01
6.23592019e-01 -1.33589566e+00 8.76530766e-01 1.14130795e+00
6.11721218e-01 6.80433929e-01 5.64770937e-01 -9.08487380e-01
-1.56718957e+00 -1.10793877e+00 1.17176986e+00 -1.65354431e-01
9.72169638e-01 -4.89217043e-01 -6.72397196e-01 5.83900571e-01
1.58063069e-01 -4.21503603e-01 3.94881070e-01 -8.92312229e-02
-5.01386881e-01 -3.29888873e-02 -2.15503186e-01 8.78050148e-01
7.33094573e-01 -6.73505545e-01 -7.95718729e-01 6.24217726e-02
1.00819707e+00 -4.09284174e-01 -6.73239291e-01 3.41774344e-01
5.66749096e-01 -6.39032304e-01 9.33154166e-01 -2.36180767e-01
5.74767530e-01 -3.18738908e-01 9.90587249e-02 -1.46110570e+00
-2.35291019e-01 -9.40862894e-01 -7.83935010e-01 1.41359675e+00
3.77001286e-01 -5.36943316e-01 4.93323028e-01 -1.75364688e-01
6.78875074e-02 -8.36775482e-01 -3.84967804e-01 -6.94257796e-01
-2.22429752e-01 -3.13502669e-01 8.12451422e-01 6.50247395e-01
-2.86511481e-01 8.16374063e-01 -8.25853765e-01 -1.00642659e-01
4.02472049e-01 6.13941729e-01 9.21693683e-01 -1.13269639e+00
-4.06347692e-01 -4.34759051e-01 2.15246201e-01 -1.49666953e+00
5.53891286e-02 -9.75611925e-01 6.12922683e-02 -1.48618138e+00
3.12687069e-01 -2.11281732e-01 -2.94772416e-01 5.63735485e-01
-6.67776108e-01 -1.27308115e-01 3.73719186e-01 5.15447378e-01
-3.73496115e-01 7.68806219e-01 1.26000774e+00 -2.52339721e-01
-2.96239406e-01 1.35144398e-01 -8.30747664e-01 5.27795851e-01
4.55874056e-01 -5.01759887e-01 -4.06827301e-01 1.24381222e-01
5.70646405e-01 -1.18448911e-02 1.96406305e-01 -7.21533537e-01
2.63565421e-01 -3.16812485e-01 2.19764411e-01 -1.01853633e+00
1.10860683e-01 -6.47630394e-01 3.28234397e-02 3.65147948e-01
-3.33532244e-01 8.03568065e-02 7.04538524e-02 5.90786397e-01
-2.97963828e-01 -2.82851249e-01 2.84041315e-01 -1.88302204e-01
-9.26039398e-01 6.56802893e-01 -4.83095795e-01 1.37389824e-01
6.60074592e-01 1.74676746e-01 -9.94414166e-02 -3.98633450e-01
-1.04020320e-01 4.99980105e-03 -1.35858012e-02 6.28084362e-01
6.85591936e-01 -1.27367902e+00 -8.63249898e-01 9.55667943e-02
4.37316686e-01 -1.76890805e-01 5.72697148e-02 7.66842484e-01
-1.33719683e-01 7.00242221e-01 1.64893299e-01 -1.17500909e-02
-1.00115454e+00 7.07667291e-01 -2.68025368e-01 -1.07739723e+00
-6.32016480e-01 6.80652738e-01 3.04228634e-01 -2.35361263e-01
3.09757106e-02 -6.08089864e-01 -3.83469284e-01 5.59361279e-01
7.51201630e-01 1.58166066e-01 2.60585368e-01 -4.32680488e-01
7.26439506e-02 2.88111657e-01 -6.23540878e-01 -1.20339461e-01
1.08144581e+00 -1.03570512e-02 -3.77731532e-01 3.74885976e-01
1.06025541e+00 3.03305894e-01 -8.41772318e-01 -7.21753120e-01
1.18945591e-01 -7.67049566e-02 2.52375156e-01 -7.33011961e-01
-5.94603360e-01 6.48084044e-01 -8.44936073e-03 2.37499282e-01
1.24695539e+00 5.33926636e-02 1.35818470e+00 4.18221802e-01
3.17351460e-01 -1.12321234e+00 1.57101735e-01 7.65504897e-01
7.78860331e-01 -8.24164331e-01 9.95277241e-02 -4.04754817e-01
-4.85314876e-01 1.24502468e+00 2.28575304e-01 1.36064082e-01
5.00569165e-01 1.47099078e-01 -2.69794822e-01 -8.21425673e-03
-8.64849806e-01 -4.15504277e-01 7.99986243e-01 -6.39718920e-02
2.17240706e-01 1.32082731e-01 -5.56158721e-01 7.57915080e-01
-4.86915022e-01 -5.33537865e-02 2.67501444e-01 9.52150881e-01
-5.20080209e-01 -1.34005642e+00 -2.77951449e-01 6.16730988e-01
-6.86653137e-01 -5.96848607e-01 -6.37026250e-01 5.56769550e-01
-8.54908600e-02 7.08794355e-01 -2.80996442e-01 -6.33566141e-01
3.80842462e-02 2.80692637e-01 2.82925010e-01 -6.30625010e-01
-6.06877029e-01 4.64209616e-01 -6.10618033e-02 -1.09715566e-01
-1.76632196e-01 -2.11597785e-01 -9.85952973e-01 -2.00762630e-01
-4.30191308e-01 5.38305759e-01 3.38567168e-01 1.03093100e+00
2.60595977e-01 6.29025042e-01 9.05772269e-01 -4.21096623e-01
-7.19302952e-01 -1.10941887e+00 -5.02173841e-01 1.55153275e-01
1.85333475e-01 -2.69995391e-01 -2.41414141e-02 2.05915123e-01] | [12.315957069396973, 8.8900146484375] |
3e20234b-9bb6-41ab-a567-09e1aff14fd6 | deep-echo-state-networks-with-uncertainty | 1806.10728 | null | http://arxiv.org/abs/1806.10728v2 | http://arxiv.org/pdf/1806.10728v2.pdf | Deep Echo State Networks with Uncertainty Quantification for Spatio-Temporal Forecasting | Long-lead forecasting for spatio-temporal systems can often entail complex
nonlinear dynamics that are difficult to specify it a priori. Current
statistical methodologies for modeling these processes are often highly
parameterized and thus, challenging to implement from a computational
perspective. One potential parsimonious solution to this problem is a method
from the dynamical systems and engineering literature referred to as an echo
state network (ESN). ESN models use so-called {\it reservoir computing} to
efficiently compute recurrent neural network (RNN) forecasts. Moreover,
so-called "deep" models have recently been shown to be successful at predicting
high-dimensional complex nonlinear processes, particularly those with multiple
spatial and temporal scales of variability (such as we often find in
spatio-temporal environmental data). Here we introduce a deep ensemble ESN
(D-EESN) model. We present two versions of this model for spatio-temporal
processes that both produce forecasts and associated measures of uncertainty.
The first approach utilizes a bootstrap ensemble framework and the second is
developed within a hierarchical Bayesian framework (BD-EESN). This more general
hierarchical Bayesian framework naturally accommodates non-Gaussian data types
and multiple levels of uncertainties. The methodology is first applied to a
data set simulated from a novel non-Gaussian multiscale Lorenz-96 dynamical
system simulation model and then to a long-lead United States (U.S.) soil
moisture forecasting application. | ['Christopher K. Wikle', 'Patrick L. McDermott'] | 2018-06-28 | null | null | null | null | ['spatio-temporal-forecasting'] | ['time-series'] | [-1.09746486e-01 -4.18233782e-01 5.24765909e-01 -5.73476590e-02
-4.36143219e-01 -5.82924306e-01 1.12264609e+00 -5.04782125e-02
-8.34496096e-02 1.08608294e+00 8.59564394e-02 -7.45490670e-01
-5.06402373e-01 -9.31813180e-01 -3.92375410e-01 -9.84792888e-01
-4.60958689e-01 4.63541329e-01 1.19674332e-01 -4.28193450e-01
2.01125920e-01 7.61066139e-01 -1.76734102e+00 -3.39820623e-01
8.96812797e-01 8.56861055e-01 1.55959517e-01 8.50712657e-01
5.64900646e-03 3.91694367e-01 -3.22709054e-01 1.86604694e-01
1.11343116e-01 -2.09039986e-01 -2.34998137e-01 -2.20979393e-01
-2.25325420e-01 -4.55100909e-02 9.68071222e-02 7.98950851e-01
4.74311054e-01 4.56301332e-01 9.71689880e-01 -9.33391094e-01
-6.21404536e-02 2.06363067e-01 -1.91819385e-01 3.88641655e-01
-9.20436606e-02 6.98236376e-02 3.62156868e-01 -1.00371671e+00
2.90385634e-01 1.39052153e+00 1.09163415e+00 -1.42944440e-01
-1.78788912e+00 -3.88530731e-01 -1.05759874e-01 -2.65005171e-01
-1.41982627e+00 -4.28402096e-01 4.65231031e-01 -1.14488351e+00
1.06877756e+00 6.34202287e-02 5.99068165e-01 1.08829176e+00
6.76376045e-01 1.27409935e-01 1.74216187e+00 -9.71920416e-02
6.67990923e-01 -1.06489964e-01 1.98478669e-01 4.31435406e-02
2.73357511e-01 8.33692670e-01 -2.89448440e-01 -6.10170424e-01
6.48677170e-01 -1.82635784e-01 -3.19762602e-02 -4.60524186e-02
-9.32179928e-01 1.02464926e+00 3.49226333e-02 4.81708676e-01
-8.52410674e-01 1.89634129e-01 3.33019972e-01 2.76112527e-01
1.02233827e+00 -1.04240319e-02 -5.99747837e-01 -2.12773606e-01
-1.44830108e+00 5.85473537e-01 1.04545653e+00 3.89280319e-01
5.59193015e-01 5.65376580e-01 1.47867665e-01 5.30366838e-01
6.04306936e-01 9.63158011e-01 1.42296597e-01 -8.44784558e-01
9.28705093e-03 -2.33512521e-01 6.33127153e-01 -1.24034905e+00
-5.55283427e-01 -4.91642743e-01 -1.44567060e+00 4.79866773e-01
2.20524177e-01 -3.68012458e-01 -7.67743230e-01 1.54178584e+00
3.65897387e-01 4.38549280e-01 3.61959994e-01 3.28142941e-01
2.15874672e-01 1.28581262e+00 4.54110801e-02 -5.15418708e-01
1.04104495e+00 -1.91570178e-01 -8.87118459e-01 8.08783546e-02
1.56253830e-01 -6.02189898e-01 2.74406284e-01 4.34204698e-01
-7.69799471e-01 -4.04363960e-01 -8.04709971e-01 6.64062560e-01
-8.16003799e-01 -1.58948973e-01 2.14487478e-01 6.00528717e-01
-1.40914929e+00 9.58890438e-01 -1.31774819e+00 -2.50196457e-01
-3.98476213e-01 3.54866907e-02 6.52659163e-02 4.23377842e-01
-1.62209535e+00 1.29452980e+00 4.61328477e-01 7.14986205e-01
-9.08007145e-01 -5.20096481e-01 -7.46751666e-01 1.08066656e-01
5.15883900e-02 -5.25691807e-01 1.07344866e+00 -4.58134562e-01
-1.76061749e+00 3.57160419e-02 -3.59432787e-01 -3.89571846e-01
5.16816020e-01 1.12583123e-01 -5.94920814e-01 -2.06857100e-01
-1.09329507e-01 -1.72657594e-02 1.01948214e+00 -1.28147948e+00
-2.80402124e-01 -2.48406678e-01 -6.02879226e-01 -1.30431145e-01
4.01204109e-01 6.91500306e-02 5.87206304e-01 -8.58029246e-01
3.33647043e-01 -1.18764448e+00 -5.28195500e-01 -4.85307157e-01
-1.77444652e-01 -1.08307153e-01 6.69972479e-01 -9.04256284e-01
1.20731926e+00 -1.76957643e+00 2.29458824e-01 5.44550657e-01
-1.65227413e-01 2.66609818e-01 1.12902038e-01 9.58226085e-01
-6.91174641e-02 2.18214929e-01 -7.52664566e-01 -3.51695001e-01
6.25068173e-02 5.12532413e-01 -7.65830874e-01 3.64545792e-01
3.07453901e-01 2.53478140e-01 -7.53766775e-01 4.04604152e-02
4.29104477e-01 7.85977721e-01 -8.71610641e-02 3.56564447e-02
-8.10227841e-02 6.83139563e-01 -3.24227154e-01 3.35268259e-01
8.09647858e-01 -1.51774183e-01 -3.85282049e-03 2.22684547e-01
-7.05311835e-01 -1.39442608e-01 -1.63098669e+00 8.40348899e-01
-5.16805172e-01 5.58795869e-01 3.15558314e-01 -8.33680093e-01
1.05009151e+00 6.26906872e-01 2.94397444e-01 5.55244647e-02
-1.85102180e-01 5.50982714e-01 5.75769180e-03 -3.28934193e-01
7.48609245e-01 -4.90384579e-01 1.71452910e-01 2.92586297e-01
-4.02910896e-02 -3.97309214e-01 3.78173701e-02 -9.27451700e-02
8.07920933e-01 3.31322849e-01 4.36855167e-01 -8.93562019e-01
4.53716427e-01 -1.66589975e-01 6.15898609e-01 8.55019748e-01
1.74317788e-02 2.42702276e-01 4.74728942e-01 -4.42017972e-01
-1.12513459e+00 -9.30398226e-01 -6.94230735e-01 6.52952015e-01
-4.98171240e-01 -1.83433339e-01 -1.78509638e-01 2.80138344e-01
2.05904871e-01 8.26911151e-01 -6.41134143e-01 2.92722434e-01
-1.64113700e-01 -1.15252137e+00 4.92890477e-01 2.59476811e-01
1.21059455e-01 -9.39135432e-01 -6.13838255e-01 8.60201836e-01
1.07305616e-01 -8.15176547e-01 4.49917138e-01 3.77796710e-01
-1.14570177e+00 -3.84802997e-01 -7.33253419e-01 7.42796808e-02
7.03901798e-02 -3.84538770e-01 1.13852823e+00 -6.17133617e-01
1.33307964e-01 3.05359125e-01 -2.15506360e-01 -4.62088794e-01
-7.43069351e-01 -2.47300476e-01 4.95527387e-01 4.46137190e-02
7.95452595e-02 -1.00231564e+00 -3.44718248e-01 2.87341237e-01
-1.09794986e+00 -1.32672206e-01 2.81732053e-01 9.65332806e-01
3.76668006e-01 2.65927255e-01 7.61536181e-01 -3.80654752e-01
8.53331566e-01 -9.37842011e-01 -1.10514486e+00 1.63077623e-01
-7.43186414e-01 1.74507245e-01 4.53404069e-01 -2.92390347e-01
-1.07332790e+00 -1.16216160e-01 -3.66407186e-02 -2.73201346e-01
-3.37000728e-01 1.28004777e+00 4.56490129e-01 1.03180282e-01
4.63615507e-01 4.68824714e-01 -2.25421842e-02 -3.58404219e-01
4.36713547e-02 5.76452732e-01 3.36874038e-01 -7.95833945e-01
6.70536816e-01 2.82288253e-01 4.07862484e-01 -1.22424400e+00
-1.76120669e-01 -2.44466409e-01 -5.60871243e-01 -4.39780235e-01
7.51460493e-01 -8.43914747e-01 -5.78750134e-01 9.72484231e-01
-1.28338683e+00 -4.87532377e-01 -6.52253851e-02 7.58412182e-01
-5.14718950e-01 2.64466256e-02 -6.71101451e-01 -1.62323451e+00
-9.48010236e-02 -9.55173492e-01 1.05167449e+00 2.09472179e-02
-3.08566205e-02 -1.43087459e+00 6.13512814e-01 -6.08326972e-01
1.11644459e+00 6.91210985e-01 7.64438748e-01 -3.10619503e-01
-1.22701161e-01 -7.55134374e-02 -3.11580468e-02 4.16514099e-01
-5.74920028e-02 5.69039881e-01 -9.55491781e-01 -2.52264559e-01
4.00645941e-01 1.20114185e-01 7.72533178e-01 8.47288191e-01
4.43451077e-01 -8.75832886e-02 -1.48631662e-01 3.23700458e-01
1.56514692e+00 3.20630044e-01 4.28625524e-01 -2.54733339e-02
2.94817328e-01 9.08280373e-01 1.92474887e-01 7.40441561e-01
3.03249061e-01 3.95635843e-01 2.02143505e-01 2.25270033e-01
7.45902717e-01 1.62993208e-01 4.37747985e-01 1.23713827e+00
-4.64630604e-01 -2.61630118e-01 -1.46680808e+00 4.54462141e-01
-1.90683901e+00 -1.21991074e+00 -5.26035070e-01 2.09933829e+00
6.79304659e-01 -6.16287142e-02 -1.88292935e-01 2.33331608e-04
8.04917157e-01 2.90679842e-01 -4.83326972e-01 -5.79393685e-01
-2.23050892e-01 1.57446921e-01 4.66318756e-01 7.23596632e-01
-1.19098771e+00 5.47823191e-01 6.33643341e+00 5.93515933e-01
-9.86296654e-01 1.20674215e-01 4.41532344e-01 4.54037368e-01
-2.24504277e-01 1.80872753e-01 -9.72999096e-01 5.60941458e-01
1.82862151e+00 -1.29002929e-01 2.64837533e-01 4.80666131e-01
8.19914877e-01 -5.08842230e-01 -6.11078918e-01 6.48595750e-01
-5.52223027e-01 -1.17982483e+00 -3.74533385e-01 3.88822794e-01
8.42772543e-01 2.54265368e-01 -7.76254162e-02 3.68698716e-01
7.47483790e-01 -1.10231686e+00 4.65008885e-01 1.35852015e+00
5.03330350e-01 -7.06905186e-01 8.81731987e-01 5.96631765e-01
-1.22090220e+00 1.42543912e-01 -2.37232715e-01 -1.15634307e-01
6.42245770e-01 1.09677005e+00 -1.82761461e-01 6.78605080e-01
8.11874688e-01 8.12727690e-01 -3.97162810e-02 8.71414244e-01
2.86315054e-01 8.33151579e-01 -8.79245937e-01 4.93120439e-02
5.41746736e-01 -6.59705579e-01 9.92904484e-01 1.15080202e+00
1.05017483e+00 2.89022952e-01 -5.23233190e-02 8.57901156e-01
8.04937601e-01 -2.49418452e-01 -8.68299246e-01 3.53996679e-02
4.55692708e-01 1.00505817e+00 -5.83667576e-01 -4.01557028e-01
-3.07777710e-02 3.08630705e-01 -2.89906204e-01 6.25687361e-01
-4.95688975e-01 -7.94751570e-02 6.64116800e-01 -4.35726121e-02
3.82046521e-01 -7.18235850e-01 1.66056696e-02 -1.13421738e+00
-3.30080122e-01 -6.09224379e-01 -1.10556800e-02 -1.01248705e+00
-1.69238973e+00 8.42006266e-01 5.08658588e-01 -1.28386271e+00
-7.56021976e-01 -6.81298494e-01 -5.41135728e-01 1.51167905e+00
-1.35470355e+00 -7.94132531e-01 -9.75512993e-03 4.25297588e-01
1.00913182e-01 -9.54258665e-02 1.10680485e+00 -8.00767392e-02
-5.03156900e-01 -5.60325980e-01 1.07874537e+00 -5.31731427e-01
3.89700770e-01 -1.37839782e+00 4.37475502e-01 8.99194121e-01
-5.72270811e-01 6.17214441e-01 1.19917262e+00 -8.71164560e-01
-1.04768991e+00 -1.07482731e+00 9.41062629e-01 -2.92982817e-01
1.10998368e+00 -1.87818095e-01 -1.17664468e+00 3.99194956e-01
1.09884515e-01 8.35444629e-02 5.44986606e-01 4.80352864e-02
-2.63134222e-02 9.74449068e-02 -1.07597256e+00 2.88684189e-01
2.04742894e-01 -5.98761678e-01 -6.42694235e-01 1.73678771e-01
2.22456604e-01 -2.90515795e-02 -1.20833874e+00 7.12934554e-01
5.41731060e-01 -8.92440975e-01 5.57847738e-01 -4.34695750e-01
2.62537807e-01 -4.70412165e-01 -5.61319709e-01 -1.78452194e+00
-2.74236888e-01 -7.94514656e-01 -1.96570933e-01 1.14824510e+00
2.01515928e-01 -1.15706730e+00 1.01374932e-01 6.75137877e-01
5.04142046e-03 -3.88682187e-01 -1.22153878e+00 -1.03275681e+00
4.91499931e-01 -6.12489939e-01 4.24435139e-01 9.88169193e-01
-4.69088346e-01 -8.50380361e-02 -5.41930676e-01 5.47868073e-01
7.10624933e-01 -8.60582218e-02 3.94506842e-01 -1.84680700e+00
-1.79799736e-01 -4.92126942e-01 -2.70721525e-01 -4.86742556e-01
1.65274277e-01 -3.38138074e-01 3.73324305e-01 -1.11171615e+00
-4.57414031e-01 -5.88931561e-01 -1.48303941e-01 -4.13937829e-02
7.93195739e-02 -2.78457642e-01 -2.03726396e-01 3.28670055e-01
4.46328223e-01 8.39886010e-01 5.02581835e-01 3.13382804e-01
-2.32673779e-01 2.38842130e-01 2.25454152e-01 6.49674892e-01
7.77518153e-01 -6.75536454e-01 -2.51024097e-01 1.74706638e-01
5.29100180e-01 8.27436805e-01 6.08192086e-01 -1.02245772e+00
2.08437473e-01 -2.49169216e-01 2.05252811e-01 -9.88038957e-01
2.58098572e-01 -7.66574383e-01 9.14083362e-01 3.90673757e-01
-1.99000333e-02 5.30332923e-01 4.50601459e-01 8.34483206e-01
-5.44423699e-01 -1.97802693e-01 7.85645843e-01 -5.41458800e-02
-4.12996411e-01 1.40545279e-01 -1.15477765e+00 -3.97830516e-01
8.00354064e-01 1.35675445e-01 -1.90631121e-01 -5.84722579e-01
-1.28010070e+00 1.37781903e-01 1.47598475e-01 -7.79048353e-02
2.04280376e-01 -1.08965874e+00 -9.51398313e-01 4.17703204e-02
-3.08743536e-01 -2.00870812e-01 4.96202141e-01 1.02172673e+00
-4.66200709e-01 4.22939301e-01 1.06606921e-02 -8.10052693e-01
-5.59413254e-01 2.42764205e-01 6.58442855e-01 -5.58553755e-01
-2.67357022e-01 4.35516536e-01 -1.67548686e-01 -6.76498890e-01
-4.10875261e-01 -5.34063399e-01 -2.15728119e-01 5.74718177e-01
3.79276812e-01 5.66407502e-01 7.70537779e-02 -9.78168070e-01
-2.02548012e-01 3.90306741e-01 6.98195577e-01 -6.72102094e-01
1.60995865e+00 -3.38845342e-01 -4.90260333e-01 1.31237602e+00
7.57346213e-01 -4.24509168e-01 -1.35072207e+00 -3.03057283e-01
3.21963876e-01 5.36620989e-02 3.00864488e-01 -6.20452583e-01
-5.70629060e-01 1.04778397e+00 7.47329354e-01 8.17942441e-01
8.85085762e-01 -7.25182354e-01 1.74588546e-01 4.25871551e-01
3.20835590e-01 -9.54397261e-01 -6.61546469e-01 1.00580418e+00
1.14910674e+00 -1.07203341e+00 -8.23185891e-02 1.53402552e-01
-3.49304855e-01 1.37526262e+00 -7.59075508e-02 -3.01998913e-01
1.55935371e+00 4.71421272e-01 -2.58538306e-01 7.04627763e-03
-9.78420079e-01 -1.32088244e-01 1.55375436e-01 2.36218810e-01
2.46049225e-01 3.44843864e-01 -1.81325704e-01 3.81279796e-01
-3.32342461e-02 9.37958956e-02 5.63798130e-01 9.69053209e-01
-2.79590607e-01 -8.37478817e-01 -7.49851763e-01 4.59927499e-01
-3.46213758e-01 -3.04318041e-01 4.28272367e-01 6.47325933e-01
-2.64810920e-01 9.77258980e-01 1.01790048e-01 -1.86880857e-01
9.58480909e-02 3.74181986e-01 -1.45920962e-01 -3.28482926e-01
-5.69104314e-01 3.25964957e-01 7.05089942e-02 -2.93398052e-01
-6.76062226e-01 -1.10651684e+00 -5.55467963e-01 -6.13222182e-01
-2.86458075e-01 3.95181417e-01 1.06048369e+00 1.07197928e+00
3.51142913e-01 4.71548706e-01 6.38536453e-01 -1.63568389e+00
-7.82768846e-01 -1.30945623e+00 -1.09977233e+00 -3.00193608e-01
5.32432556e-01 -8.79004836e-01 -7.59533405e-01 -1.05102763e-01] | [6.56524658203125, 3.309788227081299] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.